]>
Commit | Line | Data |
---|---|---|
d3968363 ML |
1 | #!/usr/bin/perl -w |
2 | # | |
3 | # This tool is copyright (c) 2005, Martin Langhoff. | |
4 | # It is released under the Gnu Public License, version 2. | |
5 | # | |
6 | # The basic idea is to walk the output of tla abrowse, | |
7 | # fetch the changesets and apply them. | |
8 | # | |
241b5967 | 9 | |
d3968363 ML |
10 | =head1 Invocation |
11 | ||
42f4570c EW |
12 | git-archimport [ -h ] [ -v ] [ -o ] [ -a ] [ -f ] [ -T ] |
13 | [ -D depth] [ -t tempdir ] <archive>/<branch> [ <archive>/<branch> ] | |
d3968363 | 14 | |
241b5967 ML |
15 | Imports a project from one or more Arch repositories. It will follow branches |
16 | and repositories within the namespaces defined by the <archive/branch> | |
82e5a82f | 17 | parameters supplied. If it cannot find the remote branch a merge comes from |
241b5967 ML |
18 | it will just import it as a regular commit. If it can find it, it will mark it |
19 | as a merge whenever possible. | |
d3968363 | 20 | |
241b5967 | 21 | See man (1) git-archimport for more details. |
d3968363 | 22 | |
241b5967 | 23 | =head1 TODO |
d3968363 | 24 | |
241b5967 | 25 | - create tag objects instead of ref tags |
d3968363 | 26 | - audit shell-escaping of filenames |
241b5967 ML |
27 | - hide our private tags somewhere smarter |
28 | - find a way to make "cat *patches | patch" safe even when patchfiles are missing newlines | |
3e525e67 EW |
29 | - sort and apply patches by graphing ancestry relations instead of just |
30 | relying in dates supplied in the changeset itself. | |
31 | tla ancestry-graph -m could be helpful here... | |
d3968363 ML |
32 | |
33 | =head1 Devel tricks | |
34 | ||
35 | Add print in front of the shell commands invoked via backticks. | |
36 | ||
22ff00fc EW |
37 | =head1 Devel Notes |
38 | ||
39 | There are several places where Arch and git terminology are intermixed | |
40 | and potentially confused. | |
41 | ||
42 | The notion of a "branch" in git is approximately equivalent to | |
43 | a "archive/category--branch--version" in Arch. Also, it should be noted | |
44 | that the "--branch" portion of "archive/category--branch--version" is really | |
45 | optional in Arch although not many people (nor tools!) seem to know this. | |
46 | This means that "archive/category--version" is also a valid "branch" | |
47 | in git terms. | |
48 | ||
49 | We always refer to Arch names by their fully qualified variant (which | |
50 | means the "archive" name is prefixed. | |
51 | ||
52 | For people unfamiliar with Arch, an "archive" is the term for "repository", | |
53 | and can contain multiple, unrelated branches. | |
54 | ||
d3968363 ML |
55 | =cut |
56 | ||
57 | use strict; | |
58 | use warnings; | |
59 | use Getopt::Std; | |
42f44b08 | 60 | use File::Temp qw(tempdir); |
f88961a8 | 61 | use File::Path qw(mkpath rmtree); |
d3968363 | 62 | use File::Basename qw(basename dirname); |
d3968363 ML |
63 | use Data::Dumper qw/ Dumper /; |
64 | use IPC::Open2; | |
65 | ||
66 | $SIG{'PIPE'}="IGNORE"; | |
67 | $ENV{'TZ'}="UTC"; | |
68 | ||
1d4710d0 ML |
69 | my $git_dir = $ENV{"GIT_DIR"} || ".git"; |
70 | $ENV{"GIT_DIR"} = $git_dir; | |
a7fb51d3 | 71 | my $ptag_dir = "$git_dir/archimport/tags"; |
1d4710d0 | 72 | |
3e525e67 | 73 | our($opt_h,$opt_f,$opt_v,$opt_T,$opt_t,$opt_D,$opt_a,$opt_o); |
d3968363 ML |
74 | |
75 | sub usage() { | |
76 | print STDERR <<END; | |
77 | Usage: ${\basename $0} # fetch/update GIT from Arch | |
