gitweb/gitweb.perl

   1 #!/usr/bin/perl
   2
   3 # gitweb - simple web interface to track changes in git repositories
   4 #
   5 # (C) 2005-2006, Kay Sievers <kay.sievers@vrfy.org>
   6 # (C) 2005, Christian Gierke
   7 #
   8 # This program is licensed under the GPLv2
   9
  10 use 5.008;
  11 use strict;
  12 use warnings;
  13 # handle ACL in file access tests
  14 use filetest 'access';
  15 use CGI qw(:standard :escapeHTML -nosticky);
  16 use CGI::Util qw(unescape);
  17 use CGI::Carp qw(fatalsToBrowser set_message);
  18 use Encode;
  19 use Fcntl ':mode';
  20 use File::Find qw();
  21 use File::Basename qw(basename);
  22 use Time::HiRes qw(gettimeofday tv_interval);
  23 use Digest::MD5 qw(md5_hex);
  24
  25 binmode STDOUT, ':utf8';
  26
  27 if (!defined($CGI::VERSION) || $CGI::VERSION < 4.08) {
  28         eval 'sub CGI::multi_param { CGI::param(@_) }'
  29 }
  30
  31 our $t0 = [ gettimeofday() ];
  32 our $number_of_git_cmds = 0;
  33
  34 BEGIN {
  35         CGI->compile() if $ENV{'MOD_PERL'};
  36 }
  37
  38 our $version = "++GIT_VERSION++";
  39
  40 our ($my_url, $my_uri, $base_url, $path_info, $home_link);
  41 sub evaluate_uri {
  42         our $cgi;
  43
  44         our $my_url = $cgi->url();
  45         our $my_uri = $cgi->url(-absolute => 1);
  46
  47         # Base URL for relative URLs in gitweb ($logo, $favicon, ...),
  48         # needed and used only for URLs with nonempty PATH_INFO
  49         our $base_url = $my_url;
  50
  51         # When the script is used as DirectoryIndex, the URL does not contain the name
  52         # of the script file itself, and $cgi->url() fails to strip PATH_INFO, so we
  53         # have to do it ourselves. We make $path_info global because it's also used
  54         # later on.
  55         #
  56         # Another issue with the script being the DirectoryIndex is that the resulting
  57         # $my_url data is not the full script URL: this is good, because we want
  58         # generated links to keep implying the script name if it wasn't explicitly
  59         # indicated in the URL we're handling, but it means that $my_url cannot be used
  60         # as base URL.
  61         # Therefore, if we needed to strip PATH_INFO, then we know that we have
  62         # to build the base URL ourselves:
  63         our $path_info = decode_utf8($ENV{"PATH_INFO"});
  64         if ($path_info) {
  65                 # $path_info has already been URL-decoded by the web server, but
  66                 # $my_url and $my_uri have not. URL-decode them so we can properly
  67                 # strip $path_info.
  68                 $my_url = unescape($my_url);
  69                 $my_uri = unescape($my_uri);
  70                 if ($my_url =~ s,\Q$path_info\E$,, &&
  71                     $my_uri =~ s,\Q$path_info\E$,, &&
  72                     defined $ENV{'SCRIPT_NAME'}) {
  73                         $base_url = $cgi->url(-base => 1) . $ENV{'SCRIPT_NAME'};
  74                 }
  75         }
  76
  77         # target of the home link on top of all pages
  78         our $home_link = $my_uri || "/";
  79 }
  80
  81 # core git executable to use
  82 # this can just be "git" if your webserver has a sensible PATH
  83 our $GIT = "++GIT_BINDIR++/git";
  84
  85 # absolute fs-path which will be prepended to the project path
  86 #our $projectroot = "/pub/scm";
  87 our $projectroot = "++GITWEB_PROJECTROOT++";
  88
  89 # fs traversing limit for getting project list
  90 # the number is relative to the projectroot
  91 our $project_maxdepth = "++GITWEB_PROJECT_MAXDEPTH++";
  92
  93 # string of the home link on top of all pages
  94 our $home_link_str = "++GITWEB_HOME_LINK_STR++";
  95
  96 # extra breadcrumbs preceding the home link
  97 our @extra_breadcrumbs = ();
  98
  99 # name of your site or organization to appear in page titles
 100 # replace this with something more descriptive for clearer bookmarks
 101 our $site_name = "++GITWEB_SITENAME++"
 102                  || ($ENV{'SERVER_NAME'} || "Untitled") . " Git";
 103
 104 # html snippet to include in the <head> section of each page
 105 our $site_html_head_string = "++GITWEB_SITE_HTML_HEAD_STRING++";
 106 # filename of html text to include at top of each page
 107 our $site_header = "++GITWEB_SITE_HEADER++";
 108 # html text to include at home page
 109 our $home_text = "++GITWEB_HOMETEXT++";
 110 # filename of html text to include at bottom of each page
 111 our $site_footer = "++GITWEB_SITE_FOOTER++";
 112
 113 # URI of stylesheets
 114 our @stylesheets = ("++GITWEB_CSS++");
 115 # URI of a single stylesheet, which can be overridden in GITWEB_CONFIG.
 116 our $stylesheet = undef;
 117 # URI of GIT logo (72x27 size)
 118 our $logo = "++GITWEB_LOGO++";
 119 # URI of GIT favicon, assumed to be image/png type
 120 our $favicon = "++GITWEB_FAVICON++";
 121 # URI of gitweb.js (JavaScript code for gitweb)
 122 our $javascript = "++GITWEB_JS++";
 123
 124 # URI and label (title) of GIT logo link
 125 #our $logo_url = "http://www.kernel.org/pub/software/scm/git/docs/";
 126 #our $logo_label = "git documentation";
 127 our $logo_url = "http://git-scm.com/";
 128 our $logo_label = "git homepage";
 129
 130 # source of projects list
 131 our $projects_list = "++GITWEB_LIST++";
 132
 133 # the width (in characters) of the projects list "Description" column
 134 our $projects_list_description_width = 25;
 135
 136 # group projects by category on the projects list
 137 # (enabled if this variable evaluates to true)
 138 our $projects_list_group_categories = 0;
 139
 140 # default category if none specified
 141 # (leave the empty string for no category)
 142 our $project_list_default_category = "";
 143
 144 # default order of projects list
 145 # valid values are none, project, descr, owner, and age
 146 our $default_projects_order = "project";
 147
 148 # show repository only if this file exists
 149 # (only effective if this variable evaluates to true)
 150 our $export_ok = "++GITWEB_EXPORT_OK++";
 151
 152 # don't generate age column on the projects list page
 153 our $omit_age_column = 0;
 154
 155 # don't generate information about owners of repositories
 156 our $omit_owner=0;
 157
 158 # show repository only if this subroutine returns true
 159 # when given the path to the project, for example:
 160 #    sub { return -e "$_[0]/git-daemon-export-ok"; }
 161 our $export_auth_hook = undef;
 162
 163 # only allow viewing of repositories also shown on the overview page
 164 our $strict_export = "++GITWEB_STRICT_EXPORT++";
 165
 166 # list of git base URLs used for URL to where fetch project from,
 167 # i.e. full URL is "$git_base_url/$project"
 168 our @git_base_url_list = grep { $_ ne '' } ("++GITWEB_BASE_URL++");
 169
 170 # default blob_plain mimetype and default charset for text/plain blob
 171 our $default_blob_plain_mimetype = 'text/plain';
 172 our $default_text_plain_charset  = undef;
 173
 174 # file to use for guessing MIME types before trying /etc/mime.types
 175 # (relative to the current git repository)
 176 our $mimetypes_file = undef;
 177
 178 # assume this charset if line contains non-UTF-8 characters;
 179 # it should be valid encoding (see Encoding::Supported(3pm) for list),
 180 # for which encoding all byte sequences are valid, for example
 181 # 'iso-8859-1' aka 'latin1' (it is decoded without checking, so it
 182 # could be even 'utf-8' for the old behavior)
 183 our $fallback_encoding = 'latin1';
 184
 185 # rename detection options for git-diff and git-diff-tree
 186 # - default is '-M', with the cost proportional to
 187 #   (number of removed files) * (number of new files).
 188 # - more costly is '-C' (which implies '-M'), with the cost proportional to
 189 #   (number of changed files + number of removed files) * (number of new files)
 190 # - even more costly is '-C', '--find-copies-harder' with cost
 191 #   (number of files in the original tree) * (number of new files)
 192 # - one might want to include '-B' option, e.g. '-B', '-M'
 193 our @diff_opts = ('-M'); # taken from git_commit
 194
 195 # Disables features that would allow repository owners to inject script into
 196 # the gitweb domain.
 197 our $prevent_xss = 0;
 198
 199 # Path to the highlight executable to use (must be the one from
 200 # http://www.andre-simon.de due to assumptions about parameters and output).
 201 # Useful if highlight is not installed on your webserver's PATH.
 202 # [Default: highlight]
 203 our $highlight_bin = "++HIGHLIGHT_BIN++";
 204
 205 # information about snapshot formats that gitweb is capable of serving
 206 our %known_snapshot_formats = (
 207         # name => {
 208         #       'display' => display name,
 209         #       'type' => mime type,
 210         #       'suffix' => filename suffix,
 211         #       'format' => --format for git-archive,
 212         #       'compressor' => [compressor command and arguments]
 213         #                       (array reference, optional)
 214         #       'disabled' => boolean (optional)}
 215         #
 216         'tgz' => {
 217                 'display' => 'tar.gz',
 218                 'type' => 'application/x-gzip',
 219                 'suffix' => '.tar.gz',
 220                 'format' => 'tar',
 221                 'compressor' => ['gzip', '-n']},
 222
 223         'tbz2' => {
 224                 'display' => 'tar.bz2',
 225                 'type' => 'application/x-bzip2',
 226                 'suffix' => '.tar.bz2',
 227                 'format' => 'tar',
 228                 'compressor' => ['bzip2']},
 229
 230         'txz' => {
 231                 'display' => 'tar.xz',
 232                 'type' => 'application/x-xz',
 233                 'suffix' => '.tar.xz',
 234                 'format' => 'tar',
 235                 'compressor' => ['xz'],
 236                 'disabled' => 1},
 237
 238         'zip' => {
 239                 'display' => 'zip',
 240                 'type' => 'application/x-zip',
 241                 'suffix' => '.zip',
 242                 'format' => 'zip'},
 243 );
 244
 245 # Aliases so we understand old gitweb.snapshot values in repository
 246 # configuration.
 247 our %known_snapshot_format_aliases = (
 248         'gzip'  => 'tgz',
 249         'bzip2' => 'tbz2',
 250         'xz'    => 'txz',
 251
 252         # backward compatibility: legacy gitweb config support
 253         'x-gzip' => undef, 'gz' => undef,
 254         'x-bzip2' => undef, 'bz2' => undef,
 255         'x-zip' => undef, '' => undef,
 256 );
 257
 258 # Pixel sizes for icons and avatars. If the default font sizes or lineheights
 259 # are changed, it may be appropriate to change these values too via
 260 # $GITWEB_CONFIG.
 261 our %avatar_size = (
 262         'default' => 16,
 263         'double'  => 32
 264 );
 265
 266 # Used to set the maximum load that we will still respond to gitweb queries.
 267 # If server load exceed this value then return "503 server busy" error.
 268 # If gitweb cannot determined server load, it is taken to be 0.
 269 # Leave it undefined (or set to 'undef') to turn off load checking.
 270 our $maxload = 300;
 271
 272 # configuration for 'highlight' (http://www.andre-simon.de/)
 273 # match by basename
 274 our %highlight_basename = (
 275         #'Program' => 'py',
 276         #'Library' => 'py',
 277         'SConstruct' => 'py', # SCons equivalent of Makefile
 278         'Makefile' => 'make',
 279 );
 280 # match by extension
 281 our %highlight_ext = (
 282         # main extensions, defining name of syntax;
 283         # see files in /usr/share/highlight/langDefs/ directory
 284         (map { $_ => $_ } qw(py rb java css js tex bib xml awk bat ini spec tcl sql)),
 285         # alternate extensions, see /etc/highlight/filetypes.conf
 286         (map { $_ => 'c'   } qw(c h)),
 287         (map { $_ => 'sh'  } qw(sh bash zsh ksh)),
 288         (map { $_ => 'cpp' } qw(cpp cxx c++ cc)),
 289         (map { $_ => 'php' } qw(php php3 php4 php5 phps)),
 290         (map { $_ => 'pl'  } qw(pl perl pm)), # perhaps also 'cgi'
 291         (map { $_ => 'make'} qw(make mak mk)),
 292         (map { $_ => 'xml' } qw(xml xhtml html htm)),
 293 );
 294
 295 # You define site-wide feature defaults here; override them with
 296 # $GITWEB_CONFIG as necessary.
 297 our %feature = (
 298         # feature => {
 299         #       'sub' => feature-sub (subroutine),
 300         #       'override' => allow-override (boolean),
 301         #       'default' => [ default options...] (array reference)}
 302         #
 303         # if feature is overridable (it means that allow-override has true value),
 304         # then feature-sub will be called with default options as parameters;
 305         # return value of feature-sub indicates if to enable specified feature
 306         #
 307         # if there is no 'sub' key (no feature-sub), then feature cannot be
 308         # overridden
 309         #
 310         # use gitweb_get_feature(<feature>) to retrieve the <feature> value
 311         # (an array) or gitweb_check_feature(<feature>) to check if <feature>
 312         # is enabled
 313
 314         # Enable the 'blame' blob view, showing the last commit that modified
 315         # each line in the file. This can be very CPU-intensive.
 316
 317         # To enable system wide have in $GITWEB_CONFIG
 318         # $feature{'blame'}{'default'} = [1];
 319         # To have project specific config enable override in $GITWEB_CONFIG
 320         # $feature{'blame'}{'override'} = 1;
 321         # and in project config gitweb.blame = 0|1;
 322         'blame' => {
 323                 'sub' => sub { feature_bool('blame', @_) },
 324                 'override' => 0,
 325                 'default' => [0]},
 326
 327         # Enable the 'snapshot' link, providing a compressed archive of any
 328         # tree. This can potentially generate high traffic if you have large
 329         # project.
 330
 331         # Value is a list of formats defined in %known_snapshot_formats that
 332         # you wish to offer.
 333         # To disable system wide have in $GITWEB_CONFIG
 334         # $feature{'snapshot'}{'default'} = [];
 335         # To have project specific config enable override in $GITWEB_CONFIG
 336         # $feature{'snapshot'}{'override'} = 1;
 337         # and in project config, a comma-separated list of formats or "none"
 338         # to disable.  Example: gitweb.snapshot = tbz2,zip;
 339         'snapshot' => {
 340                 'sub' => \&feature_snapshot,
 341                 'override' => 0,
 342                 'default' => ['tgz']},
 343
 344         # Enable text search, which will list the commits which match author,
 345         # committer or commit text to a given string.  Enabled by default.
 346         # Project specific override is not supported.
 347         #
 348         # Note that this controls all search features, which means that if
 349         # it is disabled, then 'grep' and 'pickaxe' search would also be
 350         # disabled.
 351         'search' => {
 352                 'override' => 0,
 353                 'default' => [1]},
 354
 355         # Enable grep search, which will list the files in currently selected
 356         # tree containing the given string. Enabled by default. This can be
 357         # potentially CPU-intensive, of course.
 358         # Note that you need to have 'search' feature enabled too.
 359
 360         # To enable system wide have in $GITWEB_CONFIG
 361         # $feature{'grep'}{'default'} = [1];
 362         # To have project specific config enable override in $GITWEB_CONFIG
 363         # $feature{'grep'}{'override'} = 1;
 364         # and in project config gitweb.grep = 0|1;
 365         'grep' => {
 366                 'sub' => sub { feature_bool('grep', @_) },
 367                 'override' => 0,
 368                 'default' => [1]},
 369
 370         # Enable the pickaxe search, which will list the commits that modified
 371         # a given string in a file. This can be practical and quite faster
 372         # alternative to 'blame', but still potentially CPU-intensive.
 373         # Note that you need to have 'search' feature enabled too.
 374
 375         # To enable system wide have in $GITWEB_CONFIG
 376         # $feature{'pickaxe'}{'default'} = [1];
 377         # To have project specific config enable override in $GITWEB_CONFIG
 378         # $feature{'pickaxe'}{'override'} = 1;
 379         # and in project config gitweb.pickaxe = 0|1;
 380         'pickaxe' => {
 381                 'sub' => sub { feature_bool('pickaxe', @_) },
 382                 'override' => 0,
 383                 'default' => [1]},
 384
 385         # Enable showing size of blobs in a 'tree' view, in a separate
 386         # column, similar to what 'ls -l' does.  This cost a bit of IO.
 387
 388         # To disable system wide have in $GITWEB_CONFIG
 389         # $feature{'show-sizes'}{'default'} = [0];
 390         # To have project specific config enable override in $GITWEB_CONFIG
 391         # $feature{'show-sizes'}{'override'} = 1;
 392         # and in project config gitweb.showsizes = 0|1;
 393         'show-sizes' => {
 394                 'sub' => sub { feature_bool('showsizes', @_) },
 395                 'override' => 0,
 396                 'default' => [1]},
 397
 398         # Make gitweb use an alternative format of the URLs which can be
 399         # more readable and natural-looking: project name is embedded
 400         # directly in the path and the query string contains other
 401         # auxiliary information. All gitweb installations recognize
 402         # URL in either format; this configures in which formats gitweb
 403         # generates links.
 404
 405         # To enable system wide have in $GITWEB_CONFIG
 406         # $feature{'pathinfo'}{'default'} = [1];
 407         # Project specific override is not supported.
 408
 409         # Note that you will need to change the default location of CSS,
 410         # favicon, logo and possibly other files to an absolute URL. Also,
 411         # if gitweb.cgi serves as your indexfile, you will need to force
 412         # $my_uri to contain the script name in your $GITWEB_CONFIG.
 413         'pathinfo' => {
 414                 'override' => 0,
 415                 'default' => [0]},
 416
 417         # Make gitweb consider projects in project root subdirectories
 418         # to be forks of existing projects. Given project $projname.git,
 419         # projects matching $projname/*.git will not be shown in the main
 420         # projects list, instead a '+' mark will be added to $projname
 421         # there and a 'forks' view will be enabled for the project, listing
 422         # all the forks. If project list is taken from a file, forks have
 423         # to be listed after the main project.
 424
 425         # To enable system wide have in $GITWEB_CONFIG
 426         # $feature{'forks'}{'default'} = [1];
 427         # Project specific override is not supported.
 428         'forks' => {
 429                 'override' => 0,
 430                 'default' => [0]},
 431
 432         # Insert custom links to the action bar of all project pages.
 433         # This enables you mainly to link to third-party scripts integrating
 434         # into gitweb; e.g. git-browser for graphical history representation
 435         # or custom web-based repository administration interface.
 436
 437         # The 'default' value consists of a list of triplets in the form
 438         # (label, link, position) where position is the label after which
 439         # to insert the link and link is a format string where %n expands
 440         # to the project name, %f to the project path within the filesystem,
 441         # %h to the current hash (h gitweb parameter) and %b to the current
 442         # hash base (hb gitweb parameter); %% expands to %.
 443
 444         # To enable system wide have in $GITWEB_CONFIG e.g.
 445         # $feature{'actions'}{'default'} = [('graphiclog',
 446         #       '/git-browser/by-commit.html?r=%n', 'summary')];
 447         # Project specific override is not supported.
 448         'actions' => {
 449                 'override' => 0,
 450                 'default' => []},
 451
 452         # Allow gitweb scan project content tags of project repository,
 453         # and display the popular Web 2.0-ish "tag cloud" near the projects
 454         # list.  Note that this is something COMPLETELY different from the
 455         # normal Git tags.
 456
 457         # gitweb by itself can show existing tags, but it does not handle
 458         # tagging itself; you need to do it externally, outside gitweb.
 459         # The format is described in git_get_project_ctags() subroutine.
 460         # You may want to install the HTML::TagCloud Perl module to get
 461         # a pretty tag cloud instead of just a list of tags.
 462
 463         # To enable system wide have in $GITWEB_CONFIG
 464         # $feature{'ctags'}{'default'} = [1];
 465         # Project specific override is not supported.
 466
 467         # In the future whether ctags editing is enabled might depend
 468         # on the value, but using 1 should always mean no editing of ctags.
 469         'ctags' => {
 470                 'override' => 0,
 471                 'default' => [0]},
 472
 473         # The maximum number of patches in a patchset generated in patch
 474         # view. Set this to 0 or undef to disable patch view, or to a
 475         # negative number to remove any limit.
 476
 477         # To disable system wide have in $GITWEB_CONFIG
 478         # $feature{'patches'}{'default'} = [0];
 479         # To have project specific config enable override in $GITWEB_CONFIG
 480         # $feature{'patches'}{'override'} = 1;
 481         # and in project config gitweb.patches = 0|n;
 482         # where n is the maximum number of patches allowed in a patchset.
 483         'patches' => {
 484                 'sub' => \&feature_patches,
 485                 'override' => 0,
 486                 'default' => [16]},
 487
 488         # Avatar support. When this feature is enabled, views such as
 489         # shortlog or commit will display an avatar associated with
 490         # the email of the committer(s) and/or author(s).
 491
 492         # Currently available providers are gravatar and picon.
 493         # If an unknown provider is specified, the feature is disabled.
 494
 495         # Picon currently relies on the indiana.edu database.
 496
 497         # To enable system wide have in $GITWEB_CONFIG
 498         # $feature{'avatar'}{'default'} = ['<provider>'];
 499         # where <provider> is either gravatar or picon.
 500         # To have project specific config enable override in $GITWEB_CONFIG
 501         # $feature{'avatar'}{'override'} = 1;
 502         # and in project config gitweb.avatar = <provider>;
 503         'avatar' => {
 504                 'sub' => \&feature_avatar,
 505                 'override' => 0,
 506                 'default' => ['']},
 507
 508         # Enable displaying how much time and how many git commands
 509         # it took to generate and display page.  Disabled by default.
 510         # Project specific override is not supported.
 511         'timed' => {
 512                 'override' => 0,
 513                 'default' => [0]},
 514
 515         # Enable turning some links into links to actions which require
 516         # JavaScript to run (like 'blame_incremental').  Not enabled by
 517         # default.  Project specific override is currently not supported.
 518         'javascript-actions' => {
 519                 'override' => 0,
 520                 'default' => [0]},
 521
 522         # Enable and configure ability to change common timezone for dates
 523         # in gitweb output via JavaScript.  Enabled by default.
 524         # Project specific override is not supported.
 525         'javascript-timezone' => {
 526                 'override' => 0,
 527                 'default' => [
 528                         'local',     # default timezone: 'utc', 'local', or '(-|+)HHMM' format,
 529                                      # or undef to turn off this feature
 530                         'gitweb_tz', # name of cookie where to store selected timezone
 531                         'datetime',  # CSS class used to mark up dates for manipulation
 532                 ]},
 533
 534         # Syntax highlighting support. This is based on Daniel Svensson's
 535         # and Sham Chukoury's work in gitweb-xmms2.git.
 536         # It requires the 'highlight' program present in $PATH,
 537         # and therefore is disabled by default.
 538
 539         # To enable system wide have in $GITWEB_CONFIG
 540         # $feature{'highlight'}{'default'} = [1];
 541
 542         'highlight' => {
 543                 'sub' => sub { feature_bool('highlight', @_) },
 544                 'override' => 0,
 545                 'default' => [0]},
 546
 547         # Enable displaying of remote heads in the heads list
 548
 549         # To enable system wide have in $GITWEB_CONFIG
 550         # $feature{'remote_heads'}{'default'} = [1];
 551         # To have project specific config enable override in $GITWEB_CONFIG
 552         # $feature{'remote_heads'}{'override'} = 1;
 553         # and in project config gitweb.remoteheads = 0|1;
 554         'remote_heads' => {
 555                 'sub' => sub { feature_bool('remote_heads', @_) },
 556                 'override' => 0,
 557                 'default' => [0]},
 558
 559         # Enable showing branches under other refs in addition to heads
 560
 561         # To set system wide extra branch refs have in $GITWEB_CONFIG
 562         # $feature{'extra-branch-refs'}{'default'} = ['dirs', 'of', 'choice'];
 563         # To have project specific config enable override in $GITWEB_CONFIG
 564         # $feature{'extra-branch-refs'}{'override'} = 1;
 565         # and in project config gitweb.extrabranchrefs = dirs of choice
 566         # Every directory is separated with whitespace.
 567
 568         'extra-branch-refs' => {
 569                 'sub' => \&feature_extra_branch_refs,
 570                 'override' => 0,
 571                 'default' => []},
 572
 573         # Redact e-mail addresses.
 574
 575         # To enable system wide have in $GITWEB_CONFIG
 576         # $feature{'email-privacy'}{'default'} = [1];
 577         'email-privacy' => {
 578                 'sub' => sub { feature_bool('email-privacy', @_) },
 579                 'override' => 1,
 580                 'default' => [0]},
 581 );
 582
 583 sub gitweb_get_feature {
 584         my ($name) = @_;
 585         return unless exists $feature{$name};
 586         my ($sub, $override, @defaults) = (
 587                 $feature{$name}{'sub'},
 588                 $feature{$name}{'override'},
 589                 @{$feature{$name}{'default'}});
 590         # project specific override is possible only if we have project
 591         our $git_dir; # global variable, declared later
 592         if (!$override || !defined $git_dir) {
 593                 return @defaults;
 594         }
 595         if (!defined $sub) {
 596                 warn "feature $name is not overridable";
 597                 return @defaults;
 598         }
 599         return $sub->(@defaults);
 600 }
 601
 602 # A wrapper to check if a given feature is enabled.
 603 # With this, you can say
 604 #
 605 #   my $bool_feat = gitweb_check_feature('bool_feat');
 606 #   gitweb_check_feature('bool_feat') or somecode;
 607 #
 608 # instead of
 609 #
 610 #   my ($bool_feat) = gitweb_get_feature('bool_feat');
 611 #   (gitweb_get_feature('bool_feat'))[0] or somecode;
 612 #
 613 sub gitweb_check_feature {
 614         return (gitweb_get_feature(@_))[0];
 615 }
 616
 617
 618 sub feature_bool {
 619         my $key = shift;
 620         my ($val) = git_get_project_config($key, '--bool');
 621
 622         if (!defined $val) {
 623                 return ($_[0]);
 624         } elsif ($val eq 'true') {
 625                 return (1);
 626         } elsif ($val eq 'false') {
 627                 return (0);
 628         }
 629 }
 630
 631 sub feature_snapshot {
 632         my (@fmts) = @_;
 633
 634         my ($val) = git_get_project_config('snapshot');
 635
 636         if ($val) {
 637                 @fmts = ($val eq 'none' ? () : split /\s*[,\s]\s*/, $val);
 638         }
 639
 640         return @fmts;
 641 }
 642
 643 sub feature_patches {
 644         my @val = (git_get_project_config('patches', '--int'));
 645
 646         if (@val) {
 647                 return @val;
 648         }
 649
 650         return ($_[0]);
 651 }
 652
 653 sub feature_avatar {
 654         my @val = (git_get_project_config('avatar'));
 655
 656         return @val ? @val : @_;
 657 }
 658
 659 sub feature_extra_branch_refs {
 660         my (@branch_refs) = @_;
 661         my $values = git_get_project_config('extrabranchrefs');
 662
 663         if ($values) {
 664                 $values = config_to_multi ($values);
 665                 @branch_refs = ();
 666                 foreach my $value (@{$values}) {
 667                         push @branch_refs, split /\s+/, $value;
 668                 }
 669         }
 670
 671         return @branch_refs;
 672 }
 673
 674 # checking HEAD file with -e is fragile if the repository was
 675 # initialized long time ago (i.e. symlink HEAD) and was pack-ref'ed
 676 # and then pruned.
 677 sub check_head_link {
 678         my ($dir) = @_;
 679         my $headfile = "$dir/HEAD";
 680         return ((-e $headfile) ||
 681                 (-l $headfile && readlink($headfile) =~ /^refs\/heads\//));
 682 }
 683
 684 sub check_export_ok {
 685         my ($dir) = @_;
 686         return (check_head_link($dir) &&
 687                 (!$export_ok || -e "$dir/$export_ok") &&
 688                 (!$export_auth_hook || $export_auth_hook->($dir)));
 689 }
 690
 691 # process alternate names for backward compatibility
 692 # filter out unsupported (unknown) snapshot formats
 693 sub filter_snapshot_fmts {
 694         my @fmts = @_;
 695
 696         @fmts = map {
 697                 exists $known_snapshot_format_aliases{$_} ?
 698                        $known_snapshot_format_aliases{$_} : $_} @fmts;
 699         @fmts = grep {
 700                 exists $known_snapshot_formats{$_} &&
 701                 !$known_snapshot_formats{$_}{'disabled'}} @fmts;
 702 }
 703
 704 sub filter_and_validate_refs {
 705         my @refs = @_;
 706         my %unique_refs = ();
 707
 708         foreach my $ref (@refs) {
 709                 die_error(500, "Invalid ref '$ref' in 'extra-branch-refs' feature") unless (is_valid_ref_format($ref));
 710                 # 'heads' are added implicitly in get_branch_refs().
 711                 $unique_refs{$ref} = 1 if ($ref ne 'heads');
 712         }
 713         return sort keys %unique_refs;
 714 }
 715
 716 # If it is set to code reference, it is code that it is to be run once per
 717 # request, allowing updating configurations that change with each request,
 718 # while running other code in config file only once.
 719 #
 720 # Otherwise, if it is false then gitweb would process config file only once;
 721 # if it is true then gitweb config would be run for each request.
 722 our $per_request_config = 1;
 723
 724 # read and parse gitweb config file given by its parameter.
 725 # returns true on success, false on recoverable error, allowing
 726 # to chain this subroutine, using first file that exists.
 727 # dies on errors during parsing config file, as it is unrecoverable.
 728 sub read_config_file {
 729         my $filename = shift;
 730         return unless defined $filename;
 731         # die if there are errors parsing config file
 732         if (-e $filename) {
 733                 do $filename;
 734                 die $@ if $@;
 735                 return 1;
 736         }
 737         return;
 738 }
 739
 740 our ($GITWEB_CONFIG, $GITWEB_CONFIG_SYSTEM, $GITWEB_CONFIG_COMMON);
 741 sub evaluate_gitweb_config {
 742         our $GITWEB_CONFIG = $ENV{'GITWEB_CONFIG'} || "++GITWEB_CONFIG++";
 743         our $GITWEB_CONFIG_SYSTEM = $ENV{'GITWEB_CONFIG_SYSTEM'} || "++GITWEB_CONFIG_SYSTEM++";
 744         our $GITWEB_CONFIG_COMMON = $ENV{'GITWEB_CONFIG_COMMON'} || "++GITWEB_CONFIG_COMMON++";
 745
 746         # Protect against duplications of file names, to not read config twice.
 747         # Only one of $GITWEB_CONFIG and $GITWEB_CONFIG_SYSTEM is used, so
 748         # there possibility of duplication of filename there doesn't matter.
 749         $GITWEB_CONFIG = ""        if ($GITWEB_CONFIG eq $GITWEB_CONFIG_COMMON);
 750         $GITWEB_CONFIG_SYSTEM = "" if ($GITWEB_CONFIG_SYSTEM eq $GITWEB_CONFIG_COMMON);
 751
 752         # Common system-wide settings for convenience.
 753         # Those settings can be overridden by GITWEB_CONFIG or GITWEB_CONFIG_SYSTEM.
 754         read_config_file($GITWEB_CONFIG_COMMON);
 755
 756         # Use first config file that exists.  This means use the per-instance
 757         # GITWEB_CONFIG if exists, otherwise use GITWEB_SYSTEM_CONFIG.
 758         read_config_file($GITWEB_CONFIG) and return;
 759         read_config_file($GITWEB_CONFIG_SYSTEM);
 760 }
 761
 762 # Get loadavg of system, to compare against $maxload.
 763 # Currently it requires '/proc/loadavg' present to get loadavg;
 764 # if it is not present it returns 0, which means no load checking.
 765 sub get_loadavg {
 766         if( -e '/proc/loadavg' ){
 767                 open my $fd, '<', '/proc/loadavg'
 768                         or return 0;
 769                 my @load = split(/\s+/, scalar <$fd>);
 770                 close $fd;
 771
 772                 # The first three columns measure CPU and IO utilization of the last one,
 773                 # five, and 10 minute periods.  The fourth column shows the number of
 774                 # currently running processes and the total number of processes in the m/n
 775                 # format.  The last column displays the last process ID used.
 776                 return $load[0] || 0;
 777         }
 778         # additional checks for load average should go here for things that don't export
 779         # /proc/loadavg
 780
 781         return 0;
 782 }
 783
 784 # version of the core git binary
 785 our $git_version;
 786 sub evaluate_git_version {
 787         our $git_version = qx("$GIT" --version) =~ m/git version (.*)$/ ? $1 : "unknown";
 788         $number_of_git_cmds++;
 789 }
 790
 791 sub check_loadavg {
 792         if (defined $maxload && get_loadavg() > $maxload) {
 793                 die_error(503, "The load average on the server is too high");
 794         }
 795 }
 796
 797 # ======================================================================
 798 # input validation and dispatch
 799
 800 # Various hash size-related values.
 801 my $sha1_len = 40;
 802 my $sha256_extra_len = 24;
 803 my $sha256_len = $sha1_len + $sha256_extra_len;
 804
 805 # A regex matching $len hex characters. $len may be a range (e.g. 7,64).
 806 sub oid_nlen_regex {
 807         my $len = shift;
 808         my $hchr = qr/[0-9a-fA-F]/;
 809         return qr/(?:(?:$hchr){$len})/;
 810 }
 811
 812 # A regex matching two sets of $nlen hex characters, prefixed by the literal
 813 # string $prefix and with the literal string $infix between them.
 814 sub oid_nlen_prefix_infix_regex {
 815         my $nlen = shift;
 816         my $prefix = shift;
 817         my $infix = shift;
 818
 819         my $rx = oid_nlen_regex($nlen);
 820
 821         return qr/^\Q$prefix\E$rx\Q$infix\E$rx$/;
 822 }
 823
 824 # A regex matching a valid object ID.
 825 our $oid_regex;
 826 {
 827         my $x = oid_nlen_regex($sha1_len);
 828         my $y = oid_nlen_regex($sha256_extra_len);
 829         $oid_regex = qr/(?:$x(?:$y)?)/;
 830 }
 831
 832 # input parameters can be collected from a variety of sources (presently, CGI
 833 # and PATH_INFO), so we define an %input_params hash that collects them all
 834 # together during validation: this allows subsequent uses (e.g. href()) to be
 835 # agnostic of the parameter origin
 836
 837 our %input_params = ();
 838
 839 # input parameters are stored with the long parameter name as key. This will
 840 # also be used in the href subroutine to convert parameters to their CGI
 841 # equivalent, and since the href() usage is the most frequent one, we store
 842 # the name -> CGI key mapping here, instead of the reverse.
 843 #
 844 # XXX: Warning: If you touch this, check the search form for updating,
 845 # too.
 846
 847 our @cgi_param_mapping = (
 848         project => "p",
 849         action => "a",
 850         file_name => "f",
 851         file_parent => "fp",
 852         hash => "h",
 853         hash_parent => "hp",
 854         hash_base => "hb",
 855         hash_parent_base => "hpb",
 856         page => "pg",
 857         order => "o",
 858         searchtext => "s",
 859         searchtype => "st",
 860         snapshot_format => "sf",
 861         extra_options => "opt",
 862         search_use_regexp => "sr",
 863         ctag => "by_tag",
 864         diff_style => "ds",
 865         project_filter => "pf",
 866         # this must be last entry (for manipulation from JavaScript)
 867         javascript => "js"
 868 );
 869 our %cgi_param_mapping = @cgi_param_mapping;
 870
 871 # we will also need to know the possible actions, for validation
 872 our %actions = (
 873         "blame" => \&git_blame,
 874         "blame_incremental" => \&git_blame_incremental,
 875         "blame_data" => \&git_blame_data,
 876         "blobdiff" => \&git_blobdiff,
 877         "blobdiff_plain" => \&git_blobdiff_plain,
 878         "blob" => \&git_blob,
 879         "blob_plain" => \&git_blob_plain,
 880         "commitdiff" => \&git_commitdiff,
 881         "commitdiff_plain" => \&git_commitdiff_plain,
 882         "commit" => \&git_commit,
 883         "forks" => \&git_forks,
 884         "heads" => \&git_heads,
 885         "history" => \&git_history,
 886         "log" => \&git_log,
 887         "patch" => \&git_patch,
 888         "patches" => \&git_patches,
 889         "remotes" => \&git_remotes,
 890         "rss" => \&git_rss,
 891         "atom" => \&git_atom,
 892         "search" => \&git_search,
 893         "search_help" => \&git_search_help,
 894         "shortlog" => \&git_shortlog,
 895         "summary" => \&git_summary,
 896         "tag" => \&git_tag,
 897         "tags" => \&git_tags,
 898         "tree" => \&git_tree,
 899         "snapshot" => \&git_snapshot,
 900         "object" => \&git_object,
 901         # those below don't need $project
 902         "opml" => \&git_opml,
 903         "project_list" => \&git_project_list,
 904         "project_index" => \&git_project_index,
 905 );
 906
 907 # finally, we have the hash of allowed extra_options for the commands that
 908 # allow them
 909 our %allowed_options = (
 910         "--no-merges" => [ qw(rss atom log shortlog history) ],
 911 );
 912
 913 # fill %input_params with the CGI parameters. All values except for 'opt'
 914 # should be single values, but opt can be an array. We should probably
 915 # build an array of parameters that can be multi-valued, but since for the time
 916 # being it's only this one, we just single it out
 917 sub evaluate_query_params {
 918         our $cgi;
 919
 920         while (my ($name, $symbol) = each %cgi_param_mapping) {
 921                 if ($symbol eq 'opt') {
 922                         $input_params{$name} = [ map { decode_utf8($_) } $cgi->multi_param($symbol) ];
 923                 } else {
 924                         $input_params{$name} = decode_utf8($cgi->param($symbol));
 925                 }
 926         }
 927 }
 928
 929 # now read PATH_INFO and update the parameter list for missing parameters
 930 sub evaluate_path_info {
 931         return if defined $input_params{'project'};
 932         return if !$path_info;
 933         $path_info =~ s,^/+,,;
 934         return if !$path_info;
 935
 936         # find which part of PATH_INFO is project
 937         my $project = $path_info;
 938         $project =~ s,/+$,,;
 939         while ($project && !check_head_link("$projectroot/$project")) {
 940                 $project =~ s,/*[^/]*$,,;
 941         }
 942         return unless $project;
 943         $input_params{'project'} = $project;
 944
 945         # do not change any parameters if an action is given using the query string
 946         return if $input_params{'action'};
 947         $path_info =~ s,^\Q$project\E/*,,;
 948
 949         # next, check if we have an action
 950         my $action = $path_info;
 951         $action =~ s,/.*$,,;
 952         if (exists $actions{$action}) {
 953                 $path_info =~ s,^$action/*,,;
 954                 $input_params{'action'} = $action;
 955         }
 956
 957         # list of actions that want hash_base instead of hash, but can have no
 958         # pathname (f) parameter
 959         my @wants_base = (
 960                 'tree',
 961                 'history',
 962         );
 963
 964         # we want to catch, among others
 965         # [$hash_parent_base[:$file_parent]..]$hash_parent[:$file_name]
 966         my ($parentrefname, $parentpathname, $refname, $pathname) =
 967                 ($path_info =~ /^(?:(.+?)(?::(.+))?\.\.)?([^:]+?)?(?::(.+))?$/);
 968
 969         # first, analyze the 'current' part
 970         if (defined $pathname) {
 971                 # we got "branch:filename" or "branch:dir/"
 972                 # we could use git_get_type(branch:pathname), but:
 973                 # - it needs $git_dir
 974                 # - it does a git() call
 975                 # - the convention of terminating directories with a slash
 976                 #   makes it superfluous
 977                 # - embedding the action in the PATH_INFO would make it even
 978                 #   more superfluous
 979                 $pathname =~ s,^/+,,;
 980                 if (!$pathname || substr($pathname, -1) eq "/") {
 981                         $input_params{'action'} ||= "tree";
 982                         $pathname =~ s,/$,,;
 983                 } else {
 984                         # the default action depends on whether we had parent info
 985                         # or not
 986                         if ($parentrefname) {
 987                                 $input_params{'action'} ||= "blobdiff_plain";
 988                         } else {
 989                                 $input_params{'action'} ||= "blob_plain";
 990                         }
 991                 }
 992                 $input_params{'hash_base'} ||= $refname;
 993                 $input_params{'file_name'} ||= $pathname;
 994         } elsif (defined $refname) {
 995                 # we got "branch". In this case we have to choose if we have to
 996                 # set hash or hash_base.
 997                 #
 998                 # Most of the actions without a pathname only want hash to be
 999                 # set, except for the ones specified in @wants_base that want
1000                 # hash_base instead. It should also be noted that hand-crafted
1001                 # links having 'history' as an action and no pathname or hash
1002                 # set will fail, but that happens regardless of PATH_INFO.
1003                 if (defined $parentrefname) {
1004                         # if there is parent let the default be 'shortlog' action
1005                         # (for http://git.example.com/repo.git/A..B links); if there
1006                         # is no parent, dispatch will detect type of object and set
1007                         # action appropriately if required (if action is not set)
1008                         $input_params{'action'} ||= "shortlog";
1009                 }
1010                 if ($input_params{'action'} &&
1011                     grep { $_ eq $input_params{'action'} } @wants_base) {
1012                         $input_params{'hash_base'} ||= $refname;
1013                 } else {
1014                         $input_params{'hash'} ||= $refname;
1015                 }
1016         }
1017
1018         # next, handle the 'parent' part, if present
1019         if (defined $parentrefname) {
1020                 # a missing pathspec defaults to the 'current' filename, allowing e.g.
1021                 # someproject/blobdiff/oldrev..newrev:/filename
1022                 if ($parentpathname) {
1023                         $parentpathname =~ s,^/+,,;
1024                         $parentpathname =~ s,/$,,;
1025                         $input_params{'file_parent'} ||= $parentpathname;
1026                 } else {
1027                         $input_params{'file_parent'} ||= $input_params{'file_name'};
1028                 }
1029                 # we assume that hash_parent_base is wanted if a path was specified,
1030                 # or if the action wants hash_base instead of hash
1031                 if (defined $input_params{'file_parent'} ||
1032                         grep { $_ eq $input_params{'action'} } @wants_base) {
1033                         $input_params{'hash_parent_base'} ||= $parentrefname;
1034                 } else {
1035                         $input_params{'hash_parent'} ||= $parentrefname;
1036                 }
1037         }
1038
1039         # for the snapshot action, we allow URLs in the form
1040         # $project/snapshot/$hash.ext
1041         # where .ext determines the snapshot and gets removed from the
1042         # passed $refname to provide the $hash.
