]>
Commit | Line | Data |
---|---|---|
eaf718f3 MU |
1 | #!/usr/bin/perl -w |
2 | ||
3 | # This tool is copyright (c) 2005, Matthias Urlichs. | |
4 | # It is released under the Gnu Public License, version 2. | |
5 | # | |
6 | # The basic idea is to pull and analyze SVN changes. | |
7 | # | |
4a91b796 | 8 | # Checking out the files is done by a single long-running SVN connection. |
eaf718f3 MU |
9 | # |
10 | # The head revision is on branch "origin" by default. | |
11 | # You can change that with the '-o' option. | |
12 | ||
eaf718f3 MU |
13 | use strict; |
14 | use warnings; | |
15 | use Getopt::Std; | |
d3cac2c9 | 16 | use File::Copy; |
eaf718f3 MU |
17 | use File::Spec; |
18 | use File::Temp qw(tempfile); | |
19 | use File::Path qw(mkpath); | |
20 | use File::Basename qw(basename dirname); | |
21 | use Time::Local; | |
22 | use IO::Pipe; | |
23 | use POSIX qw(strftime dup2); | |
24 | use IPC::Open2; | |
25 | use SVN::Core; | |
26 | use SVN::Ra; | |
27 | ||
f3ad0625 | 28 | die "Need SVN:Core 1.2.1 or better" if $SVN::Core::VERSION lt "1.2.1"; |
37dcf6de | 29 | |
eaf718f3 MU |
30 | $SIG{'PIPE'}="IGNORE"; |
31 | $ENV{'TZ'}="UTC"; | |
32 | ||
c55f3fff | 33 | our($opt_h,$opt_o,$opt_v,$opt_u,$opt_C,$opt_i,$opt_m,$opt_M,$opt_t,$opt_T, |
40006ea0 SK |
34 | $opt_b,$opt_r,$opt_I,$opt_A,$opt_s,$opt_l,$opt_d,$opt_D,$opt_S,$opt_F, |
35 | $opt_P,$opt_R); | |
eaf718f3 MU |
36 | |
37 | sub usage() { | |
38 | print STDERR <<END; | |
f3ad0625 | 39 | Usage: ${\basename $0} # fetch/update GIT from SVN |
40006ea0 | 40 | [-o branch-for-HEAD] [-h] [-v] [-l max_rev] [-R repack_each_revs] |
eaf718f3 | 41 | [-C GIT_repository] [-t tagname] [-T trunkname] [-b branchname] |
c55f3fff | 42 | [-d|-D] [-i] [-u] [-r] [-I ignorefilename] [-s start_chg] |
ec1e4689 | 43 | [-m] [-M regex] [-A author_file] [-S] [-F] [-P project_name] [SVN_URL] |
eaf718f3 MU |
44 | END |
45 | exit(1); | |
46 | } | |
47 | ||
40006ea0 | 48 | getopts("A:b:C:dDFhiI:l:mM:o:rs:t:T:SP:R:uv") or usage(); |
eaf718f3 MU |
49 | usage if $opt_h; |
50 | ||
51 | my $tag_name = $opt_t || "tags"; | |
0476786e | 52 | my $trunk_name = defined $opt_T ? $opt_T : "trunk"; |
eaf718f3 | 53 | my $branch_name = $opt_b || "branches"; |
ec1e4689 SK |
54 | my $project_name = $opt_P || ""; |
55 | $project_name = "/" . $project_name if ($project_name); | |
40006ea0 | 56 | my $repack_after = $opt_R || 1000; |
5d17d765 | 57 | my $root_pool = SVN::Pool->new_default; |
eaf718f3 | 58 | |
25f6f325 | 59 | @ARGV == 1 or @ARGV == 2 or usage(); |
eaf718f3 MU |
60 | |
61 | $opt_o ||= "origin"; | |
2fa92046 | 62 | $opt_s ||= 1; |
eaf718f3 MU |
63 | my $git_tree = $opt_C; |
64 | $git_tree ||= "."; | |
65 | ||
4a91b796 | 66 | my $svn_url = $ARGV[0]; |
25f6f325 | 67 | my $svn_dir = $ARGV[1]; |
eaf718f3 MU |
68 | |
69 | our @mergerx = (); | |
70 | if ($opt_m) { | |
65160b8b FF |
71 | my $branch_esc = quotemeta ($branch_name); |
72 | my $trunk_esc = quotemeta ($trunk_name); | |
73 | @mergerx = | |
74 | ( | |
75 | qr!\b(?:merg(?:ed?|ing))\b.*?\b((?:(?<=$branch_esc/)[\w\.\-]+)|(?:$trunk_esc))\b!i, | |
76 | qr!\b(?:from|of)\W+((?:(?<=$branch_esc/)[\w\.\-]+)|(?:$trunk_esc))\b!i, | |
77 | qr!\b(?:from|of)\W+(?:the )?([\w\.\-]+)[-\s]branch\b!i | |
78 | ); | |
eaf718f3 MU |
79 | } |
80 | if ($opt_M) { | |
65160b8b | 81 | unshift (@mergerx, qr/$opt_M/); |
eaf718f3 MU |
82 | } |
83 | ||
d3cac2c9 KW |
84 | # Absolutize filename now, since we will have chdir'ed by the time we |
85 | # get around to opening it. | |
86 | $opt_A = File::Spec->rel2abs($opt_A) if $opt_A; | |
87 | ||
36610b24 | 88 | our %users = (); |
d3cac2c9 KW |
89 | our $users_file = undef; |
90 | sub read_users($) { | |
91 | $users_file = File::Spec->rel2abs(@_); | |
92 | die "Cannot open $users_file\n" unless -f $users_file; | |
93 | open(my $authors,$users_file); | |
36610b24 KW |
94 | while(<$authors>) { |
95 | chomp; | |
80804d0a | 96 | next unless /^(\S+?)\s*=\s*(.+?)\s*<(.+)>\s*$/; |
36610b24 KW |
97 | (my $user,my $name,my $email) = ($1,$2,$3); |
98 | $users{$user} = [$name,$email]; | |
99 | } | |
100 | close($authors); | |
101 | } | |
102 | ||
eaf718f3 MU |
103 | select(STDERR); $|=1; select(STDOUT); |
104 | ||
105 | ||
106 | package SVNconn; | |
107 | # Basic SVN connection. | |
108 | # We're only interested in connecting and downloading, so ... | |
109 | ||
110 | use File::Spec; | |
111 | use File::Temp qw(tempfile); | |
112 | use POSIX qw(strftime dup2); | |
08ddd4f7 | 113 | use Fcntl qw(SEEK_SET); |
eaf718f3 MU |
114 | |
115 | sub new { | |
116 | my($what,$repo) = @_; | |