42f4570c | 78 | [ -h ] [ -v ] [ -o ] [ -a ] [ -f ] [ -T ] [ -D depth ] [ -t tempdir ] |
d3968363 ML |
79 | repository/arch-branch [ repository/arch-branch] ... |
80 | END | |
81 | exit(1); | |
82 | } | |
83 | ||
3e525e67 | 84 | getopts("fThvat:D:") or usage(); |
d3968363 ML |
85 | usage if $opt_h; |
86 | ||
87 | @ARGV >= 1 or usage(); | |
42f44b08 EW |
88 | # $arch_branches: |
89 | # values associated with keys: | |
90 | # =1 - Arch version / git 'branch' detected via abrowse on a limit | |
82e5a82f | 91 | # >1 - Arch version / git 'branch' of an auxiliary branch we've merged |
d9cb5399 PB |
92 | my %arch_branches = map { my $branch = $_; $branch =~ s/:[^:]*$//; $branch => 1 } @ARGV; |
93 | ||
94 | # $branch_name_map: | |
95 | # maps arch branches to git branch names | |
96 | my %branch_name_map = map { m/^(.*):([^:]*)$/; $1 => $2 } grep { m/:/ } @ARGV; | |
d3968363 | 97 | |
5744f277 EW |
98 | $ENV{'TMPDIR'} = $opt_t if $opt_t; # $ENV{TMPDIR} will affect tempdir() calls: |
99 | my $tmp = tempdir('git-archimport-XXXXXX', TMPDIR => 1, CLEANUP => 1); | |
127bf00f | 100 | $opt_v && print "+ Using $tmp as temporary directory\n"; |
d3968363 | 101 | |
11dbe9e8 GP |
102 | unless (-d $git_dir) { # initial import needs empty directory |
103 | opendir DIR, '.' or die "Unable to open current directory: $!\n"; | |
104 | while (my $entry = readdir DIR) { | |
105 | $entry =~ /^\.\.?$/ or | |
106 | die "Initial import needs an empty current working directory.\n" | |
107 | } | |
108 | closedir DIR | |
109 | } | |
110 | ||
d9cb5399 | 111 | my $default_archive; # default Arch archive |
42f44b08 EW |
112 | my %reachable = (); # Arch repositories we can access |
113 | my %unreachable = (); # Arch repositories we can't access :< | |
d3968363 | 114 | my @psets = (); # the collection |
b779d5f0 | 115 | my %psets = (); # the collection, by name |
3e525e67 EW |
116 | my %stats = ( # Track which strategy we used to import: |
117 | get_tag => 0, replay => 0, get_new => 0, get_delta => 0, | |
118 | simple_changeset => 0, import_or_tag => 0 | |
119 | ); | |
b779d5f0 ML |
120 | |
121 | my %rptags = (); # my reverse private tags | |
122 | # to map a SHA1 to a commitid | |
2777ef76 | 123 | my $TLA = $ENV{'ARCH_CLIENT'} || 'tla'; |
d3968363 | 124 | |
42f44b08 EW |
125 | sub do_abrowse { |
126 | my $stage = shift; | |
127 | while (my ($limit, $level) = each %arch_branches) { | |
128 | next unless $level == $stage; | |
129 | ||
130 | open ABROWSE, "$TLA abrowse -fkD --merges $limit |" | |
131 | or die "Problems with tla abrowse: $!"; | |
d3968363 | 132 | |
42f44b08 EW |
133 | my %ps = (); # the current one |
134 | my $lastseen = ''; | |
d3968363 | 135 | |
42f44b08 EW |
136 | while (<ABROWSE>) { |
137 | chomp; | |
d3968363 | 138 | |
42f44b08 EW |
139 | # first record padded w 8 spaces |
140 | if (s/^\s{8}\b//) { | |
141 | my ($id, $type) = split(m/\s+/, $_, 2); | |
142 | ||
143 | my %last_ps; | |
144 | # store the record we just captured | |
145 | if (%ps && !exists $psets{ $ps{id} }) { | |
146 | %last_ps = %ps; # break references | |
147 | push (@psets, \%last_ps); | |
148 | $psets{ $last_ps{id} } = \%last_ps; | |
149 | } | |
d3968363 | 150 | |
42f44b08 EW |
151 | my $branch = extract_versionname($id); |
152 | %ps = ( id => $id, branch => $branch ); | |
153 | if (%last_ps && ($last_ps{branch} eq $branch)) { | |
154 | $ps{parent_id} = $last_ps{id}; | |
155 | } | |
156 | ||
157 | $arch_branches{$branch} = 1; | |
158 | $lastseen = 'id'; | |