1043         #
1044         # To be able to tell that $refname includes the format extension, we
1045         # require the following two conditions to be satisfied:
1046         # - the hash input parameter MUST have been set from the $refname part
1047         #   of the URL (i.e. they must be equal)
1048         # - the snapshot format MUST NOT have been defined already (e.g. from
1049         #   CGI parameter sf)
1050         # It's also useless to try any matching unless $refname has a dot,
1051         # so we check for that too
1052         if (defined $input_params{'action'} &&
1053                 $input_params{'action'} eq 'snapshot' &&
1054                 defined $refname && index($refname, '.') != -1 &&
1055                 $refname eq $input_params{'hash'} &&
1056                 !defined $input_params{'snapshot_format'}) {
1057                 # We loop over the known snapshot formats, checking for
1058                 # extensions. Allowed extensions are both the defined suffix
1059                 # (which includes the initial dot already) and the snapshot
1060                 # format key itself, with a prepended dot
1061                 while (my ($fmt, $opt) = each %known_snapshot_formats) {
1062                         my $hash = $refname;
1063                         unless ($hash =~ s/(\Q$opt->{'suffix'}\E|\Q.$fmt\E)$//) {
1064                                 next;
1065                         }
1066                         my $sfx = $1;
1067                         # a valid suffix was found, so set the snapshot format
1068                         # and reset the hash parameter
1069                         $input_params{'snapshot_format'} = $fmt;
1070                         $input_params{'hash'} = $hash;
1071                         # we also set the format suffix to the one requested
1072                         # in the URL: this way a request for e.g. .tgz returns
1073                         # a .tgz instead of a .tar.gz
1074                         $known_snapshot_formats{$fmt}{'suffix'} = $sfx;
1075                         last;
1076                 }
1077         }
1078 }
1079
1080 our ($action, $project, $file_name, $file_parent, $hash, $hash_parent, $hash_base,
1081      $hash_parent_base, @extra_options, $page, $searchtype, $search_use_regexp,
1082      $searchtext, $search_regexp, $project_filter);
1083 sub evaluate_and_validate_params {
1084         our $action = $input_params{'action'};
1085         if (defined $action) {
1086                 if (!is_valid_action($action)) {
1087                         die_error(400, "Invalid action parameter");
1088                 }
1089         }
1090
1091         # parameters which are pathnames
1092         our $project = $input_params{'project'};
1093         if (defined $project) {
1094                 if (!is_valid_project($project)) {
1095                         undef $project;
1096                         die_error(404, "No such project");
1097                 }
1098         }
1099
1100         our $project_filter = $input_params{'project_filter'};
1101         if (defined $project_filter) {
1102                 if (!is_valid_pathname($project_filter)) {
1103                         die_error(404, "Invalid project_filter parameter");
1104                 }
1105         }
1106
1107         our $file_name = $input_params{'file_name'};
1108         if (defined $file_name) {
1109                 if (!is_valid_pathname($file_name)) {
1110                         die_error(400, "Invalid file parameter");
1111                 }
1112         }
1113
1114         our $file_parent = $input_params{'file_parent'};
1115         if (defined $file_parent) {
1116                 if (!is_valid_pathname($file_parent)) {
1117                         die_error(400, "Invalid file parent parameter");
1118                 }
1119         }
1120
1121         # parameters which are refnames
1122         our $hash = $input_params{'hash'};
1123         if (defined $hash) {
1124                 if (!is_valid_refname($hash)) {
1125                         die_error(400, "Invalid hash parameter");
1126                 }
1127         }
1128
1129         our $hash_parent = $input_params{'hash_parent'};
1130         if (defined $hash_parent) {
1131                 if (!is_valid_refname($hash_parent)) {
1132                         die_error(400, "Invalid hash parent parameter");
1133                 }
1134         }
1135
1136         our $hash_base = $input_params{'hash_base'};
1137         if (defined $hash_base) {
1138                 if (!is_valid_refname($hash_base)) {
1139                         die_error(400, "Invalid hash base parameter");
1140                 }
1141         }
1142
1143         our @extra_options = @{$input_params{'extra_options'}};
1144         # @extra_options is always defined, since it can only be (currently) set from
1145         # CGI, and $cgi->param() returns the empty array in array context if the param
1146         # is not set
1147         foreach my $opt (@extra_options) {
1148                 if (not exists $allowed_options{$opt}) {
1149                         die_error(400, "Invalid option parameter");
1150                 }
1151                 if (not grep(/^$action$/, @{$allowed_options{$opt}})) {
1152                         die_error(400, "Invalid option parameter for this action");
1153                 }
1154         }
1155
1156         our $hash_parent_base = $input_params{'hash_parent_base'};
1157         if (defined $hash_parent_base) {
1158                 if (!is_valid_refname($hash_parent_base)) {
1159                         die_error(400, "Invalid hash parent base parameter");
1160                 }
1161         }
1162
1163         # other parameters
1164         our $page = $input_params{'page'};
1165         if (defined $page) {
1166                 if ($page =~ m/[^0-9]/) {
1167                         die_error(400, "Invalid page parameter");
1168                 }
1169         }
1170
1171         our $searchtype = $input_params{'searchtype'};
1172         if (defined $searchtype) {
1173                 if ($searchtype =~ m/[^a-z]/) {
1174                         die_error(400, "Invalid searchtype parameter");
1175                 }
1176         }
1177
1178         our $search_use_regexp = $input_params{'search_use_regexp'};
1179
1180         our $searchtext = $input_params{'searchtext'};
1181         our $search_regexp = undef;
1182         if (defined $searchtext) {
1183                 if (length($searchtext) < 2) {
1184                         die_error(403, "At least two characters are required for search parameter");
1185                 }
1186                 if ($search_use_regexp) {
1187                         $search_regexp = $searchtext;
1188                         if (!eval { qr/$search_regexp/; 1; }) {
1189                                 (my $error = $@) =~ s/ at \S+ line \d+.*\n?//;
1190                                 die_error(400, "Invalid search regexp '$search_regexp'",
1191                                           esc_html($error));
1192                         }
1193                 } else {
1194                         $search_regexp = quotemeta $searchtext;
1195                 }
1196         }
1197 }
1198
1199 # path to the current git repository
1200 our $git_dir;
1201 sub evaluate_git_dir {
1202         our $git_dir = "$projectroot/$project" if $project;
1203 }
1204
1205 our (@snapshot_fmts, $git_avatar, @extra_branch_refs);
1206 sub configure_gitweb_features {
1207         # list of supported snapshot formats
1208         our @snapshot_fmts = gitweb_get_feature('snapshot');
1209         @snapshot_fmts = filter_snapshot_fmts(@snapshot_fmts);
1210
1211         our ($git_avatar) = gitweb_get_feature('avatar');
1212         $git_avatar = '' unless $git_avatar =~ /^(?:gravatar|picon)$/s;
1213
1214         our @extra_branch_refs = gitweb_get_feature('extra-branch-refs');
1215         @extra_branch_refs = filter_and_validate_refs (@extra_branch_refs);
1216 }
1217
1218 sub get_branch_refs {
1219         return ('heads', @extra_branch_refs);
1220 }
1221
1222 # custom error handler: 'die <message>' is Internal Server Error
1223 sub handle_errors_html {
1224         my $msg = shift; # it is already HTML escaped
1225
1226         # to avoid infinite loop where error occurs in die_error,
1227         # change handler to default handler, disabling handle_errors_html
1228         set_message("Error occurred when inside die_error:\n$msg");
1229
1230         # you cannot jump out of die_error when called as error handler;
1231         # the subroutine set via CGI::Carp::set_message is called _after_
1232         # HTTP headers are already written, so it cannot write them itself
1233         die_error(undef, undef, $msg, -error_handler => 1, -no_http_header => 1);
1234 }
1235 set_message(\&handle_errors_html);
1236
1237 # dispatch
1238 sub dispatch {
1239         if (!defined $action) {
1240                 if (defined $hash) {
1241                         $action = git_get_type($hash);
1242                         $action or die_error(404, "Object does not exist");
1243                 } elsif (defined $hash_base && defined $file_name) {
1244                         $action = git_get_type("$hash_base:$file_name");
1245                         $action or die_error(404, "File or directory does not exist");
1246                 } elsif (defined $project) {
1247                         $action = 'summary';
1248                 } else {
1249                         $action = 'project_list';
1250                 }
1251         }
1252         if (!defined($actions{$action})) {
1253                 die_error(400, "Unknown action");
1254         }
1255         if ($action !~ m/^(?:opml|project_list|project_index)$/ &&
1256             !$project) {
1257                 die_error(400, "Project needed");
1258         }
1259         $actions{$action}->();
1260 }
1261
1262 sub reset_timer {
1263         our $t0 = [ gettimeofday() ]
1264                 if defined $t0;
1265         our $number_of_git_cmds = 0;
1266 }
1267
1268 our $first_request = 1;
1269 sub run_request {
1270         reset_timer();
1271
1272         evaluate_uri();
1273         if ($first_request) {
1274                 evaluate_gitweb_config();
1275                 evaluate_git_version();
1276         }
1277         if ($per_request_config) {
1278                 if (ref($per_request_config) eq 'CODE') {
1279                         $per_request_config->();
1280                 } elsif (!$first_request) {
1281                         evaluate_gitweb_config();
1282                 }
1283         }
1284         check_loadavg();
1285
1286         # $projectroot and $projects_list might be set in gitweb config file
1287         $projects_list ||= $projectroot;
1288
1289         evaluate_query_params();
1290         evaluate_path_info();
1291         evaluate_and_validate_params();
1292         evaluate_git_dir();
1293
1294         configure_gitweb_features();
1295
1296         dispatch();
1297 }
1298
1299 our $is_last_request = sub { 1 };
1300 our ($pre_dispatch_hook, $post_dispatch_hook, $pre_listen_hook);
1301 our $CGI = 'CGI';
1302 our $cgi;
1303 our $FCGI_Stream_PRINT_raw = \&FCGI::Stream::PRINT;
1304 sub configure_as_fcgi {
1305         require CGI::Fast;
1306         our $CGI = 'CGI::Fast';
1307         # FCGI is not Unicode aware hence the UTF-8 encoding must be done manually.
1308         # However no encoding must be done within git_blob_plain() and git_snapshot()
1309         # which must still output in raw binary mode.
1310         no warnings 'redefine';
1311         my $enc = Encode::find_encoding('UTF-8');
1312         *FCGI::Stream::PRINT = sub {
1313                 my @OUTPUT = @_;
1314                 for (my $i = 1; $i < @_; $i++) {
1315                         $OUTPUT[$i] = $enc->encode($_[$i], Encode::FB_CROAK|Encode::LEAVE_SRC);
1316                 }
1317                 @_ = @OUTPUT;
1318                 goto $FCGI_Stream_PRINT_raw;
1319         };
1320
1321         my $request_number = 0;
1322         # let each child service 100 requests
1323         our $is_last_request = sub { ++$request_number > 100 };
1324 }
1325 sub evaluate_argv {
1326         my $script_name = $ENV{'SCRIPT_NAME'} || $ENV{'SCRIPT_FILENAME'} || __FILE__;
1327         configure_as_fcgi()
1328                 if $script_name =~ /\.fcgi$/;
1329
1330         return unless (@ARGV);
1331
1332         require Getopt::Long;
1333         Getopt::Long::GetOptions(
1334                 'fastcgi|fcgi|f' => \&configure_as_fcgi,
1335                 'nproc|n=i' => sub {
1336                         my ($arg, $val) = @_;
1337                         return unless eval { require FCGI::ProcManager; 1; };
1338                         my $proc_manager = FCGI::ProcManager->new({
1339                                 n_processes => $val,
1340                         });
1341                         our $pre_listen_hook    = sub { $proc_manager->pm_manage()        };
1342                         our $pre_dispatch_hook  = sub { $proc_manager->pm_pre_dispatch()  };
1343                         our $post_dispatch_hook = sub { $proc_manager->pm_post_dispatch() };
1344                 },
1345         );
1346 }
1347
1348 sub run {
1349         evaluate_argv();
1350
1351         $first_request = 1;
1352         $pre_listen_hook->()
1353                 if $pre_listen_hook;
1354
1355  REQUEST:
1356         while ($cgi = $CGI->new()) {
1357                 $pre_dispatch_hook->()
1358                         if $pre_dispatch_hook;
1359
1360                 run_request();
1361
1362                 $post_dispatch_hook->()
1363                         if $post_dispatch_hook;
1364                 $first_request = 0;
1365
1366                 last REQUEST if ($is_last_request->());
1367         }
1368
1369  DONE_GITWEB:
1370         1;
1371 }
1372
1373 run();
1374
1375 if (defined caller) {
1376         # wrapped in a subroutine processing requests,
1377         # e.g. mod_perl with ModPerl::Registry, or PSGI with Plack::App::WrapCGI
1378         return;
1379 } else {
1380         # pure CGI script, serving single request
1381         exit;
1382 }
1383
1384 ## ======================================================================
1385 ## action links
1386
1387 # possible values of extra options
1388 # -full => 0|1      - use absolute/full URL ($my_uri/$my_url as base)
1389 # -replay => 1      - start from a current view (replay with modifications)
1390 # -path_info => 0|1 - don't use/use path_info URL (if possible)
1391 # -anchor => ANCHOR - add #ANCHOR to end of URL, implies -replay if used alone
1392 sub href {
1393         my %params = @_;
1394         # default is to use -absolute url() i.e. $my_uri
1395         my $href = $params{-full} ? $my_url : $my_uri;
1396
1397         # implicit -replay, must be first of implicit params
1398         $params{-replay} = 1 if (keys %params == 1 && $params{-anchor});
1399
1400         $params{'project'} = $project unless exists $params{'project'};
1401
1402         if ($params{-replay}) {
1403                 while (my ($name, $symbol) = each %cgi_param_mapping) {
1404                         if (!exists $params{$name}) {
1405                                 $params{$name} = $input_params{$name};
1406                         }
1407                 }
1408         }
1409
1410         my $use_pathinfo = gitweb_check_feature('pathinfo');
1411         if (defined $params{'project'} &&
1412             (exists $params{-path_info} ? $params{-path_info} : $use_pathinfo)) {
1413                 # try to put as many parameters as possible in PATH_INFO:
1414                 #   - project name
1415                 #   - action
1416                 #   - hash_parent or hash_parent_base:/file_parent
1417                 #   - hash or hash_base:/filename
1418                 #   - the snapshot_format as an appropriate suffix
1419
1420                 # When the script is the root DirectoryIndex for the domain,
1421                 # $href here would be something like http://gitweb.example.com/
1422                 # Thus, we strip any trailing / from $href, to spare us double
1423                 # slashes in the final URL
1424                 $href =~ s,/$,,;
1425
1426                 # Then add the project name, if present
1427                 $href .= "/".esc_path_info($params{'project'});
1428                 delete $params{'project'};
1429
1430                 # since we destructively absorb parameters, we keep this
1431                 # boolean that remembers if we're handling a snapshot
1432                 my $is_snapshot = $params{'action'} eq 'snapshot';
1433
1434                 # Summary just uses the project path URL, any other action is
1435                 # added to the URL
1436                 if (defined $params{'action'}) {
1437                         $href .= "/".esc_path_info($params{'action'})
1438                                 unless $params{'action'} eq 'summary';
1439                         delete $params{'action'};
1440                 }
1441
1442                 # Next, we put hash_parent_base:/file_parent..hash_base:/file_name,
1443                 # stripping nonexistent or useless pieces
1444                 $href .= "/" if ($params{'hash_base'} || $params{'hash_parent_base'}
1445                         || $params{'hash_parent'} || $params{'hash'});
1446                 if (defined $params{'hash_base'}) {
1447                         if (defined $params{'hash_parent_base'}) {
1448                                 $href .= esc_path_info($params{'hash_parent_base'});
1449                                 # skip the file_parent if it's the same as the file_name
1450                                 if (defined $params{'file_parent'}) {
1451                                         if (defined $params{'file_name'} && $params{'file_parent'} eq $params{'file_name'}) {
1452                                                 delete $params{'file_parent'};
1453                                         } elsif ($params{'file_parent'} !~ /\.\./) {
1454                                                 $href .= ":/".esc_path_info($params{'file_parent'});
1455                                                 delete $params{'file_parent'};
1456                                         }
1457                                 }
1458                                 $href .= "..";
1459                                 delete $params{'hash_parent'};
1460                                 delete $params{'hash_parent_base'};
1461                         } elsif (defined $params{'hash_parent'}) {
1462                                 $href .= esc_path_info($params{'hash_parent'}). "..";
1463                                 delete $params{'hash_parent'};
1464                         }
1465
1466                         $href .= esc_path_info($params{'hash_base'});
1467                         if (defined $params{'file_name'} && $params{'file_name'} !~ /\.\./) {
1468                                 $href .= ":/".esc_path_info($params{'file_name'});
1469                                 delete $params{'file_name'};
1470                         }
1471                         delete $params{'hash'};
1472                         delete $params{'hash_base'};
1473                 } elsif (defined $params{'hash'}) {
1474                         $href .= esc_path_info($params{'hash'});
1475                         delete $params{'hash'};
1476                 }
1477
1478                 # If the action was a snapshot, we can absorb the
1479                 # snapshot_format parameter too
1480                 if ($is_snapshot) {
1481                         my $fmt = $params{'snapshot_format'};
1482                         # snapshot_format should always be defined when href()
1483                         # is called, but just in case some code forgets, we
1484                         # fall back to the default
1485                         $fmt ||= $snapshot_fmts[0];
1486                         $href .= $known_snapshot_formats{$fmt}{'suffix'};
1487                         delete $params{'snapshot_format'};
1488                 }
1489         }
1490
1491         # now encode the parameters explicitly
1492         my @result = ();
1493         for (my $i = 0; $i < @cgi_param_mapping; $i += 2) {
1494                 my ($name, $symbol) = ($cgi_param_mapping[$i], $cgi_param_mapping[$i+1]);
1495                 if (defined $params{$name}) {
1496                         if (ref($params{$name}) eq "ARRAY") {
1497                                 foreach my $par (@{$params{$name}}) {
1498                                         push @result, $symbol . "=" . esc_param($par);
1499                                 }
1500                         } else {
1501                                 push @result, $symbol . "=" . esc_param($params{$name});
1502                         }
1503                 }
1504         }
1505         $href .= "?" . join(';', @result) if scalar @result;
1506
1507         # final transformation: trailing spaces must be escaped (URI-encoded)
1508         $href =~ s/(\s+)$/CGI::escape($1)/e;
1509
1510         if ($params{-anchor}) {
1511                 $href .= "#".esc_param($params{-anchor});
1512         }
1513
1514         return $href;
1515 }
1516
1517
1518 ## ======================================================================
1519 ## validation, quoting/unquoting and escaping
1520
1521 sub is_valid_action {
1522         my $input = shift;
1523         return undef unless exists $actions{$input};
1524         return 1;
1525 }
1526
1527 sub is_valid_project {
1528         my $input = shift;
1529
1530         return unless defined $input;
1531         if (!is_valid_pathname($input) ||
1532                 !(-d "$projectroot/$input") ||
1533                 !check_export_ok("$projectroot/$input") ||
1534                 ($strict_export && !project_in_list($input))) {
1535                 return undef;
1536         } else {
1537                 return 1;
1538         }
1539 }
1540
1541 sub is_valid_pathname {
1542         my $input = shift;
1543
1544         return undef unless defined $input;
1545         # no '.' or '..' as elements of path, i.e. no '.' or '..'
1546         # at the beginning, at the end, and between slashes.
1547         # also this catches doubled slashes
1548         if ($input =~ m!(^|/)(|\.|\.\.)(/|$)!) {
1549                 return undef;
1550         }
1551         # no null characters
1552         if ($input =~ m!\0!) {
1553                 return undef;
1554         }
1555         return 1;
1556 }
1557
1558 sub is_valid_ref_format {
1559         my $input = shift;
1560
1561         return undef unless defined $input;
1562         # restrictions on ref name according to git-check-ref-format
1563         if ($input =~ m!(/\.|\.\.|[\000-\040\177 ~^:?*\[]|/$)!) {
1564                 return undef;
1565         }
1566         return 1;
1567 }
1568
1569 sub is_valid_refname {
1570         my $input = shift;
1571
1572         return undef unless defined $input;
1573         # textual hashes are O.K.
1574         if ($input =~ m/^$oid_regex$/) {
1575                 return 1;
1576         }
1577         # it must be correct pathname
1578         is_valid_pathname($input) or return undef;
1579         # check git-check-ref-format restrictions
1580         is_valid_ref_format($input) or return undef;
1581         return 1;
1582 }
1583
1584 # decode sequences of octets in utf8 into Perl's internal form,
1585 # which is utf-8 with utf8 flag set if needed.  gitweb writes out
1586 # in utf-8 thanks to "binmode STDOUT, ':utf8'" at beginning
1587 sub to_utf8 {
1588         my $str = shift;
1589         return undef unless defined $str;
1590
1591         if (utf8::is_utf8($str) || utf8::decode($str)) {
1592                 return $str;
1593         } else {
1594                 return decode($fallback_encoding, $str, Encode::FB_DEFAULT);
1595         }
1596 }
1597
1598 # quote unsafe chars, but keep the slash, even when it's not
1599 # correct, but quoted slashes look too horrible in bookmarks
1600 sub esc_param {
1601         my $str = shift;
1602         return undef unless defined $str;
1603         $str =~ s/([^A-Za-z0-9\-_.~()\/:@ ]+)/CGI::escape($1)/eg;
1604         $str =~ s/ /\+/g;
1605         return $str;
1606 }
1607
1608 # the quoting rules for path_info fragment are slightly different
1609 sub esc_path_info {
1610         my $str = shift;
1611         return undef unless defined $str;
1612
1613         # path_info doesn't treat '+' as space (specially), but '?' must be escaped
1614         $str =~ s/([^A-Za-z0-9\-_.~();\/;:@&= +]+)/CGI::escape($1)/eg;
1615
1616         return $str;
1617 }
1618
1619 # quote unsafe chars in whole URL, so some characters cannot be quoted
1620 sub esc_url {
1621         my $str = shift;
1622         return undef unless defined $str;
1623         $str =~ s/([^A-Za-z0-9\-_.~();\/;?:@&= ]+)/CGI::escape($1)/eg;
1624         $str =~ s/ /\+/g;
1625         return $str;
1626 }
1627
1628 # quote unsafe characters in HTML attributes
1629 sub esc_attr {
1630
1631         # for XHTML conformance escaping '"' to '&quot;' is not enough
1632         return esc_html(@_);
1633 }
1634
1635 # replace invalid utf8 character with SUBSTITUTION sequence
1636 sub esc_html {
1637         my $str = shift;
1638         my %opts = @_;
1639
1640         return undef unless defined $str;
1641
1642         $str = to_utf8($str);
1643         $str = $cgi->escapeHTML($str);
1644         if ($opts{'-nbsp'}) {
1645                 $str =~ s/ /&nbsp;/g;
1646         }
1647         $str =~ s|([[:cntrl:]])|(($1 ne "\t") ? quot_cec($1) : $1)|eg;
1648         return $str;
1649 }
1650
1651 # quote control characters and escape filename to HTML
1652 sub esc_path {
1653         my $str = shift;
1654         my %opts = @_;
1655
1656         return undef unless defined $str;
1657
1658         $str = to_utf8($str);
1659         $str = $cgi->escapeHTML($str);
1660         if ($opts{'-nbsp'}) {
1661                 $str =~ s/ /&nbsp;/g;
1662         }
1663         $str =~ s|([[:cntrl:]])|quot_cec($1)|eg;
1664         return $str;
1665 }
1666
1667 # Sanitize for use in XHTML + application/xml+xhtml (valid XML 1.0)
1668 sub sanitize {
1669         my $str = shift;
1670
1671         return undef unless defined $str;
1672
1673         $str = to_utf8($str);
1674         $str =~ s|([[:cntrl:]])|(index("\t\n\r", $1) != -1 ? $1 : quot_cec($1))|eg;
1675         return $str;
1676 }
1677
1678 # Make control characters "printable", using character escape codes (CEC)
1679 sub quot_cec {
1680         my $cntrl = shift;
1681         my %opts = @_;
1682         my %es = ( # character escape codes, aka escape sequences
1683                 "\t" => '\t',   # tab             (HT)
1684                 "\n" => '\n',   # line feed       (LF)
1685                 "\r" => '\r',   # carriage return (CR)
1686                 "\f" => '\f',   # form feed       (FF)
1687                 "\b" => '\b',   # backspace       (BS)
1688                 "\a" => '\a',   # alarm (bell)    (BEL)
1689                 "\e" => '\e',   # escape          (ESC)
1690                 "\013" => '\v', # vertical tab    (VT)
1691                 "\000" => '\0', # nul character   (NUL)
1692         );
1693         my $chr = ( (exists $es{$cntrl})
1694                     ? $es{$cntrl}
1695                     : sprintf('\%2x', ord($cntrl)) );
1696         if ($opts{-nohtml}) {
1697                 return $chr;
1698         } else {
1699                 return "<span class=\"cntrl\">$chr</span>";
1700         }
1701 }
1702
1703 # Alternatively use unicode control pictures codepoints,
1704 # Unicode "printable representation" (PR)
1705 sub quot_upr {
1706         my $cntrl = shift;
1707         my %opts = @_;
1708
1709         my $chr = sprintf('&#%04d;', 0x2400+ord($cntrl));
1710         if ($opts{-nohtml}) {
1711                 return $chr;
1712         } else {
1713                 return "<span class=\"cntrl\">$chr</span>";
1714         }
1715 }
1716
1717 # git may return quoted and escaped filenames
1718 sub unquote {
1719         my $str = shift;
1720
1721         sub unq {
1722                 my $seq = shift;
1723                 my %es = ( # character escape codes, aka escape sequences
1724                         't' => "\t",   # tab            (HT, TAB)
1725                         'n' => "\n",   # newline        (NL)
1726                         'r' => "\r",   # return         (CR)
1727                         'f' => "\f",   # form feed      (FF)
1728                         'b' => "\b",   # backspace      (BS)
1729                         'a' => "\a",   # alarm (bell)   (BEL)
1730                         'e' => "\e",   # escape         (ESC)
1731                         'v' => "\013", # vertical tab   (VT)
1732                 );
1733
1734                 if ($seq =~ m/^[0-7]{1,3}$/) {
1735                         # octal char sequence
1736                         return chr(oct($seq));
1737                 } elsif (exists $es{$seq}) {
1738                         # C escape sequence, aka character escape code
1739                         return $es{$seq};
1740                 }
1741                 # quoted ordinary character
1742                 return $seq;
1743         }
1744
1745         if ($str =~ m/^"(.*)"$/) {
1746                 # needs unquoting
1747                 $str = $1;
1748                 $str =~ s/\\([^0-7]|[0-7]{1,3})/unq($1)/eg;
1749         }
1750         return $str;
1751 }
1752
1753 # escape tabs (convert tabs to spaces)
1754 sub untabify {
1755         my $line = shift;
1756
1757         while ((my $pos = index($line, "\t")) != -1) {
1758                 if (my $count = (8 - ($pos % 8))) {
1759                         my $spaces = ' ' x $count;
1760                         $line =~ s/\t/$spaces/;
1761                 }
1762         }
1763
1764         return $line;
1765 }
1766
1767 sub project_in_list {
1768         my $project = shift;
1769         my @list = git_get_projects_list();
1770         return @list && scalar(grep { $_->{'path'} eq $project } @list);
1771 }
1772
1773 ## ----------------------------------------------------------------------
1774 ## HTML aware string manipulation
1775
1776 # Try to chop given string on a word boundary between position
1777 # $len and $len+$add_len. If there is no word boundary there,
1778 # chop at $len+$add_len. Do not chop if chopped part plus ellipsis
1779 # (marking chopped part) would be longer than given string.
1780 sub chop_str {
1781         my $str = shift;
1782         my $len = shift;
1783         my $add_len = shift || 10;
1784         my $where = shift || 'right'; # 'left' | 'center' | 'right'
1785
1786         # Make sure perl knows it is utf8 encoded so we don't
1787         # cut in the middle of a utf8 multibyte char.
1788         $str = to_utf8($str);
1789
1790         # allow only $len chars, but don't cut a word if it would fit in $add_len
1791         # if it doesn't fit, cut it if it's still longer than the dots we would add
1792         # remove chopped character entities entirely
1793
1794         # when chopping in the middle, distribute $len into left and right part
1795         # return early if chopping wouldn't make string shorter
1796         if ($where eq 'center') {
1797                 return $str if ($len + 5 >= length($str)); # filler is length 5
1798                 $len = int($len/2);
1799         } else {
1800                 return $str if ($len + 4 >= length($str)); # filler is length 4
1801         }
1802
1803         # regexps: ending and beginning with word part up to $add_len
1804         my $endre = qr/.{$len}\w{0,$add_len}/;
1805         my $begre = qr/\w{0,$add_len}.{$len}/;
1806
1807         if ($where eq 'left') {
1808                 $str =~ m/^(.*?)($begre)$/;
1809                 my ($lead, $body) = ($1, $2);
1810                 if (length($lead) > 4) {
1811                         $lead = " ...";
1812                 }
1813                 return "$lead$body";
1814
1815         } elsif ($where eq 'center') {
1816                 $str =~ m/^($endre)(.*)$/;
1817                 my ($left, $str)  = ($1, $2);
1818                 $str =~ m/^(.*?)($begre)$/;
1819                 my ($mid, $right) = ($1, $2);
1820                 if (length($mid) > 5) {
1821                         $mid = " ... ";
1822                 }
1823                 return "$left$mid$right";
1824
1825         } else {
1826                 $str =~ m/^($endre)(.*)$/;
1827                 my $body = $1;
1828                 my $tail = $2;
1829                 if (length($tail) > 4) {
1830                         $tail = "... ";
1831                 }
1832                 return "$body$tail";
1833         }
1834 }
1835
1836 # takes the same arguments as chop_str, but also wraps a <span> around the
1837 # result with a title attribute if it does get chopped. Additionally, the
1838 # string is HTML-escaped.
1839 sub chop_and_escape_str {
1840         my ($str) = @_;
1841
1842         my $chopped = chop_str(@_);
1843         $str = to_utf8($str);
1844         if ($chopped eq $str) {
1845                 return esc_html($chopped);
1846         } else {
1847                 $str =~ s/[[:cntrl:]]/?/g;
1848                 return $cgi->span({-title=>$str}, esc_html($chopped));
1849         }
1850 }
1851
1852 # Highlight selected fragments of string, using given CSS class,
1853 # and escape HTML.  It is assumed that fragments do not overlap.
1854 # Regions are passed as list of pairs (array references).
1855 #
1856 # Example: esc_html_hl_regions("foobar", "mark", [ 0, 3 ]) returns
1857 # '<span class="mark">foo</span>bar'
1858 sub esc_html_hl_regions {
1859         my ($str, $css_class, @sel) = @_;
1860         my %opts = grep { ref($_) ne 'ARRAY' } @sel;
1861         @sel     = grep { ref($_) eq 'ARRAY' } @sel;
1862         return esc_html($str, %opts) unless @sel;
1863
1864         my $out = '';
1865         my $pos = 0;
1866
1867         for my $s (@sel) {
1868                 my ($begin, $end) = @$s;
1869
1870                 # Don't create empty <span> elements.
1871                 next if $end <= $begin;
1872
1873                 my $escaped = esc_html(substr($str, $begin, $end - $begin),
1874                                        %opts);
1875
1876                 $out .= esc_html(substr($str, $pos, $begin - $pos), %opts)
1877                         if ($begin - $pos > 0);
1878                 $out .= $cgi->span({-class => $css_class}, $escaped);
1879
1880                 $pos = $end;
1881         }
1882         $out .= esc_html(substr($str, $pos), %opts)
1883                 if ($pos < length($str));
1884
1885         return $out;
1886 }
1887
1888 # return positions of beginning and end of each match
1889 sub matchpos_list {
1890         my ($str, $regexp) = @_;
1891         return unless (defined $str && defined $regexp);
1892
1893         my @matches;
1894         while ($str =~ /$regexp/g) {
1895                 push @matches, [$-[0], $+[0]];
1896         }
1897         return @matches;
1898 }
1899
1900 # highlight match (if any), and escape HTML
1901 sub esc_html_match_hl {
1902         my ($str, $regexp) = @_;
1903         return esc_html($str) unless defined $regexp;
1904
1905         my @matches = matchpos_list($str, $regexp);
1906         return esc_html($str) unless @matches;
1907
1908         return esc_html_hl_regions($str, 'match', @matches);
1909 }
1910
1911
1912 # highlight match (if any) of shortened string, and escape HTML
1913 sub esc_html_match_hl_chopped {
1914         my ($str, $chopped, $regexp) = @_;
1915         return esc_html_match_hl($str, $regexp) unless defined $chopped;
1916
1917         my @matches = matchpos_list($str, $regexp);
1918         return esc_html($chopped) unless @matches;
1919
1920         # filter matches so that we mark chopped string
1921         my $tail = "... "; # see chop_str
1922         unless ($chopped =~ s/\Q$tail\E$//) {
1923                 $tail = '';
1924         }
1925         my $chop_len = length($chopped);
1926         my $tail_len = length($tail);
1927         my @filtered;
1928
1929         for my $m (@matches) {
1930                 if ($m->[0] > $chop_len) {
1931                         push @filtered, [ $chop_len, $chop_len + $tail_len ] if ($tail_len > 0);
1932                         last;
1933                 } elsif ($m->[1] > $chop_len) {
1934                         push @filtered, [ $m->[0], $chop_len + $tail_len ];
1935                         last;
1936                 }
1937                 push @filtered, $m;
1938         }
1939
1940         return esc_html_hl_regions($chopped . $tail, 'match', @filtered);
1941 }
1942
1943 ## ----------------------------------------------------------------------
1944 ## functions returning short strings
1945
1946 # CSS class for given age value (in seconds)
1947 sub age_class {
1948         my $age = shift;
1949
1950         if (!defined $age) {
1951                 return "noage";
1952         } elsif ($age < 60*60*2) {
1953                 return "age0";
1954         } elsif ($age < 60*60*24*2) {
1955                 return "age1";
1956         } else {
1957                 return "age2";
1958         }
1959 }
1960
1961 # convert age in seconds to "nn units ago" string
1962 sub age_string {
1963         my $age = shift;
1964         my $age_str;
1965
1966         if ($age > 60*60*24*365*2) {
1967                 $age_str = (int $age/60/60/24/365);
1968                 $age_str .= " years ago";
1969         } elsif ($age > 60*60*24*(365/12)*2) {
1970                 $age_str = int $age/60/60/24/(365/12);
1971                 $age_str .= " months ago";
1972         } elsif ($age > 60*60*24*7*2) {
1973                 $age_str = int $age/60/60/24/7;
1974                 $age_str .= " weeks ago";
1975         } elsif ($age > 60*60*24*2) {
1976                 $age_str = int $age/60/60/24;
1977                 $age_str .= " days ago";
1978         } elsif ($age > 60*60*2) {
1979                 $age_str = int $age/60/60;
1980                 $age_str .= " hours ago";
1981         } elsif ($age > 60*2) {
1982                 $age_str = int $age/60;
1983                 $age_str .= " min ago";
1984         } elsif ($age > 2) {
1985                 $age_str = int $age;
1986                 $age_str .= " sec ago";
1987         } else {
1988                 $age_str .= " right now";
1989         }
1990         return $age_str;
1991 }
1992
1993 use constant {
1994         S_IFINVALID => 0030000,
1995         S_IFGITLINK => 0160000,
1996 };
1997
1998 # submodule/subproject, a commit object reference
1999 sub S_ISGITLINK {
2000         my $mode = shift;
2001
2002         return (($mode & S_IFMT) == S_IFGITLINK)
2003 }
2004
2005 # convert file mode in octal to symbolic file mode string
2006 sub mode_str {
2007         my $mode = oct shift;
2008
2009         if (S_ISGITLINK($mode)) {
2010                 return 'm---------';
2011         } elsif (S_ISDIR($mode & S_IFMT)) {
2012                 return 'drwxr-xr-x';
2013         } elsif (S_ISLNK($mode)) {
2014                 return 'lrwxrwxrwx';
2015         } elsif (S_ISREG($mode)) {
2016                 # git cares only about the executable bit
2017                 if ($mode & S_IXUSR) {
2018                         return '-rwxr-xr-x';
2019                 } else {
2020                         return '-rw-r--r--';
2021                 };
2022         } else {
2023                 return '----------';
2024         }
2025 }
2026
2027 # convert file mode in octal to file type string
2028 sub file_type {
2029         my $mode = shift;
2030
2031         if ($mode !~ m/^[0-7]+$/) {
2032                 return $mode;
2033         } else {
2034                 $mode = oct $mode;
2035         }
2036
2037         if (S_ISGITLINK($mode)) {
2038                 return "submodule";
2039         } elsif (S_ISDIR($mode & S_IFMT)) {
2040                 return "directory";
2041         } elsif (S_ISLNK($mode)) {
2042                 return "symlink";
2043         } elsif (S_ISREG($mode)) {
2044                 return "file";
2045         } else {
2046                 return "unknown";
2047         }
2048 }
2049
2050 # convert file mode in octal to file type description string
2051 sub file_type_long {
2052         my $mode = shift;
2053
2054         if ($mode !~ m/^[0-7]+$/) {
2055                 return $mode;
2056         } else {
2057                 $mode = oct $mode;
2058         }
2059
2060         if (S_ISGITLINK($mode)) {
2061                 return "submodule";
2062         } elsif (S_ISDIR($mode & S_IFMT)) {
2063                 return "directory";
2064         } elsif (S_ISLNK($mode)) {
2065                 return "symlink";
2066         } elsif (S_ISREG($mode)) {
2067                 if ($mode & S_IXUSR) {
2068                         return "executable";
2069                 } else {
2070                         return "file";
2071                 };
2072         } else {
2073                 return "unknown";
2074         }
2075 }
2076
2077
2078 ## ----------------------------------------------------------------------
2079 ## functions returning short HTML fragments, or transforming HTML fragments
2080 ## which don't belong to other sections
2081
2082 # format line of commit message.
2083 sub format_log_line_html {
2084         my $line = shift;
2085
2086         # Potentially abbreviated OID.
2087         my $regex = oid_nlen_regex("7,64");
2088
2089         $line = esc_html($line, -nbsp=>1);
2090         $line =~ s{
2091         \b
2092         (
2093             # The output of "git describe", e.g. v2.10.0-297-gf6727b0
2094             # or hadoop-20160921-113441-20-g094fb7d
2095             (?<!-) # see strbuf_check_tag_ref(). Tags can't start with -
2096             [A-Za-z0-9.-]+
2097             (?!\.) # refs can't end with ".", see check_refname_format()
2098             -g$regex
2099             |
2100             # Just a normal looking Git SHA1
2101             $regex
2102         )
2103         \b
2104     }{
2105                 $cgi->a({-href => href(action=>"object", hash=>$1),
2106                                         -class => "text"}, $1);
2107         }egx;
2108
2109         return $line;
2110 }
2111
2112 # format marker of refs pointing to given object
2113
2114 # the destination action is chosen based on object type and current context:
2115 # - for annotated tags, we choose the tag view unless it's the current view
2116 #   already, in which case we go to shortlog view
2117 # - for other refs, we keep the current view if we're in history, shortlog or
2118 #   log view, and select shortlog otherwise
2119 sub format_ref_marker {
2120         my ($refs, $id) = @_;
2121         my $markers = '';
2122
2123         if (defined $refs->{$id}) {
2124                 foreach my $ref (@{$refs->{$id}}) {
2125                         # this code exploits the fact that non-lightweight tags are the
2126                         # only indirect objects, and that they are the only objects for which
2127                         # we want to use tag instead of shortlog as action
2128                         my ($type, $name) = qw();
2129                         my $indirect = ($ref =~ s/\^\{\}$//);
2130                         # e.g. tags/v2.6.11 or heads/next
2131                         if ($ref =~ m!^(.*?)s?/(.*)$!) {
2132                                 $type = $1;
2133                                 $name = $2;
2134                         } else {
2135                                 $type = "ref";
2136                                 $name = $ref;
2137                         }
2138
2139                         my $class = $type;
2140                         $class .= " indirect" if $indirect;
2141
2142                         my $dest_action = "shortlog";
2143
2144                         if ($indirect) {
2145                                 $dest_action = "tag" unless $action eq "tag";
2146                         } elsif ($action =~ /^(history|(short)?log)$/) {
2147                                 $dest_action = $action;
2148                         }
2149
2150                         my $dest = "";
2151                         $dest .= "refs/" unless $ref =~ m!^refs/!;
2152                         $dest .= $ref;
2153
2154                         my $link = $cgi->a({
2155                                 -href => href(
2156                                         action=>$dest_action,
2157                                         hash=>$dest
2158                                 )}, esc_html($name));
2159
2160                         $markers .= " <span class=\"".esc_attr($class)."\" title=\"".esc_attr($ref)."\">" .
2161                                 $link . "</span>";
2162                 }
2163         }
2164
2165         if ($markers) {
2166                 return ' <span class="refs">'. $markers . '</span>';
2167         } else {
2168                 return "";
2169         }
2170 }
2171
2172 # format, perhaps shortened and with markers, title line
2173 sub format_subject_html {
2174         my ($long, $short, $href, $extra) = @_;
2175         $extra = '' unless defined($extra);
2176
2177         if (length($short) < length($long)) {
2178                 $long =~ s/[[:cntrl:]]/?/g;
2179                 return $cgi->a({-href => $href, -class => "list subject",
2180                                 -title => to_utf8($long)},
2181                        esc_html($short)) . $extra;
2182         } else {
2183                 return $cgi->a({-href => $href, -class => "list subject"},
2184                        esc_html($long)) . $extra;
2185         }
2186 }
2187
2188 # Rather than recomputing the url for an email multiple times, we cache it
2189 # after the first hit. This gives a visible benefit in views where the avatar
2190 # for the same email is used repeatedly (e.g. shortlog).
2191 # The cache is shared by all avatar engines (currently gravatar only), which
2192 # are free to use it as preferred. Since only one avatar engine is used for any
2193 # given page, there's no risk for cache conflicts.
2194 our %avatar_cache = ();
2195
2196 # Compute the picon url for a given email, by using the picon search service over at
2197 # http://www.cs.indiana.edu/picons/search.html
2198 sub picon_url {
2199         my $email = lc shift;
2200         if (!$avatar_cache{$email}) {
2201                 my ($user, $domain) = split('@', $email);
2202                 $avatar_cache{$email} =
2203                         "//www.cs.indiana.edu/cgi-pub/kinzler/piconsearch.cgi/" .