117 | $what=ref($what) if ref($what); | |
118 | ||
119 | my $self = {}; | |
120 | $self->{'buffer'} = ""; | |
121 | bless($self,$what); | |
122 | ||
123 | $repo =~ s#/+$##; | |
124 | $self->{'fullrep'} = $repo; | |
125 | $self->conn(); | |
126 | ||
eaf718f3 MU |
127 | return $self; |
128 | } | |
129 | ||
130 | sub conn { | |
131 | my $self = shift; | |
132 | my $repo = $self->{'fullrep'}; | |
2961e0ee EW |
133 | my $auth = SVN::Core::auth_open ([SVN::Client::get_simple_provider, |
134 | SVN::Client::get_ssl_server_trust_file_provider, | |
135 | SVN::Client::get_username_provider]); | |
5d17d765 | 136 | my $s = SVN::Ra->new(url => $repo, auth => $auth, pool => $root_pool); |
eaf718f3 MU |
137 | die "SVN connection to $repo: $!\n" unless defined $s; |
138 | $self->{'svn'} = $s; | |
139 | $self->{'repo'} = $repo; | |
140 | $self->{'maxrev'} = $s->get_latest_revnum(); | |
141 | } | |
142 | ||
143 | sub file { | |
144 | my($self,$path,$rev) = @_; | |
eaf718f3 | 145 | |
29504118 | 146 | my ($fh, $name) = tempfile('gitsvn.XXXXXX', |
eaf718f3 MU |
147 | DIR => File::Spec->tmpdir(), UNLINK => 1); |
148 | ||
2b5e63d1 | 149 | print "... $rev $path ...\n" if $opt_v; |
4802426d | 150 | my (undef, $properties); |
09c3a408 | 151 | $path =~ s#^/*##; |
5d17d765 | 152 | my $subpool = SVN::Pool::new_default_sub; |
4802426d | 153 | eval { (undef, $properties) |
5d17d765 | 154 | = $self->{'svn'}->get_file($path,$rev,$fh); }; |
25f6f325 MU |
155 | if($@) { |
156 | return undef if $@ =~ /Attempted to get checksum/; | |
157 | die $@; | |
158 | } | |
4802426d KW |
159 | my $mode; |
160 | if (exists $properties->{'svn:executable'}) { | |
08ddd4f7 HVR |
161 | $mode = '100755'; |
162 | } elsif (exists $properties->{'svn:special'}) { | |
163 | my ($special_content, $filesize); | |
164 | $filesize = tell $fh; | |
165 | seek $fh, 0, SEEK_SET; | |
166 | read $fh, $special_content, $filesize; | |
167 | if ($special_content =~ s/^link //) { | |
168 | $mode = '120000'; | |
169 | seek $fh, 0, SEEK_SET; | |
170 | truncate $fh, 0; | |
171 | print $fh $special_content; | |
172 | } else { | |
173 | die "unexpected svn:special file encountered"; | |
174 | } | |
4802426d | 175 | } else { |
08ddd4f7 | 176 | $mode = '100644'; |
4802426d | 177 | } |
eaf718f3 MU |
178 | close ($fh); |
179 | ||
4802426d | 180 | return ($name, $mode); |
eaf718f3 MU |
181 | } |
182 | ||
c55f3fff KW |
183 | sub ignore { |
184 | my($self,$path,$rev) = @_; | |
185 | ||
186 | print "... $rev $path ...\n" if $opt_v; | |
09c3a408 | 187 | $path =~ s#^/*##; |
5d17d765 | 188 | my $subpool = SVN::Pool::new_default_sub; |
c55f3fff KW |
189 | my (undef,undef,$properties) |
190 | = $self->{'svn'}->get_dir($path,$rev,undef); | |
191 | if (exists $properties->{'svn:ignore'}) { | |
192 | my ($fh, $name) = tempfile('gitsvn.XXXXXX', | |
193 | DIR => File::Spec->tmpdir(), | |
194 | UNLINK => 1); | |
195 | print $fh $properties->{'svn:ignore'}; | |
196 | close($fh); | |
197 | return $name; | |
198 | } else { | |
199 | return undef; | |
200 | } | |
201 | } | |
202 | ||
83936a29 SK |
203 | sub dir_list { |
204 | my($self,$path,$rev) = @_; | |
09c3a408 | 205 | $path =~ s#^/*##; |
5d17d765 | 206 | my $subpool = SVN::Pool::new_default_sub; |
83936a29 SK |
207 | my ($dirents,undef,$properties) |
208 | = $self->{'svn'}->get_dir($path,$rev,undef); | |
209 | return $dirents; | |
210 | } | |
211 | ||
eaf718f3 | 212 | package main; |
25f6f325 | 213 | use URI; |
eaf718f3 | 214 | |
03490804 | 215 | our $svn = $svn_url; |
25f6f325 | 216 | $svn .= "/$svn_dir" if defined $svn_dir; |
03490804 | 217 | my $svn2 = SVNconn->new($svn); |
25f6f325 | 218 | $svn = SVNconn->new($svn); |
eaf718f3 | 219 | |
25f6f325 MU |
220 | my $lwp_ua; |
221 | if($opt_d or $opt_D) { | |
222 | $svn_url = URI->new($svn_url)->canonical; | |
223 | if($opt_D) { | |
224 | $svn_dir =~ s#/*$#/#; | |
225 | } else { | |
226 | $svn_dir = ""; | |
227 | } | |
228 | if ($svn_url->scheme eq "http") { | |
229 | use LWP::UserAgent; | |
230 | $lwp_ua = LWP::UserAgent->new(keep_alive => 1, requests_redirectable => []); | |
231 | } else { | |
232 | print STDERR "Warning: not HTTP; turning off direct file access\n"; | |
233 | $opt_d=0; | |
234 | } | |
235 | } | |
eaf718f3 MU |
236 | |
237 | sub pdate($) { | |
238 | my($d) = @_; | |
239 | $d =~ m#(\d\d\d\d)-(\d\d)-(\d\d)T(\d\d):(\d\d):(\d\d)# | |
240 | or die "Unparseable date: $d\n"; | |
241 | my $y=$1; $y-=1900 if $y>1900; | |
242 | return timegm($6||0,$5,$4,$3,$2-1,$y); | |
243 | } | |
244 | ||
eaf718f3 MU |