159 | ||
160 | # deal with types (should work with baz or tla): | |
161 | if ($type =~ m/\(.*changeset\)/) { | |
162 | $ps{type} = 's'; | |
163 | } elsif ($type =~ /\(.*import\)/) { | |
164 | $ps{type} = 'i'; | |
6df896b5 | 165 | } elsif ($type =~ m/\(tag.*?(\S+\@\S+).*?\)/) { |
42f44b08 EW |
166 | $ps{type} = 't'; |
167 | # read which revision we've tagged when we parse the log | |
6df896b5 | 168 | $ps{tag} = $1; |
42f44b08 EW |
169 | } else { |
170 | warn "Unknown type $type"; | |
171 | } | |
172 | ||
173 | $arch_branches{$branch} = 1; | |
174 | $lastseen = 'id'; | |
175 | } elsif (s/^\s{10}//) { | |
176 | # 10 leading spaces or more | |
177 | # indicate commit metadata | |
178 | ||
179 | # date | |
180 | if ($lastseen eq 'id' && m/^(\d{4}-\d\d-\d\d \d\d:\d\d:\d\d)/){ | |
181 | $ps{date} = $1; | |
182 | $lastseen = 'date'; | |
183 | } elsif ($_ eq 'merges in:') { | |
184 | $ps{merges} = []; | |
185 | $lastseen = 'merges'; | |
186 | } elsif ($lastseen eq 'merges' && s/^\s{2}//) { | |
187 | my $id = $_; | |
188 | push (@{$ps{merges}}, $id); | |
189 | ||
190 | # aggressive branch finding: | |
191 | if ($opt_D) { | |
192 | my $branch = extract_versionname($id); | |
193 | my $repo = extract_reponame($branch); | |
194 | ||
195 | if (archive_reachable($repo) && | |
196 | !defined $arch_branches{$branch}) { | |
197 | $arch_branches{$branch} = $stage + 1; | |
198 | } | |
199 | } | |
200 | } else { | |
201 | warn "more metadata after merges!?: $_\n" unless /^\s*$/; | |
d3968363 | 202 | } |
d3968363 | 203 | } |
d3968363 | 204 | } |
d3968363 | 205 | |
42f44b08 EW |
206 | if (%ps && !exists $psets{ $ps{id} }) { |
207 | my %temp = %ps; # break references | |
208 | if (@psets && $psets[$#psets]{branch} eq $ps{branch}) { | |
209 | $temp{parent_id} = $psets[$#psets]{id}; | |
210 | } | |
211 | push (@psets, \%temp); | |
212 | $psets{ $temp{id} } = \%temp; | |
213 | } | |
214 | ||
215 | close ABROWSE or die "$TLA abrowse failed on $limit\n"; | |
216 | } | |
d3968363 ML |
217 | } # end foreach $root |
218 | ||
42f44b08 EW |
219 | do_abrowse(1); |
220 | my $depth = 2; | |
221 | $opt_D ||= 0; | |
222 | while ($depth <= $opt_D) { | |
223 | do_abrowse($depth); | |
224 | $depth++; | |
225 | } | |
226 | ||
d3968363 | 227 | ## Order patches by time |
42f44b08 EW |
228 | # FIXME see if we can find a more optimal way to do this by graphing |
229 | # the ancestry data and walking it, that way we won't have to rely on | |
230 | # client-supplied dates | |
d3968363 ML |
231 | @psets = sort {$a->{date}.$b->{id} cmp $b->{date}.$b->{id}} @psets; |
232 | ||
233 | #print Dumper \@psets; | |
234 | ||
235 | ## | |
236 | ## TODO cleanup irrelevant patches | |
237 | ## and put an initial import | |
238 | ## or a full tag | |
3292ae47 | 239 | my $import = 0; |
1d4710d0 | 240 | unless (-d $git_dir) { # initial import |
d3968363 ML |
241 | if ($psets[0]{type} eq 'i' || $psets[0]{type} eq 't') { |
242 | print "Starting import from $psets[0]{id}\n"; | |
5c94f87e | 243 | `git-init`; |
3292ae47 ML |
244 | die $! if $?; |
245 | $import = 1; | |
d3968363 ML |
246 | } else { |
247 | die "Need to start from an import or a tag -- cannot use $psets[0]{id}"; | |
248 | } | |
b779d5f0 ML |
249 | } else { # progressing an import |
250 | # load the rptags | |
42f44b08 | 251 | opendir(DIR, $ptag_dir) |
b779d5f0 ML |
252 | || die "can't opendir: $!"; |
253 | while (my $file = readdir(DIR)) { | |
a7fb51d3 EW |
254 | # skip non-interesting-files |