2204                         "$domain/$user/" .
2205                         "users+domains+unknown/up/single";
2206         }
2207         return $avatar_cache{$email};
2208 }
2209
2210 # Compute the gravatar url for a given email, if it's not in the cache already.
2211 # Gravatar stores only the part of the URL before the size, since that's the
2212 # one computationally more expensive. This also allows reuse of the cache for
2213 # different sizes (for this particular engine).
2214 sub gravatar_url {
2215         my $email = lc shift;
2216         my $size = shift;
2217         $avatar_cache{$email} ||=
2218                 "//www.gravatar.com/avatar/" .
2219                         md5_hex($email) . "?s=";
2220         return $avatar_cache{$email} . $size;
2221 }
2222
2223 # Insert an avatar for the given $email at the given $size if the feature
2224 # is enabled.
2225 sub git_get_avatar {
2226         my ($email, %opts) = @_;
2227         my $pre_white  = ($opts{-pad_before} ? "&nbsp;" : "");
2228         my $post_white = ($opts{-pad_after}  ? "&nbsp;" : "");
2229         $opts{-size} ||= 'default';
2230         my $size = $avatar_size{$opts{-size}} || $avatar_size{'default'};
2231         my $url = "";
2232         if ($git_avatar eq 'gravatar') {
2233                 $url = gravatar_url($email, $size);
2234         } elsif ($git_avatar eq 'picon') {
2235                 $url = picon_url($email);
2236         }
2237         # Other providers can be added by extending the if chain, defining $url
2238         # as needed. If no variant puts something in $url, we assume avatars
2239         # are completely disabled/unavailable.
2240         if ($url) {
2241                 return $pre_white .
2242                        "<img width=\"$size\" " .
2243                             "class=\"avatar\" " .
2244                             "src=\"".esc_url($url)."\" " .
2245                             "alt=\"\" " .
2246                        "/>" . $post_white;
2247         } else {
2248                 return "";
2249         }
2250 }
2251
2252 sub format_search_author {
2253         my ($author, $searchtype, $displaytext) = @_;
2254         my $have_search = gitweb_check_feature('search');
2255
2256         if ($have_search) {
2257                 my $performed = "";
2258                 if ($searchtype eq 'author') {
2259                         $performed = "authored";
2260                 } elsif ($searchtype eq 'committer') {
2261                         $performed = "committed";
2262                 }
2263
2264                 return $cgi->a({-href => href(action=>"search", hash=>$hash,
2265                                 searchtext=>$author,
2266                                 searchtype=>$searchtype), class=>"list",
2267                                 title=>"Search for commits $performed by $author"},
2268                                 $displaytext);
2269
2270         } else {
2271                 return $displaytext;
2272         }
2273 }
2274
2275 # format the author name of the given commit with the given tag
2276 # the author name is chopped and escaped according to the other
2277 # optional parameters (see chop_str).
2278 sub format_author_html {
2279         my $tag = shift;
2280         my $co = shift;
2281         my $author = chop_and_escape_str($co->{'author_name'}, @_);
2282         return "<$tag class=\"author\">" .
2283                format_search_author($co->{'author_name'}, "author",
2284                        git_get_avatar($co->{'author_email'}, -pad_after => 1) .
2285                        $author) .
2286                "</$tag>";
2287 }
2288
2289 # format git diff header line, i.e. "diff --(git|combined|cc) ..."
2290 sub format_git_diff_header_line {
2291         my $line = shift;
2292         my $diffinfo = shift;
2293         my ($from, $to) = @_;
2294
2295         if ($diffinfo->{'nparents'}) {
2296                 # combined diff
2297                 $line =~ s!^(diff (.*?) )"?.*$!$1!;
2298                 if ($to->{'href'}) {
2299                         $line .= $cgi->a({-href => $to->{'href'}, -class => "path"},
2300                                          esc_path($to->{'file'}));
2301                 } else { # file was deleted (no href)
2302                         $line .= esc_path($to->{'file'});
2303                 }
2304         } else {
2305                 # "ordinary" diff
2306                 $line =~ s!^(diff (.*?) )"?a/.*$!$1!;
2307                 if ($from->{'href'}) {
2308                         $line .= $cgi->a({-href => $from->{'href'}, -class => "path"},
2309                                          'a/' . esc_path($from->{'file'}));
2310                 } else { # file was added (no href)
2311                         $line .= 'a/' . esc_path($from->{'file'});
2312                 }
2313                 $line .= ' ';
2314                 if ($to->{'href'}) {
2315                         $line .= $cgi->a({-href => $to->{'href'}, -class => "path"},
2316                                          'b/' . esc_path($to->{'file'}));
2317                 } else { # file was deleted
2318                         $line .= 'b/' . esc_path($to->{'file'});
2319                 }
2320         }
2321
2322         return "<div class=\"diff header\">$line</div>\n";
2323 }
2324
2325 # format extended diff header line, before patch itself
2326 sub format_extended_diff_header_line {
2327         my $line = shift;
2328         my $diffinfo = shift;
2329         my ($from, $to) = @_;
2330
2331         # match <path>
2332         if ($line =~ s!^((copy|rename) from ).*$!$1! && $from->{'href'}) {
2333                 $line .= $cgi->a({-href=>$from->{'href'}, -class=>"path"},
2334                                        esc_path($from->{'file'}));
2335         }
2336         if ($line =~ s!^((copy|rename) to ).*$!$1! && $to->{'href'}) {
2337                 $line .= $cgi->a({-href=>$to->{'href'}, -class=>"path"},
2338                                  esc_path($to->{'file'}));
2339         }
2340         # match single <mode>
2341         if ($line =~ m/\s(\d{6})$/) {
2342                 $line .= '<span class="info"> (' .
2343                          file_type_long($1) .
2344                          ')</span>';
2345         }
2346         # match <hash>
2347         if ($line =~ oid_nlen_prefix_infix_regex($sha1_len, "index ", ",") |
2348             $line =~ oid_nlen_prefix_infix_regex($sha256_len, "index ", ",")) {
2349                 # can match only for combined diff
2350                 $line = 'index ';
2351                 for (my $i = 0; $i < $diffinfo->{'nparents'}; $i++) {
2352                         if ($from->{'href'}[$i]) {
2353                                 $line .= $cgi->a({-href=>$from->{'href'}[$i],
2354                                                   -class=>"hash"},
2355                                                  substr($diffinfo->{'from_id'}[$i],0,7));
2356                         } else {
2357                                 $line .= '0' x 7;
2358                         }
2359                         # separator
2360                         $line .= ',' if ($i < $diffinfo->{'nparents'} - 1);
2361                 }
2362                 $line .= '..';
2363                 if ($to->{'href'}) {
2364                         $line .= $cgi->a({-href=>$to->{'href'}, -class=>"hash"},
2365                                          substr($diffinfo->{'to_id'},0,7));
2366                 } else {
2367                         $line .= '0' x 7;
2368                 }
2369
2370         } elsif ($line =~ oid_nlen_prefix_infix_regex($sha1_len, "index ", "..") |
2371                  $line =~ oid_nlen_prefix_infix_regex($sha256_len, "index ", "..")) {
2372                 # can match only for ordinary diff
2373                 my ($from_link, $to_link);
2374                 if ($from->{'href'}) {
2375                         $from_link = $cgi->a({-href=>$from->{'href'}, -class=>"hash"},
2376                                              substr($diffinfo->{'from_id'},0,7));
2377                 } else {
2378                         $from_link = '0' x 7;
2379                 }
2380                 if ($to->{'href'}) {
2381                         $to_link = $cgi->a({-href=>$to->{'href'}, -class=>"hash"},
2382                                            substr($diffinfo->{'to_id'},0,7));
2383                 } else {
2384                         $to_link = '0' x 7;
2385                 }
2386                 my ($from_id, $to_id) = ($diffinfo->{'from_id'}, $diffinfo->{'to_id'});
2387                 $line =~ s!$from_id\.\.$to_id!$from_link..$to_link!;
2388         }
2389
2390         return $line . "<br/>\n";
2391 }
2392
2393 # format from-file/to-file diff header
2394 sub format_diff_from_to_header {
2395         my ($from_line, $to_line, $diffinfo, $from, $to, @parents) = @_;
2396         my $line;
2397         my $result = '';
2398
2399         $line = $from_line;
2400         #assert($line =~ m/^---/) if DEBUG;
2401         # no extra formatting for "^--- /dev/null"
2402         if (! $diffinfo->{'nparents'}) {
2403                 # ordinary (single parent) diff
2404                 if ($line =~ m!^--- "?a/!) {
2405                         if ($from->{'href'}) {
2406                                 $line = '--- a/' .
2407                                         $cgi->a({-href=>$from->{'href'}, -class=>"path"},
2408                                                 esc_path($from->{'file'}));
2409                         } else {
2410                                 $line = '--- a/' .
2411                                         esc_path($from->{'file'});
2412                         }
2413                 }
2414                 $result .= qq!<div class="diff from_file">$line</div>\n!;
2415
2416         } else {
2417                 # combined diff (merge commit)
2418                 for (my $i = 0; $i < $diffinfo->{'nparents'}; $i++) {
2419                         if ($from->{'href'}[$i]) {
2420                                 $line = '--- ' .
2421                                         $cgi->a({-href=>href(action=>"blobdiff",
2422                                                              hash_parent=>$diffinfo->{'from_id'}[$i],
2423                                                              hash_parent_base=>$parents[$i],
2424                                                              file_parent=>$from->{'file'}[$i],
2425                                                              hash=>$diffinfo->{'to_id'},
2426                                                              hash_base=>$hash,
2427                                                              file_name=>$to->{'file'}),
2428                                                  -class=>"path",
2429                                                  -title=>"diff" . ($i+1)},
2430                                                 $i+1) .
2431                                         '/' .
2432                                         $cgi->a({-href=>$from->{'href'}[$i], -class=>"path"},
2433                                                 esc_path($from->{'file'}[$i]));
2434                         } else {
2435                                 $line = '--- /dev/null';
2436                         }
2437                         $result .= qq!<div class="diff from_file">$line</div>\n!;
2438                 }
2439         }
2440
2441         $line = $to_line;
2442         #assert($line =~ m/^\+\+\+/) if DEBUG;
2443         # no extra formatting for "^+++ /dev/null"
2444         if ($line =~ m!^\+\+\+ "?b/!) {
2445                 if ($to->{'href'}) {
2446                         $line = '+++ b/' .
2447                                 $cgi->a({-href=>$to->{'href'}, -class=>"path"},
2448                                         esc_path($to->{'file'}));
2449                 } else {
2450                         $line = '+++ b/' .
2451                                 esc_path($to->{'file'});
2452                 }
2453         }
2454         $result .= qq!<div class="diff to_file">$line</div>\n!;
2455
2456         return $result;
2457 }
2458
2459 # create note for patch simplified by combined diff
2460 sub format_diff_cc_simplified {
2461         my ($diffinfo, @parents) = @_;
2462         my $result = '';
2463
2464         $result .= "<div class=\"diff header\">" .
2465                    "diff --cc ";
2466         if (!is_deleted($diffinfo)) {
2467                 $result .= $cgi->a({-href => href(action=>"blob",
2468                                                   hash_base=>$hash,
2469                                                   hash=>$diffinfo->{'to_id'},
2470                                                   file_name=>$diffinfo->{'to_file'}),
2471                                     -class => "path"},
2472                                    esc_path($diffinfo->{'to_file'}));
2473         } else {
2474                 $result .= esc_path($diffinfo->{'to_file'});
2475         }
2476         $result .= "</div>\n" . # class="diff header"
2477                    "<div class=\"diff nodifferences\">" .
2478                    "Simple merge" .
2479                    "</div>\n"; # class="diff nodifferences"
2480
2481         return $result;
2482 }
2483
2484 sub diff_line_class {
2485         my ($line, $from, $to) = @_;
2486
2487         # ordinary diff
2488         my $num_sign = 1;
2489         # combined diff
2490         if ($from && $to && ref($from->{'href'}) eq "ARRAY") {
2491                 $num_sign = scalar @{$from->{'href'}};
2492         }
2493
2494         my @diff_line_classifier = (
2495                 { regexp => qr/^\@\@{$num_sign} /, class => "chunk_header"},
2496                 { regexp => qr/^\\/,               class => "incomplete"  },
2497                 { regexp => qr/^ {$num_sign}/,     class => "ctx" },
2498                 # classifier for context must come before classifier add/rem,
2499                 # or we would have to use more complicated regexp, for example
2500                 # qr/(?= {0,$m}\+)[+ ]{$num_sign}/, where $m = $num_sign - 1;
2501                 { regexp => qr/^[+ ]{$num_sign}/,   class => "add" },
2502                 { regexp => qr/^[- ]{$num_sign}/,   class => "rem" },
2503         );
2504         for my $clsfy (@diff_line_classifier) {
2505                 return $clsfy->{'class'}
2506                         if ($line =~ $clsfy->{'regexp'});
2507         }
2508
2509         # fallback
2510         return "";
2511 }
2512
2513 # assumes that $from and $to are defined and correctly filled,
2514 # and that $line holds a line of chunk header for unified diff
2515 sub format_unidiff_chunk_header {
2516         my ($line, $from, $to) = @_;
2517
2518         my ($from_text, $from_start, $from_lines, $to_text, $to_start, $to_lines, $section) =
2519                 $line =~ m/^\@{2} (-(\d+)(?:,(\d+))?) (\+(\d+)(?:,(\d+))?) \@{2}(.*)$/;
2520
2521         $from_lines = 0 unless defined $from_lines;
2522         $to_lines   = 0 unless defined $to_lines;
2523
2524         if ($from->{'href'}) {
2525                 $from_text = $cgi->a({-href=>"$from->{'href'}#l$from_start",
2526                                      -class=>"list"}, $from_text);
2527         }
2528         if ($to->{'href'}) {
2529                 $to_text   = $cgi->a({-href=>"$to->{'href'}#l$to_start",
2530                                      -class=>"list"}, $to_text);
2531         }
2532         $line = "<span class=\"chunk_info\">@@ $from_text $to_text @@</span>" .
2533                 "<span class=\"section\">" . esc_html($section, -nbsp=>1) . "</span>";
2534         return $line;
2535 }
2536
2537 # assumes that $from and $to are defined and correctly filled,
2538 # and that $line holds a line of chunk header for combined diff
2539 sub format_cc_diff_chunk_header {
2540         my ($line, $from, $to) = @_;
2541
2542         my ($prefix, $ranges, $section) = $line =~ m/^(\@+) (.*?) \@+(.*)$/;
2543         my (@from_text, @from_start, @from_nlines, $to_text, $to_start, $to_nlines);
2544
2545         @from_text = split(' ', $ranges);
2546         for (my $i = 0; $i < @from_text; ++$i) {
2547                 ($from_start[$i], $from_nlines[$i]) =
2548                         (split(',', substr($from_text[$i], 1)), 0);
2549         }
2550
2551         $to_text   = pop @from_text;
2552         $to_start  = pop @from_start;
2553         $to_nlines = pop @from_nlines;
2554
2555         $line = "<span class=\"chunk_info\">$prefix ";
2556         for (my $i = 0; $i < @from_text; ++$i) {
2557                 if ($from->{'href'}[$i]) {
2558                         $line .= $cgi->a({-href=>"$from->{'href'}[$i]#l$from_start[$i]",
2559                                           -class=>"list"}, $from_text[$i]);
2560                 } else {
2561                         $line .= $from_text[$i];
2562                 }
2563                 $line .= " ";
2564         }
2565         if ($to->{'href'}) {
2566                 $line .= $cgi->a({-href=>"$to->{'href'}#l$to_start",
2567                                   -class=>"list"}, $to_text);
2568         } else {
2569                 $line .= $to_text;
2570         }
2571         $line .= " $prefix</span>" .
2572                  "<span class=\"section\">" . esc_html($section, -nbsp=>1) . "</span>";
2573         return $line;
2574 }
2575
2576 # process patch (diff) line (not to be used for diff headers),
2577 # returning HTML-formatted (but not wrapped) line.
2578 # If the line is passed as a reference, it is treated as HTML and not
2579 # esc_html()'ed.
2580 sub format_diff_line {
2581         my ($line, $diff_class, $from, $to) = @_;
2582
2583         if (ref($line)) {
2584                 $line = $$line;
2585         } else {
2586                 chomp $line;
2587                 $line = untabify($line);
2588
2589                 if ($from && $to && $line =~ m/^\@{2} /) {
2590                         $line = format_unidiff_chunk_header($line, $from, $to);
2591                 } elsif ($from && $to && $line =~ m/^\@{3}/) {
2592                         $line = format_cc_diff_chunk_header($line, $from, $to);
2593                 } else {
2594                         $line = esc_html($line, -nbsp=>1);
2595                 }
2596         }
2597
2598         my $diff_classes = "diff";
2599         $diff_classes .= " $diff_class" if ($diff_class);
2600         $line = "<div class=\"$diff_classes\">$line</div>\n";
2601
2602         return $line;
2603 }
2604
2605 # Generates undef or something like "_snapshot_" or "snapshot (_tbz2_ _zip_)",
2606 # linked.  Pass the hash of the tree/commit to snapshot.
2607 sub format_snapshot_links {
2608         my ($hash) = @_;
2609         my $num_fmts = @snapshot_fmts;
2610         if ($num_fmts > 1) {
2611                 # A parenthesized list of links bearing format names.
2612                 # e.g. "snapshot (_tar.gz_ _zip_)"
2613                 return "snapshot (" . join(' ', map
2614                         $cgi->a({
2615                                 -href => href(
2616                                         action=>"snapshot",
2617                                         hash=>$hash,
2618                                         snapshot_format=>$_
2619                                 )
2620                         }, $known_snapshot_formats{$_}{'display'})
2621                 , @snapshot_fmts) . ")";
2622         } elsif ($num_fmts == 1) {
2623                 # A single "snapshot" link whose tooltip bears the format name.
2624                 # i.e. "_snapshot_"
2625                 my ($fmt) = @snapshot_fmts;
2626                 return
2627                         $cgi->a({
2628                                 -href => href(
2629                                         action=>"snapshot",
2630                                         hash=>$hash,
2631                                         snapshot_format=>$fmt
2632                                 ),
2633                                 -title => "in format: $known_snapshot_formats{$fmt}{'display'}"
2634                         }, "snapshot");
2635         } else { # $num_fmts == 0
2636                 return undef;
2637         }
2638 }
2639
2640 ## ......................................................................
2641 ## functions returning values to be passed, perhaps after some
2642 ## transformation, to other functions; e.g. returning arguments to href()
2643
2644 # returns hash to be passed to href to generate gitweb URL
2645 # in -title key it returns description of link
2646 sub get_feed_info {
2647         my $format = shift || 'Atom';
2648         my %res = (action => lc($format));
2649         my $matched_ref = 0;
2650
2651         # feed links are possible only for project views
2652         return unless (defined $project);
2653         # some views should link to OPML, or to generic project feed,
2654         # or don't have specific feed yet (so they should use generic)
2655         return if (!$action || $action =~ /^(?:tags|heads|forks|tag|search)$/x);
2656
2657         my $branch = undef;
2658         # branches refs uses 'refs/' + $get_branch_refs()[x] + '/' prefix
2659         # (fullname) to differentiate from tag links; this also makes
2660         # possible to detect branch links
2661         for my $ref (get_branch_refs()) {
2662                 if ((defined $hash_base && $hash_base =~ m!^refs/\Q$ref\E/(.*)$!) ||
2663                     (defined $hash      && $hash      =~ m!^refs/\Q$ref\E/(.*)$!)) {
2664                         $branch = $1;
2665                         $matched_ref = $ref;
2666                         last;
2667                 }
2668         }
2669         # find log type for feed description (title)
2670         my $type = 'log';
2671         if (defined $file_name) {
2672                 $type  = "history of $file_name";
2673                 $type .= "/" if ($action eq 'tree');
2674                 $type .= " on '$branch'" if (defined $branch);
2675         } else {
2676                 $type = "log of $branch" if (defined $branch);
2677         }
2678
2679         $res{-title} = $type;
2680         $res{'hash'} = (defined $branch ? "refs/$matched_ref/$branch" : undef);
2681         $res{'file_name'} = $file_name;
2682
2683         return %res;
2684 }
2685
2686 ## ----------------------------------------------------------------------
2687 ## git utility subroutines, invoking git commands
2688
2689 # returns path to the core git executable and the --git-dir parameter as list
2690 sub git_cmd {
2691         $number_of_git_cmds++;
2692         return $GIT, '--git-dir='.$git_dir;
2693 }
2694
2695 # quote the given arguments for passing them to the shell
2696 # quote_command("command", "arg 1", "arg with ' and ! characters")
2697 # => "'command' 'arg 1' 'arg with '\'' and '\!' characters'"
2698 # Try to avoid using this function wherever possible.
2699 sub quote_command {
2700         return join(' ',
2701                 map { my $a = $_; $a =~ s/(['!])/'\\$1'/g; "'$a'" } @_ );
2702 }
2703
2704 # get HEAD ref of given project as hash
2705 sub git_get_head_hash {
2706         return git_get_full_hash(shift, 'HEAD');
2707 }
2708
2709 sub git_get_full_hash {
2710         return git_get_hash(@_);
2711 }
2712
2713 sub git_get_short_hash {
2714         return git_get_hash(@_, '--short=7');
2715 }
2716
2717 sub git_get_hash {
2718         my ($project, $hash, @options) = @_;
2719         my $o_git_dir = $git_dir;
2720         my $retval = undef;
2721         $git_dir = "$projectroot/$project";
2722         if (open my $fd, '-|', git_cmd(), 'rev-parse',
2723             '--verify', '-q', @options, $hash) {
2724                 $retval = <$fd>;
2725                 chomp $retval if defined $retval;
2726                 close $fd;
2727         }
2728         if (defined $o_git_dir) {
2729                 $git_dir = $o_git_dir;
2730         }
2731         return $retval;
2732 }
2733
2734 # get type of given object
2735 sub git_get_type {
2736         my $hash = shift;
2737
2738         open my $fd, "-|", git_cmd(), "cat-file", '-t', $hash or return;
2739         my $type = <$fd>;
2740         close $fd or return;
2741         chomp $type;
2742         return $type;
2743 }
2744
2745 # repository configuration
2746 our $config_file = '';
2747 our %config;
2748
2749 # store multiple values for single key as anonymous array reference
2750 # single values stored directly in the hash, not as [ <value> ]
2751 sub hash_set_multi {
2752         my ($hash, $key, $value) = @_;
2753
2754         if (!exists $hash->{$key}) {
2755                 $hash->{$key} = $value;
2756         } elsif (!ref $hash->{$key}) {
2757                 $hash->{$key} = [ $hash->{$key}, $value ];
2758         } else {
2759                 push @{$hash->{$key}}, $value;
2760         }
2761 }
2762
2763 # return hash of git project configuration
2764 # optionally limited to some section, e.g. 'gitweb'
2765 sub git_parse_project_config {
2766         my $section_regexp = shift;
2767         my %config;
2768
2769         local $/ = "\0";
2770
2771         open my $fh, "-|", git_cmd(), "config", '-z', '-l',
2772                 or return;
2773
2774         while (my $keyval = <$fh>) {
2775                 chomp $keyval;
2776                 my ($key, $value) = split(/\n/, $keyval, 2);
2777
2778                 hash_set_multi(\%config, $key, $value)
2779                         if (!defined $section_regexp || $key =~ /^(?:$section_regexp)\./o);
2780         }
2781         close $fh;
2782
2783         return %config;
2784 }
2785
2786 # convert config value to boolean: 'true' or 'false'
2787 # no value, number > 0, 'true' and 'yes' values are true
2788 # rest of values are treated as false (never as error)
2789 sub config_to_bool {
2790         my $val = shift;
2791
2792         return 1 if !defined $val;             # section.key
2793
2794         # strip leading and trailing whitespace
2795         $val =~ s/^\s+//;
2796         $val =~ s/\s+$//;
2797
2798         return (($val =~ /^\d+$/ && $val) ||   # section.key = 1
2799                 ($val =~ /^(?:true|yes)$/i));  # section.key = true
2800 }
2801
2802 # convert config value to simple decimal number
2803 # an optional value suffix of 'k', 'm', or 'g' will cause the value
2804 # to be multiplied by 1024, 1048576, or 1073741824
2805 sub config_to_int {
2806         my $val = shift;
2807
2808         # strip leading and trailing whitespace
2809         $val =~ s/^\s+//;
2810         $val =~ s/\s+$//;
2811
2812         if (my ($num, $unit) = ($val =~ /^([0-9]*)([kmg])$/i)) {
2813                 $unit = lc($unit);
2814                 # unknown unit is treated as 1
2815                 return $num * ($unit eq 'g' ? 1073741824 :
2816                                $unit eq 'm' ?    1048576 :
2817                                $unit eq 'k' ?       1024 : 1);
2818         }
2819         return $val;
2820 }
2821
2822 # convert config value to array reference, if needed
2823 sub config_to_multi {
2824         my $val = shift;
2825
2826         return ref($val) ? $val : (defined($val) ? [ $val ] : []);
2827 }
2828
2829 sub git_get_project_config {
2830         my ($key, $type) = @_;
2831
2832         return unless defined $git_dir;
2833
2834         # key sanity check
2835         return unless ($key);
2836         # only subsection, if exists, is case sensitive,
2837         # and not lowercased by 'git config -z -l'
2838         if (my ($hi, $mi, $lo) = ($key =~ /^([^.]*)\.(.*)\.([^.]*)$/)) {
2839                 $lo =~ s/_//g;
2840                 $key = join(".", lc($hi), $mi, lc($lo));
2841                 return if ($lo =~ /\W/ || $hi =~ /\W/);
2842         } else {
2843                 $key = lc($key);
2844                 $key =~ s/_//g;
2845                 return if ($key =~ /\W/);
2846         }
2847         $key =~ s/^gitweb\.//;
2848
2849         # type sanity check
2850         if (defined $type) {
2851                 $type =~ s/^--//;
2852                 $type = undef
2853                         unless ($type eq 'bool' || $type eq 'int');
2854         }
2855
2856         # get config
2857         if (!defined $config_file ||
2858             $config_file ne "$git_dir/config") {
2859                 %config = git_parse_project_config('gitweb');
2860                 $config_file = "$git_dir/config";
2861         }
2862
2863         # check if config variable (key) exists
2864         return unless exists $config{"gitweb.$key"};
2865
2866         # ensure given type
2867         if (!defined $type) {
2868                 return $config{"gitweb.$key"};
2869         } elsif ($type eq 'bool') {
2870                 # backward compatibility: 'git config --bool' returns true/false
2871                 return config_to_bool($config{"gitweb.$key"}) ? 'true' : 'false';
2872         } elsif ($type eq 'int') {
2873                 return config_to_int($config{"gitweb.$key"});
2874         }
2875         return $config{"gitweb.$key"};
2876 }
2877
2878 # get hash of given path at given ref
2879 sub git_get_hash_by_path {
2880         my $base = shift;
2881         my $path = shift || return undef;
2882         my $type = shift;
2883
2884         $path =~ s,/+$,,;
2885
2886         open my $fd, "-|", git_cmd(), "ls-tree", $base, "--", $path
2887                 or die_error(500, "Open git-ls-tree failed");
2888         my $line = <$fd>;
2889         close $fd or return undef;
2890
2891         if (!defined $line) {
2892                 # there is no tree or hash given by $path at $base
2893                 return undef;
2894         }
2895
2896         #'100644 blob 0fa3f3a66fb6a137f6ec2c19351ed4d807070ffa  panic.c'
2897         $line =~ m/^([0-9]+) (.+) ($oid_regex)\t/;
2898         if (defined $type && $type ne $2) {
2899                 # type doesn't match
2900                 return undef;
2901         }
2902         return $3;
2903 }
2904
2905 # get path of entry with given hash at given tree-ish (ref)
2906 # used to get 'from' filename for combined diff (merge commit) for renames
2907 sub git_get_path_by_hash {
2908         my $base = shift || return;
2909         my $hash = shift || return;
2910
2911         local $/ = "\0";
2912
2913         open my $fd, "-|", git_cmd(), "ls-tree", '-r', '-t', '-z', $base
2914                 or return undef;
2915         while (my $line = <$fd>) {
2916                 chomp $line;
2917
2918                 #'040000 tree 595596a6a9117ddba9fe379b6b012b558bac8423  gitweb'
2919                 #'100644 blob e02e90f0429be0d2a69b76571101f20b8f75530f  gitweb/README'
2920                 if ($line =~ m/(?:[0-9]+) (?:.+) $hash\t(.+)$/) {
2921                         close $fd;
2922                         return $1;
2923                 }
2924         }
2925         close $fd;
2926         return undef;
2927 }
2928
2929 ## ......................................................................
2930 ## git utility functions, directly accessing git repository
2931
2932 # get the value of config variable either from file named as the variable
2933 # itself in the repository ($GIT_DIR/$name file), or from gitweb.$name
2934 # configuration variable in the repository config file.
2935 sub git_get_file_or_project_config {
2936         my ($path, $name) = @_;
2937
2938         $git_dir = "$projectroot/$path";
2939         open my $fd, '<', "$git_dir/$name"
2940                 or return git_get_project_config($name);
2941         my $conf = <$fd>;
2942         close $fd;
2943         if (defined $conf) {
2944                 chomp $conf;
2945         }
2946         return $conf;
2947 }
2948
2949 sub git_get_project_description {
2950         my $path = shift;
2951         return git_get_file_or_project_config($path, 'description');
2952 }
2953
2954 sub git_get_project_category {
2955         my $path = shift;
2956         return git_get_file_or_project_config($path, 'category');
2957 }
2958
2959
2960 # supported formats:
2961 # * $GIT_DIR/ctags/<tagname> file (in 'ctags' subdirectory)
2962 #   - if its contents is a number, use it as tag weight,
2963 #   - otherwise add a tag with weight 1
2964 # * $GIT_DIR/ctags file, each line is a tag (with weight 1)
2965 #   the same value multiple times increases tag weight
2966 # * `gitweb.ctag' multi-valued repo config variable
2967 sub git_get_project_ctags {
2968         my $project = shift;
2969         my $ctags = {};
2970
2971         $git_dir = "$projectroot/$project";
2972         if (opendir my $dh, "$git_dir/ctags") {
2973                 my @files = grep { -f $_ } map { "$git_dir/ctags/$_" } readdir($dh);
2974                 foreach my $tagfile (@files) {
2975                         open my $ct, '<', $tagfile
2976                                 or next;
2977                         my $val = <$ct>;
2978                         chomp $val if $val;
2979                         close $ct;
2980
2981                         (my $ctag = $tagfile) =~ s#.*/##;
2982                         if ($val =~ /^\d+$/) {
2983                                 $ctags->{$ctag} = $val;
2984                         } else {
2985                                 $ctags->{$ctag} = 1;
2986                         }
2987                 }
2988                 closedir $dh;
2989
2990         } elsif (open my $fh, '<', "$git_dir/ctags") {
2991                 while (my $line = <$fh>) {
2992                         chomp $line;
2993                         $ctags->{$line}++ if $line;
2994                 }
2995                 close $fh;
2996
2997         } else {
2998                 my $taglist = config_to_multi(git_get_project_config('ctag'));
2999                 foreach my $tag (@$taglist) {
3000                         $ctags->{$tag}++;
3001                 }
3002         }
3003
3004         return $ctags;
3005 }
3006
3007 # return hash, where keys are content tags ('ctags'),
3008 # and values are sum of weights of given tag in every project
3009 sub git_gather_all_ctags {
3010         my $projects = shift;
3011         my $ctags = {};
3012
3013         foreach my $p (@$projects) {
3014                 foreach my $ct (keys %{$p->{'ctags'}}) {
3015                         $ctags->{$ct} += $p->{'ctags'}->{$ct};
3016                 }
3017         }
3018
3019         return $ctags;
3020 }
3021
3022 sub git_populate_project_tagcloud {
3023         my $ctags = shift;
3024
3025         # First, merge different-cased tags; tags vote on casing
3026         my %ctags_lc;
3027         foreach (keys %$ctags) {
3028                 $ctags_lc{lc $_}->{count} += $ctags->{$_};
3029                 if (not $ctags_lc{lc $_}->{topcount}
3030                     or $ctags_lc{lc $_}->{topcount} < $ctags->{$_}) {
3031                         $ctags_lc{lc $_}->{topcount} = $ctags->{$_};
3032                         $ctags_lc{lc $_}->{topname} = $_;
3033                 }
3034         }
3035
3036         my $cloud;
3037         my $matched = $input_params{'ctag'};
3038         if (eval { require HTML::TagCloud; 1; }) {
3039                 $cloud = HTML::TagCloud->new;
3040                 foreach my $ctag (sort keys %ctags_lc) {
3041                         # Pad the title with spaces so that the cloud looks
3042                         # less crammed.
3043                         my $title = esc_html($ctags_lc{$ctag}->{topname});
3044                         $title =~ s/ /&nbsp;/g;
3045                         $title =~ s/^/&nbsp;/g;
3046                         $title =~ s/$/&nbsp;/g;
3047                         if (defined $matched && $matched eq $ctag) {
3048                                 $title = qq(<span class="match">$title</span>);
3049                         }
3050                         $cloud->add($title, href(project=>undef, ctag=>$ctag),
3051                                     $ctags_lc{$ctag}->{count});
3052                 }
3053         } else {
3054                 $cloud = {};
3055                 foreach my $ctag (keys %ctags_lc) {
3056                         my $title = esc_html($ctags_lc{$ctag}->{topname}, -nbsp=>1);
3057                         if (defined $matched && $matched eq $ctag) {
3058                                 $title = qq(<span class="match">$title</span>);
3059                         }
3060                         $cloud->{$ctag}{count} = $ctags_lc{$ctag}->{count};
3061                         $cloud->{$ctag}{ctag} =
3062                                 $cgi->a({-href=>href(project=>undef, ctag=>$ctag)}, $title);
3063                 }
3064         }
3065         return $cloud;
3066 }
3067
3068 sub git_show_project_tagcloud {
3069         my ($cloud, $count) = @_;
3070         if (ref $cloud eq 'HTML::TagCloud') {
3071                 return $cloud->html_and_css($count);
3072         } else {
3073                 my @tags = sort { $cloud->{$a}->{'count'} <=> $cloud->{$b}->{'count'} } keys %$cloud;
3074                 return
3075                         '<div id="htmltagcloud"'.($project ? '' : ' align="center"').'>' .
3076                         join (', ', map {
3077                                 $cloud->{$_}->{'ctag'}
3078                         } splice(@tags, 0, $count)) .
3079                         '</div>';
3080         }
3081 }
3082
3083 sub git_get_project_url_list {
3084         my $path = shift;
3085
3086         $git_dir = "$projectroot/$path";
3087         open my $fd, '<', "$git_dir/cloneurl"
3088                 or return wantarray ?
3089                 @{ config_to_multi(git_get_project_config('url')) } :
3090                    config_to_multi(git_get_project_config('url'));
3091         my @git_project_url_list = map { chomp; $_ } <$fd>;
3092         close $fd;
3093
3094         return wantarray ? @git_project_url_list : \@git_project_url_list;
3095 }
3096
3097 sub git_get_projects_list {
3098         my $filter = shift || '';
3099         my $paranoid = shift;
3100         my @list;
3101
3102         if (-d $projects_list) {
3103                 # search in directory
3104                 my $dir = $projects_list;
3105                 # remove the trailing "/"
3106                 $dir =~ s!/+$!!;
3107                 my $pfxlen = length("$dir");
3108                 my $pfxdepth = ($dir =~ tr!/!!);
3109                 # when filtering, search only given subdirectory
3110                 if ($filter && !$paranoid) {
3111                         $dir .= "/$filter";
3112                         $dir =~ s!/+$!!;
3113                 }
3114
3115                 File::Find::find({
3116                         follow_fast => 1, # follow symbolic links
3117                         follow_skip => 2, # ignore duplicates
3118                         dangling_symlinks => 0, # ignore dangling symlinks, silently
3119                         wanted => sub {
3120                                 # global variables
3121                                 our $project_maxdepth;
3122                                 our $projectroot;
3123                                 # skip project-list toplevel, if we get it.
3124                                 return if (m!^[/.]$!);
3125                                 # only directories can be git repositories
3126                                 return unless (-d $_);
3127                                 # need search permission
3128                                 return unless (-x $_);
3129                                 # don't traverse too deep (Find is super slow on os x)
3130                                 # $project_maxdepth excludes depth of $projectroot
3131                                 if (($File::Find::name =~ tr!/!!) - $pfxdepth > $project_maxdepth) {
3132                                         $File::Find::prune = 1;
3133                                         return;
3134                                 }
3135
3136                                 my $path = substr($File::Find::name, $pfxlen + 1);
3137                                 # paranoidly only filter here
3138                                 if ($paranoid && $filter && $path !~ m!^\Q$filter\E/!) {
3139                                         next;
3140                                 }
3141                                 # we check related file in $projectroot
3142                                 if (check_export_ok("$projectroot/$path")) {
3143                                         push @list, { path => $path };
3144                                         $File::Find::prune = 1;
3145                                 }
3146                         },
3147                 }, "$dir");
3148
3149         } elsif (-f $projects_list) {
3150                 # read from file(url-encoded):
3151                 # 'git%2Fgit.git Linus+Torvalds'
3152                 # 'libs%2Fklibc%2Fklibc.git H.+Peter+Anvin'
3153                 # 'linux%2Fhotplug%2Fudev.git Greg+Kroah-Hartman'
3154                 open my $fd, '<', $projects_list or return;
3155         PROJECT:
3156                 while (my $line = <$fd>) {
3157                         chomp $line;
3158                         my ($path, $owner) = split ' ', $line;
3159                         $path = unescape($path);
3160                         $owner = unescape($owner);
3161                         if (!defined $path) {
3162                                 next;
3163                         }
3164                         # if $filter is rpovided, check if $path begins with $filter
3165                         if ($filter && $path !~ m!^\Q$filter\E/!) {
3166                                 next;
3167                         }
3168                         if (check_export_ok("$projectroot/$path")) {
3169                                 my $pr = {
3170                                         path => $path
3171                                 };
3172                                 if ($owner) {
3173                                         $pr->{'owner'} = to_utf8($owner);
3174                                 }
3175                                 push @list, $pr;
3176                         }
3177                 }
3178                 close $fd;
3179         }
3180         return @list;
3181 }
3182
3183 # written with help of Tree::Trie module (Perl Artistic License, GPL compatible)
3184 # as side effects it sets 'forks' field to list of forks for forked projects
3185 sub filter_forks_from_projects_list {
3186         my $projects = shift;
3187
3188         my %trie; # prefix tree of directories (path components)
3189         # generate trie out of those directories that might contain forks
3190         foreach my $pr (@$projects) {
3191                 my $path = $pr->{'path'};
3192                 $path =~ s/\.git$//;      # forks of 'repo.git' are in 'repo/' directory
3193                 next if ($path =~ m!/$!); # skip non-bare repositories, e.g. 'repo/.git'
3194                 next unless ($path);      # skip '.git' repository: tests, git-instaweb
3195                 next unless (-d "$projectroot/$path"); # containing directory exists
3196                 $pr->{'forks'} = [];      # there can be 0 or more forks of project
3197
3198                 # add to trie
3199                 my @dirs = split('/', $path);
3200                 # walk the trie, until either runs out of components or out of trie
3201                 my $ref = \%trie;
3202                 while (scalar @dirs &&
3203                        exists($ref->{$dirs[0]})) {
3204                         $ref = $ref->{shift @dirs};
3205                 }
3206                 # create rest of trie structure from rest of components
3207                 foreach my $dir (@dirs) {
3208                         $ref = $ref->{$dir} = {};
3209                 }
3210                 # create end marker, store $pr as a data
3211                 $ref->{''} = $pr if (!exists $ref->{''});
3212         }
3213
3214         # filter out forks, by finding shortest prefix match for paths
3215         my @filtered;
3216  PROJECT:
3217         foreach my $pr (@$projects) {
3218                 # trie lookup
3219                 my $ref = \%trie;
3220         DIR:
3221                 foreach my $dir (split('/', $pr->{'path'})) {
3222                         if (exists $ref->{''}) {
3223                                 # found [shortest] prefix, is a fork - skip it
3224                                 push @{$ref->{''}{'forks'}}, $pr;
3225                                 next PROJECT;
3226                         }
3227                         if (!exists $ref->{$dir}) {
3228                                 # not in trie, cannot have prefix, not a fork
3229                                 push @filtered, $pr;
3230                                 next PROJECT;
3231                         }
3232                         # If the dir is there, we just walk one step down the trie.
3233                         $ref = $ref->{$dir};
3234                 }
3235                 # we ran out of trie
3236                 # (shouldn't happen: it's either no match, or end marker)
3237                 push @filtered, $pr;
3238         }
3239
3240         return @filtered;
3241 }
3242
3243 # note: fill_project_list_info must be run first,
3244 # for 'descr_long' and 'ctags' to be filled
3245 sub search_projects_list {
3246         my ($projlist, %opts) = @_;
3247         my $tagfilter  = $opts{'tagfilter'};
3248         my $search_re = $opts{'search_regexp'};
3249
3250         return @$projlist
3251                 unless ($tagfilter || $search_re);
3252
3253         # searching projects require filling to be run before it;
3254         fill_project_list_info($projlist,
3255                                $tagfilter  ? 'ctags' : (),
3256                                $search_re ? ('path', 'descr') : ());
3257         my @projects;
3258  PROJECT:
3259         foreach my $pr (@$projlist) {
3260
3261                 if ($tagfilter) {
3262                         next unless ref($pr->{'ctags'}) eq 'HASH';
3263                         next unless
3264                                 grep { lc($_) eq lc($tagfilter) } keys %{$pr->{'ctags'}};
3265                 }
3266
3267                 if ($search_re) {
3268                         next unless
3269                                 $pr->{'path'} =~ /$search_re/ ||
3270                                 $pr->{'descr_long'} =~ /$search_re/;
3271                 }
3272
3273                 push @projects, $pr;
3274         }
3275
3276         return @projects;
3277 }
3278
3279 our $gitweb_project_owner = undef;
3280 sub git_get_project_list_from_file {
3281
3282         return if (defined $gitweb_project_owner);
3283
3284         $gitweb_project_owner = {};
3285         # read from file (url-encoded):
3286         # 'git%2Fgit.git Linus+Torvalds'
3287         # 'libs%2Fklibc%2Fklibc.git H.+Peter+Anvin'
3288         # 'linux%2Fhotplug%2Fudev.git Greg+Kroah-Hartman'
3289         if (-f $projects_list) {
3290                 open(my $fd, '<', $projects_list);
3291                 while (my $line = <$fd>) {
3292                         chomp $line;
3293                         my ($pr, $ow) = split ' ', $line;
3294                         $pr = unescape($pr);
3295                         $ow = unescape($ow);
3296                         $gitweb_project_owner->{$pr} = to_utf8($ow);
3297                 }
3298                 close $fd;
3299         }
3300 }
3301
3302 sub git_get_project_owner {
3303         my $project = shift;
3304         my $owner;
3305
3306         return undef unless $project;
3307         $git_dir = "$projectroot/$project";
3308
3309         if (!defined $gitweb_project_owner) {
3310                 git_get_project_list_from_file();
3311         }
3312
3313         if (exists $gitweb_project_owner->{$project}) {
3314                 $owner = $gitweb_project_owner->{$project};
3315         }
3316         if (!defined $owner){
3317                 $owner = git_get_project_config('owner');
3318         }
3319         if (!defined $owner) {
3320                 $owner = get_file_owner("$git_dir");
3321         }
3322
3323         return $owner;
3324 }
3325
3326 sub git_get_last_activity {
3327         my ($path) = @_;
3328         my $fd;
3329
3330         $git_dir = "$projectroot/$path";
3331         open($fd, "-|", git_cmd(), 'for-each-ref',
3332              '--format=%(committer)',
3333              '--sort=-committerdate',
3334              '--count=1',
3335              map { "refs/$_" } get_branch_refs ()) or return;
3336         my $most_recent = <$fd>;
3337         close $fd or return;
3338         if (defined $most_recent &&
3339             $most_recent =~ / (\d+) [-+][01]\d\d\d$/) {
3340                 my $timestamp = $1;
3341                 my $age = time - $timestamp;
3342                 return ($age, age_string($age));
3343         }
3344         return (undef, undef);
3345 }
3346
3347 # Implementation note: when a single remote is wanted, we cannot use 'git
3348 # remote show -n' because that command always work (assuming it's a remote URL
3349 # if it's not defined), and we cannot use 'git remote show' because that would
3350 # try to make a network roundtrip. So the only way to find if that particular
3351 # remote is defined is to walk the list provided by 'git remote -v' and stop if
3352 # and when we find what we want.