245 | sub getwd() { |
246 | my $pwd = `pwd`; | |
247 | chomp $pwd; | |
248 | return $pwd; | |
249 | } | |
250 | ||
251 | ||
252 | sub get_headref($$) { | |
253 | my $name = shift; | |
29504118 | 254 | my $git_dir = shift; |
eaf718f3 | 255 | my $sha; |
29504118 | 256 | |
eaf718f3 MU |
257 | if (open(C,"$git_dir/refs/heads/$name")) { |
258 | chomp($sha = <C>); | |
259 | close(C); | |
260 | length($sha) == 40 | |
261 | or die "Cannot get head id for $name ($sha): $!\n"; | |
262 | } | |
263 | return $sha; | |
264 | } | |
265 | ||
266 | ||
267 | -d $git_tree | |
268 | or mkdir($git_tree,0777) | |
269 | or die "Could not create $git_tree: $!"; | |
270 | chdir($git_tree); | |
271 | ||
272 | my $orig_branch = ""; | |
273 | my $forward_master = 0; | |
274 | my %branches; | |
275 | ||
276 | my $git_dir = $ENV{"GIT_DIR"} || ".git"; | |
277 | $git_dir = getwd()."/".$git_dir unless $git_dir =~ m#^/#; | |
278 | $ENV{"GIT_DIR"} = $git_dir; | |
279 | my $orig_git_index; | |
280 | $orig_git_index = $ENV{GIT_INDEX_FILE} if exists $ENV{GIT_INDEX_FILE}; | |
281 | my ($git_ih, $git_index) = tempfile('gitXXXXXX', SUFFIX => '.idx', | |
282 | DIR => File::Spec->tmpdir()); | |
283 | close ($git_ih); | |
284 | $ENV{GIT_INDEX_FILE} = $git_index; | |
285 | my $maxnum = 0; | |
286 | my $last_rev = ""; | |
287 | my $last_branch; | |
03490804 | 288 | my $current_rev = $opt_s || 1; |
eaf718f3 | 289 | unless(-d $git_dir) { |
5c94f87e | 290 | system("git-init"); |
eaf718f3 MU |
291 | die "Cannot init the GIT db at $git_tree: $?\n" if $?; |
292 | system("git-read-tree"); | |
293 | die "Cannot init an empty tree: $?\n" if $?; | |
294 | ||
295 | $last_branch = $opt_o; | |
296 | $orig_branch = ""; | |
297 | } else { | |
298 | -f "$git_dir/refs/heads/$opt_o" | |
299 | or die "Branch '$opt_o' does not exist.\n". | |
300 | "Either use the correct '-o branch' option,\n". | |
301 | "or import to a new repository.\n"; | |
302 | ||
303 | -f "$git_dir/svn2git" | |
304 | or die "'$git_dir/svn2git' does not exist.\n". | |
305 | "You need that file for incremental imports.\n"; | |
8366a10a PR |
306 | open(F, "git-symbolic-ref HEAD |") or |
307 | die "Cannot run git-symbolic-ref: $!\n"; | |
308 | chomp ($last_branch = <F>); | |
309 | $last_branch = basename($last_branch); | |
310 | close(F); | |
eaf718f3 MU |
311 | unless($last_branch) { |
312 | warn "Cannot read the last branch name: $! -- assuming 'master'\n"; | |
313 | $last_branch = "master"; | |
314 | } | |
315 | $orig_branch = $last_branch; | |
316 | $last_rev = get_headref($orig_branch, $git_dir); | |
317 | if (-f "$git_dir/SVN2GIT_HEAD") { | |
318 | die <<EOM; | |
319 | SVN2GIT_HEAD exists. | |
320 | Make sure your working directory corresponds to HEAD and remove SVN2GIT_HEAD. | |
321 | You may need to run | |
322 | ||
323 | git-read-tree -m -u SVN2GIT_HEAD HEAD | |
324 | EOM | |
325 | } | |
326 | system('cp', "$git_dir/HEAD", "$git_dir/SVN2GIT_HEAD"); | |
327 | ||
328 | $forward_master = | |
329 | $opt_o ne 'master' && -f "$git_dir/refs/heads/master" && | |
29504118 | 330 | system('cmp', '-s', "$git_dir/refs/heads/master", |
eaf718f3 MU |
331 | "$git_dir/refs/heads/$opt_o") == 0; |
332 | ||
333 | # populate index | |
334 | system('git-read-tree', $last_rev); | |
335 | die "read-tree failed: $?\n" if $?; | |
336 | ||
337 | # Get the last import timestamps | |
338 | open my $B,"<", "$git_dir/svn2git"; | |
339 | while(<$B>) { | |
340 | chomp; | |
341 | my($num,$branch,$ref) = split; | |
342 | $branches{$branch}{$num} = $ref; | |
343 | $branches{$branch}{"LAST"} = $ref; | |
03490804 | 344 | $current_rev = $num+1 if $current_rev <= $num; |
eaf718f3 MU |
345 | } |
346 | close($B); | |
347 | } | |
348 | -d $git_dir | |
349 | or die "Could not create git subdir ($git_dir).\n"; | |
350 | ||
d3cac2c9 KW |
351 | my $default_authors = "$git_dir/svn-authors"; |
352 | if ($opt_A) { | |
353 | read_users($opt_A); | |
354 | copy($opt_A,$default_authors) or die "Copy failed: $!"; | |
355 | } else { | |
356 | read_users($default_authors) if -f $default_authors; | |
357 | } | |
358 | ||
eaf718f3 MU |
359 | open BRANCHES,">>", "$git_dir/svn2git"; |
360 | ||
83936a29 SK |
361 | sub node_kind($$) { |
362 | my ($svnpath, $revision) = @_; | |
09c3a408 | 363 | $svnpath =~ s#^/*##; |
5d17d765 SS |
364 | my $subpool = SVN::Pool::new_default_sub; |
365 | my $kind = $svn->{'svn'}->check_path($svnpath,$revision); | |
cbce5d89 YAS |
366 | return $kind; |
367 | } | |
368 | ||
cbce5d89 | 369 | sub get_file($$$) { |
83936a29 | 370 | my($svnpath,$rev,$path) = @_; |