255 | next unless -f "$ptag_dir/$file"; | |
256 | ||
257 | # convert first '--' to '/' from old git-archimport to use | |
258 | # as an archivename/c--b--v private tag | |
259 | if ($file !~ m!,!) { | |
260 | my $oldfile = $file; | |
261 | $file =~ s!--!,!; | |
262 | print STDERR "converting old tag $oldfile to $file\n"; | |
263 | rename("$ptag_dir/$oldfile", "$ptag_dir/$file") or die $!; | |
264 | } | |
b779d5f0 ML |
265 | my $sha = ptag($file); |
266 | chomp $sha; | |
b779d5f0 ML |
267 | $rptags{$sha} = $file; |
268 | } | |
269 | closedir DIR; | |
d3968363 ML |
270 | } |
271 | ||
3292ae47 | 272 | # process patchsets |
22ff00fc EW |
273 | # extract the Arch repository name (Arch "archive" in Arch-speak) |
274 | sub extract_reponame { | |
275 | my $fq_cvbr = shift; # archivename/[[[[category]branch]version]revision] | |
276 | return (split(/\//, $fq_cvbr))[0]; | |
277 | } | |
278 | ||
279 | sub extract_versionname { | |
280 | my $name = shift; | |
281 | $name =~ s/--(?:patch|version(?:fix)?|base)-\d+$//; | |
282 | return $name; | |
283 | } | |
d3968363 | 284 | |
22ff00fc EW |
285 | # convert a fully-qualified revision or version to a unique dirname: |
286 | # normalperson@yhbt.net-05/mpd--uclinux--1--patch-2 | |
287 | # becomes: normalperson@yhbt.net-05,mpd--uclinux--1 | |
288 | # | |
289 | # the git notion of a branch is closer to | |
290 | # archive/category--branch--version than archive/category--branch, so we | |
291 | # use this to convert to git branch names. | |
292 | # Also, keep archive names but replace '/' with ',' since it won't require | |
293 | # subdirectories, and is safer than swapping '--' which could confuse | |
294 | # reverse-mapping when dealing with bastard branches that | |
295 | # are just archive/category--version (no --branch) | |
296 | sub tree_dirname { | |
297 | my $revision = shift; | |
298 | my $name = extract_versionname($revision); | |
299 | $name =~ s#/#,#; | |
300 | return $name; | |
301 | } | |
302 | ||
fee3365f ML |
303 | # old versions of git-archimport just use the <category--branch> part: |
304 | sub old_style_branchname { | |
305 | my $id = shift; | |
306 | my $ret = safe_pipe_capture($TLA,'parse-package-name','-p',$id); | |
307 | chomp $ret; | |
308 | return $ret; | |
309 | } | |
310 | ||
d9cb5399 PB |
311 | *git_default_branchname = $opt_o ? *old_style_branchname : *tree_dirname; |
312 | ||
313 | # retrieve default archive, since $branch_name_map keys might not include it | |
314 | sub get_default_archive { | |
315 | if (!defined $default_archive) { | |
316 | $default_archive = safe_pipe_capture($TLA,'my-default-archive'); | |
317 | chomp $default_archive; | |
318 | } | |
319 | return $default_archive; | |
320 | } | |
321 | ||
322 | sub git_branchname { | |
323 | my $revision = shift; | |
324 | my $name = extract_versionname($revision); | |
325 | ||
326 | if (exists $branch_name_map{$name}) { | |
327 | return $branch_name_map{$name}; | |
328 | ||
329 | } elsif ($name =~ m#^([^/]*)/(.*)$# | |
330 | && $1 eq get_default_archive() | |
331 | && exists $branch_name_map{$2}) { | |
332 | # the names given in the command-line lacked the archive. | |
333 | return $branch_name_map{$2}; | |
334 | ||
335 | } else { | |
336 | return git_default_branchname($revision); | |
337 | } | |
338 | } | |
22ff00fc | 339 | |
3e525e67 EW |
340 | sub process_patchset_accurate { |
341 | my $ps = shift; | |
342 | ||
343 | # switch to that branch if we're not already in that branch: | |