3353 sub git_get_remotes_list {
3354         my $wanted = shift;
3355         my %remotes = ();
3356
3357         open my $fd, '-|' , git_cmd(), 'remote', '-v';
3358         return unless $fd;
3359         while (my $remote = <$fd>) {
3360                 chomp $remote;
3361                 $remote =~ s!\t(.*?)\s+\((\w+)\)$!!;
3362                 next if $wanted and not $remote eq $wanted;
3363                 my ($url, $key) = ($1, $2);
3364
3365                 $remotes{$remote} ||= { 'heads' => () };
3366                 $remotes{$remote}{$key} = $url;
3367         }
3368         close $fd or return;
3369         return wantarray ? %remotes : \%remotes;
3370 }
3371
3372 # Takes a hash of remotes as first parameter and fills it by adding the
3373 # available remote heads for each of the indicated remotes.
3374 sub fill_remote_heads {
3375         my $remotes = shift;
3376         my @heads = map { "remotes/$_" } keys %$remotes;
3377         my @remoteheads = git_get_heads_list(undef, @heads);
3378         foreach my $remote (keys %$remotes) {
3379                 $remotes->{$remote}{'heads'} = [ grep {
3380                         $_->{'name'} =~ s!^$remote/!!
3381                         } @remoteheads ];
3382         }
3383 }
3384
3385 sub git_get_references {
3386         my $type = shift || "";
3387         my %refs;
3388         # 5dc01c595e6c6ec9ccda4f6f69c131c0dd945f8c refs/tags/v2.6.11
3389         # c39ae07f393806ccf406ef966e9a15afc43cc36a refs/tags/v2.6.11^{}
3390         open my $fd, "-|", git_cmd(), "show-ref", "--dereference",
3391                 ($type ? ("--", "refs/$type") : ()) # use -- <pattern> if $type
3392                 or return;
3393
3394         while (my $line = <$fd>) {
3395                 chomp $line;
3396                 if ($line =~ m!^($oid_regex)\srefs/($type.*)$!) {
3397                         if (defined $refs{$1}) {
3398                                 push @{$refs{$1}}, $2;
3399                         } else {
3400                                 $refs{$1} = [ $2 ];
3401                         }
3402                 }
3403         }
3404         close $fd or return;
3405         return \%refs;
3406 }
3407
3408 sub git_get_rev_name_tags {
3409         my $hash = shift || return undef;
3410
3411         open my $fd, "-|", git_cmd(), "name-rev", "--tags", $hash
3412                 or return;
3413         my $name_rev = <$fd>;
3414         close $fd;
3415
3416         if ($name_rev =~ m|^$hash tags/(.*)$|) {
3417                 return $1;
3418         } else {
3419                 # catches also '$hash undefined' output
3420                 return undef;
3421         }
3422 }
3423
3424 ## ----------------------------------------------------------------------
3425 ## parse to hash functions
3426
3427 sub parse_date {
3428         my $epoch = shift;
3429         my $tz = shift || "-0000";
3430
3431         my %date;
3432         my @months = ("Jan", "Feb", "Mar", "Apr", "May", "Jun", "Jul", "Aug", "Sep", "Oct", "Nov", "Dec");
3433         my @days = ("Sun", "Mon", "Tue", "Wed", "Thu", "Fri", "Sat");
3434         my ($sec, $min, $hour, $mday, $mon, $year, $wday, $yday) = gmtime($epoch);
3435         $date{'hour'} = $hour;
3436         $date{'minute'} = $min;
3437         $date{'mday'} = $mday;
3438         $date{'day'} = $days[$wday];
3439         $date{'month'} = $months[$mon];
3440         $date{'rfc2822'}   = sprintf "%s, %d %s %4d %02d:%02d:%02d +0000",
3441                              $days[$wday], $mday, $months[$mon], 1900+$year, $hour ,$min, $sec;
3442         $date{'mday-time'} = sprintf "%d %s %02d:%02d",
3443                              $mday, $months[$mon], $hour ,$min;
3444         $date{'iso-8601'}  = sprintf "%04d-%02d-%02dT%02d:%02d:%02dZ",
3445                              1900+$year, 1+$mon, $mday, $hour ,$min, $sec;
3446
3447         my ($tz_sign, $tz_hour, $tz_min) =
3448                 ($tz =~ m/^([-+])(\d\d)(\d\d)$/);
3449         $tz_sign = ($tz_sign eq '-' ? -1 : +1);
3450         my $local = $epoch + $tz_sign*((($tz_hour*60) + $tz_min)*60);
3451         ($sec, $min, $hour, $mday, $mon, $year, $wday, $yday) = gmtime($local);
3452         $date{'hour_local'} = $hour;
3453         $date{'minute_local'} = $min;
3454         $date{'tz_local'} = $tz;
3455         $date{'iso-tz'} = sprintf("%04d-%02d-%02d %02d:%02d:%02d %s",
3456                                   1900+$year, $mon+1, $mday,
3457                                   $hour, $min, $sec, $tz);
3458         return %date;
3459 }
3460
3461 sub hide_mailaddrs_if_private {
3462         my $line = shift;
3463         return $line unless gitweb_check_feature('email-privacy');
3464         $line =~ s/<[^@>]+@[^>]+>/<redacted>/g;
3465         return $line;
3466 }
3467
3468 sub parse_tag {
3469         my $tag_id = shift;
3470         my %tag;
3471         my @comment;
3472
3473         open my $fd, "-|", git_cmd(), "cat-file", "tag", $tag_id or return;
3474         $tag{'id'} = $tag_id;
3475         while (my $line = <$fd>) {
3476                 chomp $line;
3477                 if ($line =~ m/^object ($oid_regex)$/) {
3478                         $tag{'object'} = $1;
3479                 } elsif ($line =~ m/^type (.+)$/) {
3480                         $tag{'type'} = $1;
3481                 } elsif ($line =~ m/^tag (.+)$/) {
3482                         $tag{'name'} = $1;
3483                 } elsif ($line =~ m/^tagger (.*) ([0-9]+) (.*)$/) {
3484                         $tag{'author'} = hide_mailaddrs_if_private($1);
3485                         $tag{'author_epoch'} = $2;
3486                         $tag{'author_tz'} = $3;
3487                         if ($tag{'author'} =~ m/^([^<]+) <([^>]*)>/) {
3488                                 $tag{'author_name'}  = $1;
3489                                 $tag{'author_email'} = $2;
3490                         } else {
3491                                 $tag{'author_name'} = $tag{'author'};
3492                         }
3493                 } elsif ($line =~ m/--BEGIN/) {
3494                         push @comment, $line;
3495                         last;
3496                 } elsif ($line eq "") {
3497                         last;
3498                 }
3499         }
3500         push @comment, <$fd>;
3501         $tag{'comment'} = \@comment;
3502         close $fd or return;
3503         if (!defined $tag{'name'}) {
3504                 return
3505         };
3506         return %tag
3507 }
3508
3509 sub parse_commit_text {
3510         my ($commit_text, $withparents) = @_;
3511         my @commit_lines = split '\n', $commit_text;
3512         my %co;
3513
3514         pop @commit_lines; # Remove '\0'
3515
3516         if (! @commit_lines) {
3517                 return;
3518         }
3519
3520         my $header = shift @commit_lines;
3521         if ($header !~ m/^$oid_regex/) {
3522                 return;
3523         }
3524         ($co{'id'}, my @parents) = split ' ', $header;
3525         while (my $line = shift @commit_lines) {
3526                 last if $line eq "\n";
3527                 if ($line =~ m/^tree ($oid_regex)$/) {
3528                         $co{'tree'} = $1;
3529                 } elsif ((!defined $withparents) && ($line =~ m/^parent ($oid_regex)$/)) {
3530                         push @parents, $1;
3531                 } elsif ($line =~ m/^author (.*) ([0-9]+) (.*)$/) {
3532                         $co{'author'} = hide_mailaddrs_if_private(to_utf8($1));
3533                         $co{'author_epoch'} = $2;
3534                         $co{'author_tz'} = $3;
3535                         if ($co{'author'} =~ m/^([^<]+) <([^>]*)>/) {
3536                                 $co{'author_name'}  = $1;
3537                                 $co{'author_email'} = $2;
3538                         } else {
3539                                 $co{'author_name'} = $co{'author'};
3540                         }
3541                 } elsif ($line =~ m/^committer (.*) ([0-9]+) (.*)$/) {
3542                         $co{'committer'} = hide_mailaddrs_if_private(to_utf8($1));
3543                         $co{'committer_epoch'} = $2;
3544                         $co{'committer_tz'} = $3;
3545                         if ($co{'committer'} =~ m/^([^<]+) <([^>]*)>/) {
3546                                 $co{'committer_name'}  = $1;
3547                                 $co{'committer_email'} = $2;
3548                         } else {
3549                                 $co{'committer_name'} = $co{'committer'};
3550                         }
3551                 }
3552         }
3553         if (!defined $co{'tree'}) {
3554                 return;
3555         };
3556         $co{'parents'} = \@parents;
3557         $co{'parent'} = $parents[0];
3558
3559         foreach my $title (@commit_lines) {
3560                 $title =~ s/^    //;
3561                 if ($title ne "") {
3562                         $co{'title'} = chop_str($title, 80, 5);
3563                         # remove leading stuff of merges to make the interesting part visible
3564                         if (length($title) > 50) {
3565                                 $title =~ s/^Automatic //;
3566                                 $title =~ s/^merge (of|with) /Merge ... /i;
3567                                 if (length($title) > 50) {
3568                                         $title =~ s/(http|rsync):\/\///;
3569                                 }
3570                                 if (length($title) > 50) {
3571                                         $title =~ s/(master|www|rsync)\.//;
3572                                 }
3573                                 if (length($title) > 50) {
3574                                         $title =~ s/kernel.org:?//;
3575                                 }
3576                                 if (length($title) > 50) {
3577                                         $title =~ s/\/pub\/scm//;
3578                                 }
3579                         }
3580                         $co{'title_short'} = chop_str($title, 50, 5);
3581                         last;
3582                 }
3583         }
3584         if (! defined $co{'title'} || $co{'title'} eq "") {
3585                 $co{'title'} = $co{'title_short'} = '(no commit message)';
3586         }
3587         # remove added spaces, redact e-mail addresses if applicable.
3588         foreach my $line (@commit_lines) {
3589                 $line =~ s/^    //;
3590                 $line = hide_mailaddrs_if_private($line);
3591         }
3592         $co{'comment'} = \@commit_lines;
3593
3594         my $age = time - $co{'committer_epoch'};
3595         $co{'age'} = $age;
3596         $co{'age_string'} = age_string($age);
3597         my ($sec, $min, $hour, $mday, $mon, $year, $wday, $yday) = gmtime($co{'committer_epoch'});
3598         if ($age > 60*60*24*7*2) {
3599                 $co{'age_string_date'} = sprintf "%4i-%02u-%02i", 1900 + $year, $mon+1, $mday;
3600                 $co{'age_string_age'} = $co{'age_string'};
3601         } else {
3602                 $co{'age_string_date'} = $co{'age_string'};
3603                 $co{'age_string_age'} = sprintf "%4i-%02u-%02i", 1900 + $year, $mon+1, $mday;
3604         }
3605         return %co;
3606 }
3607
3608 sub parse_commit {
3609         my ($commit_id) = @_;
3610         my %co;
3611
3612         local $/ = "\0";
3613
3614         open my $fd, "-|", git_cmd(), "rev-list",
3615                 "--parents",
3616                 "--header",
3617                 "--max-count=1",
3618                 $commit_id,
3619                 "--",
3620                 or die_error(500, "Open git-rev-list failed");
3621         %co = parse_commit_text(<$fd>, 1);
3622         close $fd;
3623
3624         return %co;
3625 }
3626
3627 sub parse_commits {
3628         my ($commit_id, $maxcount, $skip, $filename, @args) = @_;
3629         my @cos;
3630
3631         $maxcount ||= 1;
3632         $skip ||= 0;
3633
3634         local $/ = "\0";
3635
3636         open my $fd, "-|", git_cmd(), "rev-list",
3637                 "--header",
3638                 @args,
3639                 ("--max-count=" . $maxcount),
3640                 ("--skip=" . $skip),
3641                 @extra_options,
3642                 $commit_id,
3643                 "--",
3644                 ($filename ? ($filename) : ())
3645                 or die_error(500, "Open git-rev-list failed");
3646         while (my $line = <$fd>) {
3647                 my %co = parse_commit_text($line);
3648                 push @cos, \%co;
3649         }
3650         close $fd;
3651
3652         return wantarray ? @cos : \@cos;
3653 }
3654
3655 # parse line of git-diff-tree "raw" output
3656 sub parse_difftree_raw_line {
3657         my $line = shift;
3658         my %res;
3659
3660         # ':100644 100644 03b218260e99b78c6df0ed378e59ed9205ccc96d 3b93d5e7cc7f7dd4ebed13a5cc1a4ad976fc94d8 M   ls-files.c'
3661         # ':100644 100644 7f9281985086971d3877aca27704f2aaf9c448ce bc190ebc71bbd923f2b728e505408f5e54bd073a M   rev-tree.c'
3662         if ($line =~ m/^:([0-7]{6}) ([0-7]{6}) ($oid_regex) ($oid_regex) (.)([0-9]{0,3})\t(.*)$/) {
3663                 $res{'from_mode'} = $1;
3664                 $res{'to_mode'} = $2;
3665                 $res{'from_id'} = $3;
3666                 $res{'to_id'} = $4;
3667                 $res{'status'} = $5;
3668                 $res{'similarity'} = $6;
3669                 if ($res{'status'} eq 'R' || $res{'status'} eq 'C') { # renamed or copied
3670                         ($res{'from_file'}, $res{'to_file'}) = map { unquote($_) } split("\t", $7);
3671                 } else {
3672                         $res{'from_file'} = $res{'to_file'} = $res{'file'} = unquote($7);
3673                 }
3674         }
3675         # '::100755 100755 100755 60e79ca1b01bc8b057abe17ddab484699a7f5fdb 94067cc5f73388f33722d52ae02f44692bc07490 94067cc5f73388f33722d52ae02f44692bc07490 MR git-gui/git-gui.sh'
3676         # combined diff (for merge commit)
3677         elsif ($line =~ s/^(::+)((?:[0-7]{6} )+)((?:$oid_regex )+)([a-zA-Z]+)\t(.*)$//) {
3678                 $res{'nparents'}  = length($1);
3679                 $res{'from_mode'} = [ split(' ', $2) ];
3680                 $res{'to_mode'} = pop @{$res{'from_mode'}};
3681                 $res{'from_id'} = [ split(' ', $3) ];
3682                 $res{'to_id'} = pop @{$res{'from_id'}};
3683                 $res{'status'} = [ split('', $4) ];
3684                 $res{'to_file'} = unquote($5);
3685         }
3686         # 'c512b523472485aef4fff9e57b229d9d243c967f'
3687         elsif ($line =~ m/^($oid_regex)$/) {
3688                 $res{'commit'} = $1;
3689         }
3690
3691         return wantarray ? %res : \%res;
3692 }
3693
3694 # wrapper: return parsed line of git-diff-tree "raw" output
3695 # (the argument might be raw line, or parsed info)
3696 sub parsed_difftree_line {
3697         my $line_or_ref = shift;
3698
3699         if (ref($line_or_ref) eq "HASH") {
3700                 # pre-parsed (or generated by hand)
3701                 return $line_or_ref;
3702         } else {
3703                 return parse_difftree_raw_line($line_or_ref);
3704         }
3705 }
3706
3707 # parse line of git-ls-tree output
3708 sub parse_ls_tree_line {
3709         my $line = shift;
3710         my %opts = @_;
3711         my %res;
3712
3713         if ($opts{'-l'}) {
3714                 #'100644 blob 0fa3f3a66fb6a137f6ec2c19351ed4d807070ffa   16717  panic.c'
3715                 $line =~ m/^([0-9]+) (.+) ($oid_regex) +(-|[0-9]+)\t(.+)$/s;
3716
3717                 $res{'mode'} = $1;
3718                 $res{'type'} = $2;
3719                 $res{'hash'} = $3;
3720                 $res{'size'} = $4;
3721                 if ($opts{'-z'}) {
3722                         $res{'name'} = $5;
3723                 } else {
3724                         $res{'name'} = unquote($5);
3725                 }
3726         } else {
3727                 #'100644 blob 0fa3f3a66fb6a137f6ec2c19351ed4d807070ffa  panic.c'
3728                 $line =~ m/^([0-9]+) (.+) ($oid_regex)\t(.+)$/s;
3729
3730                 $res{'mode'} = $1;
3731                 $res{'type'} = $2;
3732                 $res{'hash'} = $3;
3733                 if ($opts{'-z'}) {
3734                         $res{'name'} = $4;
3735                 } else {
3736                         $res{'name'} = unquote($4);
3737                 }
3738         }
3739
3740         return wantarray ? %res : \%res;
3741 }
3742
3743 # generates _two_ hashes, references to which are passed as 2 and 3 argument
3744 sub parse_from_to_diffinfo {
3745         my ($diffinfo, $from, $to, @parents) = @_;
3746
3747         if ($diffinfo->{'nparents'}) {
3748                 # combined diff
3749                 $from->{'file'} = [];
3750                 $from->{'href'} = [];
3751                 fill_from_file_info($diffinfo, @parents)
3752                         unless exists $diffinfo->{'from_file'};
3753                 for (my $i = 0; $i < $diffinfo->{'nparents'}; $i++) {
3754                         $from->{'file'}[$i] =
3755                                 defined $diffinfo->{'from_file'}[$i] ?
3756                                         $diffinfo->{'from_file'}[$i] :
3757                                         $diffinfo->{'to_file'};
3758                         if ($diffinfo->{'status'}[$i] ne "A") { # not new (added) file
3759                                 $from->{'href'}[$i] = href(action=>"blob",
3760                                                            hash_base=>$parents[$i],
3761                                                            hash=>$diffinfo->{'from_id'}[$i],
3762                                                            file_name=>$from->{'file'}[$i]);
3763                         } else {
3764                                 $from->{'href'}[$i] = undef;
3765                         }
3766                 }
3767         } else {
3768                 # ordinary (not combined) diff
3769                 $from->{'file'} = $diffinfo->{'from_file'};
3770                 if ($diffinfo->{'status'} ne "A") { # not new (added) file
3771                         $from->{'href'} = href(action=>"blob", hash_base=>$hash_parent,
3772                                                hash=>$diffinfo->{'from_id'},
3773                                                file_name=>$from->{'file'});
3774                 } else {
3775                         delete $from->{'href'};
3776                 }
3777         }
3778
3779         $to->{'file'} = $diffinfo->{'to_file'};
3780         if (!is_deleted($diffinfo)) { # file exists in result
3781                 $to->{'href'} = href(action=>"blob", hash_base=>$hash,
3782                                      hash=>$diffinfo->{'to_id'},
3783                                      file_name=>$to->{'file'});
3784         } else {
3785                 delete $to->{'href'};
3786         }
3787 }
3788
3789 ## ......................................................................
3790 ## parse to array of hashes functions
3791
3792 sub git_get_heads_list {
3793         my ($limit, @classes) = @_;
3794         @classes = get_branch_refs() unless @classes;
3795         my @patterns = map { "refs/$_" } @classes;
3796         my @headslist;
3797
3798         open my $fd, '-|', git_cmd(), 'for-each-ref',
3799                 ($limit ? '--count='.($limit+1) : ()),
3800                 '--sort=-HEAD', '--sort=-committerdate',
3801                 '--format=%(objectname) %(refname) %(subject)%00%(committer)',
3802                 @patterns
3803                 or return;
3804         while (my $line = <$fd>) {
3805                 my %ref_item;
3806
3807                 chomp $line;
3808                 my ($refinfo, $committerinfo) = split(/\0/, $line);
3809                 my ($hash, $name, $title) = split(' ', $refinfo, 3);
3810                 my ($committer, $epoch, $tz) =
3811                         ($committerinfo =~ /^(.*) ([0-9]+) (.*)$/);
3812                 $ref_item{'fullname'}  = $name;
3813                 my $strip_refs = join '|', map { quotemeta } get_branch_refs();
3814                 $name =~ s!^refs/($strip_refs|remotes)/!!;
3815                 $ref_item{'name'} = $name;
3816                 # for refs neither in 'heads' nor 'remotes' we want to
3817                 # show their ref dir
3818                 my $ref_dir = (defined $1) ? $1 : '';
3819                 if ($ref_dir ne '' and $ref_dir ne 'heads' and $ref_dir ne 'remotes') {
3820                     $ref_item{'name'} .= ' (' . $ref_dir . ')';
3821                 }
3822
3823                 $ref_item{'id'}    = $hash;
3824                 $ref_item{'title'} = $title || '(no commit message)';
3825                 $ref_item{'epoch'} = $epoch;
3826                 if ($epoch) {
3827                         $ref_item{'age'} = age_string(time - $ref_item{'epoch'});
3828                 } else {
3829                         $ref_item{'age'} = "unknown";
3830                 }
3831
3832                 push @headslist, \%ref_item;
3833         }
3834         close $fd;
3835
3836         return wantarray ? @headslist : \@headslist;
3837 }
3838
3839 sub git_get_tags_list {
3840         my $limit = shift;
3841         my @tagslist;
3842
3843         open my $fd, '-|', git_cmd(), 'for-each-ref',
3844                 ($limit ? '--count='.($limit+1) : ()), '--sort=-creatordate',
3845                 '--format=%(objectname) %(objecttype) %(refname) '.
3846                 '%(*objectname) %(*objecttype) %(subject)%00%(creator)',
3847                 'refs/tags'
3848                 or return;
3849         while (my $line = <$fd>) {
3850                 my %ref_item;
3851
3852                 chomp $line;
3853                 my ($refinfo, $creatorinfo) = split(/\0/, $line);
3854                 my ($id, $type, $name, $refid, $reftype, $title) = split(' ', $refinfo, 6);
3855                 my ($creator, $epoch, $tz) =
3856                         ($creatorinfo =~ /^(.*) ([0-9]+) (.*)$/);
3857                 $ref_item{'fullname'} = $name;
3858                 $name =~ s!^refs/tags/!!;
3859
3860                 $ref_item{'type'} = $type;
3861                 $ref_item{'id'} = $id;
3862                 $ref_item{'name'} = $name;
3863                 if ($type eq "tag") {
3864                         $ref_item{'subject'} = $title;
3865                         $ref_item{'reftype'} = $reftype;
3866                         $ref_item{'refid'}   = $refid;
3867                 } else {
3868                         $ref_item{'reftype'} = $type;
3869                         $ref_item{'refid'}   = $id;
3870                 }
3871
3872                 if ($type eq "tag" || $type eq "commit") {
3873                         $ref_item{'epoch'} = $epoch;
3874                         if ($epoch) {
3875                                 $ref_item{'age'} = age_string(time - $ref_item{'epoch'});
3876                         } else {
3877                                 $ref_item{'age'} = "unknown";
3878                         }
3879                 }
3880
3881                 push @tagslist, \%ref_item;
3882         }
3883         close $fd;
3884
3885         return wantarray ? @tagslist : \@tagslist;
3886 }
3887
3888 ## ----------------------------------------------------------------------
3889 ## filesystem-related functions
3890
3891 sub get_file_owner {
3892         my $path = shift;
3893
3894         my ($dev, $ino, $mode, $nlink, $st_uid, $st_gid, $rdev, $size) = stat($path);
3895         my ($name, $passwd, $uid, $gid, $quota, $comment, $gcos, $dir, $shell) = getpwuid($st_uid);
3896         if (!defined $gcos) {
3897                 return undef;
3898         }
3899         my $owner = $gcos;
3900         $owner =~ s/[,;].*$//;
3901         return to_utf8($owner);
3902 }
3903
3904 # assume that file exists
3905 sub insert_file {
3906         my $filename = shift;
3907
3908         open my $fd, '<', $filename;
3909         print map { to_utf8($_) } <$fd>;
3910         close $fd;
3911 }
3912
3913 ## ......................................................................
3914 ## mimetype related functions
3915
3916 sub mimetype_guess_file {
3917         my $filename = shift;
3918         my $mimemap = shift;
3919         -r $mimemap or return undef;
3920
3921         my %mimemap;
3922         open(my $mh, '<', $mimemap) or return undef;
3923         while (<$mh>) {
3924                 next if m/^#/; # skip comments
3925                 my ($mimetype, @exts) = split(/\s+/);
3926                 foreach my $ext (@exts) {
3927                         $mimemap{$ext} = $mimetype;
3928                 }
3929         }
3930         close($mh);
3931
3932         $filename =~ /\.([^.]*)$/;
3933         return $mimemap{$1};
3934 }
3935
3936 sub mimetype_guess {
3937         my $filename = shift;
3938         my $mime;
3939         $filename =~ /\./ or return undef;
3940
3941         if ($mimetypes_file) {
3942                 my $file = $mimetypes_file;
3943                 if ($file !~ m!^/!) { # if it is relative path
3944                         # it is relative to project
3945                         $file = "$projectroot/$project/$file";
3946                 }
3947                 $mime = mimetype_guess_file($filename, $file);
3948         }
3949         $mime ||= mimetype_guess_file($filename, '/etc/mime.types');
3950         return $mime;
3951 }
3952
3953 sub blob_mimetype {
3954         my $fd = shift;
3955         my $filename = shift;
3956
3957         if ($filename) {
3958                 my $mime = mimetype_guess($filename);
3959                 $mime and return $mime;
3960         }
3961
3962         # just in case
3963         return $default_blob_plain_mimetype unless $fd;
3964
3965         if (-T $fd) {
3966                 return 'text/plain';
3967         } elsif (! $filename) {
3968                 return 'application/octet-stream';
3969         } elsif ($filename =~ m/\.png$/i) {
3970                 return 'image/png';
3971         } elsif ($filename =~ m/\.gif$/i) {
3972                 return 'image/gif';
3973         } elsif ($filename =~ m/\.jpe?g$/i) {
3974                 return 'image/jpeg';
3975         } else {
3976                 return 'application/octet-stream';
3977         }
3978 }
3979
3980 sub blob_contenttype {
3981         my ($fd, $file_name, $type) = @_;
3982
3983         $type ||= blob_mimetype($fd, $file_name);
3984         if ($type eq 'text/plain' && defined $default_text_plain_charset) {
3985                 $type .= "; charset=$default_text_plain_charset";
3986         }
3987
3988         return $type;
3989 }
3990
3991 # guess file syntax for syntax highlighting; return undef if no highlighting
3992 # the name of syntax can (in the future) depend on syntax highlighter used
3993 sub guess_file_syntax {
3994         my ($highlight, $file_name) = @_;
3995         return undef unless ($highlight && defined $file_name);
3996         my $basename = basename($file_name, '.in');
3997         return $highlight_basename{$basename}
3998                 if exists $highlight_basename{$basename};
3999
4000         $basename =~ /\.([^.]*)$/;
4001         my $ext = $1 or return undef;
4002         return $highlight_ext{$ext}
4003                 if exists $highlight_ext{$ext};
4004
4005         return undef;
4006 }
4007
4008 # run highlighter and return FD of its output,
4009 # or return original FD if no highlighting
4010 sub run_highlighter {
4011         my ($fd, $highlight, $syntax) = @_;
4012         return $fd unless ($highlight);
4013
4014         close $fd;
4015         my $syntax_arg = (defined $syntax) ? "--syntax $syntax" : "--force";
4016         open $fd, quote_command(git_cmd(), "cat-file", "blob", $hash)." | ".
4017                   quote_command($^X, '-CO', '-MEncode=decode,FB_DEFAULT', '-pse',
4018                     '$_ = decode($fe, $_, FB_DEFAULT) if !utf8::decode($_);',
4019                     '--', "-fe=$fallback_encoding")." | ".
4020                   quote_command($highlight_bin).
4021                   " --replace-tabs=8 --fragment $syntax_arg |"
4022                 or die_error(500, "Couldn't open file or run syntax highlighter");
4023         return $fd;
4024 }
4025
4026 ## ======================================================================
4027 ## functions printing HTML: header, footer, error page
4028
4029 sub get_page_title {
4030         my $title = to_utf8($site_name);
4031
4032         unless (defined $project) {
4033                 if (defined $project_filter) {
4034                         $title .= " - projects in '" . esc_path($project_filter) . "'";
4035                 }
4036                 return $title;
4037         }
4038         $title .= " - " . to_utf8($project);
4039
4040         return $title unless (defined $action);
4041         $title .= "/$action"; # $action is US-ASCII (7bit ASCII)
4042
4043         return $title unless (defined $file_name);
4044         $title .= " - " . esc_path($file_name);
4045         if ($action eq "tree" && $file_name !~ m|/$|) {
4046                 $title .= "/";
4047         }
4048
4049         return $title;
4050 }
4051
4052 sub get_content_type_html {
4053         # require explicit support from the UA if we are to send the page as
4054         # 'application/xhtml+xml', otherwise send it as plain old 'text/html'.
4055         # we have to do this because MSIE sometimes globs '*/*', pretending to
4056         # support xhtml+xml but choking when it gets what it asked for.
4057         if (defined $cgi->http('HTTP_ACCEPT') &&
4058             $cgi->http('HTTP_ACCEPT') =~ m/(,|;|\s|^)application\/xhtml\+xml(,|;|\s|$)/ &&
4059             $cgi->Accept('application/xhtml+xml') != 0) {
4060                 return 'application/xhtml+xml';
4061         } else {
4062                 return 'text/html';
4063         }
4064 }
4065
4066 sub print_feed_meta {
4067         if (defined $project) {
4068                 my %href_params = get_feed_info();
4069                 if (!exists $href_params{'-title'}) {
4070                         $href_params{'-title'} = 'log';
4071                 }
4072
4073                 foreach my $format (qw(RSS Atom)) {
4074                         my $type = lc($format);
4075                         my %link_attr = (
4076                                 '-rel' => 'alternate',
4077                                 '-title' => esc_attr("$project - $href_params{'-title'} - $format feed"),
4078                                 '-type' => "application/$type+xml"
4079                         );
4080
4081                         $href_params{'extra_options'} = undef;
4082                         $href_params{'action'} = $type;
4083                         $link_attr{'-href'} = esc_attr(href(%href_params));
4084                         print "<link ".
4085                               "rel=\"$link_attr{'-rel'}\" ".
4086                               "title=\"$link_attr{'-title'}\" ".
4087                               "href=\"$link_attr{'-href'}\" ".
4088                               "type=\"$link_attr{'-type'}\" ".
4089                               "/>\n";
4090
4091                         $href_params{'extra_options'} = '--no-merges';
4092                         $link_attr{'-href'} = esc_attr(href(%href_params));
4093                         $link_attr{'-title'} .= ' (no merges)';
4094                         print "<link ".
4095                               "rel=\"$link_attr{'-rel'}\" ".
4096                               "title=\"$link_attr{'-title'}\" ".
4097                               "href=\"$link_attr{'-href'}\" ".
4098                               "type=\"$link_attr{'-type'}\" ".
4099                               "/>\n";
4100                 }
4101
4102         } else {
4103                 printf('<link rel="alternate" title="%s projects list" '.
4104                        'href="%s" type="text/plain; charset=utf-8" />'."\n",
4105                        esc_attr($site_name),
4106                        esc_attr(href(project=>undef, action=>"project_index")));
4107                 printf('<link rel="alternate" title="%s projects feeds" '.
4108                        'href="%s" type="text/x-opml" />'."\n",
4109                        esc_attr($site_name),
4110                        esc_attr(href(project=>undef, action=>"opml")));
4111         }
4112 }
4113
4114 sub print_header_links {
4115         my $status = shift;
4116
4117         # print out each stylesheet that exist, providing backwards capability
4118         # for those people who defined $stylesheet in a config file
4119         if (defined $stylesheet) {
4120                 print '<link rel="stylesheet" type="text/css" href="'.esc_url($stylesheet).'"/>'."\n";
4121         } else {
4122                 foreach my $stylesheet (@stylesheets) {
4123                         next unless $stylesheet;
4124                         print '<link rel="stylesheet" type="text/css" href="'.esc_url($stylesheet).'"/>'."\n";
4125                 }
4126         }
4127         print_feed_meta()
4128                 if ($status eq '200 OK');
4129         if (defined $favicon) {
4130                 print qq(<link rel="shortcut icon" href=").esc_url($favicon).qq(" type="image/png" />\n);
4131         }
4132 }
4133
4134 sub print_nav_breadcrumbs_path {
4135         my $dirprefix = undef;
4136         while (my $part = shift) {
4137                 $dirprefix .= "/" if defined $dirprefix;
4138                 $dirprefix .= $part;
4139                 print $cgi->a({-href => href(project => undef,
4140                                              project_filter => $dirprefix,
4141                                              action => "project_list")},
4142                               esc_html($part)) . " / ";
4143         }
4144 }
4145
4146 sub print_nav_breadcrumbs {
4147         my %opts = @_;
4148
4149         for my $crumb (@extra_breadcrumbs, [ $home_link_str => $home_link ]) {
4150                 print $cgi->a({-href => esc_url($crumb->[1])}, $crumb->[0]) . " / ";
4151         }
4152         if (defined $project) {
4153                 my @dirname = split '/', $project;
4154                 my $projectbasename = pop @dirname;
4155                 print_nav_breadcrumbs_path(@dirname);
4156                 print $cgi->a({-href => href(action=>"summary")}, esc_html($projectbasename));
4157                 if (defined $action) {
4158                         my $action_print = $action ;
4159                         if (defined $opts{-action_extra}) {
4160                                 $action_print = $cgi->a({-href => href(action=>$action)},
4161                                         $action);
4162                         }
4163                         print " / $action_print";
4164                 }
4165                 if (defined $opts{-action_extra}) {
4166                         print " / $opts{-action_extra}";
4167                 }
4168                 print "\n";
4169         } elsif (defined $project_filter) {
4170                 print_nav_breadcrumbs_path(split '/', $project_filter);
4171         }
4172 }
4173
4174 sub print_search_form {
4175         if (!defined $searchtext) {
4176                 $searchtext = "";
4177         }
4178         my $search_hash;
4179         if (defined $hash_base) {
4180                 $search_hash = $hash_base;
4181         } elsif (defined $hash) {
4182                 $search_hash = $hash;
4183         } else {
4184                 $search_hash = "HEAD";
4185         }
4186         my $action = $my_uri;
4187         my $use_pathinfo = gitweb_check_feature('pathinfo');
4188         if ($use_pathinfo) {
4189                 $action .= "/".esc_url($project);
4190         }
4191         print $cgi->start_form(-method => "get", -action => $action) .
4192               "<div class=\"search\">\n" .
4193               (!$use_pathinfo &&
4194               $cgi->input({-name=>"p", -value=>$project, -type=>"hidden"}) . "\n") .
4195               $cgi->input({-name=>"a", -value=>"search", -type=>"hidden"}) . "\n" .
4196               $cgi->input({-name=>"h", -value=>$search_hash, -type=>"hidden"}) . "\n" .
4197               $cgi->popup_menu(-name => 'st', -default => 'commit',
4198                                -values => ['commit', 'grep', 'author', 'committer', 'pickaxe']) .
4199               " " . $cgi->a({-href => href(action=>"search_help"),
4200                              -title => "search help" }, "?") . " search:\n",
4201               $cgi->textfield(-name => "s", -value => $searchtext, -override => 1) . "\n" .
4202               "<span title=\"Extended regular expression\">" .
4203               $cgi->checkbox(-name => 'sr', -value => 1, -label => 're',
4204                              -checked => $search_use_regexp) .
4205               "</span>" .
4206               "</div>" .
4207               $cgi->end_form() . "\n";
4208 }
4209
4210 sub git_header_html {
4211         my $status = shift || "200 OK";
4212         my $expires = shift;
4213         my %opts = @_;
4214
4215         my $title = get_page_title();
4216         print $cgi->header(-type=>get_content_type_html(), -charset => 'utf-8',
4217                            -status=> $status, -expires => $expires)
4218                 unless ($opts{'-no_http_header'});
4219         my $mod_perl_version = $ENV{'MOD_PERL'} ? " $ENV{'MOD_PERL'}" : '';
4220         print <<EOF;
4221 <?xml version="1.0" encoding="utf-8"?>
4222 <!DOCTYPE html [
4223         <!ENTITY nbsp "&#xA0;">
4224         <!ENTITY sdot "&#x22C5;">
4225 ]>
4226 <html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en-US" lang="en-US">
4227 <!-- git web interface version $version, (C) 2005-2006, Kay Sievers <kay.sievers\@vrfy.org>, Christian Gierke -->
4228 <!-- git core binaries version $git_version -->
4229 <head>
4230 <meta name="generator" content="gitweb/$version git/$git_version$mod_perl_version"/>
4231 <meta name="robots" content="index, nofollow"/>
4232 <title>$title</title>
4233 EOF
4234         # the stylesheet, favicon etc urls won't work correctly with path_info
4235         # unless we set the appropriate base URL
4236         if ($ENV{'PATH_INFO'}) {
4237                 print "<base href=\"".esc_url($base_url)."\" />\n";
4238         }
4239         print_header_links($status);
4240
4241         if (defined $site_html_head_string) {
4242                 print to_utf8($site_html_head_string);
4243         }
4244
4245         print "</head>\n" .
4246               "<body>\n";
4247
4248         if (defined $site_header && -f $site_header) {
4249                 insert_file($site_header);
4250         }
4251
4252         print "<div class=\"page_header\">\n";
4253         if (defined $logo) {
4254                 print $cgi->a({-href => esc_url($logo_url),
4255                                -title => $logo_label},
4256                               $cgi->img({-src => esc_url($logo),
4257                                          -width => 72, -height => 27,
4258                                          -alt => "git",
4259                                          -class => "logo"}));
4260         }
4261         print_nav_breadcrumbs(%opts);
4262         print "</div>\n";
4263
4264         my $have_search = gitweb_check_feature('search');
4265         if (defined $project && $have_search) {
4266                 print_search_form();
4267         }
4268 }
4269
4270 sub git_footer_html {
4271         my $feed_class = 'rss_logo';
4272
4273         print "<div class=\"page_footer\">\n";
4274         if (defined $project) {
4275                 my $descr = git_get_project_description($project);
4276                 if (defined $descr) {
4277                         print "<div class=\"page_footer_text\">" . esc_html($descr) . "</div>\n";
4278                 }
4279
4280                 my %href_params = get_feed_info();
4281                 if (!%href_params) {
4282                         $feed_class .= ' generic';
4283                 }
4284                 $href_params{'-title'} ||= 'log';
4285
4286                 foreach my $format (qw(RSS Atom)) {
4287                         $href_params{'action'} = lc($format);
4288                         print $cgi->a({-href => href(%href_params),
4289                                       -title => "$href_params{'-title'} $format feed",
4290                                       -class => $feed_class}, $format)."\n";
4291                 }
4292
4293         } else {
4294                 print $cgi->a({-href => href(project=>undef, action=>"opml",
4295                                              project_filter => $project_filter),
4296                               -class => $feed_class}, "OPML") . " ";
4297                 print $cgi->a({-href => href(project=>undef, action=>"project_index",
4298                                              project_filter => $project_filter),
4299                               -class => $feed_class}, "TXT") . "\n";
4300         }
4301         print "</div>\n"; # class="page_footer"
4302
4303         if (defined $t0 && gitweb_check_feature('timed')) {
4304                 print "<div id=\"generating_info\">\n";
4305                 print 'This page took '.
4306                       '<span id="generating_time" class="time_span">'.
4307                       tv_interval($t0, [ gettimeofday() ]).
4308                       ' seconds </span>'.
4309                       ' and '.
4310                       '<span id="generating_cmd">'.
4311                       $number_of_git_cmds.
4312                       '</span> git commands '.
4313                       " to generate.\n";
4314                 print "</div>\n"; # class="page_footer"
4315         }
4316
4317         if (defined $site_footer && -f $site_footer) {
4318                 insert_file($site_footer);
4319         }
4320
4321         print qq!<script type="text/javascript" src="!.esc_url($javascript).qq!"></script>\n!;
4322         if (defined $action &&
4323             $action eq 'blame_incremental') {
4324                 print qq!<script type="text/javascript">\n!.
4325                       qq!startBlame("!. esc_attr(href(action=>"blame_data", -replay=>1)) .qq!",\n!.
4326                       qq!           "!. esc_attr(href()) .qq!");\n!.
4327                       qq!</script>\n!;
4328         } else {
4329                 my ($jstimezone, $tz_cookie, $datetime_class) =
4330                         gitweb_get_feature('javascript-timezone');
4331
4332                 print qq!<script type="text/javascript">\n!.
4333                       qq!window.onload = function () {\n!;
4334                 if (gitweb_check_feature('javascript-actions')) {
4335                         print qq!       fixLinks();\n!;
4336                 }
4337                 if ($jstimezone && $tz_cookie && $datetime_class) {
4338                         print qq!       var tz_cookie = { name: '$tz_cookie', expires: 14, path: '/' };\n!. # in days
4339                               qq!       onloadTZSetup('$jstimezone', tz_cookie, '$datetime_class');\n!;
4340                 }
4341                 print qq!};\n!.