cbce5d89 | 371 | |
eaf718f3 | 372 | # now get it |
4802426d | 373 | my ($name,$mode); |
25f6f325 MU |
374 | if($opt_d) { |
375 | my($req,$res); | |
376 | ||
377 | # /svn/!svn/bc/2/django/trunk/django-docs/build.py | |
378 | my $url=$svn_url->clone(); | |
379 | $url->path($url->path."/!svn/bc/$rev/$svn_dir$svnpath"); | |
0090a618 | 380 | print "... $path...\n" if $opt_v; |
25f6f325 MU |
381 | $req = HTTP::Request->new(GET => $url); |
382 | $res = $lwp_ua->request($req); | |
383 | if ($res->is_success) { | |
384 | my $fh; | |
29504118 JH |
385 | ($fh, $name) = tempfile('gitsvn.XXXXXX', |
386 | DIR => File::Spec->tmpdir(), UNLINK => 1); | |
25f6f325 MU |
387 | print $fh $res->content; |
388 | close($fh) or die "Could not write $name: $!\n"; | |
389 | } else { | |
390 | return undef if $res->code == 301; # directory? | |
391 | die $res->status_line." at $url\n"; | |
392 | } | |
4802426d | 393 | $mode = '0644'; # can't obtain mode via direct http request? |
25f6f325 | 394 | } else { |
4802426d | 395 | ($name,$mode) = $svn->file("$svnpath",$rev); |
25f6f325 MU |
396 | return undef unless defined $name; |
397 | } | |
eaf718f3 | 398 | |
7ae0dc01 JH |
399 | my $pid = open(my $F, '-|'); |
400 | die $! unless defined $pid; | |
401 | if (!$pid) { | |
402 | exec("git-hash-object", "-w", $name) | |
eaf718f3 | 403 | or die "Cannot create object: $!\n"; |
7ae0dc01 | 404 | } |
eaf718f3 MU |
405 | my $sha = <$F>; |
406 | chomp $sha; | |
407 | close $F; | |
22dcbb75 | 408 | unlink $name; |
eaf718f3 MU |
409 | return [$mode, $sha, $path]; |
410 | } | |
411 | ||
c55f3fff | 412 | sub get_ignore($$$$$) { |
83936a29 | 413 | my($new,$old,$rev,$path,$svnpath) = @_; |
c55f3fff KW |
414 | |
415 | return unless $opt_I; | |
c55f3fff KW |
416 | my $name = $svn->ignore("$svnpath",$rev); |
417 | if ($path eq '/') { | |
418 | $path = $opt_I; | |
419 | } else { | |
420 | $path = File::Spec->catfile($path,$opt_I); | |
421 | } | |
422 | if (defined $name) { | |
423 | my $pid = open(my $F, '-|'); | |
424 | die $! unless defined $pid; | |
425 | if (!$pid) { | |
426 | exec("git-hash-object", "-w", $name) | |
427 | or die "Cannot create object: $!\n"; | |
428 | } | |
429 | my $sha = <$F>; | |
430 | chomp $sha; | |
431 | close $F; | |
432 | unlink $name; | |
433 | push(@$new,['0644',$sha,$path]); | |
83936a29 | 434 | } elsif (defined $old) { |
c55f3fff KW |
435 | push(@$old,$path); |
436 | } | |
437 | } | |
438 | ||
ec1e4689 SK |
439 | sub project_path($$) |
440 | { | |
441 | my ($path, $project) = @_; | |
442 | ||
443 | $path = "/".$path unless ($path =~ m#^\/#) ; | |
444 | return $1 if ($path =~ m#^$project\/(.*)$#); | |
445 | ||
446 | $path =~ s#\.#\\\.#g; | |
447 | $path =~ s#\+#\\\+#g; | |
448 | return "/" if ($project =~ m#^$path.*$#); | |
449 | ||
450 | return undef; | |
451 | } | |
452 | ||
eaf718f3 MU |
453 | sub split_path($$) { |
454 | my($rev,$path) = @_; | |
455 | my $branch; | |
456 | ||
457 | if($path =~ s#^/\Q$tag_name\E/([^/]+)/?##) { | |
458 | $branch = "/$1"; | |
459 | } elsif($path =~ s#^/\Q$trunk_name\E/?##) { | |
460 | $branch = "/"; | |
461 | } elsif($path =~ s#^/\Q$branch_name\E/([^/]+)/?##) { | |
462 | $branch = $1; | |
463 | } else { | |
c2c07a5c YAS |
464 | my %no_error = ( |
465 | "/" => 1, | |
466 | "/$tag_name" => 1, | |
467 | "/$branch_name" => 1 | |
468 | ); | |
469 | print STDERR "$rev: Unrecognized path: $path\n" unless (defined $no_error{$path}); | |
eaf718f3 MU |
470 | return () |
471 | } | |
ec1e4689 SK |
472 | if ($path eq "") { |
473 | $path = "/"; | |
474 | } elsif ($project_name) { | |
475 | $path = project_path($path, $project_name); | |
476 | } | |
eaf718f3 MU |
477 | return ($branch,$path); |
478 | } | |
479 | ||
8168373f YAS |
480 | sub branch_rev($$) { |
481 | ||
482 | my ($srcbranch,$uptorev) = @_; | |
483 | ||
484 | my $bbranches = $branches{$srcbranch}; | |
485 | my @revs = reverse sort { ($a eq 'LAST' ? 0 : $a) <=> ($b eq 'LAST' ? 0 : $b) } keys %$bbranches; | |
486 | my $therev; | |
487 | foreach my $arev(@revs) { | |
488 | next if ($arev eq 'LAST'); | |
489 | if ($arev <= $uptorev) { | |
490 | $therev = $arev; | |
491 | last; | |
492 | } | |
493 | } | |
494 | return $therev; | |
495 | } | |
496 | ||
83936a29 SK |
497 | sub expand_svndir($$$); |
498 | ||
499 | sub expand_svndir($$$) | |
500 | { | |
501 | my ($svnpath, $rev, $path) = @_; | |
502 | my @list; | |
503 | get_ignore(\@list, undef, $rev, $path, $svnpath); | |
504 | my $dirents = $svn->dir_list($svnpath, $rev); | |
505 | foreach my $p(keys %$dirents) { | |