344 | if (-e "$git_dir/refs/heads/$ps->{branch}") { | |
345 | system('git-checkout','-f',$ps->{branch}) == 0 or die "$! $?\n"; | |
d3968363 | 346 | |
3e525e67 EW |
347 | # remove any old stuff that got leftover: |
348 | my $rm = safe_pipe_capture('git-ls-files','--others','-z'); | |
349 | rmtree(split(/\0/,$rm)) if $rm; | |
3292ae47 | 350 | } |
3e525e67 EW |
351 | |
352 | # Apply the import/changeset/merge into the working tree | |
353 | my $dir = sync_to_ps($ps); | |
354 | # read the new log entry: | |
355 | my @commitlog = safe_pipe_capture($TLA,'cat-log','-d',$dir,$ps->{id}); | |
356 | die "Error in cat-log: $!" if $?; | |
357 | chomp @commitlog; | |
358 | ||
359 | # grab variables we want from the log, new fields get added to $ps: | |
360 | # (author, date, email, summary, message body ...) | |
361 | parselog($ps, \@commitlog); | |
362 | ||
363 | if ($ps->{id} =~ /--base-0$/ && $ps->{id} ne $psets[0]{id}) { | |
364 | # this should work when importing continuations | |
365 | if ($ps->{tag} && (my $branchpoint = eval { ptag($ps->{tag}) })) { | |
366 | ||
367 | # find where we are supposed to branch from | |
d9cb5399 PB |
368 | if (! -e "$git_dir/refs/heads/$ps->{branch}") { |
369 | system('git-branch',$ps->{branch},$branchpoint) == 0 or die "$! $?\n"; | |
370 | ||
371 | # We trust Arch with the fact that this is just a tag, | |
372 | # and it does not affect the state of the tree, so | |
373 | # we just tag and move on. If the user really wants us | |
374 | # to consolidate more branches into one, don't tag because | |
375 | # the tag name would be already taken. | |
376 | tag($ps->{id}, $branchpoint); | |
377 | ptag($ps->{id}, $branchpoint); | |
378 | print " * Tagged $ps->{id} at $branchpoint\n"; | |
379 | } | |
380 | system('git-checkout','-f',$ps->{branch}) == 0 or die "$! $?\n"; | |
381 | ||
3e525e67 EW |
382 | # remove any old stuff that got leftover: |
383 | my $rm = safe_pipe_capture('git-ls-files','--others','-z'); | |
384 | rmtree(split(/\0/,$rm)) if $rm; | |
3e525e67 EW |
385 | return 0; |
386 | } else { | |
387 | warn "Tagging from unknown id unsupported\n" if $ps->{tag}; | |
388 | } | |
389 | # allow multiple bases/imports here since Arch supports cherry-picks | |
390 | # from unrelated trees | |
391 | } | |
392 | ||
393 | # update the index with all the changes we got | |
3ff903bf EW |
394 | system('git-diff-files --name-only -z | '. |
395 | 'git-update-index --remove -z --stdin') == 0 or die "$! $?\n"; | |
3e525e67 EW |
396 | system('git-ls-files --others -z | '. |
397 | 'git-update-index --add -z --stdin') == 0 or die "$! $?\n"; | |
3e525e67 EW |
398 | return 1; |
399 | } | |
37f15d50 | 400 | |
3e525e67 EW |
401 | # the native changeset processing strategy. This is very fast, but |
402 | # does not handle permissions or any renames involving directories | |
403 | sub process_patchset_fast { | |
404 | my $ps = shift; | |
d3968363 ML |
405 | # |
406 | # create the branch if needed | |
407 | # | |
3292ae47 ML |
408 | if ($ps->{type} eq 'i' && !$import) { |
409 | die "Should not have more than one 'Initial import' per GIT import: $ps->{id}"; | |
d3968363 ML |
410 | } |
411 | ||
3292ae47 | 412 | unless ($import) { # skip for import |
1d4710d0 | 413 | if ( -e "$git_dir/refs/heads/$ps->{branch}") { |
d3968363 | 414 | # we know about this branch |
f88961a8 | 415 | system('git-checkout',$ps->{branch}); |
d3968363 ML |
416 | } else { |
417 | # new branch! we need to verify a few things | |