4342                       qq!</script>\n!;
4343         }
4344
4345         print "</body>\n" .
4346               "</html>";
4347 }
4348
4349 # die_error(<http_status_code>, <error_message>[, <detailed_html_description>])
4350 # Example: die_error(404, 'Hash not found')
4351 # By convention, use the following status codes (as defined in RFC 2616):
4352 # 400: Invalid or missing CGI parameters, or
4353 #      requested object exists but has wrong type.
4354 # 403: Requested feature (like "pickaxe" or "snapshot") not enabled on
4355 #      this server or project.
4356 # 404: Requested object/revision/project doesn't exist.
4357 # 500: The server isn't configured properly, or
4358 #      an internal error occurred (e.g. failed assertions caused by bugs), or
4359 #      an unknown error occurred (e.g. the git binary died unexpectedly).
4360 # 503: The server is currently unavailable (because it is overloaded,
4361 #      or down for maintenance).  Generally, this is a temporary state.
4362 sub die_error {
4363         my $status = shift || 500;
4364         my $error = esc_html(shift) || "Internal Server Error";
4365         my $extra = shift;
4366         my %opts = @_;
4367
4368         my %http_responses = (
4369                 400 => '400 Bad Request',
4370                 403 => '403 Forbidden',
4371                 404 => '404 Not Found',
4372                 500 => '500 Internal Server Error',
4373                 503 => '503 Service Unavailable',
4374         );
4375         git_header_html($http_responses{$status}, undef, %opts);
4376         print <<EOF;
4377 <div class="page_body">
4378 <br /><br />
4379 $status - $error
4380 <br />
4381 EOF
4382         if (defined $extra) {
4383                 print "<hr />\n" .
4384                       "$extra\n";
4385         }
4386         print "</div>\n";
4387
4388         git_footer_html();
4389         goto DONE_GITWEB
4390                 unless ($opts{'-error_handler'});
4391 }
4392
4393 ## ----------------------------------------------------------------------
4394 ## functions printing or outputting HTML: navigation
4395
4396 sub git_print_page_nav {
4397         my ($current, $suppress, $head, $treehead, $treebase, $extra) = @_;
4398         $extra = '' if !defined $extra; # pager or formats
4399
4400         my @navs = qw(summary shortlog log commit commitdiff tree);
4401         if ($suppress) {
4402                 @navs = grep { $_ ne $suppress } @navs;
4403         }
4404
4405         my %arg = map { $_ => {action=>$_} } @navs;
4406         if (defined $head) {
4407                 for (qw(commit commitdiff)) {
4408                         $arg{$_}{'hash'} = $head;
4409                 }
4410                 if ($current =~ m/^(tree | log | shortlog | commit | commitdiff | search)$/x) {
4411                         for (qw(shortlog log)) {
4412                                 $arg{$_}{'hash'} = $head;
4413                         }
4414                 }
4415         }
4416
4417         $arg{'tree'}{'hash'} = $treehead if defined $treehead;
4418         $arg{'tree'}{'hash_base'} = $treebase if defined $treebase;
4419
4420         my @actions = gitweb_get_feature('actions');
4421         my %repl = (
4422                 '%' => '%',
4423                 'n' => $project,         # project name
4424                 'f' => $git_dir,         # project path within filesystem
4425                 'h' => $treehead || '',  # current hash ('h' parameter)
4426                 'b' => $treebase || '',  # hash base ('hb' parameter)
4427         );
4428         while (@actions) {
4429                 my ($label, $link, $pos) = splice(@actions,0,3);
4430                 # insert
4431                 @navs = map { $_ eq $pos ? ($_, $label) : $_ } @navs;
4432                 # munch munch
4433                 $link =~ s/%([%nfhb])/$repl{$1}/g;
4434                 $arg{$label}{'_href'} = $link;
4435         }
4436
4437         print "<div class=\"page_nav\">\n" .
4438                 (join " | ",
4439                  map { $_ eq $current ?
4440                        $_ : $cgi->a({-href => ($arg{$_}{_href} ? $arg{$_}{_href} : href(%{$arg{$_}}))}, "$_")
4441                  } @navs);
4442         print "<br/>\n$extra<br/>\n" .
4443               "</div>\n";
4444 }
4445
4446 # returns a submenu for the navigation of the refs views (tags, heads,
4447 # remotes) with the current view disabled and the remotes view only
4448 # available if the feature is enabled
4449 sub format_ref_views {
4450         my ($current) = @_;
4451         my @ref_views = qw{tags heads};
4452         push @ref_views, 'remotes' if gitweb_check_feature('remote_heads');
4453         return join " | ", map {
4454                 $_ eq $current ? $_ :
4455                 $cgi->a({-href => href(action=>$_)}, $_)
4456         } @ref_views
4457 }
4458
4459 sub format_paging_nav {
4460         my ($action, $page, $has_next_link) = @_;
4461         my $paging_nav;
4462
4463
4464         if ($page > 0) {
4465                 $paging_nav .=
4466                         $cgi->a({-href => href(-replay=>1, page=>undef)}, "first") .
4467                         " &sdot; " .
4468                         $cgi->a({-href => href(-replay=>1, page=>$page-1),
4469                                  -accesskey => "p", -title => "Alt-p"}, "prev");
4470         } else {
4471                 $paging_nav .= "first &sdot; prev";
4472         }
4473
4474         if ($has_next_link) {
4475                 $paging_nav .= " &sdot; " .
4476                         $cgi->a({-href => href(-replay=>1, page=>$page+1),
4477                                  -accesskey => "n", -title => "Alt-n"}, "next");
4478         } else {
4479                 $paging_nav .= " &sdot; next";
4480         }
4481
4482         return $paging_nav;
4483 }
4484
4485 ## ......................................................................
4486 ## functions printing or outputting HTML: div
4487
4488 sub git_print_header_div {
4489         my ($action, $title, $hash, $hash_base) = @_;
4490         my %args = ();
4491
4492         $args{'action'} = $action;
4493         $args{'hash'} = $hash if $hash;
4494         $args{'hash_base'} = $hash_base if $hash_base;
4495
4496         print "<div class=\"header\">\n" .
4497               $cgi->a({-href => href(%args), -class => "title"},
4498               $title ? $title : $action) .
4499               "\n</div>\n";
4500 }
4501
4502 sub format_repo_url {
4503         my ($name, $url) = @_;
4504         return "<tr class=\"metadata_url\"><td>$name</td><td>$url</td></tr>\n";
4505 }
4506
4507 # Group output by placing it in a DIV element and adding a header.
4508 # Options for start_div() can be provided by passing a hash reference as the
4509 # first parameter to the function.
4510 # Options to git_print_header_div() can be provided by passing an array
4511 # reference. This must follow the options to start_div if they are present.
4512 # The content can be a scalar, which is output as-is, a scalar reference, which
4513 # is output after html escaping, an IO handle passed either as *handle or
4514 # *handle{IO}, or a function reference. In the latter case all following
4515 # parameters will be taken as argument to the content function call.
4516 sub git_print_section {
4517         my ($div_args, $header_args, $content);
4518         my $arg = shift;
4519         if (ref($arg) eq 'HASH') {
4520                 $div_args = $arg;
4521                 $arg = shift;
4522         }
4523         if (ref($arg) eq 'ARRAY') {
4524                 $header_args = $arg;
4525                 $arg = shift;
4526         }
4527         $content = $arg;
4528
4529         print $cgi->start_div($div_args);
4530         git_print_header_div(@$header_args);
4531
4532         if (ref($content) eq 'CODE') {
4533                 $content->(@_);
4534         } elsif (ref($content) eq 'SCALAR') {
4535                 print esc_html($$content);
4536         } elsif (ref($content) eq 'GLOB' or ref($content) eq 'IO::Handle') {
4537                 print <$content>;
4538         } elsif (!ref($content) && defined($content)) {
4539                 print $content;
4540         }
4541
4542         print $cgi->end_div;
4543 }
4544
4545 sub format_timestamp_html {
4546         my $date = shift;
4547         my $strtime = $date->{'rfc2822'};
4548
4549         my (undef, undef, $datetime_class) =
4550                 gitweb_get_feature('javascript-timezone');
4551         if ($datetime_class) {
4552                 $strtime = qq!<span class="$datetime_class">$strtime</span>!;
4553         }
4554
4555         my $localtime_format = '(%02d:%02d %s)';
4556         if ($date->{'hour_local'} < 6) {
4557                 $localtime_format = '(<span class="atnight">%02d:%02d</span> %s)';
4558         }
4559         $strtime .= ' ' .
4560                     sprintf($localtime_format,
4561                             $date->{'hour_local'}, $date->{'minute_local'}, $date->{'tz_local'});
4562
4563         return $strtime;
4564 }
4565
4566 # Outputs the author name and date in long form
4567 sub git_print_authorship {
4568         my $co = shift;
4569         my %opts = @_;
4570         my $tag = $opts{-tag} || 'div';
4571         my $author = $co->{'author_name'};
4572
4573         my %ad = parse_date($co->{'author_epoch'}, $co->{'author_tz'});
4574         print "<$tag class=\"author_date\">" .
4575               format_search_author($author, "author", esc_html($author)) .
4576               " [".format_timestamp_html(\%ad)."]".
4577               git_get_avatar($co->{'author_email'}, -pad_before => 1) .
4578               "</$tag>\n";
4579 }
4580
4581 # Outputs table rows containing the full author or committer information,
4582 # in the format expected for 'commit' view (& similar).
4583 # Parameters are a commit hash reference, followed by the list of people
4584 # to output information for. If the list is empty it defaults to both
4585 # author and committer.
4586 sub git_print_authorship_rows {
4587         my $co = shift;
4588         # too bad we can't use @people = @_ || ('author', 'committer')
4589         my @people = @_;
4590         @people = ('author', 'committer') unless @people;
4591         foreach my $who (@people) {
4592                 my %wd = parse_date($co->{"${who}_epoch"}, $co->{"${who}_tz"});
4593                 print "<tr><td>$who</td><td>" .
4594                       format_search_author($co->{"${who}_name"}, $who,
4595                                            esc_html($co->{"${who}_name"})) . " " .
4596                       format_search_author($co->{"${who}_email"}, $who,
4597                                            esc_html("<" . $co->{"${who}_email"} . ">")) .
4598                       "</td><td rowspan=\"2\">" .
4599                       git_get_avatar($co->{"${who}_email"}, -size => 'double') .
4600                       "</td></tr>\n" .
4601                       "<tr>" .
4602                       "<td></td><td>" .
4603                       format_timestamp_html(\%wd) .
4604                       "</td>" .
4605                       "</tr>\n";
4606         }
4607 }
4608
4609 sub git_print_page_path {
4610         my $name = shift;
4611         my $type = shift;
4612         my $hb = shift;
4613
4614
4615         print "<div class=\"page_path\">";
4616         print $cgi->a({-href => href(action=>"tree", hash_base=>$hb),
4617                       -title => 'tree root'}, to_utf8("[$project]"));
4618         print " / ";
4619         if (defined $name) {
4620                 my @dirname = split '/', $name;
4621                 my $basename = pop @dirname;
4622                 my $fullname = '';
4623
4624                 foreach my $dir (@dirname) {
4625                         $fullname .= ($fullname ? '/' : '') . $dir;
4626                         print $cgi->a({-href => href(action=>"tree", file_name=>$fullname,
4627                                                      hash_base=>$hb),
4628                                       -title => $fullname}, esc_path($dir));
4629                         print " / ";
4630                 }
4631                 if (defined $type && $type eq 'blob') {
4632                         print $cgi->a({-href => href(action=>"blob_plain", file_name=>$file_name,
4633                                                      hash_base=>$hb),
4634                                       -title => $name}, esc_path($basename));
4635                 } elsif (defined $type && $type eq 'tree') {
4636                         print $cgi->a({-href => href(action=>"tree", file_name=>$file_name,
4637                                                      hash_base=>$hb),
4638                                       -title => $name}, esc_path($basename));
4639                         print " / ";
4640                 } else {
4641                         print esc_path($basename);
4642                 }
4643         }
4644         print "<br/></div>\n";
4645 }
4646
4647 sub git_print_log {
4648         my $log = shift;
4649         my %opts = @_;
4650
4651         if ($opts{'-remove_title'}) {
4652                 # remove title, i.e. first line of log
4653                 shift @$log;
4654         }
4655         # remove leading empty lines
4656         while (defined $log->[0] && $log->[0] eq "") {
4657                 shift @$log;
4658         }
4659
4660         # print log
4661         my $skip_blank_line = 0;
4662         foreach my $line (@$log) {
4663                 if ($line =~ m/^\s*([A-Z][-A-Za-z]*-([Bb]y|[Tt]o)|C[Cc]|(Clos|Fix)es): /) {
4664                         if (! $opts{'-remove_signoff'}) {
4665                                 print "<span class=\"signoff\">" . esc_html($line) . "</span><br/>\n";
4666                                 $skip_blank_line = 1;
4667                         }
4668                         next;
4669                 }
4670
4671                 if ($line =~ m,\s*([a-z]*link): (https?://\S+),i) {
4672                         if (! $opts{'-remove_signoff'}) {
4673                                 print "<span class=\"signoff\">" . esc_html($1) . ": " .
4674                                         "<a href=\"" . esc_html($2) . "\">" . esc_html($2) . "</a>" .
4675                                         "</span><br/>\n";
4676                                 $skip_blank_line = 1;
4677                         }
4678                         next;
4679                 }
4680
4681                 # print only one empty line
4682                 # do not print empty line after signoff
4683                 if ($line eq "") {
4684                         next if ($skip_blank_line);
4685                         $skip_blank_line = 1;
4686                 } else {
4687                         $skip_blank_line = 0;
4688                 }
4689
4690                 print format_log_line_html($line) . "<br/>\n";
4691         }
4692
4693         if ($opts{'-final_empty_line'}) {
4694                 # end with single empty line
4695                 print "<br/>\n" unless $skip_blank_line;
4696         }
4697 }
4698
4699 # return link target (what link points to)
4700 sub git_get_link_target {
4701         my $hash = shift;
4702         my $link_target;
4703
4704         # read link
4705         open my $fd, "-|", git_cmd(), "cat-file", "blob", $hash
4706                 or return;
4707         {
4708                 local $/ = undef;
4709                 $link_target = <$fd>;
4710         }
4711         close $fd
4712                 or return;
4713
4714         return $link_target;
4715 }
4716
4717 # given link target, and the directory (basedir) the link is in,
4718 # return target of link relative to top directory (top tree);
4719 # return undef if it is not possible (including absolute links).
4720 sub normalize_link_target {
4721         my ($link_target, $basedir) = @_;
4722
4723         # absolute symlinks (beginning with '/') cannot be normalized
4724         return if (substr($link_target, 0, 1) eq '/');
4725
4726         # normalize link target to path from top (root) tree (dir)
4727         my $path;
4728         if ($basedir) {
4729                 $path = $basedir . '/' . $link_target;
4730         } else {
4731                 # we are in top (root) tree (dir)
4732                 $path = $link_target;
4733         }
4734
4735         # remove //, /./, and /../
4736         my @path_parts;
4737         foreach my $part (split('/', $path)) {
4738                 # discard '.' and ''
4739                 next if (!$part || $part eq '.');
4740                 # handle '..'
4741                 if ($part eq '..') {
4742                         if (@path_parts) {
4743                                 pop @path_parts;
4744                         } else {
4745                                 # link leads outside repository (outside top dir)
4746                                 return;
4747                         }
4748                 } else {
4749                         push @path_parts, $part;
4750                 }
4751         }
4752         $path = join('/', @path_parts);
4753
4754         return $path;
4755 }
4756
4757 # print tree entry (row of git_tree), but without encompassing <tr> element
4758 sub git_print_tree_entry {
4759         my ($t, $basedir, $hash_base, $have_blame) = @_;
4760
4761         my %base_key = ();
4762         $base_key{'hash_base'} = $hash_base if defined $hash_base;
4763
4764         # The format of a table row is: mode list link.  Where mode is
4765         # the mode of the entry, list is the name of the entry, an href,
4766         # and link is the action links of the entry.
4767
4768         print "<td class=\"mode\">" . mode_str($t->{'mode'}) . "</td>\n";
4769         if (exists $t->{'size'}) {
4770                 print "<td class=\"size\">$t->{'size'}</td>\n";
4771         }
4772         if ($t->{'type'} eq "blob") {
4773                 print "<td class=\"list\">" .
4774                         $cgi->a({-href => href(action=>"blob", hash=>$t->{'hash'},
4775                                                file_name=>"$basedir$t->{'name'}", %base_key),
4776                                 -class => "list"}, esc_path($t->{'name'}));
4777                 if (S_ISLNK(oct $t->{'mode'})) {
4778                         my $link_target = git_get_link_target($t->{'hash'});
4779                         if ($link_target) {
4780                                 my $norm_target = normalize_link_target($link_target, $basedir);
4781                                 if (defined $norm_target) {
4782                                         print " -> " .
4783                                               $cgi->a({-href => href(action=>"object", hash_base=>$hash_base,
4784                                                                      file_name=>$norm_target),
4785                                                        -title => $norm_target}, esc_path($link_target));
4786                                 } else {
4787                                         print " -> " . esc_path($link_target);
4788                                 }
4789                         }
4790                 }
4791                 print "</td>\n";
4792                 print "<td class=\"link\">";
4793                 print $cgi->a({-href => href(action=>"blob", hash=>$t->{'hash'},
4794                                              file_name=>"$basedir$t->{'name'}", %base_key)},
4795                               "blob");
4796                 if ($have_blame) {
4797                         print " | " .
4798                               $cgi->a({-href => href(action=>"blame", hash=>$t->{'hash'},
4799                                                      file_name=>"$basedir$t->{'name'}", %base_key)},
4800                                       "blame");
4801                 }
4802                 if (defined $hash_base) {
4803                         print " | " .
4804                               $cgi->a({-href => href(action=>"history", hash_base=>$hash_base,
4805                                                      hash=>$t->{'hash'}, file_name=>"$basedir$t->{'name'}")},
4806                                       "history");
4807                 }
4808                 print " | " .
4809                         $cgi->a({-href => href(action=>"blob_plain", hash_base=>$hash_base,
4810                                                file_name=>"$basedir$t->{'name'}")},
4811                                 "raw");
4812                 print "</td>\n";
4813
4814         } elsif ($t->{'type'} eq "tree") {
4815                 print "<td class=\"list\">";
4816                 print $cgi->a({-href => href(action=>"tree", hash=>$t->{'hash'},
4817                                              file_name=>"$basedir$t->{'name'}",
4818                                              %base_key)},
4819                               esc_path($t->{'name'}));
4820                 print "</td>\n";
4821                 print "<td class=\"link\">";
4822                 print $cgi->a({-href => href(action=>"tree", hash=>$t->{'hash'},
4823                                              file_name=>"$basedir$t->{'name'}",
4824                                              %base_key)},
4825                               "tree");
4826                 if (defined $hash_base) {
4827                         print " | " .
4828                               $cgi->a({-href => href(action=>"history", hash_base=>$hash_base,
4829                                                      file_name=>"$basedir$t->{'name'}")},
4830                                       "history");
4831                 }
4832                 print "</td>\n";
4833         } else {
4834                 # unknown object: we can only present history for it
4835                 # (this includes 'commit' object, i.e. submodule support)
4836                 print "<td class=\"list\">" .
4837                       esc_path($t->{'name'}) .
4838                       "</td>\n";
4839                 print "<td class=\"link\">";
4840                 if (defined $hash_base) {
4841                         print $cgi->a({-href => href(action=>"history",
4842                                                      hash_base=>$hash_base,
4843                                                      file_name=>"$basedir$t->{'name'}")},
4844                                       "history");
4845                 }
4846                 print "</td>\n";
4847         }
4848 }
4849
4850 ## ......................................................................
4851 ## functions printing large fragments of HTML
4852
4853 # get pre-image filenames for merge (combined) diff
4854 sub fill_from_file_info {
4855         my ($diff, @parents) = @_;
4856
4857         $diff->{'from_file'} = [ ];
4858         $diff->{'from_file'}[$diff->{'nparents'} - 1] = undef;
4859         for (my $i = 0; $i < $diff->{'nparents'}; $i++) {
4860                 if ($diff->{'status'}[$i] eq 'R' ||
4861                     $diff->{'status'}[$i] eq 'C') {
4862                         $diff->{'from_file'}[$i] =
4863                                 git_get_path_by_hash($parents[$i], $diff->{'from_id'}[$i]);
4864                 }
4865         }
4866
4867         return $diff;
4868 }
4869
4870 # is current raw difftree line of file deletion
4871 sub is_deleted {
4872         my $diffinfo = shift;
4873
4874         return $diffinfo->{'to_id'} eq ('0' x 40) || $diffinfo->{'to_id'} eq ('0' x 64);
4875 }
4876
4877 # does patch correspond to [previous] difftree raw line
4878 # $diffinfo  - hashref of parsed raw diff format
4879 # $patchinfo - hashref of parsed patch diff format
4880 #              (the same keys as in $diffinfo)
4881 sub is_patch_split {
4882         my ($diffinfo, $patchinfo) = @_;
4883
4884         return defined $diffinfo && defined $patchinfo
4885                 && $diffinfo->{'to_file'} eq $patchinfo->{'to_file'};
4886 }
4887
4888
4889 sub git_difftree_body {
4890         my ($difftree, $hash, @parents) = @_;
4891         my ($parent) = $parents[0];
4892         my $have_blame = gitweb_check_feature('blame');
4893         print "<div class=\"list_head\">\n";
4894         if ($#{$difftree} > 10) {
4895                 print(($#{$difftree} + 1) . " files changed:\n");
4896         }
4897         print "</div>\n";
4898
4899         print "<table class=\"" .
4900               (@parents > 1 ? "combined " : "") .
4901               "diff_tree\">\n";
4902
4903         # header only for combined diff in 'commitdiff' view
4904         my $has_header = @$difftree && @parents > 1 && $action eq 'commitdiff';
4905         if ($has_header) {
4906                 # table header
4907                 print "<thead><tr>\n" .
4908                        "<th></th><th></th>\n"; # filename, patchN link
4909                 for (my $i = 0; $i < @parents; $i++) {
4910                         my $par = $parents[$i];
4911                         print "<th>" .
4912                               $cgi->a({-href => href(action=>"commitdiff",
4913                                                      hash=>$hash, hash_parent=>$par),
4914                                        -title => 'commitdiff to parent number ' .
4915                                                   ($i+1) . ': ' . substr($par,0,7)},
4916                                       $i+1) .
4917                               "&nbsp;</th>\n";
4918                 }
4919                 print "</tr></thead>\n<tbody>\n";
4920         }
4921
4922         my $alternate = 1;
4923         my $patchno = 0;
4924         foreach my $line (@{$difftree}) {
4925                 my $diff = parsed_difftree_line($line);
4926
4927                 if ($alternate) {
4928                         print "<tr class=\"dark\">\n";
4929                 } else {
4930                         print "<tr class=\"light\">\n";
4931                 }
4932                 $alternate ^= 1;
4933
4934                 if (exists $diff->{'nparents'}) { # combined diff
4935
4936                         fill_from_file_info($diff, @parents)
4937                                 unless exists $diff->{'from_file'};
4938
4939                         if (!is_deleted($diff)) {
4940                                 # file exists in the result (child) commit
4941                                 print "<td>" .
4942                                       $cgi->a({-href => href(action=>"blob", hash=>$diff->{'to_id'},
4943                                                              file_name=>$diff->{'to_file'},
4944                                                              hash_base=>$hash),
4945                                               -class => "list"}, esc_path($diff->{'to_file'})) .
4946                                       "</td>\n";
4947                         } else {
4948                                 print "<td>" .
4949                                       esc_path($diff->{'to_file'}) .
4950                                       "</td>\n";
4951                         }
4952
4953                         if ($action eq 'commitdiff') {
4954                                 # link to patch
4955                                 $patchno++;
4956                                 print "<td class=\"link\">" .
4957                                       $cgi->a({-href => href(-anchor=>"patch$patchno")},
4958                                               "patch") .
4959                                       " | " .
4960                                       "</td>\n";
4961                         }
4962
4963                         my $has_history = 0;
4964                         my $not_deleted = 0;
4965                         for (my $i = 0; $i < $diff->{'nparents'}; $i++) {
4966                                 my $hash_parent = $parents[$i];
4967                                 my $from_hash = $diff->{'from_id'}[$i];
4968                                 my $from_path = $diff->{'from_file'}[$i];
4969                                 my $status = $diff->{'status'}[$i];
4970
4971                                 $has_history ||= ($status ne 'A');
4972                                 $not_deleted ||= ($status ne 'D');
4973
4974                                 if ($status eq 'A') {
4975                                         print "<td  class=\"link\" align=\"right\"> | </td>\n";
4976                                 } elsif ($status eq 'D') {
4977                                         print "<td class=\"link\">" .
4978                                               $cgi->a({-href => href(action=>"blob",
4979                                                                      hash_base=>$hash,
4980                                                                      hash=>$from_hash,
4981                                                                      file_name=>$from_path)},
4982                                                       "blob" . ($i+1)) .
4983                                               " | </td>\n";
4984                                 } else {
4985                                         if ($diff->{'to_id'} eq $from_hash) {
4986                                                 print "<td class=\"link nochange\">";
4987                                         } else {
4988                                                 print "<td class=\"link\">";
4989                                         }
4990                                         print $cgi->a({-href => href(action=>"blobdiff",
4991                                                                      hash=>$diff->{'to_id'},
4992                                                                      hash_parent=>$from_hash,
4993                                                                      hash_base=>$hash,
4994                                                                      hash_parent_base=>$hash_parent,
4995                                                                      file_name=>$diff->{'to_file'},
4996                                                                      file_parent=>$from_path)},
4997                                                       "diff" . ($i+1)) .
4998                                               " | </td>\n";
4999                                 }
5000                         }
5001
5002                         print "<td class=\"link\">";
5003                         if ($not_deleted) {
5004                                 print $cgi->a({-href => href(action=>"blob",
5005                                                              hash=>$diff->{'to_id'},
5006                                                              file_name=>$diff->{'to_file'},
5007                                                              hash_base=>$hash)},
5008                                               "blob");
5009                                 print " | " if ($has_history);
5010                         }
5011                         if ($has_history) {
5012                                 print $cgi->a({-href => href(action=>"history",
5013                                                              file_name=>$diff->{'to_file'},
5014                                                              hash_base=>$hash)},
5015                                               "history");
5016                         }
5017                         print "</td>\n";
5018
5019                         print "</tr>\n";
5020                         next; # instead of 'else' clause, to avoid extra indent
5021                 }
5022                 # else ordinary diff
5023
5024                 my ($to_mode_oct, $to_mode_str, $to_file_type);
5025                 my ($from_mode_oct, $from_mode_str, $from_file_type);
5026                 if ($diff->{'to_mode'} ne ('0' x 6)) {
5027                         $to_mode_oct = oct $diff->{'to_mode'};
5028                         if (S_ISREG($to_mode_oct)) { # only for regular file
5029                                 $to_mode_str = sprintf("%04o", $to_mode_oct & 0777); # permission bits
5030                         }
5031                         $to_file_type = file_type($diff->{'to_mode'});
5032                 }
5033                 if ($diff->{'from_mode'} ne ('0' x 6)) {
5034                         $from_mode_oct = oct $diff->{'from_mode'};
5035                         if (S_ISREG($from_mode_oct)) { # only for regular file
5036                                 $from_mode_str = sprintf("%04o", $from_mode_oct & 0777); # permission bits
5037                         }
5038                         $from_file_type = file_type($diff->{'from_mode'});
5039                 }
5040
5041                 if ($diff->{'status'} eq "A") { # created
5042                         my $mode_chng = "<span class=\"file_status new\">[new $to_file_type";
5043                         $mode_chng   .= " with mode: $to_mode_str" if $to_mode_str;
5044                         $mode_chng   .= "]</span>";
5045                         print "<td>";
5046                         print $cgi->a({-href => href(action=>"blob", hash=>$diff->{'to_id'},
5047                                                      hash_base=>$hash, file_name=>$diff->{'file'}),
5048                                       -class => "list"}, esc_path($diff->{'file'}));
5049                         print "</td>\n";
5050                         print "<td>$mode_chng</td>\n";
5051                         print "<td class=\"link\">";
5052                         if ($action eq 'commitdiff') {
5053                                 # link to patch
5054                                 $patchno++;
5055                                 print $cgi->a({-href => href(-anchor=>"patch$patchno")},
5056                                               "patch") .
5057                                       " | ";
5058                         }
5059                         print $cgi->a({-href => href(action=>"blob", hash=>$diff->{'to_id'},
5060                                                      hash_base=>$hash, file_name=>$diff->{'file'})},
5061                                       "blob");
5062                         print "</td>\n";
5063
5064                 } elsif ($diff->{'status'} eq "D") { # deleted
5065                         my $mode_chng = "<span class=\"file_status deleted\">[deleted $from_file_type]</span>";
5066                         print "<td>";
5067                         print $cgi->a({-href => href(action=>"blob", hash=>$diff->{'from_id'},
5068                                                      hash_base=>$parent, file_name=>$diff->{'file'}),
5069                                        -class => "list"}, esc_path($diff->{'file'}));
5070                         print "</td>\n";
5071                         print "<td>$mode_chng</td>\n";
5072                         print "<td class=\"link\">";
5073                         if ($action eq 'commitdiff') {
5074                                 # link to patch
5075                                 $patchno++;
5076                                 print $cgi->a({-href => href(-anchor=>"patch$patchno")},
5077                                               "patch") .
5078                                       " | ";
5079                         }
5080                         print $cgi->a({-href => href(action=>"blob", hash=>$diff->{'from_id'},
5081                                                      hash_base=>$parent, file_name=>$diff->{'file'})},
5082                                       "blob") . " | ";
5083                         if ($have_blame) {
5084                                 print $cgi->a({-href => href(action=>"blame", hash_base=>$parent,
5085                                                              file_name=>$diff->{'file'})},
5086                                               "blame") . " | ";
5087                         }
5088                         print $cgi->a({-href => href(action=>"history", hash_base=>$parent,
5089                                                      file_name=>$diff->{'file'})},
5090                                       "history");
5091                         print "</td>\n";
5092
5093                 } elsif ($diff->{'status'} eq "M" || $diff->{'status'} eq "T") { # modified, or type changed
5094                         my $mode_chnge = "";
5095                         if ($diff->{'from_mode'} != $diff->{'to_mode'}) {
5096                                 $mode_chnge = "<span class=\"file_status mode_chnge\">[changed";
5097                                 if ($from_file_type ne $to_file_type) {
5098                                         $mode_chnge .= " from $from_file_type to $to_file_type";
5099                                 }
5100                                 if (($from_mode_oct & 0777) != ($to_mode_oct & 0777)) {
5101                                         if ($from_mode_str && $to_mode_str) {
5102                                                 $mode_chnge .= " mode: $from_mode_str->$to_mode_str";
5103                                         } elsif ($to_mode_str) {
5104                                                 $mode_chnge .= " mode: $to_mode_str";
5105                                         }
5106                                 }
5107                                 $mode_chnge .= "]</span>\n";
5108                         }
5109                         print "<td>";
5110                         print $cgi->a({-href => href(action=>"blob", hash=>$diff->{'to_id'},
5111                                                      hash_base=>$hash, file_name=>$diff->{'file'}),
5112                                       -class => "list"}, esc_path($diff->{'file'}));
5113                         print "</td>\n";
5114                         print "<td>$mode_chnge</td>\n";
5115                         print "<td class=\"link\">";
5116                         if ($action eq 'commitdiff') {
5117                                 # link to patch
5118                                 $patchno++;
5119                                 print $cgi->a({-href => href(-anchor=>"patch$patchno")},
5120                                               "patch") .
5121                                       " | ";
5122                         } elsif ($diff->{'to_id'} ne $diff->{'from_id'}) {
5123                                 # "commit" view and modified file (not onlu mode changed)
5124                                 print $cgi->a({-href => href(action=>"blobdiff",
5125                                                              hash=>$diff->{'to_id'}, hash_parent=>$diff->{'from_id'},
5126                                                              hash_base=>$hash, hash_parent_base=>$parent,
5127                                                              file_name=>$diff->{'file'})},
5128                                               "diff") .
5129                                       " | ";
5130                         }
5131                         print $cgi->a({-href => href(action=>"blob", hash=>$diff->{'to_id'},
5132                                                      hash_base=>$hash, file_name=>$diff->{'file'})},
5133                                        "blob") . " | ";
5134                         if ($have_blame) {
5135                                 print $cgi->a({-href => href(action=>"blame", hash_base=>$hash,
5136                                                              file_name=>$diff->{'file'})},
5137                                               "blame") . " | ";
5138                         }
5139                         print $cgi->a({-href => href(action=>"history", hash_base=>$hash,
5140                                                      file_name=>$diff->{'file'})},
5141                                       "history");
5142                         print "</td>\n";
5143
5144                 } elsif ($diff->{'status'} eq "R" || $diff->{'status'} eq "C") { # renamed or copied
5145                         my %status_name = ('R' => 'moved', 'C' => 'copied');
5146                         my $nstatus = $status_name{$diff->{'status'}};
5147                         my $mode_chng = "";
5148                         if ($diff->{'from_mode'} != $diff->{'to_mode'}) {
5149                                 # mode also for directories, so we cannot use $to_mode_str
5150                                 $mode_chng = sprintf(", mode: %04o", $to_mode_oct & 0777);
5151                         }
5152                         print "<td>" .
5153                               $cgi->a({-href => href(action=>"blob", hash_base=>$hash,
5154                                                      hash=>$diff->{'to_id'}, file_name=>$diff->{'to_file'}),
5155                                       -class => "list"}, esc_path($diff->{'to_file'})) . "</td>\n" .
5156                               "<td><span class=\"file_status $nstatus\">[$nstatus from " .
5157                               $cgi->a({-href => href(action=>"blob", hash_base=>$parent,
5158                                                      hash=>$diff->{'from_id'}, file_name=>$diff->{'from_file'}),
5159                                       -class => "list"}, esc_path($diff->{'from_file'})) .
5160                               " with " . (int $diff->{'similarity'}) . "% similarity$mode_chng]</span></td>\n" .
5161                               "<td class=\"link\">";
5162                         if ($action eq 'commitdiff') {
5163                                 # link to patch
5164                                 $patchno++;
5165                                 print $cgi->a({-href => href(-anchor=>"patch$patchno")},
5166                                               "patch") .
5167                                       " | ";
5168                         } elsif ($diff->{'to_id'} ne $diff->{'from_id'}) {
5169                                 # "commit" view and modified file (not only pure rename or copy)
5170                                 print $cgi->a({-href => href(action=>"blobdiff",
5171                                                              hash=>$diff->{'to_id'}, hash_parent=>$diff->{'from_id'},
5172                                                              hash_base=>$hash, hash_parent_base=>$parent,
5173                                                              file_name=>$diff->{'to_file'}, file_parent=>$diff->{'from_file'})},
5174                                               "diff") .
5175                                       " | ";
5176                         }
5177                         print $cgi->a({-href => href(action=>"blob", hash=>$diff->{'to_id'},
5178                                                      hash_base=>$parent, file_name=>$diff->{'to_file'})},
5179                                       "blob") . " | ";
5180                         if ($have_blame) {
5181                                 print $cgi->a({-href => href(action=>"blame", hash_base=>$hash,
5182                                                              file_name=>$diff->{'to_file'})},
5183                                               "blame") . " | ";
5184                         }
5185                         print $cgi->a({-href => href(action=>"history", hash_base=>$hash,
5186                                                     file_name=>$diff->{'to_file'})},
5187                                       "history");
5188                         print "</td>\n";
5189
5190                 } # we should not encounter Unmerged (U) or Unknown (X) status
5191                 print "</tr>\n";
5192         }
5193         print "</tbody>" if $has_header;
5194         print "</table>\n";
5195 }
5196
5197 # Print context lines and then rem/add lines in a side-by-side manner.
5198 sub print_sidebyside_diff_lines {
5199         my ($ctx, $rem, $add) = @_;
5200
5201         # print context block before add/rem block
5202         if (@$ctx) {
5203                 print join '',
5204                         '<div class="chunk_block ctx">',
5205                                 '<div class="old">',
5206                                 @$ctx,
5207                                 '</div>',
5208                                 '<div class="new">',
5209                                 @$ctx,
5210                                 '</div>',
5211                         '</div>';
5212         }
5213
5214         if (!@$add) {
5215                 # pure removal
5216                 print join '',
5217                         '<div class="chunk_block rem">',
5218                                 '<div class="old">',
5219                                 @$rem,
5220                                 '</div>',
5221                         '</div>';
5222         } elsif (!@$rem) {
5223                 # pure addition
5224                 print join '',
5225                         '<div class="chunk_block add">',
5226                                 '<div class="new">',
5227                                 @$add,
5228                                 '</div>',
5229                         '</div>';
5230         } else {
5231                 print join '',
5232                         '<div class="chunk_block chg">',
5233                                 '<div class="old">',
5234                                 @$rem,
5235                                 '</div>',
5236                                 '<div class="new">',
5237                                 @$add,
5238                                 '</div>',
5239                         '</div>';
5240         }
5241 }
5242
5243 # Print context lines and then rem/add lines in inline manner.
5244 sub print_inline_diff_lines {
5245         my ($ctx, $rem, $add) = @_;
5246
5247         print @$ctx, @$rem, @$add;
5248 }
5249
5250 # Format removed and added line, mark changed part and HTML-format them.
5251 # Implementation is based on contrib/diff-highlight
5252 sub format_rem_add_lines_pair {
5253         my ($rem, $add, $num_parents) = @_;
5254
5255         # We need to untabify lines before split()'ing them;
5256         # otherwise offsets would be invalid.
5257         chomp $rem;
5258         chomp $add;
5259         $rem = untabify($rem);
5260         $add = untabify($add);
5261
5262         my @rem = split(//, $rem);
5263         my @add = split(//, $add);
5264         my ($esc_rem, $esc_add);
5265         # Ignore leading +/- characters for each parent.
5266         my ($prefix_len, $suffix_len) = ($num_parents, 0);
5267         my ($prefix_has_nonspace, $suffix_has_nonspace);
5268
5269         my $shorter = (@rem < @add) ? @rem : @add;
5270         while ($prefix_len < $shorter) {
5271                 last if ($rem[$prefix_len] ne $add[$prefix_len]);
5272
5273                 $prefix_has_nonspace = 1 if ($rem[$prefix_len] !~ /\s/);
5274                 $prefix_len++;
5275         }
5276
5277         while ($prefix_len + $suffix_len < $shorter) {
5278                 last if ($rem[-1 - $suffix_len] ne $add[-1 - $suffix_len]);
5279
5280                 $suffix_has_nonspace = 1 if ($rem[-1 - $suffix_len] !~ /\s/);
5281                 $suffix_len++;
5282         }
5283
5284         # Mark lines that are different from each other, but have some common
5285         # part that isn't whitespace.  If lines are completely different, don't
5286         # mark them because that would make output unreadable, especially if
5287         # diff consists of multiple lines.
5288         if ($prefix_has_nonspace || $suffix_has_nonspace) {
5289                 $esc_rem = esc_html_hl_regions($rem, 'marked',
5290                         [$prefix_len, @rem - $suffix_len], -nbsp=>1);
5291                 $esc_add = esc_html_hl_regions($add, 'marked',
5292                         [$prefix_len, @add - $suffix_len], -nbsp=>1);
5293         } else {
5294                 $esc_rem = esc_html($rem, -nbsp=>1);
5295                 $esc_add = esc_html($add, -nbsp=>1);
5296         }
5297
5298         return format_diff_line(\$esc_rem, 'rem'),
5299                format_diff_line(\$esc_add, 'add');
5300 }
5301
5302 # HTML-format diff context, removed and added lines.
5303 sub format_ctx_rem_add_lines {
5304         my ($ctx, $rem, $add, $num_parents) = @_;
5305         my (@new_ctx, @new_rem, @new_add);
5306         my $can_highlight = 0;
5307         my $is_combined = ($num_parents > 1);
5308
5309         # Highlight if every removed line has a corresponding added line.
5310         if (@$add > 0 && @$add == @$rem) {
5311                 $can_highlight = 1;
5312
5313                 # Highlight lines in combined diff only if the chunk contains
5314                 # diff between the same version, e.g.
5315                 #
5316                 #    - a
5317                 #   -  b
5318                 #    + c
5319                 #   +  d
5320                 #
5321                 # Otherwise the highlighting would be confusing.
5322                 if ($is_combined) {
5323                         for (my $i = 0; $i < @$add; $i++) {
5324                                 my $prefix_rem = substr($rem->[$i], 0, $num_parents);
5325                                 my $prefix_add = substr($add->[$i], 0, $num_parents);
5326
5327                                 $prefix_rem =~ s/-/+/g;
5328
5329                                 if ($prefix_rem ne $prefix_add) {
5330                                         $can_highlight = 0;
5331                                         last;
5332                                 }
5333                         }
5334                 }
5335         }
5336
5337         if ($can_highlight) {
5338                 for (my $i = 0; $i < @$add; $i++) {
5339                         my ($line_rem, $line_add) = format_rem_add_lines_pair(
5340                                 $rem->[$i], $add->[$i], $num_parents);
5341                         push @new_rem, $line_rem;
5342                         push @new_add, $line_add;
5343                 }
5344         } else {
5345                 @new_rem = map { format_diff_line($_, 'rem') } @$rem;
5346                 @new_add = map { format_diff_line($_, 'add') } @$add;
5347         }
5348
5349         @new_ctx = map { format_diff_line($_, 'ctx') } @$ctx;
5350
5351         return (\@new_ctx, \@new_rem, \@new_add);
5352 }
5353
5354 # Print context lines and then rem/add lines.
5355 sub print_diff_lines {
5356         my ($ctx, $rem, $add, $diff_style, $num_parents) = @_;
5357         my $is_combined = $num_parents > 1;
5358
5359         ($ctx, $rem, $add) = format_ctx_rem_add_lines($ctx, $rem, $add,
5360                 $num_parents);
5361
5362         if ($diff_style eq 'sidebyside' && !$is_combined) {
5363                 print_sidebyside_diff_lines($ctx, $rem, $add);
5364         } else {
5365                 # default 'inline' style and unknown styles
5366                 print_inline_diff_lines($ctx, $rem, $add);
5367         }
5368 }
5369
5370 sub print_diff_chunk {
5371         my ($diff_style, $num_parents, $from, $to, @chunk) = @_;
5372         my (@ctx, @rem, @add);
5373
5374         # The class of the previous line.