506 | my $kind = node_kind($svnpath.'/'.$p, $rev); | |
507 | if ($kind eq $SVN::Node::file) { | |
508 | my $f = get_file($svnpath.'/'.$p, $rev, $path.'/'.$p); | |
509 | push(@list, $f) if $f; | |
510 | } elsif ($kind eq $SVN::Node::dir) { | |
511 | push(@list, | |
512 | expand_svndir($svnpath.'/'.$p, $rev, $path.'/'.$p)); | |
513 | } | |
514 | } | |
515 | return @list; | |
516 | } | |
517 | ||
109fc2b9 | 518 | sub copy_path($$$$$$$$) { |
0090a618 MU |
519 | # Somebody copied a whole subdirectory. |
520 | # We need to find the index entries from the old version which the | |
521 | # SVN log entry points to, and add them to the new place. | |
522 | ||
109fc2b9 | 523 | my($newrev,$newbranch,$path,$oldpath,$rev,$node_kind,$new,$parents) = @_; |
0090a618 | 524 | |
8168373f | 525 | my($srcbranch,$srcpath) = split_path($rev,$oldpath); |
83936a29 SK |
526 | unless(defined $srcbranch && defined $srcpath) { |
527 | print "Path not found when copying from $oldpath @ $rev.\n". | |
528 | "Will try to copy from original SVN location...\n" | |
529 | if $opt_v; | |
530 | push (@$new, expand_svndir($oldpath, $rev, $path)); | |
4b1ca25e MU |
531 | return; |
532 | } | |
8168373f YAS |
533 | my $therev = branch_rev($srcbranch, $rev); |
534 | my $gitrev = $branches{$srcbranch}{$therev}; | |
0090a618 MU |
535 | unless($gitrev) { |
536 | print STDERR "$newrev:$newbranch: could not find $oldpath \@ $rev\n"; | |
537 | return; | |
538 | } | |
109fc2b9 YAS |
539 | if ($srcbranch ne $newbranch) { |
540 | push(@$parents, $branches{$srcbranch}{'LAST'}); | |
541 | } | |
8168373f YAS |
542 | print "$newrev:$newbranch:$path: copying from $srcbranch:$srcpath @ $rev\n" if $opt_v; |
543 | if ($node_kind eq $SVN::Node::dir) { | |
83936a29 | 544 | $srcpath =~ s#/*$#/#; |
8168373f | 545 | } |
a6080a0a | 546 | |
7ae0dc01 JH |
547 | my $pid = open my $f,'-|'; |
548 | die $! unless defined $pid; | |
549 | if (!$pid) { | |
550 | exec("git-ls-tree","-r","-z",$gitrev,$srcpath) | |
551 | or die $!; | |
552 | } | |
0090a618 MU |
553 | local $/ = "\0"; |
554 | while(<$f>) { | |
555 | chomp; | |
556 | my($m,$p) = split(/\t/,$_,2); | |
557 | my($mode,$type,$sha1) = split(/ /,$m); | |
558 | next if $type ne "blob"; | |
8168373f YAS |
559 | if ($node_kind eq $SVN::Node::dir) { |
560 | $p = $path . substr($p,length($srcpath)-1); | |
561 | } else { | |
562 | $p = $path; | |
563 | } | |
a6080a0a | 564 | push(@$new,[$mode,$sha1,$p]); |
0090a618 | 565 | } |
29504118 | 566 | close($f) or |
0090a618 MU |
567 | print STDERR "$newrev:$newbranch: could not list files in $oldpath \@ $rev\n"; |
568 | } | |
569 | ||
eaf718f3 MU |
570 | sub commit { |
571 | my($branch, $changed_paths, $revision, $author, $date, $message) = @_; | |
ae35b304 SK |
572 | my($committer_name,$committer_email,$dest); |
573 | my($author_name,$author_email); | |
109fc2b9 | 574 | my(@old,@new,@parents); |
eaf718f3 | 575 | |
35c636ec | 576 | if (not defined $author or $author eq "") { |
ae35b304 | 577 | $committer_name = $committer_email = "unknown"; |
d3cac2c9 KW |
578 | } elsif (defined $users_file) { |
579 | die "User $author is not listed in $users_file\n" | |
36610b24 | 580 | unless exists $users{$author}; |
ae35b304 | 581 | ($committer_name,$committer_email) = @{$users{$author}}; |
2b5e63d1 | 582 | } elsif ($author =~ /^(.*?)\s+<(.*)>$/) { |
ae35b304 | 583 | ($committer_name, $committer_email) = ($1, $2); |
eaf718f3 MU |
584 | } else { |
585 | $author =~ s/^<(.*)>$/$1/; | |
ae35b304 SK |
586 | $committer_name = $committer_email = $author; |
587 | } | |
588 | ||
7b40e7d1 | 589 | if ($opt_F && $message =~ /From:\s+(.*?)\s+<(.*)>\s*\n/) { |
ae35b304 | 590 | ($author_name, $author_email) = ($1, $2); |
7b40e7d1 AW |
591 | print "Author from From: $1 <$2>\n" if ($opt_v);; |
592 | } elsif ($opt_S && $message =~ /Signed-off-by:\s+(.*?)\s+<(.*)>\s*\n/) { | |
593 | ($author_name, $author_email) = ($1, $2); | |
594 | print "Author from Signed-off-by: $1 <$2>\n" if ($opt_v);; | |
ae35b304 SK |
595 | } else { |
596 | $author_name = $committer_name; | |
597 | $author_email = $committer_email; | |
eaf718f3 | 598 | } |
ae35b304 | 599 | |
eaf718f3 MU |
600 | $date = pdate($date); |
601 | ||
602 | my $tag; | |
603 | my $parent; | |
604 | if($branch eq "/") { # trunk | |
605 | $parent = $opt_o; | |
606 | } elsif($branch =~ m#^/(.+)#) { # tag | |
607 | $tag = 1; | |
608 | $parent = $1; | |
609 | } else { # "normal" branch | |
610 | # nothing to do | |
611 | $parent = $branch; | |
612 | } | |
613 | $dest = $parent; | |
614 | ||