418 | die "Branch on a non-tag!" unless $ps->{type} eq 't'; | |
419 | my $branchpoint = ptag($ps->{tag}); | |
420 | die "Tagging from unknown id unsupported: $ps->{tag}" | |
421 | unless $branchpoint; | |
422 | ||
423 | # find where we are supposed to branch from | |
d9cb5399 PB |
424 | if (! -e "$git_dir/refs/heads/$ps->{branch}") { |
425 | system('git-branch',$ps->{branch},$branchpoint) == 0 or die "$! $?\n"; | |
426 | ||
427 | # We trust Arch with the fact that this is just a tag, | |
428 | # and it does not affect the state of the tree, so | |
429 | # we just tag and move on. If the user really wants us | |
430 | # to consolidate more branches into one, don't tag because | |
431 | # the tag name would be already taken. | |
432 | tag($ps->{id}, $branchpoint); | |
433 | ptag($ps->{id}, $branchpoint); | |
434 | print " * Tagged $ps->{id} at $branchpoint\n"; | |
435 | } | |
436 | system('git-checkout',$ps->{branch}) == 0 or die "$! $?\n"; | |
3e525e67 | 437 | return 0; |
d3968363 ML |
438 | } |
439 | die $! if $?; | |
440 | } | |
441 | ||
d3968363 ML |
442 | # |
443 | # Apply the import/changeset/merge into the working tree | |
444 | # | |
445 | if ($ps->{type} eq 'i' || $ps->{type} eq 't') { | |
d3968363 | 446 | apply_import($ps) or die $!; |
3e525e67 | 447 | $stats{import_or_tag}++; |
3292ae47 | 448 | $import=0; |
d3968363 ML |
449 | } elsif ($ps->{type} eq 's') { |
450 | apply_cset($ps); | |
3e525e67 | 451 | $stats{simple_changeset}++; |
d3968363 ML |
452 | } |
453 | ||
454 | # | |
455 | # prepare update git's index, based on what arch knows | |
456 | # about the pset, resolve parents, etc | |
457 | # | |
d3968363 | 458 | |
6df896b5 | 459 | my @commitlog = safe_pipe_capture($TLA,'cat-archive-log',$ps->{id}); |
d3968363 ML |
460 | die "Error in cat-archive-log: $!" if $?; |
461 | ||
6df896b5 | 462 | parselog($ps,\@commitlog); |
d3968363 ML |
463 | |
464 | # imports don't give us good info | |
465 | # on added files. Shame on them | |
6df896b5 | 466 | if ($ps->{type} eq 'i' || $ps->{type} eq 't') { |
6df896b5 EW |
467 | system('git-ls-files --deleted -z | '. |
468 | 'git-update-index --remove -z --stdin') == 0 or die "$! $?\n"; | |
3ff903bf EW |
469 | system('git-ls-files --others -z | '. |
470 | 'git-update-index --add -z --stdin') == 0 or die "$! $?\n"; | |
d3968363 ML |
471 | } |
472 | ||
6df896b5 | 473 | # TODO: handle removed_directories and renamed_directories: |
3ff903bf | 474 | |
6df896b5 EW |
475 | if (my $del = $ps->{removed_files}) { |
476 | unlink @$del; | |
d3968363 ML |
477 | while (@$del) { |
478 | my @slice = splice(@$del, 0, 100); | |
6df896b5 EW |
479 | system('git-update-index','--remove','--',@slice) == 0 or |
480 | die "Error in git-update-index --remove: $! $?\n"; | |
d3968363 ML |
481 | } |
482 | } | |
6df896b5 EW |
483 | |
484 | if (my $ren = $ps->{renamed_files}) { # renamed | |
d3968363 ML |
485 | if (@$ren % 2) { |
486 | die "Odd number of entries in rename!?"; | |
487 | } | |
6df896b5 | 488 | |
d3968363 | 489 | while (@$ren) { |
6df896b5 EW |
490 | my $from = shift @$ren; |
491 | my $to = shift @$ren; | |
d3968363 ML |
492 | |
493 | unless (-d dirname($to)) { | |
494 | mkpath(dirname($to)); # will die on err | |
495 | } | |
3e525e67 | 496 | # print "moving $from $to"; |
6df896b5 EW |
497 | rename($from, $to) or die "Error renaming '$from' '$to': $!\n"; |
498 | system('git-update-index','--remove','--',$from) == 0 or | |