5375         my $prev_class = '';
5376
5377         return unless @chunk;
5378
5379         # incomplete last line might be among removed or added lines,
5380         # or both, or among context lines: find which
5381         for (my $i = 1; $i < @chunk; $i++) {
5382                 if ($chunk[$i][0] eq 'incomplete') {
5383                         $chunk[$i][0] = $chunk[$i-1][0];
5384                 }
5385         }
5386
5387         # guardian
5388         push @chunk, ["", ""];
5389
5390         foreach my $line_info (@chunk) {
5391                 my ($class, $line) = @$line_info;
5392
5393                 # print chunk headers
5394                 if ($class && $class eq 'chunk_header') {
5395                         print format_diff_line($line, $class, $from, $to);
5396                         next;
5397                 }
5398
5399                 ## print from accumulator when have some add/rem lines or end
5400                 # of chunk (flush context lines), or when have add and rem
5401                 # lines and new block is reached (otherwise add/rem lines could
5402                 # be reordered)
5403                 if (!$class || ((@rem || @add) && $class eq 'ctx') ||
5404                     (@rem && @add && $class ne $prev_class)) {
5405                         print_diff_lines(\@ctx, \@rem, \@add,
5406                                          $diff_style, $num_parents);
5407                         @ctx = @rem = @add = ();
5408                 }
5409
5410                 ## adding lines to accumulator
5411                 # guardian value
5412                 last unless $line;
5413                 # rem, add or change
5414                 if ($class eq 'rem') {
5415                         push @rem, $line;
5416                 } elsif ($class eq 'add') {
5417                         push @add, $line;
5418                 }
5419                 # context line
5420                 if ($class eq 'ctx') {
5421                         push @ctx, $line;
5422                 }
5423
5424                 $prev_class = $class;
5425         }
5426 }
5427
5428 sub git_patchset_body {
5429         my ($fd, $diff_style, $difftree, $hash, @hash_parents) = @_;
5430         my ($hash_parent) = $hash_parents[0];
5431
5432         my $is_combined = (@hash_parents > 1);
5433         my $patch_idx = 0;
5434         my $patch_number = 0;
5435         my $patch_line;
5436         my $diffinfo;
5437         my $to_name;
5438         my (%from, %to);
5439         my @chunk; # for side-by-side diff
5440
5441         print "<div class=\"patchset\">\n";
5442
5443         # skip to first patch
5444         while ($patch_line = <$fd>) {
5445                 chomp $patch_line;
5446
5447                 last if ($patch_line =~ m/^diff /);
5448         }
5449
5450  PATCH:
5451         while ($patch_line) {
5452
5453                 # parse "git diff" header line
5454                 if ($patch_line =~ m/^diff --git (\"(?:[^\\\"]*(?:\\.[^\\\"]*)*)\"|[^ "]*) (.*)$/) {
5455                         # $1 is from_name, which we do not use
5456                         $to_name = unquote($2);
5457                         $to_name =~ s!^b/!!;
5458                 } elsif ($patch_line =~ m/^diff --(cc|combined) ("?.*"?)$/) {
5459                         # $1 is 'cc' or 'combined', which we do not use
5460                         $to_name = unquote($2);
5461                 } else {
5462                         $to_name = undef;
5463                 }
5464
5465                 # check if current patch belong to current raw line
5466                 # and parse raw git-diff line if needed
5467                 if (is_patch_split($diffinfo, { 'to_file' => $to_name })) {
5468                         # this is continuation of a split patch
5469                         print "<div class=\"patch cont\">\n";
5470                 } else {
5471                         # advance raw git-diff output if needed
5472                         $patch_idx++ if defined $diffinfo;
5473
5474                         # read and prepare patch information
5475                         $diffinfo = parsed_difftree_line($difftree->[$patch_idx]);
5476
5477                         # compact combined diff output can have some patches skipped
5478                         # find which patch (using pathname of result) we are at now;
5479                         if ($is_combined) {
5480                                 while ($to_name ne $diffinfo->{'to_file'}) {
5481                                         print "<div class=\"patch\" id=\"patch". ($patch_idx+1) ."\">\n" .
5482                                               format_diff_cc_simplified($diffinfo, @hash_parents) .
5483                                               "</div>\n";  # class="patch"
5484
5485                                         $patch_idx++;
5486                                         $patch_number++;
5487
5488                                         last if $patch_idx > $#$difftree;
5489                                         $diffinfo = parsed_difftree_line($difftree->[$patch_idx]);
5490                                 }
5491                         }
5492
5493                         # modifies %from, %to hashes
5494                         parse_from_to_diffinfo($diffinfo, \%from, \%to, @hash_parents);
5495
5496                         # this is first patch for raw difftree line with $patch_idx index
5497                         # we index @$difftree array from 0, but number patches from 1
5498                         print "<div class=\"patch\" id=\"patch". ($patch_idx+1) ."\">\n";
5499                 }
5500
5501                 # git diff header
5502                 #assert($patch_line =~ m/^diff /) if DEBUG;
5503                 #assert($patch_line !~ m!$/$!) if DEBUG; # is chomp-ed
5504                 $patch_number++;
5505                 # print "git diff" header
5506                 print format_git_diff_header_line($patch_line, $diffinfo,
5507                                                   \%from, \%to);
5508
5509                 # print extended diff header
5510                 print "<div class=\"diff extended_header\">\n";
5511         EXTENDED_HEADER:
5512                 while ($patch_line = <$fd>) {
5513                         chomp $patch_line;
5514
5515                         last EXTENDED_HEADER if ($patch_line =~ m/^--- |^diff /);
5516
5517                         print format_extended_diff_header_line($patch_line, $diffinfo,
5518                                                                \%from, \%to);
5519                 }
5520                 print "</div>\n"; # class="diff extended_header"
5521
5522                 # from-file/to-file diff header
5523                 if (! $patch_line) {
5524                         print "</div>\n"; # class="patch"
5525                         last PATCH;
5526                 }
5527                 next PATCH if ($patch_line =~ m/^diff /);
5528                 #assert($patch_line =~ m/^---/) if DEBUG;
5529
5530                 my $last_patch_line = $patch_line;
5531                 $patch_line = <$fd>;
5532                 chomp $patch_line;
5533                 #assert($patch_line =~ m/^\+\+\+/) if DEBUG;
5534
5535                 print format_diff_from_to_header($last_patch_line, $patch_line,
5536                                                  $diffinfo, \%from, \%to,
5537                                                  @hash_parents);
5538
5539                 # the patch itself
5540         LINE:
5541                 while ($patch_line = <$fd>) {
5542                         chomp $patch_line;
5543
5544                         next PATCH if ($patch_line =~ m/^diff /);
5545
5546                         my $class = diff_line_class($patch_line, \%from, \%to);
5547
5548                         if ($class eq 'chunk_header') {
5549                                 print_diff_chunk($diff_style, scalar @hash_parents, \%from, \%to, @chunk);
5550                                 @chunk = ();
5551                         }
5552
5553                         push @chunk, [ $class, $patch_line ];
5554                 }
5555
5556         } continue {
5557                 if (@chunk) {
5558                         print_diff_chunk($diff_style, scalar @hash_parents, \%from, \%to, @chunk);
5559                         @chunk = ();
5560                 }
5561                 print "</div>\n"; # class="patch"
5562         }
5563
5564         # for compact combined (--cc) format, with chunk and patch simplification
5565         # the patchset might be empty, but there might be unprocessed raw lines
5566         for (++$patch_idx if $patch_number > 0;
5567              $patch_idx < @$difftree;
5568              ++$patch_idx) {
5569                 # read and prepare patch information
5570                 $diffinfo = parsed_difftree_line($difftree->[$patch_idx]);
5571
5572                 # generate anchor for "patch" links in difftree / whatchanged part
5573                 print "<div class=\"patch\" id=\"patch". ($patch_idx+1) ."\">\n" .
5574                       format_diff_cc_simplified($diffinfo, @hash_parents) .
5575                       "</div>\n";  # class="patch"
5576
5577                 $patch_number++;
5578         }
5579
5580         if ($patch_number == 0) {
5581                 if (@hash_parents > 1) {
5582                         print "<div class=\"diff nodifferences\">Trivial merge</div>\n";
5583                 } else {
5584                         print "<div class=\"diff nodifferences\">No differences found</div>\n";
5585                 }
5586         }
5587
5588         print "</div>\n"; # class="patchset"
5589 }
5590
5591 # . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5592
5593 sub git_project_search_form {
5594         my ($searchtext, $search_use_regexp) = @_;
5595
5596         my $limit = '';
5597         if ($project_filter) {
5598                 $limit = " in '$project_filter/'";
5599         }
5600
5601         print "<div class=\"projsearch\">\n";
5602         print $cgi->start_form(-method => 'get', -action => $my_uri) .
5603               $cgi->hidden(-name => 'a', -value => 'project_list')  . "\n";
5604         print $cgi->hidden(-name => 'pf', -value => $project_filter). "\n"
5605                 if (defined $project_filter);
5606         print $cgi->textfield(-name => 's', -value => $searchtext,
5607                               -title => "Search project by name and description$limit",
5608                               -size => 60) . "\n" .
5609               "<span title=\"Extended regular expression\">" .
5610               $cgi->checkbox(-name => 'sr', -value => 1, -label => 're',
5611                              -checked => $search_use_regexp) .
5612               "</span>\n" .
5613               $cgi->submit(-name => 'btnS', -value => 'Search') .
5614               $cgi->end_form() . "\n" .
5615               $cgi->a({-href => href(project => undef, searchtext => undef,
5616                                      project_filter => $project_filter)},
5617                       esc_html("List all projects$limit")) . "<br />\n";
5618         print "</div>\n";
5619 }
5620
5621 # entry for given @keys needs filling if at least one of keys in list
5622 # is not present in %$project_info
5623 sub project_info_needs_filling {
5624         my ($project_info, @keys) = @_;
5625
5626         # return List::MoreUtils::any { !exists $project_info->{$_} } @keys;
5627         foreach my $key (@keys) {
5628                 if (!exists $project_info->{$key}) {
5629                         return 1;
5630                 }
5631         }
5632         return;
5633 }
5634
5635 # fills project list info (age, description, owner, category, forks, etc.)
5636 # for each project in the list, removing invalid projects from
5637 # returned list, or fill only specified info.
5638 #
5639 # Invalid projects are removed from the returned list if and only if you
5640 # ask 'age' or 'age_string' to be filled, because they are the only fields
5641 # that run unconditionally git command that requires repository, and
5642 # therefore do always check if project repository is invalid.
5643 #
5644 # USAGE:
5645 # * fill_project_list_info(\@project_list, 'descr_long', 'ctags')
5646 #   ensures that 'descr_long' and 'ctags' fields are filled
5647 # * @project_list = fill_project_list_info(\@project_list)
5648 #   ensures that all fields are filled (and invalid projects removed)
5649 #
5650 # NOTE: modifies $projlist, but does not remove entries from it
5651 sub fill_project_list_info {
5652         my ($projlist, @wanted_keys) = @_;
5653         my @projects;
5654         my $filter_set = sub { return @_; };
5655         if (@wanted_keys) {
5656                 my %wanted_keys = map { $_ => 1 } @wanted_keys;
5657                 $filter_set = sub { return grep { $wanted_keys{$_} } @_; };
5658         }
5659
5660         my $show_ctags = gitweb_check_feature('ctags');
5661  PROJECT:
5662         foreach my $pr (@$projlist) {
5663                 if (project_info_needs_filling($pr, $filter_set->('age', 'age_string'))) {
5664                         my (@activity) = git_get_last_activity($pr->{'path'});
5665                         unless (@activity) {
5666                                 next PROJECT;
5667                         }
5668                         ($pr->{'age'}, $pr->{'age_string'}) = @activity;
5669                 }
5670                 if (project_info_needs_filling($pr, $filter_set->('descr', 'descr_long'))) {
5671                         my $descr = git_get_project_description($pr->{'path'}) || "";
5672                         $descr = to_utf8($descr);
5673                         $pr->{'descr_long'} = $descr;
5674                         $pr->{'descr'} = chop_str($descr, $projects_list_description_width, 5);
5675                 }
5676                 if (project_info_needs_filling($pr, $filter_set->('owner'))) {
5677                         $pr->{'owner'} = git_get_project_owner("$pr->{'path'}") || "";
5678                 }
5679                 if ($show_ctags &&
5680                     project_info_needs_filling($pr, $filter_set->('ctags'))) {
5681                         $pr->{'ctags'} = git_get_project_ctags($pr->{'path'});
5682                 }
5683                 if ($projects_list_group_categories &&
5684                     project_info_needs_filling($pr, $filter_set->('category'))) {
5685                         my $cat = git_get_project_category($pr->{'path'}) ||
5686                                                            $project_list_default_category;
5687                         $pr->{'category'} = to_utf8($cat);
5688                 }
5689
5690                 push @projects, $pr;
5691         }
5692
5693         return @projects;
5694 }
5695
5696 sub sort_projects_list {
5697         my ($projlist, $order) = @_;
5698
5699         sub order_str {
5700                 my $key = shift;
5701                 return sub { $a->{$key} cmp $b->{$key} };
5702         }
5703
5704         sub order_num_then_undef {
5705                 my $key = shift;
5706                 return sub {
5707                         defined $a->{$key} ?
5708                                 (defined $b->{$key} ? $a->{$key} <=> $b->{$key} : -1) :
5709                                 (defined $b->{$key} ? 1 : 0)
5710                 };
5711         }
5712
5713         my %orderings = (
5714                 project => order_str('path'),
5715                 descr => order_str('descr_long'),
5716                 owner => order_str('owner'),
5717                 age => order_num_then_undef('age'),
5718         );
5719
5720         my $ordering = $orderings{$order};
5721         return defined $ordering ? sort $ordering @$projlist : @$projlist;
5722 }
5723
5724 # returns a hash of categories, containing the list of project
5725 # belonging to each category
5726 sub build_projlist_by_category {
5727         my ($projlist, $from, $to) = @_;
5728         my %categories;
5729
5730         $from = 0 unless defined $from;
5731         $to = $#$projlist if (!defined $to || $#$projlist < $to);
5732
5733         for (my $i = $from; $i <= $to; $i++) {
5734                 my $pr = $projlist->[$i];
5735                 push @{$categories{ $pr->{'category'} }}, $pr;
5736         }
5737
5738         return wantarray ? %categories : \%categories;
5739 }
5740
5741 # print 'sort by' <th> element, generating 'sort by $name' replay link
5742 # if that order is not selected
5743 sub print_sort_th {
5744         print format_sort_th(@_);
5745 }
5746
5747 sub format_sort_th {
5748         my ($name, $order, $header) = @_;
5749         my $sort_th = "";
5750         $header ||= ucfirst($name);
5751
5752         if ($order eq $name) {
5753                 $sort_th .= "<th>$header</th>\n";
5754         } else {
5755                 $sort_th .= "<th>" .
5756                             $cgi->a({-href => href(-replay=>1, order=>$name),
5757                                      -class => "header"}, $header) .
5758                             "</th>\n";
5759         }
5760
5761         return $sort_th;
5762 }
5763
5764 sub git_project_list_rows {
5765         my ($projlist, $from, $to, $check_forks) = @_;
5766
5767         $from = 0 unless defined $from;
5768         $to = $#$projlist if (!defined $to || $#$projlist < $to);
5769
5770         my $alternate = 1;
5771         for (my $i = $from; $i <= $to; $i++) {
5772                 my $pr = $projlist->[$i];
5773
5774                 if ($alternate) {
5775                         print "<tr class=\"dark\">\n";
5776                 } else {
5777                         print "<tr class=\"light\">\n";
5778                 }
5779                 $alternate ^= 1;
5780
5781                 if ($check_forks) {
5782                         print "<td>";
5783                         if ($pr->{'forks'}) {
5784                                 my $nforks = scalar @{$pr->{'forks'}};
5785                                 if ($nforks > 0) {
5786                                         print $cgi->a({-href => href(project=>$pr->{'path'}, action=>"forks"),
5787                                                        -title => "$nforks forks"}, "+");
5788                                 } else {
5789                                         print $cgi->span({-title => "$nforks forks"}, "+");
5790                                 }
5791                         }
5792                         print "</td>\n";
5793                 }
5794                 print "<td>" . $cgi->a({-href => href(project=>$pr->{'path'}, action=>"summary"),
5795                                         -class => "list"},
5796                                        esc_html_match_hl($pr->{'path'}, $search_regexp)) .
5797                       "</td>\n" .
5798                       "<td>" . $cgi->a({-href => href(project=>$pr->{'path'}, action=>"summary"),
5799                                         -class => "list",
5800                                         -title => $pr->{'descr_long'}},
5801                                         $search_regexp
5802                                         ? esc_html_match_hl_chopped($pr->{'descr_long'},
5803                                                                     $pr->{'descr'}, $search_regexp)
5804                                         : esc_html($pr->{'descr'})) .
5805                       "</td>\n";
5806                 unless ($omit_owner) {
5807                         print "<td><i>" . chop_and_escape_str($pr->{'owner'}, 15) . "</i></td>\n";
5808                 }
5809                 unless ($omit_age_column) {
5810                         print "<td class=\"". age_class($pr->{'age'}) . "\">" .
5811                             (defined $pr->{'age_string'} ? $pr->{'age_string'} : "No commits") . "</td>\n";
5812                 }
5813                 print"<td class=\"link\">" .
5814                       $cgi->a({-href => href(project=>$pr->{'path'}, action=>"summary")}, "summary")   . " | " .
5815                       $cgi->a({-href => href(project=>$pr->{'path'}, action=>"shortlog")}, "shortlog") . " | " .
5816                       $cgi->a({-href => href(project=>$pr->{'path'}, action=>"log")}, "log") . " | " .
5817                       $cgi->a({-href => href(project=>$pr->{'path'}, action=>"tree")}, "tree") .
5818                       ($pr->{'forks'} ? " | " . $cgi->a({-href => href(project=>$pr->{'path'}, action=>"forks")}, "forks") : '') .
5819                       "</td>\n" .
5820                       "</tr>\n";
5821         }
5822 }
5823
5824 sub git_project_list_body {
5825         # actually uses global variable $project
5826         my ($projlist, $order, $from, $to, $extra, $no_header) = @_;
5827         my @projects = @$projlist;
5828
5829         my $check_forks = gitweb_check_feature('forks');
5830         my $show_ctags  = gitweb_check_feature('ctags');
5831         my $tagfilter = $show_ctags ? $input_params{'ctag'} : undef;
5832         $check_forks = undef
5833                 if ($tagfilter || $search_regexp);
5834
5835         # filtering out forks before filling info allows to do less work
5836         @projects = filter_forks_from_projects_list(\@projects)
5837                 if ($check_forks);
5838         # search_projects_list pre-fills required info
5839         @projects = search_projects_list(\@projects,
5840                                          'search_regexp' => $search_regexp,
5841                                          'tagfilter'  => $tagfilter)
5842                 if ($tagfilter || $search_regexp);
5843         # fill the rest
5844         my @all_fields = ('descr', 'descr_long', 'ctags', 'category');
5845         push @all_fields, ('age', 'age_string') unless($omit_age_column);
5846         push @all_fields, 'owner' unless($omit_owner);
5847         @projects = fill_project_list_info(\@projects, @all_fields);
5848
5849         $order ||= $default_projects_order;
5850         $from = 0 unless defined $from;
5851         $to = $#projects if (!defined $to || $#projects < $to);
5852
5853         # short circuit
5854         if ($from > $to) {
5855                 print "<center>\n".
5856                       "<b>No such projects found</b><br />\n".
5857                       "Click ".$cgi->a({-href=>href(project=>undef)},"here")." to view all projects<br />\n".
5858                       "</center>\n<br />\n";
5859                 return;
5860         }
5861
5862         @projects = sort_projects_list(\@projects, $order);
5863
5864         if ($show_ctags) {
5865                 my $ctags = git_gather_all_ctags(\@projects);
5866                 my $cloud = git_populate_project_tagcloud($ctags);
5867                 print git_show_project_tagcloud($cloud, 64);
5868         }
5869
5870         print "<table class=\"project_list\">\n";
5871         unless ($no_header) {
5872                 print "<tr>\n";
5873                 if ($check_forks) {
5874                         print "<th></th>\n";
5875                 }
5876                 print_sort_th('project', $order, 'Project');
5877                 print_sort_th('descr', $order, 'Description');
5878                 print_sort_th('owner', $order, 'Owner') unless $omit_owner;
5879                 print_sort_th('age', $order, 'Last Change') unless $omit_age_column;
5880                 print "<th></th>\n" . # for links
5881                       "</tr>\n";
5882         }
5883
5884         if ($projects_list_group_categories) {
5885                 # only display categories with projects in the $from-$to window
5886                 @projects = sort {$a->{'category'} cmp $b->{'category'}} @projects[$from..$to];
5887                 my %categories = build_projlist_by_category(\@projects, $from, $to);
5888                 foreach my $cat (sort keys %categories) {
5889                         unless ($cat eq "") {
5890                                 print "<tr>\n";
5891                                 if ($check_forks) {
5892                                         print "<td></td>\n";
5893                                 }
5894                                 print "<td class=\"category\" colspan=\"5\">".esc_html($cat)."</td>\n";
5895                                 print "</tr>\n";
5896                         }
5897
5898                         git_project_list_rows($categories{$cat}, undef, undef, $check_forks);
5899                 }
5900         } else {
5901                 git_project_list_rows(\@projects, $from, $to, $check_forks);
5902         }
5903
5904         if (defined $extra) {
5905                 print "<tr>\n";
5906                 if ($check_forks) {
5907                         print "<td></td>\n";
5908                 }
5909                 print "<td colspan=\"5\">$extra</td>\n" .
5910                       "</tr>\n";
5911         }
5912         print "</table>\n";
5913 }
5914
5915 sub git_log_body {
5916         # uses global variable $project
5917         my ($commitlist, $from, $to, $refs, $extra) = @_;
5918
5919         $from = 0 unless defined $from;
5920         $to = $#{$commitlist} if (!defined $to || $#{$commitlist} < $to);
5921
5922         for (my $i = 0; $i <= $to; $i++) {
5923                 my %co = %{$commitlist->[$i]};
5924                 next if !%co;
5925                 my $commit = $co{'id'};
5926                 my $ref = format_ref_marker($refs, $commit);
5927                 git_print_header_div('commit',
5928                                "<span class=\"age\">$co{'age_string'}</span>" .
5929                                esc_html($co{'title'}) . $ref,
5930                                $commit);
5931                 print "<div class=\"title_text\">\n" .
5932                       "<div class=\"log_link\">\n" .
5933                       $cgi->a({-href => href(action=>"commit", hash=>$commit)}, "commit") .
5934                       " | " .
5935                       $cgi->a({-href => href(action=>"commitdiff", hash=>$commit)}, "commitdiff") .
5936                       " | " .
5937                       $cgi->a({-href => href(action=>"tree", hash=>$commit, hash_base=>$commit)}, "tree") .
5938                       "<br/>\n" .
5939                       "</div>\n";
5940                       git_print_authorship(\%co, -tag => 'span');
5941                       print "<br/>\n</div>\n";
5942
5943                 print "<div class=\"log_body\">\n";
5944                 git_print_log($co{'comment'}, -final_empty_line=> 1);
5945                 print "</div>\n";
5946         }
5947         if ($extra) {
5948                 print "<div class=\"page_nav\">\n";
5949                 print "$extra\n";
5950                 print "</div>\n";
5951         }
5952 }
5953
5954 sub git_shortlog_body {
5955         # uses global variable $project
5956         my ($commitlist, $from, $to, $refs, $extra) = @_;
5957
5958         $from = 0 unless defined $from;
5959         $to = $#{$commitlist} if (!defined $to || $#{$commitlist} < $to);
5960
5961         print "<table class=\"shortlog\">\n";
5962         my $alternate = 1;
5963         for (my $i = $from; $i <= $to; $i++) {
5964                 my %co = %{$commitlist->[$i]};
5965                 my $commit = $co{'id'};
5966                 my $ref = format_ref_marker($refs, $commit);
5967                 if ($alternate) {
5968                         print "<tr class=\"dark\">\n";
5969                 } else {
5970                         print "<tr class=\"light\">\n";
5971                 }
5972                 $alternate ^= 1;
5973                 # git_summary() used print "<td><i>$co{'age_string'}</i></td>\n" .
5974                 print "<td title=\"$co{'age_string_age'}\"><i>$co{'age_string_date'}</i></td>\n" .
5975                       format_author_html('td', \%co, 10) . "<td>";
5976                 print format_subject_html($co{'title'}, $co{'title_short'},
5977                                           href(action=>"commit", hash=>$commit), $ref);
5978                 print "</td>\n" .
5979                       "<td class=\"link\">" .
5980                       $cgi->a({-href => href(action=>"commit", hash=>$commit)}, "commit") . " | " .
5981                       $cgi->a({-href => href(action=>"commitdiff", hash=>$commit)}, "commitdiff") . " | " .
5982                       $cgi->a({-href => href(action=>"tree", hash=>$commit, hash_base=>$commit)}, "tree");
5983                 my $snapshot_links = format_snapshot_links($commit);
5984                 if (defined $snapshot_links) {
5985                         print " | " . $snapshot_links;
5986                 }
5987                 print "</td>\n" .
5988                       "</tr>\n";
5989         }
5990         if (defined $extra) {
5991                 print "<tr>\n" .
5992                       "<td colspan=\"4\">$extra</td>\n" .
5993                       "</tr>\n";
5994         }
5995         print "</table>\n";
5996 }
5997
5998 sub git_history_body {
5999         # Warning: assumes constant type (blob or tree) during history
6000         my ($commitlist, $from, $to, $refs, $extra,
6001             $file_name, $file_hash, $ftype) = @_;
6002
6003         $from = 0 unless defined $from;
6004         $to = $#{$commitlist} unless (defined $to && $to <= $#{$commitlist});
6005
6006         print "<table class=\"history\">\n";
6007         my $alternate = 1;
6008         for (my $i = $from; $i <= $to; $i++) {
6009                 my %co = %{$commitlist->[$i]};
6010                 if (!%co) {
6011                         next;
6012                 }
6013                 my $commit = $co{'id'};
6014
6015                 my $ref = format_ref_marker($refs, $commit);
6016
6017                 if ($alternate) {
6018                         print "<tr class=\"dark\">\n";
6019                 } else {
6020                         print "<tr class=\"light\">\n";
6021                 }
6022                 $alternate ^= 1;
6023                 print "<td title=\"$co{'age_string_age'}\"><i>$co{'age_string_date'}</i></td>\n" .
6024         # shortlog:   format_author_html('td', \%co, 10)
6025                       format_author_html('td', \%co, 15, 3) . "<td>";
6026                 # originally git_history used chop_str($co{'title'}, 50)
6027                 print format_subject_html($co{'title'}, $co{'title_short'},
6028                                           href(action=>"commit", hash=>$commit), $ref);
6029                 print "</td>\n" .
6030                       "<td class=\"link\">" .
6031                       $cgi->a({-href => href(action=>$ftype, hash_base=>$commit, file_name=>$file_name)}, $ftype) . " | " .
6032                       $cgi->a({-href => href(action=>"commitdiff", hash=>$commit)}, "commitdiff");
6033
6034                 if ($ftype eq 'blob') {
6035                         print " | " .
6036                               $cgi->a({-href => href(action=>"blob_plain", hash_base=>$commit, file_name=>$file_name)}, "raw");
6037
6038                         my $blob_current = $file_hash;
6039                         my $blob_parent  = git_get_hash_by_path($commit, $file_name);
6040                         if (defined $blob_current && defined $blob_parent &&
6041                                         $blob_current ne $blob_parent) {
6042                                 print " | " .
6043                                         $cgi->a({-href => href(action=>"blobdiff",
6044                                                                hash=>$blob_current, hash_parent=>$blob_parent,
6045                                                                hash_base=>$hash_base, hash_parent_base=>$commit,
6046                                                                file_name=>$file_name)},
6047                                                 "diff to current");
6048                         }
6049                 }
6050                 print "</td>\n" .
6051                       "</tr>\n";
6052         }
6053         if (defined $extra) {
6054                 print "<tr>\n" .
6055                       "<td colspan=\"4\">$extra</td>\n" .
6056                       "</tr>\n";
6057         }
6058         print "</table>\n";
6059 }
6060
6061 sub git_tags_body {
6062         # uses global variable $project
6063         my ($taglist, $from, $to, $extra) = @_;
6064         $from = 0 unless defined $from;
6065         $to = $#{$taglist} if (!defined $to || $#{$taglist} < $to);
6066
6067         print "<table class=\"tags\">\n";
6068         my $alternate = 1;
6069         for (my $i = $from; $i <= $to; $i++) {
6070                 my $entry = $taglist->[$i];
6071                 my %tag = %$entry;
6072                 my $comment = $tag{'subject'};
6073                 my $comment_short;
6074                 if (defined $comment) {
6075                         $comment_short = chop_str($comment, 30, 5);
6076                 }
6077                 if ($alternate) {
6078                         print "<tr class=\"dark\">\n";
6079                 } else {
6080                         print "<tr class=\"light\">\n";
6081                 }
6082                 $alternate ^= 1;
6083                 if (defined $tag{'age'}) {
6084                         print "<td><i>$tag{'age'}</i></td>\n";
6085                 } else {
6086                         print "<td></td>\n";
6087                 }
6088                 print "<td>" .
6089                       $cgi->a({-href => href(action=>$tag{'reftype'}, hash=>$tag{'refid'}),
6090                                -class => "list name"}, esc_html($tag{'name'})) .
6091                       "</td>\n" .
6092                       "<td>";
6093                 if (defined $comment) {
6094                         print format_subject_html($comment, $comment_short,
6095                                                   href(action=>"tag", hash=>$tag{'id'}));
6096                 }
6097                 print "</td>\n" .
6098                       "<td class=\"selflink\">";
6099                 if ($tag{'type'} eq "tag") {
6100                         print $cgi->a({-href => href(action=>"tag", hash=>$tag{'id'})}, "tag");
6101                 } else {
6102                         print "&nbsp;";
6103                 }
6104                 print "</td>\n" .
6105                       "<td class=\"link\">" . " | " .
6106                       $cgi->a({-href => href(action=>$tag{'reftype'}, hash=>$tag{'refid'})}, $tag{'reftype'});
6107                 if ($tag{'reftype'} eq "commit") {
6108                         print " | " . $cgi->a({-href => href(action=>"shortlog", hash=>$tag{'fullname'})}, "shortlog") .
6109                               " | " . $cgi->a({-href => href(action=>"log", hash=>$tag{'fullname'})}, "log");
6110                 } elsif ($tag{'reftype'} eq "blob") {
6111                         print " | " . $cgi->a({-href => href(action=>"blob_plain", hash=>$tag{'refid'})}, "raw");
6112                 }
6113                 print "</td>\n" .
6114                       "</tr>";
6115         }
6116         if (defined $extra) {
6117                 print "<tr>\n" .
6118                       "<td colspan=\"5\">$extra</td>\n" .
6119                       "</tr>\n";
6120         }
6121         print "</table>\n";
6122 }
6123
6124 sub git_heads_body {
6125         # uses global variable $project
6126         my ($headlist, $head_at, $from, $to, $extra) = @_;
6127         $from = 0 unless defined $from;
6128         $to = $#{$headlist} if (!defined $to || $#{$headlist} < $to);
6129
6130         print "<table class=\"heads\">\n";
6131         my $alternate = 1;
6132         for (my $i = $from; $i <= $to; $i++) {
6133                 my $entry = $headlist->[$i];
6134                 my %ref = %$entry;
6135                 my $curr = defined $head_at && $ref{'id'} eq $head_at;
6136                 if ($alternate) {
6137                         print "<tr class=\"dark\">\n";
6138                 } else {
6139                         print "<tr class=\"light\">\n";
6140                 }
6141                 $alternate ^= 1;
6142                 print "<td><i>$ref{'age'}</i></td>\n" .
6143                       ($curr ? "<td class=\"current_head\">" : "<td>") .
6144                       $cgi->a({-href => href(action=>"shortlog", hash=>$ref{'fullname'}),
6145                                -class => "list name"},esc_html($ref{'name'})) .
6146                       "</td>\n" .
6147                       "<td class=\"link\">" .
6148                       $cgi->a({-href => href(action=>"shortlog", hash=>$ref{'fullname'})}, "shortlog") . " | " .
6149                       $cgi->a({-href => href(action=>"log", hash=>$ref{'fullname'})}, "log") . " | " .
6150                       $cgi->a({-href => href(action=>"tree", hash=>$ref{'fullname'}, hash_base=>$ref{'fullname'})}, "tree") .
6151                       "</td>\n" .
6152                       "</tr>";
6153         }
6154         if (defined $extra) {
6155                 print "<tr>\n" .
6156                       "<td colspan=\"3\">$extra</td>\n" .
6157                       "</tr>\n";
6158         }
6159         print "</table>\n";
6160 }
6161
6162 # Display a single remote block
6163 sub git_remote_block {
6164         my ($remote, $rdata, $limit, $head) = @_;
6165
6166         my $heads = $rdata->{'heads'};
6167         my $fetch = $rdata->{'fetch'};
6168         my $push = $rdata->{'push'};
6169
6170         my $urls_table = "<table class=\"projects_list\">\n" ;
6171
6172         if (defined $fetch) {
6173                 if ($fetch eq $push) {
6174                         $urls_table .= format_repo_url("URL", $fetch);
6175                 } else {
6176                         $urls_table .= format_repo_url("Fetch URL", $fetch);
6177                         $urls_table .= format_repo_url("Push URL", $push) if defined $push;
6178                 }
6179         } elsif (defined $push) {
6180                 $urls_table .= format_repo_url("Push URL", $push);
6181         } else {
6182                 $urls_table .= format_repo_url("", "No remote URL");
6183         }
6184
6185         $urls_table .= "</table>\n";
6186
6187         my $dots;
6188         if (defined $limit && $limit < @$heads) {
6189                 $dots = $cgi->a({-href => href(action=>"remotes", hash=>$remote)}, "...");
6190         }
6191
6192         print $urls_table;
6193         git_heads_body($heads, $head, 0, $limit, $dots);
6194 }
6195
6196 # Display a list of remote names with the respective fetch and push URLs
6197 sub git_remotes_list {
6198         my ($remotedata, $limit) = @_;
6199         print "<table class=\"heads\">\n";
6200         my $alternate = 1;
6201         my @remotes = sort keys %$remotedata;
6202
6203         my $limited = $limit && $limit < @remotes;
6204
6205         $#remotes = $limit - 1 if $limited;
6206
6207         while (my $remote = shift @remotes) {
6208                 my $rdata = $remotedata->{$remote};
6209                 my $fetch = $rdata->{'fetch'};
6210                 my $push = $rdata->{'push'};
6211                 if ($alternate) {
6212                         print "<tr class=\"dark\">\n";
6213                 } else {
6214                         print "<tr class=\"light\">\n";
6215                 }
6216                 $alternate ^= 1;
6217                 print "<td>" .
6218                       $cgi->a({-href=> href(action=>'remotes', hash=>$remote),
6219                                -class=> "list name"},esc_html($remote)) .
6220                       "</td>";
6221                 print "<td class=\"link\">" .
6222                       (defined $fetch ? $cgi->a({-href=> $fetch}, "fetch") : "fetch") .
6223                       " | " .
6224                       (defined $push ? $cgi->a({-href=> $push}, "push") : "push") .
6225                       "</td>";
6226
6227                 print "</tr>\n";
6228         }
6229
6230         if ($limited) {
6231                 print "<tr>\n" .
6232                       "<td colspan=\"3\">" .
6233                       $cgi->a({-href => href(action=>"remotes")}, "...") .
6234                       "</td>\n" . "</tr>\n";
6235         }
6236
6237         print "</table>";
6238 }
6239
6240 # Display remote heads grouped by remote, unless there are too many
6241 # remotes, in which case we only display the remote names
6242 sub git_remotes_body {
6243         my ($remotedata, $limit, $head) = @_;
6244         if ($limit and $limit < keys %$remotedata) {
6245                 git_remotes_list($remotedata, $limit);
6246         } else {
6247                 fill_remote_heads($remotedata);
6248                 while (my ($remote, $rdata) = each %$remotedata) {
6249                         git_print_section({-class=>"remote", -id=>$remote},
6250                                 ["remotes", $remote, $remote], sub {
6251                                         git_remote_block($remote, $rdata, $limit, $head);
6252                                 });
6253                 }
6254         }
6255 }
6256
6257 sub git_search_message {
6258         my %co = @_;
6259
6260         my $greptype;
6261         if ($searchtype eq 'commit') {
6262                 $greptype = "--grep=";
6263         } elsif ($searchtype eq 'author') {
6264                 $greptype = "--author=";
6265         } elsif ($searchtype eq 'committer') {
6266                 $greptype = "--committer=";
6267         }
6268         $greptype .= $searchtext;
6269         my @commitlist = parse_commits($hash, 101, (100 * $page), undef,
6270                                        $greptype, '--regexp-ignore-case',
6271                                        $search_use_regexp ? '--extended-regexp' : '--fixed-strings');
6272
6273         my $paging_nav = '';
6274         if ($page > 0) {
6275                 $paging_nav .=
6276                         $cgi->a({-href => href(-replay=>1, page=>undef)},
6277                                 "first") .
6278                         " &sdot; " .
6279                         $cgi->a({-href => href(-replay=>1, page=>$page-1),
6280                                  -accesskey => "p", -title => "Alt-p"}, "prev");
6281         } else {
6282                 $paging_nav .= "first &sdot; prev";
6283         }
6284         my $next_link = '';
6285         if ($#commitlist >= 100) {
6286                 $next_link =
6287                         $cgi->a({-href => href(-replay=>1, page=>$page+1),
6288                                  -accesskey => "n", -title => "Alt-n"}, "next");
6289                 $paging_nav .= " &sdot; $next_link";
6290         } else {
6291                 $paging_nav .= " &sdot; next";
6292         }
6293
6294         git_header_html();
6295
6296         git_print_page_nav('','', $hash,$co{'tree'},$hash, $paging_nav);
6297         git_print_header_div('commit', esc_html($co{'title'}), $hash);
6298         if ($page == 0 && !@commitlist) {
6299                 print "<p>No match.</p>\n";
6300         } else {
6301                 git_search_grep_body(\@commitlist, 0, 99, $next_link);
6302         }
6303
6304         git_footer_html();
6305 }
6306
6307 sub git_search_changes {
6308         my %co = @_;
6309
6310         local $/ = "\n";
6311         open my $fd, '-|', git_cmd(), '--no-pager', 'log', @diff_opts,
6312                 '--pretty=format:%H', '--no-abbrev', '--raw', "-S$searchtext",
6313                 ($search_use_regexp ? '--pickaxe-regex' : ())
6314                         or die_error(500, "Open git-log failed");
6315
6316         git_header_html();
6317
6318         git_print_page_nav('','', $hash,$co{'tree'},$hash);
6319         git_print_header_div('commit', esc_html($co{'title'}), $hash);
6320
6321         print "<table class=\"pickaxe search\">\n";
6322         my $alternate = 1;
6323         undef %co;
6324         my @files;
6325         while (my $line = <$fd>) {
6326                 chomp $line;
6327                 next unless $line;
6328
6329                 my %set = parse_difftree_raw_line($line);
6330                 if (defined $set{'commit'}) {
6331                         # finish previous commit
6332                         if (%co) {
6333                                 print "</td>\n" .
6334                                       "<td class=\"link\">" .
6335                                       $cgi->a({-href => href(action=>"commit", hash=>$co{'id'})},
6336                                               "commit") .
6337                                       " | " .
6338                                       $cgi->a({-href => href(action=>"tree", hash=>$co{'tree'},
6339                                                              hash_base=>$co{'id'})},
6340                                               "tree") .
6341                                       "</td>\n" .
6342                                       "</tr>\n";
6343                         }
6344
6345                         if ($alternate) {
6346                                 print "<tr class=\"dark\">\n";
6347                         } else {
6348                                 print "<tr class=\"light\">\n";
6349                         }
6350                         $alternate ^= 1;
6351                         %co = parse_commit($set{'commit'});
6352                         my $author = chop_and_escape_str($co{'author_name'}, 15, 5);
6353                         print "<td title=\"$co{'age_string_age'}\"><i>$co{'age_string_date'}</i></td>\n" .
6354                               "<td><i>$author</i></td>\n" .
6355                               "<td>" .
6356                               $cgi->a({-href => href(action=>"commit", hash=>$co{'id'}),
6357                                       -class => "list subject"},
6358                                       chop_and_escape_str($co{'title'}, 50) . "<br/>");
6359                 } elsif (defined $set{'to_id'}) {
6360                         next if is_deleted(\%set);
6361
6362                         print $cgi->a({-href => href(action=>"blob", hash_base=>$co{'id'},
6363                                                      hash=>$set{'to_id'}, file_name=>$set{'to_file'}),
6364                                       -class => "list"},
6365                                       "<span class=\"match\">" . esc_path($set{'file'}) . "</span>") .
6366                               "<br/>\n";
6367                 }
6368         }
6369         close $fd;
6370
6371         # finish last commit (warning: repetition!)
6372         if (%co) {
6373                 print "</td>\n" .
6374                       "<td class=\"link\">" .
6375                       $cgi->a({-href => href(action=>"commit", hash=>$co{'id'})},
6376                               "commit") .
6377                       " | " .
6378                       $cgi->a({-href => href(action=>"tree", hash=>$co{'tree'},
6379                                              hash_base=>$co{'id'})},
6380                               "tree") .
6381                       "</td>\n" .
6382                       "</tr>\n";
6383         }
6384
6385         print "</table>\n";
6386
6387         git_footer_html();
6388 }
6389
6390 sub git_search_files {
6391         my %co = @_;
6392
6393         local $/ = "\n";
6394         open my $fd, "-|", git_cmd(), 'grep', '-n', '-z',
6395                 $search_use_regexp ? ('-E', '-i') : '-F',
6396                 $searchtext, $co{'tree'}
6397                         or die_error(500, "Open git-grep failed");
6398
6399         git_header_html();
6400
6401         git_print_page_nav('','', $hash,$co{'tree'},$hash);
6402         git_print_header_div('commit', esc_html($co{'title'}), $hash);
6403
6404         print "<table class=\"grep_search\">\n";
6405         my $alternate = 1;
6406         my $matches = 0;
6407         my $lastfile = '';
6408         my $file_href;
6409         while (my $line = <$fd>) {
6410                 chomp $line;
6411                 my ($file, $lno, $ltext, $binary);
6412                 last if ($matches++ > 1000);
6413                 if ($line =~ /^Binary file (.+) matches$/) {
6414                         $file = $1;
6415                         $binary = 1;
6416                 } else {
6417                         ($file, $lno, $ltext) = split(/\0/, $line, 3);
6418                         $file =~ s/^$co{'tree'}://;
6419                 }
6420                 if ($file ne $lastfile) {
6421                         $lastfile and print "</td></tr>\n";
6422                         if ($alternate++) {
6423                                 print "<tr class=\"dark\">\n";
6424                         } else {
6425                                 print "<tr class=\"light\">\n";
6426                         }
6427                         $file_href = href(action=>"blob", hash_base=>$co{'id'},
6428                                           file_name=>$file);
6429                         print "<td class=\"list\">".