615 | my $prev = $changed_paths->{"/"}; | |
f0daa628 | 616 | if($prev and $prev->[0] eq "A") { |
eaf718f3 | 617 | delete $changed_paths->{"/"}; |
f0daa628 | 618 | my $oldpath = $prev->[1]; |
eaf718f3 MU |
619 | my $rev; |
620 | if(defined $oldpath) { | |
621 | my $p; | |
622 | ($parent,$p) = split_path($revision,$oldpath); | |
83936a29 SK |
623 | if(defined $parent) { |
624 | if($parent eq "/") { | |
625 | $parent = $opt_o; | |
626 | } else { | |
627 | $parent =~ s#^/##; # if it's a tag | |
628 | } | |
eaf718f3 MU |
629 | } |
630 | } else { | |
631 | $parent = undef; | |
632 | } | |
633 | } | |
634 | ||
635 | my $rev; | |
7ee74a99 | 636 | if($revision > $opt_s and defined $parent) { |
bf1ee636 | 637 | open(H,'-|',"git-rev-parse","--verify",$parent); |
eaf718f3 MU |
638 | $rev = <H>; |
639 | close(H) or do { | |
640 | print STDERR "$revision: cannot find commit '$parent'!\n"; | |
641 | return; | |
642 | }; | |
643 | chop $rev; | |
644 | if(length($rev) != 40) { | |
645 | print STDERR "$revision: cannot find commit '$parent'!\n"; | |
646 | return; | |
647 | } | |
648 | $rev = $branches{($parent eq $opt_o) ? "/" : $parent}{"LAST"}; | |
7ee74a99 | 649 | if($revision != $opt_s and not $rev) { |
eaf718f3 MU |
650 | print STDERR "$revision: do not know ancestor for '$parent'!\n"; |
651 | return; | |
652 | } | |
653 | } else { | |
654 | $rev = undef; | |
655 | } | |
656 | ||
f0daa628 | 657 | # if($prev and $prev->[0] eq "A") { |
eaf718f3 MU |
658 | # if(not $tag) { |
659 | # unless(open(H,"> $git_dir/refs/heads/$branch")) { | |
660 | # print STDERR "$revision: Could not create branch $branch: $!\n"; | |
661 | # $state=11; | |
662 | # next; | |
663 | # } | |
664 | # print H "$rev\n" | |
665 | # or die "Could not write branch $branch: $!"; | |
666 | # close(H) | |
667 | # or die "Could not write branch $branch: $!"; | |
668 | # } | |
669 | # } | |
670 | if(not defined $rev) { | |
671 | unlink($git_index); | |
672 | } elsif ($rev ne $last_rev) { | |
673 | print "Switching from $last_rev to $rev ($branch)\n" if $opt_v; | |
674 | system("git-read-tree", $rev); | |
675 | die "read-tree failed for $rev: $?\n" if $?; | |
676 | $last_rev = $rev; | |
677 | } | |
678 | ||
109fc2b9 YAS |
679 | push (@parents, $rev) if defined $rev; |
680 | ||
bf267d99 MU |
681 | my $cid; |
682 | if($tag and not %$changed_paths) { | |
683 | $cid = $rev; | |
684 | } else { | |
0090a618 MU |
685 | my @paths = sort keys %$changed_paths; |
686 | foreach my $path(@paths) { | |
687 | my $action = $changed_paths->{$path}; | |
688 | ||
8168373f YAS |
689 | if ($action->[0] eq "R") { |
690 | # refer to a file/tree in an earlier commit | |
691 | push(@old,$path); # remove any old stuff | |
692 | } | |
693 | if(($action->[0] eq "A") || ($action->[0] eq "R")) { | |
83936a29 | 694 | my $node_kind = node_kind($action->[3], $revision); |
e67c6625 | 695 | if ($node_kind eq $SVN::Node::file) { |
83936a29 SK |
696 | my $f = get_file($action->[3], |
697 | $revision, $path); | |
8168373f YAS |
698 | if ($f) { |
699 | push(@new,$f) if $f; | |
700 | } else { | |
701 | my $opath = $action->[3]; | |
702 | print STDERR "$revision: $branch: could not fetch '$opath'\n"; | |
703 | } | |
c55f3fff | 704 | } elsif ($node_kind eq $SVN::Node::dir) { |
e67c6625 KW |
705 | if($action->[1]) { |
706 | copy_path($revision, $branch, | |
707 | $path, $action->[1], | |
708 | $action->[2], $node_kind, | |
709 | \@new, \@parents); | |
710 | } else { | |
711 | get_ignore(\@new, \@old, $revision, | |
83936a29 | 712 | $path, $action->[3]); |
e67c6625 | 713 | } |
0090a618 | 714 | } |
bf267d99 MU |
715 | } elsif ($action->[0] eq "D") { |
716 | push(@old,$path); | |
717 | } elsif ($action->[0] eq "M") { | |
83936a29 | 718 | my $node_kind = node_kind($action->[3], $revision); |
8168373f | 719 | if ($node_kind eq $SVN::Node::file) { |
83936a29 SK |
720 | my $f = get_file($action->[3], |
721 | $revision, $path); | |
8168373f | 722 | push(@new,$f) if $f; |
c55f3fff KW |
723 | } elsif ($node_kind eq $SVN::Node::dir) { |
724 | get_ignore(\@new, \@old, $revision, | |
83936a29 | 725 | $path, $action->[3]); |
bf267d99 MU |
726 | } |
727 | } else { | |
728 | die "$revision: unknown action '".$action->[0]."' for $path\n"; | |
729 | } | |
730 | } | |
731 | ||
d9e2e127 SK |
732 | while(@old) { |
733 | my @o1; | |
734 | if(@old > 55) { | |
735 | @o1 = splice(@old,0,50); | |
736 | } else { | |
737 | @o1 = @old; | |
738 | @old = (); | |
739 | } | |
7ae0dc01 JH |
740 | my $pid = open my $F, "-|"; |
741 | die "$!" unless defined $pid; | |