499 | die "Error in git-update-index --remove: $! $?\n"; | |
500 | system('git-update-index','--add','--',$to) == 0 or | |
501 | die "Error in git-update-index --add: $! $?\n"; | |
d3968363 | 502 | } |
d3968363 | 503 | } |
6df896b5 | 504 | |
3ff903bf EW |
505 | if (my $add = $ps->{new_files}) { |
506 | while (@$add) { | |
507 | my @slice = splice(@$add, 0, 100); | |
508 | system('git-update-index','--add','--',@slice) == 0 or | |
509 | die "Error in git-update-index --add: $! $?\n"; | |
510 | } | |
511 | } | |
512 | ||
6df896b5 | 513 | if (my $mod = $ps->{modified_files}) { |
d3968363 ML |
514 | while (@$mod) { |
515 | my @slice = splice(@$mod, 0, 100); | |
6df896b5 EW |
516 | system('git-update-index','--',@slice) == 0 or |
517 | die "Error in git-update-index: $! $?\n"; | |
d3968363 ML |
518 | } |
519 | } | |
3e525e67 EW |
520 | return 1; # we successfully applied the changeset |
521 | } | |
522 | ||
523 | if ($opt_f) { | |
524 | print "Will import patchsets using the fast strategy\n", | |
525 | "Renamed directories and permission changes will be missed\n"; | |
526 | *process_patchset = *process_patchset_fast; | |
527 | } else { | |
528 | print "Using the default (accurate) import strategy.\n", | |
529 | "Things may be a bit slow\n"; | |
530 | *process_patchset = *process_patchset_accurate; | |
531 | } | |
6df896b5 | 532 | |
3e525e67 EW |
533 | foreach my $ps (@psets) { |
534 | # process patchsets | |
535 | $ps->{branch} = git_branchname($ps->{id}); | |
536 | ||
537 | # | |
538 | # ensure we have a clean state | |
539 | # | |
540 | if (my $dirty = `git-diff-files`) { | |
541 | die "Unclean tree when about to process $ps->{id} " . | |
542 | " - did we fail to commit cleanly before?\n$dirty"; | |
543 | } | |
544 | die $! if $?; | |
545 | ||
546 | # | |
547 | # skip commits already in repo | |
548 | # | |
549 | if (ptag($ps->{id})) { | |
550 | $opt_v && print " * Skipping already imported: $ps->{id}\n"; | |
10945e00 | 551 | next; |
3e525e67 EW |
552 | } |
553 | ||
554 | print " * Starting to work on $ps->{id}\n"; | |
555 | ||
556 | process_patchset($ps) or next; | |
557 | ||
215a7ad1 | 558 | # warn "errors when running git-update-index! $!"; |
3e525e67 | 559 | my $tree = `git-write-tree`; |
d3968363 ML |
560 | die "cannot write tree $!" if $?; |
561 | chomp $tree; | |
d3968363 ML |
562 | |
563 | # | |
564 | # Who's your daddy? | |
565 | # | |
566 | my @par; | |
1d4710d0 | 567 | if ( -e "$git_dir/refs/heads/$ps->{branch}") { |
f88961a8 | 568 | if (open HEAD, "<","$git_dir/refs/heads/$ps->{branch}") { |
d3968363 ML |
569 | my $p = <HEAD>; |
570 | close HEAD; | |
571 | chomp $p; | |
572 | push @par, '-p', $p; | |
573 | } else { | |
574 | if ($ps->{type} eq 's') { | |
575 | warn "Could not find the right head for the branch $ps->{branch}"; | |
576 | } | |
577 | } | |
578 | } | |
579 | ||
b779d5f0 ML |
580 | if ($ps->{merges}) { |
581 | push @par, find_parents($ps); | |
582 | } | |
d3968363 ML |
583 | |
584 | # | |
585 | # Commit, tag and clean state | |
586 | # | |
587 | $ENV{TZ} = 'GMT'; | |
588 | $ENV{GIT_AUTHOR_NAME} = $ps->{author}; | |
589 | $ENV{GIT_AUTHOR_EMAIL} = $ps->{email}; | |
590 | $ENV{GIT_AUTHOR_DATE} = $ps->{date}; | |
591 | $ENV{GIT_COMMITTER_NAME} = $ps->{author}; | |
592 | $ENV{GIT_COMMITTER_EMAIL} = $ps->{email}; | |
593 | $ENV{GIT_COMMITTER_DATE} = $ps->{date}; | |
594 | ||
6df896b5 | 595 | my $pid = open2(*READER, *WRITER,'git-commit-tree',$tree,@par) |
d3968363 | 596 | or die $!; |