6430                                 $cgi->a({-href => $file_href, -class => "list"}, esc_path($file));
6431                         print "</td><td>\n";
6432                         $lastfile = $file;
6433                 }
6434                 if ($binary) {
6435                         print "<div class=\"binary\">Binary file</div>\n";
6436                 } else {
6437                         $ltext = untabify($ltext);
6438                         if ($ltext =~ m/^(.*)($search_regexp)(.*)$/i) {
6439                                 $ltext = esc_html($1, -nbsp=>1);
6440                                 $ltext .= '<span class="match">';
6441                                 $ltext .= esc_html($2, -nbsp=>1);
6442                                 $ltext .= '</span>';
6443                                 $ltext .= esc_html($3, -nbsp=>1);
6444                         } else {
6445                                 $ltext = esc_html($ltext, -nbsp=>1);
6446                         }
6447                         print "<div class=\"pre\">" .
6448                                 $cgi->a({-href => $file_href.'#l'.$lno,
6449                                         -class => "linenr"}, sprintf('%4i', $lno)) .
6450                                 ' ' .  $ltext . "</div>\n";
6451                 }
6452         }
6453         if ($lastfile) {
6454                 print "</td></tr>\n";
6455                 if ($matches > 1000) {
6456                         print "<div class=\"diff nodifferences\">Too many matches, listing trimmed</div>\n";
6457                 }
6458         } else {
6459                 print "<div class=\"diff nodifferences\">No matches found</div>\n";
6460         }
6461         close $fd;
6462
6463         print "</table>\n";
6464
6465         git_footer_html();
6466 }
6467
6468 sub git_search_grep_body {
6469         my ($commitlist, $from, $to, $extra) = @_;
6470         $from = 0 unless defined $from;
6471         $to = $#{$commitlist} if (!defined $to || $#{$commitlist} < $to);
6472
6473         print "<table class=\"commit_search\">\n";
6474         my $alternate = 1;
6475         for (my $i = $from; $i <= $to; $i++) {
6476                 my %co = %{$commitlist->[$i]};
6477                 if (!%co) {
6478                         next;
6479                 }
6480                 my $commit = $co{'id'};
6481                 if ($alternate) {
6482                         print "<tr class=\"dark\">\n";
6483                 } else {
6484                         print "<tr class=\"light\">\n";
6485                 }
6486                 $alternate ^= 1;
6487                 print "<td title=\"$co{'age_string_age'}\"><i>$co{'age_string_date'}</i></td>\n" .
6488                       format_author_html('td', \%co, 15, 5) .
6489                       "<td>" .
6490                       $cgi->a({-href => href(action=>"commit", hash=>$co{'id'}),
6491                                -class => "list subject"},
6492                               chop_and_escape_str($co{'title'}, 50) . "<br/>");
6493                 my $comment = $co{'comment'};
6494                 foreach my $line (@$comment) {
6495                         if ($line =~ m/^(.*?)($search_regexp)(.*)$/i) {
6496                                 my ($lead, $match, $trail) = ($1, $2, $3);
6497                                 $match = chop_str($match, 70, 5, 'center');
6498                                 my $contextlen = int((80 - length($match))/2);
6499                                 $contextlen = 30 if ($contextlen > 30);
6500                                 $lead  = chop_str($lead,  $contextlen, 10, 'left');
6501                                 $trail = chop_str($trail, $contextlen, 10, 'right');
6502
6503                                 $lead  = esc_html($lead);
6504                                 $match = esc_html($match);
6505                                 $trail = esc_html($trail);
6506
6507                                 print "$lead<span class=\"match\">$match</span>$trail<br />";
6508                         }
6509                 }
6510                 print "</td>\n" .
6511                       "<td class=\"link\">" .
6512                       $cgi->a({-href => href(action=>"commit", hash=>$co{'id'})}, "commit") .
6513                       " | " .
6514                       $cgi->a({-href => href(action=>"commitdiff", hash=>$co{'id'})}, "commitdiff") .
6515                       " | " .
6516                       $cgi->a({-href => href(action=>"tree", hash=>$co{'tree'}, hash_base=>$co{'id'})}, "tree");
6517                 print "</td>\n" .
6518                       "</tr>\n";
6519         }
6520         if (defined $extra) {
6521                 print "<tr>\n" .
6522                       "<td colspan=\"3\">$extra</td>\n" .
6523                       "</tr>\n";
6524         }
6525         print "</table>\n";
6526 }
6527
6528 ## ======================================================================
6529 ## ======================================================================
6530 ## actions
6531
6532 sub git_project_list {
6533         my $order = $input_params{'order'};
6534         if (defined $order && $order !~ m/none|project|descr|owner|age/) {
6535                 die_error(400, "Unknown order parameter");
6536         }
6537
6538         my @list = git_get_projects_list($project_filter, $strict_export);
6539         if (!@list) {
6540                 die_error(404, "No projects found");
6541         }
6542
6543         git_header_html();
6544         if (defined $home_text && -f $home_text) {
6545                 print "<div class=\"index_include\">\n";
6546                 insert_file($home_text);
6547                 print "</div>\n";
6548         }
6549
6550         git_project_search_form($searchtext, $search_use_regexp);
6551         git_project_list_body(\@list, $order);
6552         git_footer_html();
6553 }
6554
6555 sub git_forks {
6556         my $order = $input_params{'order'};
6557         if (defined $order && $order !~ m/none|project|descr|owner|age/) {
6558                 die_error(400, "Unknown order parameter");
6559         }
6560
6561         my $filter = $project;
6562         $filter =~ s/\.git$//;
6563         my @list = git_get_projects_list($filter);
6564         if (!@list) {
6565                 die_error(404, "No forks found");
6566         }
6567
6568         git_header_html();
6569         git_print_page_nav('','');
6570         git_print_header_div('summary', "$project forks");
6571         git_project_list_body(\@list, $order);
6572         git_footer_html();
6573 }
6574
6575 sub git_project_index {
6576         my @projects = git_get_projects_list($project_filter, $strict_export);
6577         if (!@projects) {
6578                 die_error(404, "No projects found");
6579         }
6580
6581         print $cgi->header(
6582                 -type => 'text/plain',
6583                 -charset => 'utf-8',
6584                 -content_disposition => 'inline; filename="index.aux"');
6585
6586         foreach my $pr (@projects) {
6587                 if (!exists $pr->{'owner'}) {
6588                         $pr->{'owner'} = git_get_project_owner("$pr->{'path'}");
6589                 }
6590
6591                 my ($path, $owner) = ($pr->{'path'}, $pr->{'owner'});
6592                 # quote as in CGI::Util::encode, but keep the slash, and use '+' for ' '
6593                 $path  =~ s/([^a-zA-Z0-9_.\-\/ ])/sprintf("%%%02X", ord($1))/eg;
6594                 $owner =~ s/([^a-zA-Z0-9_.\-\/ ])/sprintf("%%%02X", ord($1))/eg;
6595                 $path  =~ s/ /\+/g;
6596                 $owner =~ s/ /\+/g;
6597
6598                 print "$path $owner\n";
6599         }
6600 }
6601
6602 sub git_summary {
6603         my $descr = git_get_project_description($project) || "none";
6604         my %co = parse_commit("HEAD");
6605         my %cd = %co ? parse_date($co{'committer_epoch'}, $co{'committer_tz'}) : ();
6606         my $head = $co{'id'};
6607         my $remote_heads = gitweb_check_feature('remote_heads');
6608
6609         my $owner = git_get_project_owner($project);
6610
6611         my $refs = git_get_references();
6612         # These get_*_list functions return one more to allow us to see if
6613         # there are more ...
6614         my @taglist  = git_get_tags_list(16);
6615         my @headlist = git_get_heads_list(16);
6616         my %remotedata = $remote_heads ? git_get_remotes_list() : ();
6617         my @forklist;
6618         my $check_forks = gitweb_check_feature('forks');
6619
6620         if ($check_forks) {
6621                 # find forks of a project
6622                 my $filter = $project;
6623                 $filter =~ s/\.git$//;
6624                 @forklist = git_get_projects_list($filter);
6625                 # filter out forks of forks
6626                 @forklist = filter_forks_from_projects_list(\@forklist)
6627                         if (@forklist);
6628         }
6629
6630         git_header_html();
6631         git_print_page_nav('summary','', $head);
6632
6633         print "<div class=\"title\">&nbsp;</div>\n";
6634         print "<table class=\"projects_list\">\n" .
6635               "<tr id=\"metadata_desc\"><td>description</td><td>" . esc_html($descr) . "</td></tr>\n";
6636         if ($owner and not $omit_owner) {
6637                 print  "<tr id=\"metadata_owner\"><td>owner</td><td>" . esc_html($owner) . "</td></tr>\n";
6638         }
6639         if (defined $cd{'rfc2822'}) {
6640                 print "<tr id=\"metadata_lchange\"><td>last change</td>" .
6641                       "<td>".format_timestamp_html(\%cd)."</td></tr>\n";
6642         }
6643
6644         # use per project git URL list in $projectroot/$project/cloneurl
6645         # or make project git URL from git base URL and project name
6646         my $url_tag = "URL";
6647         my @url_list = git_get_project_url_list($project);
6648         @url_list = map { "$_/$project" } @git_base_url_list unless @url_list;
6649         foreach my $git_url (@url_list) {
6650                 next unless $git_url;
6651                 print format_repo_url($url_tag, $git_url);
6652                 $url_tag = "";
6653         }
6654
6655         # Tag cloud
6656         my $show_ctags = gitweb_check_feature('ctags');
6657         if ($show_ctags) {
6658                 my $ctags = git_get_project_ctags($project);
6659                 if (%$ctags) {
6660                         # without ability to add tags, don't show if there are none
6661                         my $cloud = git_populate_project_tagcloud($ctags);
6662                         print "<tr id=\"metadata_ctags\">" .
6663                               "<td>content tags</td>" .
6664                               "<td>".git_show_project_tagcloud($cloud, 48)."</td>" .
6665                               "</tr>\n";
6666                 }
6667         }
6668
6669         print "</table>\n";
6670
6671         # If XSS prevention is on, we don't include README.html.
6672         # TODO: Allow a readme in some safe format.
6673         if (!$prevent_xss && -s "$projectroot/$project/README.html") {
6674                 print "<div class=\"title\">readme</div>\n" .
6675                       "<div class=\"readme\">\n";
6676                 insert_file("$projectroot/$project/README.html");
6677                 print "\n</div>\n"; # class="readme"
6678         }
6679
6680         # we need to request one more than 16 (0..15) to check if
6681         # those 16 are all
6682         my @commitlist = $head ? parse_commits($head, 17) : ();
6683         if (@commitlist) {
6684                 git_print_header_div('shortlog');
6685                 git_shortlog_body(\@commitlist, 0, 15, $refs,
6686                                   $#commitlist <=  15 ? undef :
6687                                   $cgi->a({-href => href(action=>"shortlog")}, "..."));
6688         }
6689
6690         if (@taglist) {
6691                 git_print_header_div('tags');
6692                 git_tags_body(\@taglist, 0, 15,
6693                               $#taglist <=  15 ? undef :
6694                               $cgi->a({-href => href(action=>"tags")}, "..."));
6695         }
6696
6697         if (@headlist) {
6698                 git_print_header_div('heads');
6699                 git_heads_body(\@headlist, $head, 0, 15,
6700                                $#headlist <= 15 ? undef :
6701                                $cgi->a({-href => href(action=>"heads")}, "..."));
6702         }
6703
6704         if (%remotedata) {
6705                 git_print_header_div('remotes');
6706                 git_remotes_body(\%remotedata, 15, $head);
6707         }
6708
6709         if (@forklist) {
6710                 git_print_header_div('forks');
6711                 git_project_list_body(\@forklist, 'age', 0, 15,
6712                                       $#forklist <= 15 ? undef :
6713                                       $cgi->a({-href => href(action=>"forks")}, "..."),
6714                                       'no_header');
6715         }
6716
6717         git_footer_html();
6718 }
6719
6720 sub git_tag {
6721         my %tag = parse_tag($hash);
6722
6723         if (! %tag) {
6724                 die_error(404, "Unknown tag object");
6725         }
6726
6727         my $head = git_get_head_hash($project);
6728         git_header_html();
6729         git_print_page_nav('','', $head,undef,$head);
6730         git_print_header_div('commit', esc_html($tag{'name'}), $hash);
6731         print "<div class=\"title_text\">\n" .
6732               "<table class=\"object_header\">\n" .
6733               "<tr>\n" .
6734               "<td>object</td>\n" .
6735               "<td>" . $cgi->a({-class => "list", -href => href(action=>$tag{'type'}, hash=>$tag{'object'})},
6736                                $tag{'object'}) . "</td>\n" .
6737               "<td class=\"link\">" . $cgi->a({-href => href(action=>$tag{'type'}, hash=>$tag{'object'})},
6738                                               $tag{'type'}) . "</td>\n" .
6739               "</tr>\n";
6740         if (defined($tag{'author'})) {
6741                 git_print_authorship_rows(\%tag, 'author');
6742         }
6743         print "</table>\n\n" .
6744               "</div>\n";
6745         print "<div class=\"page_body\">";
6746         my $comment = $tag{'comment'};
6747         foreach my $line (@$comment) {
6748                 chomp $line;
6749                 print esc_html($line, -nbsp=>1) . "<br/>\n";
6750         }
6751         print "</div>\n";
6752         git_footer_html();
6753 }
6754
6755 sub git_blame_common {
6756         my $format = shift || 'porcelain';
6757         if ($format eq 'porcelain' && $input_params{'javascript'}) {
6758                 $format = 'incremental';
6759                 $action = 'blame_incremental'; # for page title etc
6760         }
6761
6762         # permissions
6763         gitweb_check_feature('blame')
6764                 or die_error(403, "Blame view not allowed");
6765
6766         # error checking
6767         die_error(400, "No file name given") unless $file_name;
6768         $hash_base ||= git_get_head_hash($project);
6769         die_error(404, "Couldn't find base commit") unless $hash_base;
6770         my %co = parse_commit($hash_base)
6771                 or die_error(404, "Commit not found");
6772         my $ftype = "blob";
6773         if (!defined $hash) {
6774                 $hash = git_get_hash_by_path($hash_base, $file_name, "blob")
6775                         or die_error(404, "Error looking up file");
6776         } else {
6777                 $ftype = git_get_type($hash);
6778                 if ($ftype !~ "blob") {
6779                         die_error(400, "Object is not a blob");
6780                 }
6781         }
6782
6783         my $fd;
6784         if ($format eq 'incremental') {
6785                 # get file contents (as base)
6786                 open $fd, "-|", git_cmd(), 'cat-file', 'blob', $hash
6787                         or die_error(500, "Open git-cat-file failed");
6788         } elsif ($format eq 'data') {
6789                 # run git-blame --incremental
6790                 open $fd, "-|", git_cmd(), "blame", "--incremental",
6791                         $hash_base, "--", $file_name
6792                         or die_error(500, "Open git-blame --incremental failed");
6793         } else {
6794                 # run git-blame --porcelain
6795                 open $fd, "-|", git_cmd(), "blame", '-p',
6796                         $hash_base, '--', $file_name
6797                         or die_error(500, "Open git-blame --porcelain failed");
6798         }
6799         binmode $fd, ':utf8';
6800
6801         # incremental blame data returns early
6802         if ($format eq 'data') {
6803                 print $cgi->header(
6804                         -type=>"text/plain", -charset => "utf-8",
6805                         -status=> "200 OK");
6806                 local $| = 1; # output autoflush
6807                 while (my $line = <$fd>) {
6808                         print to_utf8($line);
6809                 }
6810                 close $fd
6811                         or print "ERROR $!\n";
6812
6813                 print 'END';
6814                 if (defined $t0 && gitweb_check_feature('timed')) {
6815                         print ' '.
6816                               tv_interval($t0, [ gettimeofday() ]).
6817                               ' '.$number_of_git_cmds;
6818                 }
6819                 print "\n";
6820
6821                 return;
6822         }
6823
6824         # page header
6825         git_header_html();
6826         my $formats_nav =
6827                 $cgi->a({-href => href(action=>"blob", -replay=>1)},
6828                         "blob") .
6829                 " | ";
6830         if ($format eq 'incremental') {
6831                 $formats_nav .=
6832                         $cgi->a({-href => href(action=>"blame", javascript=>0, -replay=>1)},
6833                                 "blame") . " (non-incremental)";
6834         } else {
6835                 $formats_nav .=
6836                         $cgi->a({-href => href(action=>"blame_incremental", -replay=>1)},
6837                                 "blame") . " (incremental)";
6838         }
6839         $formats_nav .=
6840                 " | " .
6841                 $cgi->a({-href => href(action=>"history", -replay=>1)},
6842                         "history") .
6843                 " | " .
6844                 $cgi->a({-href => href(action=>$action, file_name=>$file_name)},
6845                         "HEAD");
6846         git_print_page_nav('','', $hash_base,$co{'tree'},$hash_base, $formats_nav);
6847         git_print_header_div('commit', esc_html($co{'title'}), $hash_base);
6848         git_print_page_path($file_name, $ftype, $hash_base);
6849
6850         # page body
6851         if ($format eq 'incremental') {
6852                 print "<noscript>\n<div class=\"error\"><center><b>\n".
6853                       "This page requires JavaScript to run.\n Use ".
6854                       $cgi->a({-href => href(action=>'blame',javascript=>0,-replay=>1)},
6855                               'this page').
6856                       " instead.\n".
6857                       "</b></center></div>\n</noscript>\n";
6858
6859                 print qq!<div id="progress_bar" style="width: 100%; background-color: yellow"></div>\n!;
6860         }
6861
6862         print qq!<div class="page_body">\n!;
6863         print qq!<div id="progress_info">... / ...</div>\n!
6864                 if ($format eq 'incremental');
6865         print qq!<table id="blame_table" class="blame" width="100%">\n!.
6866               #qq!<col width="5.5em" /><col width="2.5em" /><col width="*" />\n!.
6867               qq!<thead>\n!.
6868               qq!<tr><th>Commit</th><th>Line</th><th>Data</th></tr>\n!.
6869               qq!</thead>\n!.
6870               qq!<tbody>\n!;
6871
6872         my @rev_color = qw(light dark);
6873         my $num_colors = scalar(@rev_color);
6874         my $current_color = 0;
6875
6876         if ($format eq 'incremental') {
6877                 my $color_class = $rev_color[$current_color];
6878
6879                 #contents of a file
6880                 my $linenr = 0;
6881         LINE:
6882                 while (my $line = <$fd>) {
6883                         chomp $line;
6884                         $linenr++;
6885
6886                         print qq!<tr id="l$linenr" class="$color_class">!.
6887                               qq!<td class="sha1"><a href=""> </a></td>!.
6888                               qq!<td class="linenr">!.
6889                               qq!<a class="linenr" href="">$linenr</a></td>!;
6890                         print qq!<td class="pre">! . esc_html($line) . "</td>\n";
6891                         print qq!</tr>\n!;
6892                 }
6893
6894         } else { # porcelain, i.e. ordinary blame
6895                 my %metainfo = (); # saves information about commits
6896
6897                 # blame data
6898         LINE:
6899                 while (my $line = <$fd>) {
6900                         chomp $line;
6901                         # the header: <SHA-1> <src lineno> <dst lineno> [<lines in group>]
6902                         # no <lines in group> for subsequent lines in group of lines
6903                         my ($full_rev, $orig_lineno, $lineno, $group_size) =
6904                            ($line =~ /^($oid_regex) (\d+) (\d+)(?: (\d+))?$/);
6905                         if (!exists $metainfo{$full_rev}) {
6906                                 $metainfo{$full_rev} = { 'nprevious' => 0 };
6907                         }
6908                         my $meta = $metainfo{$full_rev};
6909                         my $data;
6910                         while ($data = <$fd>) {
6911                                 chomp $data;
6912                                 last if ($data =~ s/^\t//); # contents of line
6913                                 if ($data =~ /^(\S+)(?: (.*))?$/) {
6914                                         $meta->{$1} = $2 unless exists $meta->{$1};
6915                                 }
6916                                 if ($data =~ /^previous /) {
6917                                         $meta->{'nprevious'}++;
6918                                 }
6919                         }
6920                         my $short_rev = substr($full_rev, 0, 8);
6921                         my $author = $meta->{'author'};
6922                         my %date =
6923                                 parse_date($meta->{'author-time'}, $meta->{'author-tz'});
6924                         my $date = $date{'iso-tz'};
6925                         if ($group_size) {
6926                                 $current_color = ($current_color + 1) % $num_colors;
6927                         }
6928                         my $tr_class = $rev_color[$current_color];
6929                         $tr_class .= ' boundary' if (exists $meta->{'boundary'});
6930                         $tr_class .= ' no-previous' if ($meta->{'nprevious'} == 0);
6931                         $tr_class .= ' multiple-previous' if ($meta->{'nprevious'} > 1);
6932                         print "<tr id=\"l$lineno\" class=\"$tr_class\">\n";
6933                         if ($group_size) {
6934                                 print "<td class=\"sha1\"";
6935                                 print " title=\"". esc_html($author) . ", $date\"";
6936                                 print " rowspan=\"$group_size\"" if ($group_size > 1);
6937                                 print ">";
6938                                 print $cgi->a({-href => href(action=>"commit",
6939                                                              hash=>$full_rev,
6940                                                              file_name=>$file_name)},
6941                                               esc_html($short_rev));
6942                                 if ($group_size >= 2) {
6943                                         my @author_initials = ($author =~ /\b([[:upper:]])\B/g);
6944                                         if (@author_initials) {
6945                                                 print "<br />" .
6946                                                       esc_html(join('', @author_initials));
6947                                                 #           or join('.', ...)
6948                                         }
6949                                 }
6950                                 print "</td>\n";
6951                         }
6952                         # 'previous' <sha1 of parent commit> <filename at commit>
6953                         if (exists $meta->{'previous'} &&
6954                             $meta->{'previous'} =~ /^($oid_regex) (.*)$/) {
6955                                 $meta->{'parent'} = $1;
6956                                 $meta->{'file_parent'} = unquote($2);
6957                         }
6958                         my $linenr_commit =
6959                                 exists($meta->{'parent'}) ?
6960                                 $meta->{'parent'} : $full_rev;
6961                         my $linenr_filename =
6962                                 exists($meta->{'file_parent'}) ?
6963                                 $meta->{'file_parent'} : unquote($meta->{'filename'});
6964                         my $blamed = href(action => 'blame',
6965                                           file_name => $linenr_filename,
6966                                           hash_base => $linenr_commit);
6967                         print "<td class=\"linenr\">";
6968                         print $cgi->a({ -href => "$blamed#l$orig_lineno",
6969                                         -class => "linenr" },
6970                                       esc_html($lineno));
6971                         print "</td>";
6972                         print "<td class=\"pre\">" . esc_html($data) . "</td>\n";
6973                         print "</tr>\n";
6974                 } # end while
6975
6976         }
6977
6978         # footer
6979         print "</tbody>\n".
6980               "</table>\n"; # class="blame"
6981         print "</div>\n";   # class="blame_body"
6982         close $fd
6983                 or print "Reading blob failed\n";
6984
6985         git_footer_html();
6986 }
6987
6988 sub git_blame {
6989         git_blame_common();
6990 }
6991
6992 sub git_blame_incremental {
6993         git_blame_common('incremental');
6994 }
6995
6996 sub git_blame_data {
6997         git_blame_common('data');
6998 }
6999
7000 sub git_tags {
7001         my $head = git_get_head_hash($project);
7002         git_header_html();
7003         git_print_page_nav('','', $head,undef,$head,format_ref_views('tags'));
7004         git_print_header_div('summary', $project);
7005
7006         my @tagslist = git_get_tags_list();
7007         if (@tagslist) {
7008                 git_tags_body(\@tagslist);
7009         }
7010         git_footer_html();
7011 }
7012
7013 sub git_heads {
7014         my $head = git_get_head_hash($project);
7015         git_header_html();
7016         git_print_page_nav('','', $head,undef,$head,format_ref_views('heads'));
7017         git_print_header_div('summary', $project);
7018
7019         my @headslist = git_get_heads_list();
7020         if (@headslist) {
7021                 git_heads_body(\@headslist, $head);
7022         }
7023         git_footer_html();
7024 }
7025
7026 # used both for single remote view and for list of all the remotes
7027 sub git_remotes {
7028         gitweb_check_feature('remote_heads')
7029                 or die_error(403, "Remote heads view is disabled");
7030
7031         my $head = git_get_head_hash($project);
7032         my $remote = $input_params{'hash'};
7033
7034         my $remotedata = git_get_remotes_list($remote);
7035         die_error(500, "Unable to get remote information") unless defined $remotedata;
7036
7037         unless (%$remotedata) {
7038                 die_error(404, defined $remote ?
7039                         "Remote $remote not found" :
7040                         "No remotes found");
7041         }
7042
7043         git_header_html(undef, undef, -action_extra => $remote);
7044         git_print_page_nav('', '',  $head, undef, $head,
7045                 format_ref_views($remote ? '' : 'remotes'));
7046
7047         fill_remote_heads($remotedata);
7048         if (defined $remote) {
7049                 git_print_header_div('remotes', "$remote remote for $project");
7050                 git_remote_block($remote, $remotedata->{$remote}, undef, $head);
7051         } else {
7052                 git_print_header_div('summary', "$project remotes");
7053                 git_remotes_body($remotedata, undef, $head);
7054         }
7055
7056         git_footer_html();
7057 }
7058
7059 sub git_blob_plain {
7060         my $type = shift;
7061         my $expires;
7062
7063         if (!defined $hash) {
7064                 if (defined $file_name) {
7065                         my $base = $hash_base || git_get_head_hash($project);
7066                         $hash = git_get_hash_by_path($base, $file_name, "blob")
7067                                 or die_error(404, "Cannot find file");
7068                 } else {
7069                         die_error(400, "No file name defined");
7070                 }
7071         } elsif ($hash =~ m/^$oid_regex$/) {
7072                 # blobs defined by non-textual hash id's can be cached
7073                 $expires = "+1d";
7074         }
7075
7076         open my $fd, "-|", git_cmd(), "cat-file", "blob", $hash
7077                 or die_error(500, "Open git-cat-file blob '$hash' failed");
7078
7079         # content-type (can include charset)
7080         $type = blob_contenttype($fd, $file_name, $type);
7081
7082         # "save as" filename, even when no $file_name is given
7083         my $save_as = "$hash";
7084         if (defined $file_name) {
7085                 $save_as = $file_name;
7086         } elsif ($type =~ m/^text\//) {
7087                 $save_as .= '.txt';
7088         }
7089
7090         # With XSS prevention on, blobs of all types except a few known safe
7091         # ones are served with "Content-Disposition: attachment" to make sure
7092         # they don't run in our security domain.  For certain image types,
7093         # blob view writes an <img> tag referring to blob_plain view, and we
7094         # want to be sure not to break that by serving the image as an
7095         # attachment (though Firefox 3 doesn't seem to care).
7096         my $sandbox = $prevent_xss &&
7097                 $type !~ m!^(?:text/[a-z]+|image/(?:gif|png|jpeg))(?:[ ;]|$)!;
7098
7099         # serve text/* as text/plain
7100         if ($prevent_xss &&
7101             ($type =~ m!^text/[a-z]+\b(.*)$! ||
7102              ($type =~ m!^[a-z]+/[a-z]\+xml\b(.*)$! && -T $fd))) {
7103                 my $rest = $1;
7104                 $rest = defined $rest ? $rest : '';
7105                 $type = "text/plain$rest";
7106         }
7107
7108         print $cgi->header(
7109                 -type => $type,
7110                 -expires => $expires,
7111                 -content_disposition =>
7112                         ($sandbox ? 'attachment' : 'inline')
7113                         . '; filename="' . $save_as . '"');
7114         local $/ = undef;
7115         local *FCGI::Stream::PRINT = $FCGI_Stream_PRINT_raw;
7116         binmode STDOUT, ':raw';
7117         print <$fd>;
7118         binmode STDOUT, ':utf8'; # as set at the beginning of gitweb.cgi
7119         close $fd;
7120 }
7121
7122 sub git_blob {
7123         my $expires;
7124
7125         if (!defined $hash) {
7126                 if (defined $file_name) {
7127                         my $base = $hash_base || git_get_head_hash($project);
7128                         $hash = git_get_hash_by_path($base, $file_name, "blob")
7129                                 or die_error(404, "Cannot find file");
7130                 } else {
7131                         die_error(400, "No file name defined");
7132                 }
7133         } elsif ($hash =~ m/^$oid_regex$/) {
7134                 # blobs defined by non-textual hash id's can be cached
7135                 $expires = "+1d";
7136         }
7137
7138         my $have_blame = gitweb_check_feature('blame');
7139         open my $fd, "-|", git_cmd(), "cat-file", "blob", $hash
7140                 or die_error(500, "Couldn't cat $file_name, $hash");
7141         my $mimetype = blob_mimetype($fd, $file_name);
7142         # use 'blob_plain' (aka 'raw') view for files that cannot be displayed
7143         if ($mimetype !~ m!^(?:text/|image/(?:gif|png|jpeg)$)! && -B $fd) {
7144                 close $fd;
7145                 return git_blob_plain($mimetype);
7146         }
7147         # we can have blame only for text/* mimetype
7148         $have_blame &&= ($mimetype =~ m!^text/!);
7149
7150         my $highlight = gitweb_check_feature('highlight');
7151         my $syntax = guess_file_syntax($highlight, $file_name);
7152         $fd = run_highlighter($fd, $highlight, $syntax);
7153
7154         git_header_html(undef, $expires);
7155         my $formats_nav = '';
7156         if (defined $hash_base && (my %co = parse_commit($hash_base))) {
7157                 if (defined $file_name) {
7158                         if ($have_blame) {
7159                                 $formats_nav .=
7160                                         $cgi->a({-href => href(action=>"blame", -replay=>1)},
7161                                                 "blame") .
7162                                         " | ";
7163                         }
7164                         $formats_nav .=
7165                                 $cgi->a({-href => href(action=>"history", -replay=>1)},
7166                                         "history") .
7167                                 " | " .
7168                                 $cgi->a({-href => href(action=>"blob_plain", -replay=>1)},
7169                                         "raw") .
7170                                 " | " .
7171                                 $cgi->a({-href => href(action=>"blob",
7172                                                        hash_base=>"HEAD", file_name=>$file_name)},
7173                                         "HEAD");
7174                 } else {
7175                         $formats_nav .=
7176                                 $cgi->a({-href => href(action=>"blob_plain", -replay=>1)},
7177                                         "raw");
7178                 }
7179                 git_print_page_nav('','', $hash_base,$co{'tree'},$hash_base, $formats_nav);
7180                 git_print_header_div('commit', esc_html($co{'title'}), $hash_base);
7181         } else {
7182                 print "<div class=\"page_nav\">\n" .
7183                       "<br/><br/></div>\n" .
7184                       "<div class=\"title\">".esc_html($hash)."</div>\n";
7185         }
7186         git_print_page_path($file_name, "blob", $hash_base);
7187         print "<div class=\"page_body\">\n";
7188         if ($mimetype =~ m!^image/!) {
7189                 print qq!<img class="blob" type="!.esc_attr($mimetype).qq!"!;
7190                 if ($file_name) {
7191                         print qq! alt="!.esc_attr($file_name).qq!" title="!.esc_attr($file_name).qq!"!;
7192                 }
7193                 print qq! src="! .
7194                       esc_attr(href(action=>"blob_plain", hash=>$hash,
7195                            hash_base=>$hash_base, file_name=>$file_name)) .
7196                       qq!" />\n!;
7197         } else {
7198                 my $nr;
7199                 while (my $line = <$fd>) {
7200                         chomp $line;
7201                         $nr++;
7202                         $line = untabify($line);
7203                         printf qq!<div class="pre"><a id="l%i" href="%s#l%i" class="linenr">%4i</a> %s</div>\n!,
7204                                $nr, esc_attr(href(-replay => 1)), $nr, $nr,
7205                                $highlight ? sanitize($line) : esc_html($line, -nbsp=>1);
7206                 }
7207         }
7208         close $fd
7209                 or print "Reading blob failed.\n";
7210         print "</div>";
7211         git_footer_html();
7212 }
7213
7214 sub git_tree {
7215         if (!defined $hash_base) {
7216                 $hash_base = "HEAD";
7217         }
7218         if (!defined $hash) {
7219                 if (defined $file_name) {
7220                         $hash = git_get_hash_by_path($hash_base, $file_name, "tree");
7221                 } else {
7222                         $hash = $hash_base;
7223                 }
7224         }
7225         die_error(404, "No such tree") unless defined($hash);
7226
7227         my $show_sizes = gitweb_check_feature('show-sizes');
7228         my $have_blame = gitweb_check_feature('blame');
7229
7230         my @entries = ();
7231         {
7232                 local $/ = "\0";
7233                 open my $fd, "-|", git_cmd(), "ls-tree", '-z',
7234                         ($show_sizes ? '-l' : ()), @extra_options, $hash
7235                         or die_error(500, "Open git-ls-tree failed");
7236                 @entries = map { chomp; $_ } <$fd>;
7237                 close $fd
7238                         or die_error(404, "Reading tree failed");
7239         }
7240
7241         my $refs = git_get_references();
7242         my $ref = format_ref_marker($refs, $hash_base);
7243         git_header_html();
7244         my $basedir = '';
7245         if (defined $hash_base && (my %co = parse_commit($hash_base))) {
7246                 my @views_nav = ();
7247                 if (defined $file_name) {
7248                         push @views_nav,
7249                                 $cgi->a({-href => href(action=>"history", -replay=>1)},
7250                                         "history"),
7251                                 $cgi->a({-href => href(action=>"tree",
7252                                                        hash_base=>"HEAD", file_name=>$file_name)},
7253                                         "HEAD"),
7254                 }
7255                 my $snapshot_links = format_snapshot_links($hash);
7256                 if (defined $snapshot_links) {
7257                         # FIXME: Should be available when we have no hash base as well.
7258                         push @views_nav, $snapshot_links;
7259                 }
7260                 git_print_page_nav('tree','', $hash_base, undef, undef,
7261                                    join(' | ', @views_nav));
7262                 git_print_header_div('commit', esc_html($co{'title'}) . $ref, $hash_base);
7263         } else {
7264                 undef $hash_base;
7265                 print "<div class=\"page_nav\">\n";
7266                 print "<br/><br/></div>\n";
7267                 print "<div class=\"title\">".esc_html($hash)."</div>\n";
7268         }
7269         if (defined $file_name) {
7270                 $basedir = $file_name;
7271                 if ($basedir ne '' && substr($basedir, -1) ne '/') {
7272                         $basedir .= '/';
7273                 }
7274                 git_print_page_path($file_name, 'tree', $hash_base);
7275         }
7276         print "<div class=\"page_body\">\n";
7277         print "<table class=\"tree\">\n";
7278         my $alternate = 1;
7279         # '..' (top directory) link if possible
7280         if (defined $hash_base &&
7281             defined $file_name && $file_name =~ m![^/]+$!) {
7282                 if ($alternate) {
7283                         print "<tr class=\"dark\">\n";
7284                 } else {
7285                         print "<tr class=\"light\">\n";
7286                 }
7287                 $alternate ^= 1;
7288
7289                 my $up = $file_name;
7290                 $up =~ s!/?[^/]+$!!;
7291                 undef $up unless $up;
7292                 # based on git_print_tree_entry
7293                 print '<td class="mode">' . mode_str('040000') . "</td>\n";
7294                 print '<td class="size">&nbsp;</td>'."\n" if $show_sizes;
7295                 print '<td class="list">';
7296                 print $cgi->a({-href => href(action=>"tree",
7297                                              hash_base=>$hash_base,
7298                                              file_name=>$up)},
7299                               "..");
7300                 print "</td>\n";
7301                 print "<td class=\"link\"></td>\n";
7302
7303                 print "</tr>\n";
7304         }
7305         foreach my $line (@entries) {
7306                 my %t = parse_ls_tree_line($line, -z => 1, -l => $show_sizes);
7307
7308                 if ($alternate) {
7309                         print "<tr class=\"dark\">\n";
7310                 } else {
7311                         print "<tr class=\"light\">\n";
7312                 }
7313                 $alternate ^= 1;
7314
7315                 git_print_tree_entry(\%t, $basedir, $hash_base, $have_blame);
7316
7317                 print "</tr>\n";
7318         }
7319         print "</table>\n" .
7320               "</div>";
7321         git_footer_html();
7322 }
7323
7324 sub sanitize_for_filename {
7325     my $name = shift;
7326
7327     $name =~ s!/!-!g;
7328     $name =~ s/[^[:alnum:]_.-]//g;
7329
7330     return $name;
7331 }
7332
7333 sub snapshot_name {
7334         my ($project, $hash) = @_;
7335
7336         # path/to/project.git  -> project
7337         # path/to/project/.git -> project
7338         my $name = to_utf8($project);
7339         $name =~ s,([^/])/*\.git$,$1,;
7340         $name = sanitize_for_filename(basename($name));
7341
7342         my $ver = $hash;
7343         if ($hash =~ /^[0-9a-fA-F]+$/) {
7344                 # shorten SHA-1 hash
7345                 my $full_hash = git_get_full_hash($project, $hash);
7346                 if ($full_hash =~ /^$hash/ && length($hash) > 7) {
7347                         $ver = git_get_short_hash($project, $hash);
7348                 }
7349         } elsif ($hash =~ m!^refs/tags/(.*)$!) {
7350                 # tags don't need shortened SHA-1 hash
7351                 $ver = $1;
7352         } else {
7353                 # branches and other need shortened SHA-1 hash
7354                 my $strip_refs = join '|', map { quotemeta } get_branch_refs();
7355                 if ($hash =~ m!^refs/($strip_refs|remotes)/(.*)$!) {
7356                         my $ref_dir = (defined $1) ? $1 : '';
7357                         $ver = $2;
7358
7359                         $ref_dir = sanitize_for_filename($ref_dir);
7360                         # for refs neither in heads nor remotes we want to
7361                         # add a ref dir to archive name
7362                         if ($ref_dir ne '' and $ref_dir ne 'heads' and $ref_dir ne 'remotes') {
7363                                 $ver = $ref_dir . '-' . $ver;
7364                         }
7365                 }
7366                 $ver .= '-' . git_get_short_hash($project, $hash);
7367         }
7368         # special case of sanitization for filename - we change
7369         # slashes to dots instead of dashes
7370         # in case of hierarchical branch names
7371         $ver =~ s!/!.!g;
7372         $ver =~ s/[^[:alnum:]_.-]//g;
7373
7374         # name = project-version_string
7375         $name = "$name-$ver";
7376
7377         return wantarray ? ($name, $name) : $name;
7378 }
7379
7380 sub exit_if_unmodified_since {
7381         my ($latest_epoch) = @_;
7382         our $cgi;
7383
7384         my $if_modified = $cgi->http('IF_MODIFIED_SINCE');
7385         if (defined $if_modified) {
7386                 my $since;
7387                 if (eval { require HTTP::Date; 1; }) {
7388                         $since = HTTP::Date::str2time($if_modified);
7389                 } elsif (eval { require Time::ParseDate; 1; }) {
7390                         $since = Time::ParseDate::parsedate($if_modified, GMT => 1);
7391                 }
7392                 if (defined $since && $latest_epoch <= $since) {
7393                         my %latest_date = parse_date($latest_epoch);
7394                         print $cgi->header(
7395                                 -last_modified => $latest_date{'rfc2822'},
7396                                 -status => '304 Not Modified');
7397                         goto DONE_GITWEB;
7398                 }
7399         }
7400 }
7401
7402 sub git_snapshot {
7403         my $format = $input_params{'snapshot_format'};
7404         if (!@snapshot_fmts) {
7405                 die_error(403, "Snapshots not allowed");
7406         }
7407         # default to first supported snapshot format
7408         $format ||= $snapshot_fmts[0];
7409         if ($format !~ m/^[a-z0-9]+$/) {
7410                 die_error(400, "Invalid snapshot format parameter");
7411         } elsif (!exists($known_snapshot_formats{$format})) {
7412                 die_error(400, "Unknown snapshot format");
7413         } elsif ($known_snapshot_formats{$format}{'disabled'}) {
7414                 die_error(403, "Snapshot format not allowed");
7415         } elsif (!grep($_ eq $format, @snapshot_fmts)) {
7416                 die_error(403, "Unsupported snapshot format");
7417         }
7418
7419         my $type = git_get_type("$hash^{}");
7420         if (!$type) {
7421                 die_error(404, 'Object does not exist');
7422         }  elsif ($type eq 'blob') {
7423                 die_error(400, 'Object is not a tree-ish');
7424         }
7425
7426         my ($name, $prefix) = snapshot_name($project, $hash);
7427         my $filename = "$name$known_snapshot_formats{$format}{'suffix'}";
7428
7429         my %co = parse_commit($hash);
7430         exit_if_unmodified_since($co{'committer_epoch'}) if %co;
7431
7432         my $cmd = quote_command(
7433                 git_cmd(), 'archive',
7434                 "--format=$known_snapshot_formats{$format}{'format'}",
7435                 "--prefix=$prefix/", $hash);
7436         if (exists $known_snapshot_formats{$format}{'compressor'}) {
7437                 $cmd .= ' | ' . quote_command(@{$known_snapshot_formats{$format}{'compressor'}});
7438         }
7439
7440         $filename =~ s/(["\\])/\\$1/g;
7441         my %latest_date;
7442         if (%co) {
7443                 %latest_date = parse_date($co{'committer_epoch'}, $co{'committer_tz'});
7444         }
7445
7446         print $cgi->header(
7447                 -type => $known_snapshot_formats{$format}{'type'},
7448                 -content_disposition => 'inline; filename="' . $filename . '"',
7449                 %co ? (-last_modified => $latest_date{'rfc2822'}) : (),
7450                 -status => '200 OK');
7451
7452         open my $fd, "-|", $cmd
7453                 or die_error(500, "Execute git-archive failed");
7454         local *FCGI::Stream::PRINT = $FCGI_Stream_PRINT_raw;
7455         binmode STDOUT, ':raw';
7456         print <$fd>;
7457         binmode STDOUT, ':utf8'; # as set at the beginning of gitweb.cgi
7458         close $fd;
7459 }
7460
7461 sub git_log_generic {
7462         my ($fmt_name, $body_subr, $base, $parent, $file_name, $file_hash) = @_;
7463
7464         my $head = git_get_head_hash($project);
7465         if (!defined $base) {
7466                 $base = $head;
7467         }
7468         if (!defined $page) {
7469                 $page = 0;
7470         }
7471         my $refs = git_get_references();
7472
7473         my $commit_hash = $base;
7474         if (defined $parent) {
7475                 $commit_hash = "$parent..$base";
7476         }
7477         my @commitlist =
7478                 parse_commits($commit_hash, 101, (100 * $page),
7479                               defined $file_name ? ($file_name, "--full-history") : ());
7480
7481         my $ftype;
7482         if (!defined $file_hash && defined $file_name) {
7483                 # some commits could have deleted file in question,
7484                 # and not have it in tree, but one of them has to have it
7485                 for (my $i = 0; $i < @commitlist; $i++) {
7486                         $file_hash = git_get_hash_by_path($commitlist[$i]{'id'}, $file_name);
7487                         last if defined $file_hash;
7488                 }
7489         }
7490         if (defined $file_hash) {
7491                 $ftype = git_get_type($file_hash);
7492         }
7493         if (defined $file_name && !defined $ftype) {
7494                 die_error(500, "Unknown type of object");
7495         }
7496         my %co;
7497         if (defined $file_name) {
7498                 %co = parse_commit($base)
7499                         or die_error(404, "Unknown commit object");
7500         }
7501
7502
7503         my $paging_nav = format_paging_nav($fmt_name, $page, $#commitlist >= 100);
7504         my $next_link = '';
7505         if ($#commitlist >= 100) {
7506                 $next_link =
7507                         $cgi->a({-href => href(-replay=>1, page=>$page+1),
7508                                  -accesskey => "n", -title => "Alt-n"}, "next");
7509         }
7510         my $patch_max = gitweb_get_feature('patches');
7511         if ($patch_max && !defined $file_name &&
7512                 !gitweb_check_feature('email-privacy')) {
7513                 if ($patch_max < 0 || @commitlist <= $patch_max) {
7514                         $paging_nav .= " &sdot; " .