742 | if (!$pid) { | |
743 | exec("git-ls-files", "-z", @o1) or die $!; | |
744 | } | |
d9e2e127 | 745 | @o1 = (); |
0090a618 | 746 | local $/ = "\0"; |
eaf718f3 MU |
747 | while(<$F>) { |
748 | chomp; | |
d9e2e127 | 749 | push(@o1,$_); |
bf267d99 MU |
750 | } |
751 | close($F); | |
752 | ||
d9e2e127 | 753 | while(@o1) { |
bf267d99 | 754 | my @o2; |
d9e2e127 SK |
755 | if(@o1 > 55) { |
756 | @o2 = splice(@o1,0,50); | |
bf267d99 | 757 | } else { |
d9e2e127 SK |
758 | @o2 = @o1; |
759 | @o1 = (); | |
bf267d99 MU |
760 | } |
761 | system("git-update-index","--force-remove","--",@o2); | |
762 | die "Cannot remove files: $?\n" if $?; | |
eaf718f3 | 763 | } |
eaf718f3 | 764 | } |
bf267d99 MU |
765 | while(@new) { |
766 | my @n2; | |
767 | if(@new > 12) { | |
768 | @n2 = splice(@new,0,10); | |
eaf718f3 | 769 | } else { |
bf267d99 MU |
770 | @n2 = @new; |
771 | @new = (); | |
eaf718f3 | 772 | } |
bf267d99 MU |
773 | system("git-update-index","--add", |
774 | (map { ('--cacheinfo', @$_) } @n2)); | |
775 | die "Cannot add files: $?\n" if $?; | |
eaf718f3 | 776 | } |
eaf718f3 | 777 | |
bf267d99 MU |
778 | my $pid = open(C,"-|"); |
779 | die "Cannot fork: $!" unless defined $pid; | |
780 | unless($pid) { | |
781 | exec("git-write-tree"); | |
782 | die "Cannot exec git-write-tree: $!\n"; | |
eaf718f3 | 783 | } |
bf267d99 MU |
784 | chomp(my $tree = <C>); |
785 | length($tree) == 40 | |
786 | or die "Cannot get tree id ($tree): $!\n"; | |
787 | close(C) | |
788 | or die "Error running git-write-tree: $?\n"; | |
789 | print "Tree ID $tree\n" if $opt_v; | |
790 | ||
791 | my $pr = IO::Pipe->new() or die "Cannot open pipe: $!\n"; | |
792 | my $pw = IO::Pipe->new() or die "Cannot open pipe: $!\n"; | |
793 | $pid = fork(); | |
794 | die "Fork: $!\n" unless defined $pid; | |
795 | unless($pid) { | |
796 | $pr->writer(); | |
797 | $pw->reader(); | |
798 | open(OUT,">&STDOUT"); | |
799 | dup2($pw->fileno(),0); | |
800 | dup2($pr->fileno(),1); | |
801 | $pr->close(); | |
802 | $pw->close(); | |
803 | ||
804 | my @par = (); | |
bf267d99 MU |
805 | |
806 | # loose detection of merges | |
807 | # based on the commit msg | |
808 | foreach my $rx (@mergerx) { | |
809 | if ($message =~ $rx) { | |
810 | my $mparent = $1; | |
811 | if ($mparent eq 'HEAD') { $mparent = $opt_o }; | |
812 | if ( -e "$git_dir/refs/heads/$mparent") { | |
813 | $mparent = get_headref($mparent, $git_dir); | |
109fc2b9 | 814 | push (@parents, $mparent); |
bf267d99 MU |
815 | print OUT "Merge parent branch: $mparent\n" if $opt_v; |
816 | } | |
29504118 | 817 | } |
bf267d99 | 818 | } |
109fc2b9 YAS |
819 | my %seen_parents = (); |
820 | my @unique_parents = grep { ! $seen_parents{$_} ++ } @parents; | |
821 | foreach my $bparent (@unique_parents) { | |
822 | push @par, '-p', $bparent; | |
823 | print OUT "Merge parent branch: $bparent\n" if $opt_v; | |
824 | } | |
eaf718f3 | 825 | |
bf267d99 MU |
826 | exec("env", |
827 | "GIT_AUTHOR_NAME=$author_name", | |
828 | "GIT_AUTHOR_EMAIL=$author_email", | |
829 | "GIT_AUTHOR_DATE=".strftime("+0000 %Y-%m-%d %H:%M:%S",gmtime($date)), | |
ae35b304 SK |
830 | "GIT_COMMITTER_NAME=$committer_name", |
831 | "GIT_COMMITTER_EMAIL=$committer_email", | |
bf267d99 MU |
832 | "GIT_COMMITTER_DATE=".strftime("+0000 %Y-%m-%d %H:%M:%S",gmtime($date)), |
833 | "git-commit-tree", $tree,@par); | |
834 | die "Cannot exec git-commit-tree: $!\n"; | |
835 | } | |
836 | $pw->writer(); | |
837 | $pr->reader(); | |
eaf718f3 | 838 | |
bf267d99 | 839 | $message =~ s/[\s\n]+\z//; |
0a48a344 | 840 | $message = "r$revision: $message" if $opt_r; |
eaf718f3 | 841 | |
bf267d99 MU |
842 | print $pw "$message\n" |
843 | or die "Error writing to git-commit-tree: $!\n"; | |
844 | $pw->close(); | |
eaf718f3 | 845 | |
bf267d99 MU |
846 | print "Committed change $revision:$branch ".strftime("%Y-%m-%d %H:%M:%S",gmtime($date)).")\n" if $opt_v; |
847 | chomp($cid = <$pr>); | |
848 | length($cid) == 40 | |
849 | or die "Cannot get commit id ($cid): $!\n"; | |
850 | print "Commit ID $cid\n" if $opt_v; | |
851 | $pr->close(); | |
eaf718f3 | 852 | |
bf267d99 MU |
853 | waitpid($pid,0); |
854 | die "Error running git-commit-tree: $?\n" if $?; | |
855 | } | |
eaf718f3 | 856 | |
a16db4f4 YAS |
857 | if (not defined $cid) { |
858 | $cid = $branches{"/"}{"LAST"}; | |
859 | } | |
860 | ||
f02b3eba MU |
861 | if(not defined $dest) { |
862 | print "... no known parent\n" if $opt_v; | |
863 | } elsif(not $tag) { | |
eaf718f3 | 864 | print "Writing to refs/heads/$dest\n" if $opt_v; |
29504118 | 865 | open(C,">$git_dir/refs/heads/$dest") and |
eaf718f3 MU |
866 | print C ("$cid\n") and |
867 | close(C) | |
868 | or die "Cannot write branch $dest for update: $!\n"; | |
eaf718f3 | 869 | } |
eaf718f3 | 870 | |
47ee8ed2 | 871 | if ($tag) { |
eaf718f3 MU |
872 | $last_rev = "-" if %$changed_paths; |
873 | # the tag was 'complex', i.e. did not refer to a "real" revision | |
29504118 | 874 | |
f02b3eba | 875 | $dest =~ tr/_/\./ if $opt_u; |
eaf718f3 | 876 | |
ee834cf0 | 877 | system('git-tag', '-f', $dest, $cid) == 0 |
47ee8ed2 | 878 | or die "Cannot create tag $dest: $!\n"; |
eaf718f3 | 879 | |
f02b3eba | 880 | print "Created tag '$dest' on '$branch'\n" if $opt_v; |
eaf718f3 | 881 | } |
e7e477df MU |
882 | $branches{$branch}{"LAST"} = $cid; |
883 | $branches{$branch}{$revision} = $cid; | |
884 | $last_rev = $cid; | |
885 | print BRANCHES "$revision $branch $cid\n"; | |
886 | print "DONE: $revision $dest $cid\n" if $opt_v; | |
eaf718f3 MU |
887 | } |
888 | ||
03490804 MU |
889 | sub commit_all { |
890 | # Recursive use of the SVN connection does not work | |
891 | local $svn = $svn2; | |
892 | ||
5d17d765 | 893 | my ($changed_paths, $revision, $author, $date, $message) = @_; |
f0daa628 MU |
894 | my %p; |
895 | while(my($path,$action) = each %$changed_paths) { | |
0090a618 | 896 | $p{$path} = [ $action->action,$action->copyfrom_path, $action->copyfrom_rev, $path ]; |
f0daa628 MU |
897 | } |
898 | $changed_paths = \%p; | |
f0daa628 | 899 | |
eaf718f3 MU |
900 | my %done; |
901 | my @col; | |
902 | my $pref; | |
903 | my $branch; | |
904 | ||
905 | while(my($path,$action) = each %$changed_paths) { | |
906 | ($branch,$path) = split_path($revision,$path); | |
907 | next if not defined $branch; | |
ec1e4689 | 908 | next if not defined $path; |
eaf718f3 MU |
909 | $done{$branch}{$path} = $action; |
910 | } | |
911 | while(($branch,$changed_paths) = each %done) { | |
912 | commit($branch, $changed_paths, $revision, $author, $date, $message); | |
913 | } | |
914 | } | |
915 | ||
03490804 | 916 | $opt_l = $svn->{'maxrev'} if not defined $opt_l or $opt_l > $svn->{'maxrev'}; |
988eece4 | 917 | |
a7cfb4a4 | 918 | if ($opt_l < $current_rev) { |
988eece4 ML |
919 | print "Up to date: no new revisions to fetch!\n" if $opt_v; |
920 | unlink("$git_dir/SVN2GIT_HEAD"); | |
921 | exit; | |
922 | } | |
923 | ||
40006ea0 | 924 | print "Processing from $current_rev to $opt_l ...\n" if $opt_v; |
03490804 | 925 | |
40006ea0 | 926 | my $from_rev; |
69216777 | 927 | my $to_rev = $current_rev - 1; |
40006ea0 | 928 | |
5d17d765 | 929 | my $subpool = SVN::Pool::new_default_sub; |
40006ea0 | 930 | while ($to_rev < $opt_l) { |
5d17d765 | 931 | $subpool->clear; |
69216777 | 932 | $from_rev = $to_rev + 1; |
40006ea0 SK |
933 | $to_rev = $from_rev + $repack_after; |
934 | $to_rev = $opt_l if $opt_l < $to_rev; | |
935 | print "Fetching from $from_rev to $to_rev ...\n" if $opt_v; | |
5d17d765 | 936 | $svn->{'svn'}->get_log("/",$from_rev,$to_rev,0,1,1,\&commit_all); |
40006ea0 SK |
937 | my $pid = fork(); |
938 | die "Fork: $!\n" unless defined $pid; | |
939 | unless($pid) { | |
940 | exec("git-repack", "-d") | |
941 | or die "Cannot repack: $!\n"; | |
942 | } | |
943 | waitpid($pid, 0); | |
944 | } | |
eaf718f3 MU |
945 | |
946 | ||
947 | unlink($git_index); | |
948 | ||
949 | if (defined $orig_git_index) { | |
950 | $ENV{GIT_INDEX_FILE} = $orig_git_index; | |
951 | } else { | |
952 | delete $ENV{GIT_INDEX_FILE}; | |
953 | } | |
954 | ||
955 | # Now switch back to the branch we were in before all of this happened | |
956 | if($orig_branch) { | |
3ef378a6 | 957 | print "DONE\n" if $opt_v and (not defined $opt_l or $opt_l > 0); |
eaf718f3 MU |
958 | system("cp","$git_dir/refs/heads/$opt_o","$git_dir/refs/heads/master") |
959 | if $forward_master; | |
960 | unless ($opt_i) { | |
961 | system('git-read-tree', '-m', '-u', 'SVN2GIT_HEAD', 'HEAD'); | |
962 | die "read-tree failed: $?\n" if $?; | |
963 | } | |
964 | } else { | |
965 | $orig_branch = "master"; | |
3ef378a6 | 966 | print "DONE; creating $orig_branch branch\n" if $opt_v and (not defined $opt_l or $opt_l > 0); |
eaf718f3 MU |
967 | system("cp","$git_dir/refs/heads/$opt_o","$git_dir/refs/heads/master") |
968 | unless -f "$git_dir/refs/heads/master"; | |
8366a10a | 969 | system('git-update-ref', 'HEAD', "$orig_branch"); |
eaf718f3 MU |
970 | unless ($opt_i) { |
971 | system('git checkout'); | |
972 | die "checkout failed: $?\n" if $?; | |
973 | } | |
974 | } | |
975 | unlink("$git_dir/SVN2GIT_HEAD"); | |
976 | close(BRANCHES); |