7515                                 $cgi->a({-href => href(action=>"patches", -replay=>1)},
7516                                         "patches");
7517                 }
7518         }
7519
7520         git_header_html();
7521         git_print_page_nav($fmt_name,'', $hash,$hash,$hash, $paging_nav);
7522         if (defined $file_name) {
7523                 git_print_header_div('commit', esc_html($co{'title'}), $base);
7524         } else {
7525                 git_print_header_div('summary', $project)
7526         }
7527         git_print_page_path($file_name, $ftype, $hash_base)
7528                 if (defined $file_name);
7529
7530         $body_subr->(\@commitlist, 0, 99, $refs, $next_link,
7531                      $file_name, $file_hash, $ftype);
7532
7533         git_footer_html();
7534 }
7535
7536 sub git_log {
7537         git_log_generic('log', \&git_log_body,
7538                         $hash, $hash_parent);
7539 }
7540
7541 sub git_commit {
7542         $hash ||= $hash_base || "HEAD";
7543         my %co = parse_commit($hash)
7544             or die_error(404, "Unknown commit object");
7545
7546         my $parent  = $co{'parent'};
7547         my $parents = $co{'parents'}; # listref
7548
7549         # we need to prepare $formats_nav before any parameter munging
7550         my $formats_nav;
7551         if (!defined $parent) {
7552                 # --root commitdiff
7553                 $formats_nav .= '(initial)';
7554         } elsif (@$parents == 1) {
7555                 # single parent commit
7556                 $formats_nav .=
7557                         '(parent: ' .
7558                         $cgi->a({-href => href(action=>"commit",
7559                                                hash=>$parent)},
7560                                 esc_html(substr($parent, 0, 7))) .
7561                         ')';
7562         } else {
7563                 # merge commit
7564                 $formats_nav .=
7565                         '(merge: ' .
7566                         join(' ', map {
7567                                 $cgi->a({-href => href(action=>"commit",
7568                                                        hash=>$_)},
7569                                         esc_html(substr($_, 0, 7)));
7570                         } @$parents ) .
7571                         ')';
7572         }
7573         if (gitweb_check_feature('patches') && @$parents <= 1 &&
7574                 !gitweb_check_feature('email-privacy')) {
7575                 $formats_nav .= " | " .
7576                         $cgi->a({-href => href(action=>"patch", -replay=>1)},
7577                                 "patch");
7578         }
7579
7580         if (!defined $parent) {
7581                 $parent = "--root";
7582         }
7583         my @difftree;
7584         open my $fd, "-|", git_cmd(), "diff-tree", '-r', "--no-commit-id",
7585                 @diff_opts,
7586                 (@$parents <= 1 ? $parent : '-c'),
7587                 $hash, "--"
7588                 or die_error(500, "Open git-diff-tree failed");
7589         @difftree = map { chomp; $_ } <$fd>;
7590         close $fd or die_error(404, "Reading git-diff-tree failed");
7591
7592         # non-textual hash id's can be cached
7593         my $expires;
7594         if ($hash =~ m/^$oid_regex$/) {
7595                 $expires = "+1d";
7596         }
7597         my $refs = git_get_references();
7598         my $ref = format_ref_marker($refs, $co{'id'});
7599
7600         git_header_html(undef, $expires);
7601         git_print_page_nav('commit', '',
7602                            $hash, $co{'tree'}, $hash,
7603                            $formats_nav);
7604
7605         if (defined $co{'parent'}) {
7606                 git_print_header_div('commitdiff', esc_html($co{'title'}) . $ref, $hash);
7607         } else {
7608                 git_print_header_div('tree', esc_html($co{'title'}) . $ref, $co{'tree'}, $hash);
7609         }
7610         print "<div class=\"title_text\">\n" .
7611               "<table class=\"object_header\">\n";
7612         git_print_authorship_rows(\%co);
7613         print "<tr><td>commit</td><td class=\"sha1\">$co{'id'}</td></tr>\n";
7614         print "<tr>" .
7615               "<td>tree</td>" .
7616               "<td class=\"sha1\">" .
7617               $cgi->a({-href => href(action=>"tree", hash=>$co{'tree'}, hash_base=>$hash),
7618                        class => "list"}, $co{'tree'}) .
7619               "</td>" .
7620               "<td class=\"link\">" .
7621               $cgi->a({-href => href(action=>"tree", hash=>$co{'tree'}, hash_base=>$hash)},
7622                       "tree");
7623         my $snapshot_links = format_snapshot_links($hash);
7624         if (defined $snapshot_links) {
7625                 print " | " . $snapshot_links;
7626         }
7627         print "</td>" .
7628               "</tr>\n";
7629
7630         foreach my $par (@$parents) {
7631                 print "<tr>" .
7632                       "<td>parent</td>" .
7633                       "<td class=\"sha1\">" .
7634                       $cgi->a({-href => href(action=>"commit", hash=>$par),
7635                                class => "list"}, $par) .
7636                       "</td>" .
7637                       "<td class=\"link\">" .
7638                       $cgi->a({-href => href(action=>"commit", hash=>$par)}, "commit") .
7639                       " | " .
7640                       $cgi->a({-href => href(action=>"commitdiff", hash=>$hash, hash_parent=>$par)}, "diff") .
7641                       "</td>" .
7642                       "</tr>\n";
7643         }
7644         print "</table>".
7645               "</div>\n";
7646
7647         print "<div class=\"page_body\">\n";
7648         git_print_log($co{'comment'});
7649         print "</div>\n";
7650
7651         git_difftree_body(\@difftree, $hash, @$parents);
7652
7653         git_footer_html();
7654 }
7655
7656 sub git_object {
7657         # object is defined by:
7658         # - hash or hash_base alone
7659         # - hash_base and file_name
7660         my $type;
7661
7662         # - hash or hash_base alone
7663         if ($hash || ($hash_base && !defined $file_name)) {
7664                 my $object_id = $hash || $hash_base;
7665
7666                 open my $fd, "-|", quote_command(
7667                         git_cmd(), 'cat-file', '-t', $object_id) . ' 2> /dev/null'
7668                         or die_error(404, "Object does not exist");
7669                 $type = <$fd>;
7670                 defined $type && chomp $type;
7671                 close $fd
7672                         or die_error(404, "Object does not exist");
7673
7674         # - hash_base and file_name
7675         } elsif ($hash_base && defined $file_name) {
7676                 $file_name =~ s,/+$,,;
7677
7678                 system(git_cmd(), "cat-file", '-e', $hash_base) == 0
7679                         or die_error(404, "Base object does not exist");
7680
7681                 # here errors should not happen
7682                 open my $fd, "-|", git_cmd(), "ls-tree", $hash_base, "--", $file_name
7683                         or die_error(500, "Open git-ls-tree failed");
7684                 my $line = <$fd>;
7685                 close $fd;
7686
7687                 #'100644 blob 0fa3f3a66fb6a137f6ec2c19351ed4d807070ffa  panic.c'
7688                 unless ($line && $line =~ m/^([0-9]+) (.+) ($oid_regex)\t/) {
7689                         die_error(404, "File or directory for given base does not exist");
7690                 }
7691                 $type = $2;
7692                 $hash = $3;
7693         } else {
7694                 die_error(400, "Not enough information to find object");
7695         }
7696
7697         print $cgi->redirect(-uri => href(action=>$type, -full=>1,
7698                                           hash=>$hash, hash_base=>$hash_base,
7699                                           file_name=>$file_name),
7700                              -status => '302 Found');
7701 }
7702
7703 sub git_blobdiff {
7704         my $format = shift || 'html';
7705         my $diff_style = $input_params{'diff_style'} || 'inline';
7706
7707         my $fd;
7708         my @difftree;
7709         my %diffinfo;
7710         my $expires;
7711
7712         # preparing $fd and %diffinfo for git_patchset_body
7713         # new style URI
7714         if (defined $hash_base && defined $hash_parent_base) {
7715                 if (defined $file_name) {
7716                         # read raw output
7717                         open $fd, "-|", git_cmd(), "diff-tree", '-r', @diff_opts,
7718                                 $hash_parent_base, $hash_base,
7719                                 "--", (defined $file_parent ? $file_parent : ()), $file_name
7720                                 or die_error(500, "Open git-diff-tree failed");
7721                         @difftree = map { chomp; $_ } <$fd>;
7722                         close $fd
7723                                 or die_error(404, "Reading git-diff-tree failed");
7724                         @difftree
7725                                 or die_error(404, "Blob diff not found");
7726
7727                 } elsif (defined $hash &&
7728                          $hash =~ $oid_regex) {
7729                         # try to find filename from $hash
7730
7731                         # read filtered raw output
7732                         open $fd, "-|", git_cmd(), "diff-tree", '-r', @diff_opts,
7733                                 $hash_parent_base, $hash_base, "--"
7734                                 or die_error(500, "Open git-diff-tree failed");
7735                         @difftree =
7736                                 # ':100644 100644 03b21826... 3b93d5e7... M     ls-files.c'
7737                                 # $hash == to_id
7738                                 grep { /^:[0-7]{6} [0-7]{6} $oid_regex $hash/ }
7739                                 map { chomp; $_ } <$fd>;
7740                         close $fd
7741                                 or die_error(404, "Reading git-diff-tree failed");
7742                         @difftree
7743                                 or die_error(404, "Blob diff not found");
7744
7745                 } else {
7746                         die_error(400, "Missing one of the blob diff parameters");
7747                 }
7748
7749                 if (@difftree > 1) {
7750                         die_error(400, "Ambiguous blob diff specification");
7751                 }
7752
7753                 %diffinfo = parse_difftree_raw_line($difftree[0]);
7754                 $file_parent ||= $diffinfo{'from_file'} || $file_name;
7755                 $file_name   ||= $diffinfo{'to_file'};
7756
7757                 $hash_parent ||= $diffinfo{'from_id'};
7758                 $hash        ||= $diffinfo{'to_id'};
7759
7760                 # non-textual hash id's can be cached
7761                 if ($hash_base =~ m/^$oid_regex$/ &&
7762                     $hash_parent_base =~ m/^$oid_regex$/) {
7763                         $expires = '+1d';
7764                 }
7765
7766                 # open patch output
7767                 open $fd, "-|", git_cmd(), "diff-tree", '-r', @diff_opts,
7768                         '-p', ($format eq 'html' ? "--full-index" : ()),
7769                         $hash_parent_base, $hash_base,
7770                         "--", (defined $file_parent ? $file_parent : ()), $file_name
7771                         or die_error(500, "Open git-diff-tree failed");
7772         }
7773
7774         # old/legacy style URI -- not generated anymore since 1.4.3.
7775         if (!%diffinfo) {
7776                 die_error('404 Not Found', "Missing one of the blob diff parameters")
7777         }
7778
7779         # header
7780         if ($format eq 'html') {
7781                 my $formats_nav =
7782                         $cgi->a({-href => href(action=>"blobdiff_plain", -replay=>1)},
7783                                 "raw");
7784                 $formats_nav .= diff_style_nav($diff_style);
7785                 git_header_html(undef, $expires);
7786                 if (defined $hash_base && (my %co = parse_commit($hash_base))) {
7787                         git_print_page_nav('','', $hash_base,$co{'tree'},$hash_base, $formats_nav);
7788                         git_print_header_div('commit', esc_html($co{'title'}), $hash_base);
7789                 } else {
7790                         print "<div class=\"page_nav\"><br/>$formats_nav<br/></div>\n";
7791                         print "<div class=\"title\">".esc_html("$hash vs $hash_parent")."</div>\n";
7792                 }
7793                 if (defined $file_name) {
7794                         git_print_page_path($file_name, "blob", $hash_base);
7795                 } else {
7796                         print "<div class=\"page_path\"></div>\n";
7797                 }
7798
7799         } elsif ($format eq 'plain') {
7800                 print $cgi->header(
7801                         -type => 'text/plain',
7802                         -charset => 'utf-8',
7803                         -expires => $expires,
7804                         -content_disposition => 'inline; filename="' . "$file_name" . '.patch"');
7805
7806                 print "X-Git-Url: " . $cgi->self_url() . "\n\n";
7807
7808         } else {
7809                 die_error(400, "Unknown blobdiff format");
7810         }
7811
7812         # patch
7813         if ($format eq 'html') {
7814                 print "<div class=\"page_body\">\n";
7815
7816                 git_patchset_body($fd, $diff_style,
7817                                   [ \%diffinfo ], $hash_base, $hash_parent_base);
7818                 close $fd;
7819
7820                 print "</div>\n"; # class="page_body"
7821                 git_footer_html();
7822
7823         } else {
7824                 while (my $line = <$fd>) {
7825                         $line =~ s!a/($hash|$hash_parent)!'a/'.esc_path($diffinfo{'from_file'})!eg;
7826                         $line =~ s!b/($hash|$hash_parent)!'b/'.esc_path($diffinfo{'to_file'})!eg;
7827
7828                         print $line;
7829
7830                         last if $line =~ m!^\+\+\+!;
7831                 }
7832                 local $/ = undef;
7833                 print <$fd>;
7834                 close $fd;
7835         }
7836 }
7837
7838 sub git_blobdiff_plain {
7839         git_blobdiff('plain');
7840 }
7841
7842 # assumes that it is added as later part of already existing navigation,
7843 # so it returns "| foo | bar" rather than just "foo | bar"
7844 sub diff_style_nav {
7845         my ($diff_style, $is_combined) = @_;
7846         $diff_style ||= 'inline';
7847
7848         return "" if ($is_combined);
7849
7850         my @styles = (inline => 'inline', 'sidebyside' => 'side by side');
7851         my %styles = @styles;
7852         @styles =
7853                 @styles[ map { $_ * 2 } 0..$#styles/2 ];
7854
7855         return join '',
7856                 map { " | ".$_ }
7857                 map {
7858                         $_ eq $diff_style ? $styles{$_} :
7859                         $cgi->a({-href => href(-replay=>1, diff_style => $_)}, $styles{$_})
7860                 } @styles;
7861 }
7862
7863 sub git_commitdiff {
7864         my %params = @_;
7865         my $format = $params{-format} || 'html';
7866         my $diff_style = $input_params{'diff_style'} || 'inline';
7867
7868         my ($patch_max) = gitweb_get_feature('patches');
7869         if ($format eq 'patch') {
7870                 die_error(403, "Patch view not allowed") unless $patch_max;
7871         }
7872
7873         $hash ||= $hash_base || "HEAD";
7874         my %co = parse_commit($hash)
7875             or die_error(404, "Unknown commit object");
7876
7877         # choose format for commitdiff for merge
7878         if (! defined $hash_parent && @{$co{'parents'}} > 1) {
7879                 $hash_parent = '--cc';
7880         }
7881         # we need to prepare $formats_nav before almost any parameter munging
7882         my $formats_nav;
7883         if ($format eq 'html') {
7884                 $formats_nav =
7885                         $cgi->a({-href => href(action=>"commitdiff_plain", -replay=>1)},
7886                                 "raw");
7887                 if ($patch_max && @{$co{'parents'}} <= 1 &&
7888                         !gitweb_check_feature('email-privacy')) {
7889                         $formats_nav .= " | " .
7890                                 $cgi->a({-href => href(action=>"patch", -replay=>1)},
7891                                         "patch");
7892                 }
7893                 $formats_nav .= diff_style_nav($diff_style, @{$co{'parents'}} > 1);
7894
7895                 if (defined $hash_parent &&
7896                     $hash_parent ne '-c' && $hash_parent ne '--cc') {
7897                         # commitdiff with two commits given
7898                         my $hash_parent_short = $hash_parent;
7899                         if ($hash_parent =~ m/^$oid_regex$/) {
7900                                 $hash_parent_short = substr($hash_parent, 0, 7);
7901                         }
7902                         $formats_nav .=
7903                                 ' (from';
7904                         for (my $i = 0; $i < @{$co{'parents'}}; $i++) {
7905                                 if ($co{'parents'}[$i] eq $hash_parent) {
7906                                         $formats_nav .= ' parent ' . ($i+1);
7907                                         last;
7908                                 }
7909                         }
7910                         $formats_nav .= ': ' .
7911                                 $cgi->a({-href => href(-replay=>1,
7912                                                        hash=>$hash_parent, hash_base=>undef)},
7913                                         esc_html($hash_parent_short)) .
7914                                 ')';
7915                 } elsif (!$co{'parent'}) {
7916                         # --root commitdiff
7917                         $formats_nav .= ' (initial)';
7918                 } elsif (scalar @{$co{'parents'}} == 1) {
7919                         # single parent commit
7920                         $formats_nav .=
7921                                 ' (parent: ' .
7922                                 $cgi->a({-href => href(-replay=>1,
7923                                                        hash=>$co{'parent'}, hash_base=>undef)},
7924                                         esc_html(substr($co{'parent'}, 0, 7))) .
7925                                 ')';
7926                 } else {
7927                         # merge commit
7928                         if ($hash_parent eq '--cc') {
7929                                 $formats_nav .= ' | ' .
7930                                         $cgi->a({-href => href(-replay=>1,
7931                                                                hash=>$hash, hash_parent=>'-c')},
7932                                                 'combined');
7933                         } else { # $hash_parent eq '-c'
7934                                 $formats_nav .= ' | ' .
7935                                         $cgi->a({-href => href(-replay=>1,
7936                                                                hash=>$hash, hash_parent=>'--cc')},
7937                                                 'compact');
7938                         }
7939                         $formats_nav .=
7940                                 ' (merge: ' .
7941                                 join(' ', map {
7942                                         $cgi->a({-href => href(-replay=>1,
7943                                                                hash=>$_, hash_base=>undef)},
7944                                                 esc_html(substr($_, 0, 7)));
7945                                 } @{$co{'parents'}} ) .
7946                                 ')';
7947                 }
7948         }
7949
7950         my $hash_parent_param = $hash_parent;
7951         if (!defined $hash_parent_param) {
7952                 # --cc for multiple parents, --root for parentless
7953                 $hash_parent_param =
7954                         @{$co{'parents'}} > 1 ? '--cc' : $co{'parent'} || '--root';
7955         }
7956
7957         # read commitdiff
7958         my $fd;
7959         my @difftree;
7960         if ($format eq 'html') {
7961                 open $fd, "-|", git_cmd(), "diff-tree", '-r', @diff_opts,
7962                         "--no-commit-id", "--patch-with-raw", "--full-index",
7963                         $hash_parent_param, $hash, "--"
7964                         or die_error(500, "Open git-diff-tree failed");
7965
7966                 while (my $line = <$fd>) {
7967                         chomp $line;
7968                         # empty line ends raw part of diff-tree output
7969                         last unless $line;
7970                         push @difftree, scalar parse_difftree_raw_line($line);
7971                 }
7972
7973         } elsif ($format eq 'plain') {
7974                 open $fd, "-|", git_cmd(), "diff-tree", '-r', @diff_opts,
7975                         '-p', $hash_parent_param, $hash, "--"
7976                         or die_error(500, "Open git-diff-tree failed");
7977         } elsif ($format eq 'patch') {
7978                 # For commit ranges, we limit the output to the number of
7979                 # patches specified in the 'patches' feature.
7980                 # For single commits, we limit the output to a single patch,
7981                 # diverging from the git-format-patch default.
7982                 my @commit_spec = ();
7983                 if ($hash_parent) {
7984                         if ($patch_max > 0) {
7985                                 push @commit_spec, "-$patch_max";
7986                         }
7987                         push @commit_spec, '-n', "$hash_parent..$hash";
7988                 } else {
7989                         if ($params{-single}) {
7990                                 push @commit_spec, '-1';
7991                         } else {
7992                                 if ($patch_max > 0) {
7993                                         push @commit_spec, "-$patch_max";
7994                                 }
7995                                 push @commit_spec, "-n";
7996                         }
7997                         push @commit_spec, '--root', $hash;
7998                 }
7999                 open $fd, "-|", git_cmd(), "format-patch", @diff_opts,
8000                         '--encoding=utf8', '--stdout', @commit_spec
8001                         or die_error(500, "Open git-format-patch failed");
8002         } else {
8003                 die_error(400, "Unknown commitdiff format");
8004         }
8005
8006         # non-textual hash id's can be cached
8007         my $expires;
8008         if ($hash =~ m/^$oid_regex$/) {
8009                 $expires = "+1d";
8010         }
8011
8012         # write commit message
8013         if ($format eq 'html') {
8014                 my $refs = git_get_references();
8015                 my $ref = format_ref_marker($refs, $co{'id'});
8016
8017                 git_header_html(undef, $expires);
8018                 git_print_page_nav('commitdiff','', $hash,$co{'tree'},$hash, $formats_nav);
8019                 git_print_header_div('commit', esc_html($co{'title'}) . $ref, $hash);
8020                 print "<div class=\"title_text\">\n" .
8021                       "<table class=\"object_header\">\n";
8022                 git_print_authorship_rows(\%co);
8023                 print "</table>".
8024                       "</div>\n";
8025                 print "<div class=\"page_body\">\n";
8026                 if (@{$co{'comment'}} > 1) {
8027                         print "<div class=\"log\">\n";
8028                         git_print_log($co{'comment'}, -final_empty_line=> 1, -remove_title => 1);
8029                         print "</div>\n"; # class="log"
8030                 }
8031
8032         } elsif ($format eq 'plain') {
8033                 my $refs = git_get_references("tags");
8034                 my $tagname = git_get_rev_name_tags($hash);
8035                 my $filename = basename($project) . "-$hash.patch";
8036
8037                 print $cgi->header(
8038                         -type => 'text/plain',
8039                         -charset => 'utf-8',
8040                         -expires => $expires,
8041                         -content_disposition => 'inline; filename="' . "$filename" . '"');
8042                 my %ad = parse_date($co{'author_epoch'}, $co{'author_tz'});
8043                 print "From: " . to_utf8($co{'author'}) . "\n";
8044                 print "Date: $ad{'rfc2822'} ($ad{'tz_local'})\n";
8045                 print "Subject: " . to_utf8($co{'title'}) . "\n";
8046
8047                 print "X-Git-Tag: $tagname\n" if $tagname;
8048                 print "X-Git-Url: " . $cgi->self_url() . "\n\n";
8049
8050                 foreach my $line (@{$co{'comment'}}) {
8051                         print to_utf8($line) . "\n";
8052                 }
8053                 print "---\n\n";
8054         } elsif ($format eq 'patch') {
8055                 my $filename = basename($project) . "-$hash.patch";
8056
8057                 print $cgi->header(
8058                         -type => 'text/plain',
8059                         -charset => 'utf-8',
8060                         -expires => $expires,
8061                         -content_disposition => 'inline; filename="' . "$filename" . '"');
8062         }
8063
8064         # write patch
8065         if ($format eq 'html') {
8066                 my $use_parents = !defined $hash_parent ||
8067                         $hash_parent eq '-c' || $hash_parent eq '--cc';
8068                 git_difftree_body(\@difftree, $hash,
8069                                   $use_parents ? @{$co{'parents'}} : $hash_parent);
8070                 print "<br/>\n";
8071
8072                 git_patchset_body($fd, $diff_style,
8073                                   \@difftree, $hash,
8074                                   $use_parents ? @{$co{'parents'}} : $hash_parent);
8075                 close $fd;
8076                 print "</div>\n"; # class="page_body"
8077                 git_footer_html();
8078
8079         } elsif ($format eq 'plain') {
8080                 local $/ = undef;
8081                 print <$fd>;
8082                 close $fd
8083                         or print "Reading git-diff-tree failed\n";
8084         } elsif ($format eq 'patch') {
8085                 local $/ = undef;
8086                 print <$fd>;
8087                 close $fd
8088                         or print "Reading git-format-patch failed\n";
8089         }
8090 }
8091
8092 sub git_commitdiff_plain {
8093         git_commitdiff(-format => 'plain');
8094 }
8095
8096 # format-patch-style patches
8097 sub git_patch {
8098         git_commitdiff(-format => 'patch', -single => 1);
8099 }
8100
8101 sub git_patches {
8102         git_commitdiff(-format => 'patch');
8103 }
8104
8105 sub git_history {
8106         git_log_generic('history', \&git_history_body,
8107                         $hash_base, $hash_parent_base,
8108                         $file_name, $hash);
8109 }
8110
8111 sub git_search {
8112         $searchtype ||= 'commit';
8113
8114         # check if appropriate features are enabled
8115         gitweb_check_feature('search')
8116                 or die_error(403, "Search is disabled");
8117         if ($searchtype eq 'pickaxe') {
8118                 # pickaxe may take all resources of your box and run for several minutes
8119                 # with every query - so decide by yourself how public you make this feature
8120                 gitweb_check_feature('pickaxe')
8121                         or die_error(403, "Pickaxe search is disabled");
8122         }
8123         if ($searchtype eq 'grep') {
8124                 # grep search might be potentially CPU-intensive, too
8125                 gitweb_check_feature('grep')
8126                         or die_error(403, "Grep search is disabled");
8127         }
8128
8129         if (!defined $searchtext) {
8130                 die_error(400, "Text field is empty");
8131         }
8132         if (!defined $hash) {
8133                 $hash = git_get_head_hash($project);
8134         }
8135         my %co = parse_commit($hash);
8136         if (!%co) {
8137                 die_error(404, "Unknown commit object");
8138         }
8139         if (!defined $page) {
8140                 $page = 0;
8141         }
8142
8143         if ($searchtype eq 'commit' ||
8144             $searchtype eq 'author' ||
8145             $searchtype eq 'committer') {
8146                 git_search_message(%co);
8147         } elsif ($searchtype eq 'pickaxe') {
8148                 git_search_changes(%co);
8149         } elsif ($searchtype eq 'grep') {
8150                 git_search_files(%co);
8151         } else {
8152                 die_error(400, "Unknown search type");
8153         }
8154 }
8155
8156 sub git_search_help {
8157         git_header_html();
8158         git_print_page_nav('','', $hash,$hash,$hash);
8159         print <<EOT;
8160 <p><strong>Pattern</strong> is by default a normal string that is matched precisely (but without
8161 regard to case, except in the case of pickaxe). However, when you check the <em>re</em> checkbox,
8162 the pattern entered is recognized as the POSIX extended
8163 <a href="https://en.wikipedia.org/wiki/Regular_expression">regular expression</a> (also case
8164 insensitive).</p>
8165 <dl>
8166 <dt><b>commit</b></dt>
8167 <dd>The commit messages and authorship information will be scanned for the given pattern.</dd>
8168 EOT
8169         my $have_grep = gitweb_check_feature('grep');
8170         if ($have_grep) {
8171                 print <<EOT;
8172 <dt><b>grep</b></dt>
8173 <dd>All files in the currently selected tree (HEAD unless you are explicitly browsing
8174     a different one) are searched for the given pattern. On large trees, this search can take
8175 a while and put some strain on the server, so please use it with some consideration. Note that
8176 due to git-grep peculiarity, currently if regexp mode is turned off, the matches are
8177 case-sensitive.</dd>
8178 EOT
8179         }
8180         print <<EOT;
8181 <dt><b>author</b></dt>
8182 <dd>Name and e-mail of the change author and date of birth of the patch will be scanned for the given pattern.</dd>
8183 <dt><b>committer</b></dt>
8184 <dd>Name and e-mail of the committer and date of commit will be scanned for the given pattern.</dd>
8185 EOT
8186         my $have_pickaxe = gitweb_check_feature('pickaxe');
8187         if ($have_pickaxe) {
8188                 print <<EOT;
8189 <dt><b>pickaxe</b></dt>
8190 <dd>All commits that caused the string to appear or disappear from any file (changes that
8191 added, removed or "modified" the string) will be listed. This search can take a while and
8192 takes a lot of strain on the server, so please use it wisely. Note that since you may be
8193 interested even in changes just changing the case as well, this search is case sensitive.</dd>
8194 EOT
8195         }
8196         print "</dl>\n";
8197         git_footer_html();
8198 }
8199
8200 sub git_shortlog {
8201         git_log_generic('shortlog', \&git_shortlog_body,
8202                         $hash, $hash_parent);
8203 }
8204
8205 ## ......................................................................
8206 ## feeds (RSS, Atom; OPML)
8207
8208 sub git_feed {
8209         my $format = shift || 'atom';
8210         my $have_blame = gitweb_check_feature('blame');
8211
8212         # Atom: http://www.atomenabled.org/developers/syndication/
8213         # RSS:  http://www.notestips.com/80256B3A007F2692/1/NAMO5P9UPQ
8214         if ($format ne 'rss' && $format ne 'atom') {
8215                 die_error(400, "Unknown web feed format");
8216         }
8217
8218         # log/feed of current (HEAD) branch, log of given branch, history of file/directory
8219         my $head = $hash || 'HEAD';
8220         my @commitlist = parse_commits($head, 150, 0, $file_name);
8221
8222         my %latest_commit;
8223         my %latest_date;
8224         my $content_type = "application/$format+xml";
8225         if (defined $cgi->http('HTTP_ACCEPT') &&
8226                  $cgi->Accept('text/xml') > $cgi->Accept($content_type)) {
8227                 # browser (feed reader) prefers text/xml
8228                 $content_type = 'text/xml';
8229         }
8230         if (defined($commitlist[0])) {
8231                 %latest_commit = %{$commitlist[0]};
8232                 my $latest_epoch = $latest_commit{'committer_epoch'};
8233                 exit_if_unmodified_since($latest_epoch);
8234                 %latest_date = parse_date($latest_epoch, $latest_commit{'committer_tz'});
8235         }
8236         print $cgi->header(
8237                 -type => $content_type,
8238                 -charset => 'utf-8',
8239                 %latest_date ? (-last_modified => $latest_date{'rfc2822'}) : (),
8240                 -status => '200 OK');
8241
8242         # Optimization: skip generating the body if client asks only
8243         # for Last-Modified date.
8244         return if ($cgi->request_method() eq 'HEAD');
8245
8246         # header variables
8247         my $title = "$site_name - $project/$action";
8248         my $feed_type = 'log';
8249         if (defined $hash) {
8250                 $title .= " - '$hash'";
8251                 $feed_type = 'branch log';
8252                 if (defined $file_name) {
8253                         $title .= " :: $file_name";
8254                         $feed_type = 'history';
8255                 }
8256         } elsif (defined $file_name) {
8257                 $title .= " - $file_name";
8258                 $feed_type = 'history';
8259         }
8260         $title .= " $feed_type";
8261         $title = esc_html($title);
8262         my $descr = git_get_project_description($project);
8263         if (defined $descr) {
8264                 $descr = esc_html($descr);
8265         } else {
8266                 $descr = "$project " .
8267                          ($format eq 'rss' ? 'RSS' : 'Atom') .
8268                          " feed";
8269         }
8270         my $owner = git_get_project_owner($project);
8271         $owner = esc_html($owner);
8272
8273         #header
8274         my $alt_url;
8275         if (defined $file_name) {
8276                 $alt_url = href(-full=>1, action=>"history", hash=>$hash, file_name=>$file_name);
8277         } elsif (defined $hash) {
8278                 $alt_url = href(-full=>1, action=>"log", hash=>$hash);
8279         } else {
8280                 $alt_url = href(-full=>1, action=>"summary");
8281         }
8282         $alt_url = esc_attr($alt_url);
8283         print qq!<?xml version="1.0" encoding="utf-8"?>\n!;
8284         if ($format eq 'rss') {
8285                 print <<XML;
8286 <rss version="2.0" xmlns:content="http://purl.org/rss/1.0/modules/content/">
8287 <channel>
8288 XML
8289                 print "<title>$title</title>\n" .
8290                       "<link>$alt_url</link>\n" .
8291                       "<description>$descr</description>\n" .
8292                       "<language>en</language>\n" .
8293                       # project owner is responsible for 'editorial' content
8294                       "<managingEditor>$owner</managingEditor>\n";
8295                 if (defined $logo || defined $favicon) {
8296                         # prefer the logo to the favicon, since RSS
8297                         # doesn't allow both
8298                         my $img = esc_url($logo || $favicon);
8299                         print "<image>\n" .
8300                               "<url>$img</url>\n" .
8301                               "<title>$title</title>\n" .
8302                               "<link>$alt_url</link>\n" .
8303                               "</image>\n";
8304                 }
8305                 if (%latest_date) {
8306                         print "<pubDate>$latest_date{'rfc2822'}</pubDate>\n";
8307                         print "<lastBuildDate>$latest_date{'rfc2822'}</lastBuildDate>\n";
8308                 }
8309                 print "<generator>gitweb v.$version/$git_version</generator>\n";
8310         } elsif ($format eq 'atom') {
8311                 print <<XML;
8312 <feed xmlns="http://www.w3.org/2005/Atom">
8313 XML
8314                 print "<title>$title</title>\n" .
8315                       "<subtitle>$descr</subtitle>\n" .
8316                       '<link rel="alternate" type="text/html" href="' .
8317                       $alt_url . '" />' . "\n" .
8318                       '<link rel="self" type="' . $content_type . '" href="' .
8319                       $cgi->self_url() . '" />' . "\n" .
8320                       "<id>" . esc_url(href(-full=>1)) . "</id>\n" .
8321                       # use project owner for feed author
8322                       "<author><name>$owner</name></author>\n";
8323                 if (defined $favicon) {
8324                         print "<icon>" . esc_url($favicon) . "</icon>\n";
8325                 }
8326                 if (defined $logo) {
8327                         # not twice as wide as tall: 72 x 27 pixels
8328                         print "<logo>" . esc_url($logo) . "</logo>\n";
8329                 }
8330                 if (! %latest_date) {
8331                         # dummy date to keep the feed valid until commits trickle in:
8332                         print "<updated>1970-01-01T00:00:00Z</updated>\n";
8333                 } else {
8334                         print "<updated>$latest_date{'iso-8601'}</updated>\n";
8335                 }
8336                 print "<generator version='$version/$git_version'>gitweb</generator>\n";
8337         }
8338
8339         # contents
8340         for (my $i = 0; $i <= $#commitlist; $i++) {
8341                 my %co = %{$commitlist[$i]};
8342                 my $commit = $co{'id'};
8343                 # we read 150, we always show 30 and the ones more recent than 48 hours
8344                 if (($i >= 20) && ((time - $co{'author_epoch'}) > 48*60*60)) {
8345                         last;
8346                 }
8347                 my %cd = parse_date($co{'author_epoch'}, $co{'author_tz'});
8348
8349                 # get list of changed files
8350                 open my $fd, "-|", git_cmd(), "diff-tree", '-r', @diff_opts,
8351                         $co{'parent'} || "--root",
8352                         $co{'id'}, "--", (defined $file_name ? $file_name : ())
8353                         or next;
8354                 my @difftree = map { chomp; $_ } <$fd>;
8355                 close $fd
8356                         or next;
8357
8358                 # print element (entry, item)
8359                 my $co_url = href(-full=>1, action=>"commitdiff", hash=>$commit);
8360                 if ($format eq 'rss') {
8361                         print "<item>\n" .
8362                               "<title>" . esc_html($co{'title'}) . "</title>\n" .
8363                               "<author>" . esc_html($co{'author'}) . "</author>\n" .
8364                               "<pubDate>$cd{'rfc2822'}</pubDate>\n" .
8365                               "<guid isPermaLink=\"true\">$co_url</guid>\n" .
8366                               "<link>" . esc_html($co_url) . "</link>\n" .
8367                               "<description>" . esc_html($co{'title'}) . "</description>\n" .
8368                               "<content:encoded>" .
8369                               "<![CDATA[\n";
8370                 } elsif ($format eq 'atom') {
8371                         print "<entry>\n" .
8372                               "<title type=\"html\">" . esc_html($co{'title'}) . "</title>\n" .
8373                               "<updated>$cd{'iso-8601'}</updated>\n" .
8374                               "<author>\n" .
8375                               "  <name>" . esc_html($co{'author_name'}) . "</name>\n";
8376                         if ($co{'author_email'}) {
8377                                 print "  <email>" . esc_html($co{'author_email'}) . "</email>\n";
8378                         }
8379                         print "</author>\n" .
8380                               # use committer for contributor
8381                               "<contributor>\n" .
8382                               "  <name>" . esc_html($co{'committer_name'}) . "</name>\n";
8383                         if ($co{'committer_email'}) {
8384                                 print "  <email>" . esc_html($co{'committer_email'}) . "</email>\n";
8385                         }
8386                         print "</contributor>\n" .
8387                               "<published>$cd{'iso-8601'}</published>\n" .
8388                               "<link rel=\"alternate\" type=\"text/html\" href=\"" . esc_attr($co_url) . "\" />\n" .
8389                               "<id>" . esc_html($co_url) . "</id>\n" .
8390                               "<content type=\"xhtml\" xml:base=\"" . esc_url($my_url) . "\">\n" .
8391                               "<div xmlns=\"http://www.w3.org/1999/xhtml\">\n";
8392                 }
8393                 my $comment = $co{'comment'};
8394                 print "<pre>\n";
8395                 foreach my $line (@$comment) {
8396                         $line = esc_html($line);
8397                         print "$line\n";
8398                 }
8399                 print "</pre><ul>\n";
8400                 foreach my $difftree_line (@difftree) {
8401                         my %difftree = parse_difftree_raw_line($difftree_line);
8402                         next if !$difftree{'from_id'};
8403
8404                         my $file = $difftree{'file'} || $difftree{'to_file'};
8405
8406                         print "<li>" .
8407                               "[" .
8408                               $cgi->a({-href => href(-full=>1, action=>"blobdiff",
8409                                                      hash=>$difftree{'to_id'}, hash_parent=>$difftree{'from_id'},
8410                                                      hash_base=>$co{'id'}, hash_parent_base=>$co{'parent'},
8411                                                      file_name=>$file, file_parent=>$difftree{'from_file'}),
8412                                       -title => "diff"}, 'D');
8413                         if ($have_blame) {
8414                                 print $cgi->a({-href => href(-full=>1, action=>"blame",
8415                                                              file_name=>$file, hash_base=>$commit),
8416                                               -title => "blame"}, 'B');
8417                         }
8418                         # if this is not a feed of a file history
8419                         if (!defined $file_name || $file_name ne $file) {
8420                                 print $cgi->a({-href => href(-full=>1, action=>"history",
8421                                                              file_name=>$file, hash=>$commit),
8422                                               -title => "history"}, 'H');
8423                         }
8424                         $file = esc_path($file);
8425                         print "] ".
8426                               "$file</li>\n";
8427                 }
8428                 if ($format eq 'rss') {
8429                         print "</ul>]]>\n" .
8430                               "</content:encoded>\n" .
8431                               "</item>\n";
8432                 } elsif ($format eq 'atom') {
8433                         print "</ul>\n</div>\n" .
8434                               "</content>\n" .
8435                               "</entry>\n";
8436                 }
8437         }
8438
8439         # end of feed
8440         if ($format eq 'rss') {
8441                 print "</channel>\n</rss>\n";
8442         } elsif ($format eq 'atom') {
8443                 print "</feed>\n";
8444         }
8445 }
8446
8447 sub git_rss {
8448         git_feed('rss');
8449 }
8450
8451 sub git_atom {
8452         git_feed('atom');
8453 }
8454
8455 sub git_opml {
8456         my @list = git_get_projects_list($project_filter, $strict_export);
8457         if (!@list) {
8458                 die_error(404, "No projects found");
8459         }
8460
8461         print $cgi->header(
8462                 -type => 'text/xml',
8463                 -charset => 'utf-8',
8464                 -content_disposition => 'inline; filename="opml.xml"');
8465
8466         my $title = esc_html($site_name);
8467         my $filter = " within subdirectory ";
8468         if (defined $project_filter) {
8469                 $filter .= esc_html($project_filter);
8470         } else {
8471                 $filter = "";
8472         }
8473         print <<XML;
8474 <?xml version="1.0" encoding="utf-8"?>
8475 <opml version="1.0">
8476 <head>
8477   <title>$title OPML Export$filter</title>
8478 </head>
8479 <body>
8480 <outline text="git RSS feeds">
8481 XML
8482
8483         foreach my $pr (@list) {
8484                 my %proj = %$pr;
8485                 my $head = git_get_head_hash($proj{'path'});
8486                 if (!defined $head) {
8487                         next;
8488                 }
8489                 $git_dir = "$projectroot/$proj{'path'}";
8490                 my %co = parse_commit($head);
8491                 if (!%co) {
8492                         next;
8493                 }
8494
8495                 my $path = esc_html(chop_str($proj{'path'}, 25, 5));
8496                 my $rss  = esc_attr(href('project' => $proj{'path'}, 'action' => 'rss', -full => 1));
8497                 my $html = esc_attr(href('project' => $proj{'path'}, 'action' => 'summary', -full => 1));
8498                 print "<outline type=\"rss\" text=\"$path\" title=\"$path\" xmlUrl=\"$rss\" htmlUrl=\"$html\"/>\n";
8499         }
8500         print <<XML;
8501 </outline>
8502 </body>
8503 </opml>
8504 XML
8505 }