a94f457e | 597 | print WRITER $ps->{summary},"\n\n"; |
6df896b5 EW |
598 | print WRITER $ps->{message},"\n"; |
599 | ||
600 | # make it easy to backtrack and figure out which Arch revision this was: | |
601 | print WRITER 'git-archimport-id: ',$ps->{id},"\n"; | |
602 | ||
d3968363 ML |
603 | close WRITER; |
604 | my $commitid = <READER>; # read | |
605 | chomp $commitid; | |
606 | close READER; | |
607 | waitpid $pid,0; # close; | |
608 | ||
609 | if (length $commitid != 40) { | |
610 | die "Something went wrong with the commit! $! $commitid"; | |
611 | } | |
612 | # | |
613 | # Update the branch | |
614 | # | |
f88961a8 | 615 | open HEAD, ">","$git_dir/refs/heads/$ps->{branch}"; |
d3968363 ML |
616 | print HEAD $commitid; |
617 | close HEAD; | |
8366a10a | 618 | system('git-update-ref', 'HEAD', "$ps->{branch}"); |
d3968363 ML |
619 | |
620 | # tag accordingly | |
621 | ptag($ps->{id}, $commitid); # private tag | |
622 | if ($opt_T || $ps->{type} eq 't' || $ps->{type} eq 'i') { | |
623 | tag($ps->{id}, $commitid); | |
624 | } | |
625 | print " * Committed $ps->{id}\n"; | |
626 | print " + tree $tree\n"; | |
627 | print " + commit $commitid\n"; | |
b779d5f0 | 628 | $opt_v && print " + commit date is $ps->{date} \n"; |
f88961a8 | 629 | $opt_v && print " + parents: ",join(' ',@par),"\n"; |
3e525e67 EW |
630 | } |
631 | ||
632 | if ($opt_v) { | |
633 | foreach (sort keys %stats) { | |
634 | print" $_: $stats{$_}\n"; | |
635 | } | |
636 | } | |
637 | exit 0; | |
638 | ||
639 | # used by the accurate strategy: | |
640 | sub sync_to_ps { | |
641 | my $ps = shift; | |
642 | my $tree_dir = $tmp.'/'.tree_dirname($ps->{id}); | |
643 | ||
644 | $opt_v && print "sync_to_ps($ps->{id}) method: "; | |
645 | ||
646 | if (-d $tree_dir) { | |
647 | if ($ps->{type} eq 't') { | |
648 | $opt_v && print "get (tag)\n"; | |
649 | # looks like a tag-only or (worse,) a mixed tags/changeset branch, | |
650 | # can't rely on replay to work correctly on these | |
651 | rmtree($tree_dir); | |
652 | safe_pipe_capture($TLA,'get','--no-pristine',$ps->{id},$tree_dir); | |
653 | $stats{get_tag}++; | |
654 | } else { | |
655 | my $tree_id = arch_tree_id($tree_dir); | |
656 | if ($ps->{parent_id} && ($ps->{parent_id} eq $tree_id)) { | |
657 | # the common case (hopefully) | |
658 | $opt_v && print "replay\n"; | |
659 | safe_pipe_capture($TLA,'replay','-d',$tree_dir,$ps->{id}); | |
660 | $stats{replay}++; | |
661 | } else { | |
662 | # getting one tree is usually faster than getting two trees | |
663 | # and applying the delta ... | |
664 | rmtree($tree_dir); | |
665 | $opt_v && print "apply-delta\n"; | |
666 | safe_pipe_capture($TLA,'get','--no-pristine', | |
667 | $ps->{id},$tree_dir); | |
668 | $stats{get_delta}++; | |
669 | } | |
670 | } | |
671 | } else { | |
672 | # new branch work | |
673 | $opt_v && print "get (new tree)\n"; | |
674 | safe_pipe_capture($TLA,'get','--no-pristine',$ps->{id},$tree_dir); | |
675 | $stats{get_new}++; | |
676 | } | |
677 | ||
678 | # added -I flag to rsync since we're going to fast! AIEEEEE!!!! | |
679 | system('rsync','-aI','--delete','--exclude',$git_dir, | |
680 | # '--exclude','.arch-inventory', | |
681 | '--exclude','.arch-ids','--exclude','{arch}', | |
682 | '--exclude','+*','--exclude',',*', | |
683 | "$tree_dir/",'./') == 0 or die "Cannot rsync $tree_dir: $! $?"; | |
684 | return $tree_dir; | |
d3968363 ML |
685 | } |
686 | ||
d3968363 ML |
687 | sub apply_import { |
688 | my $ps = shift; | |
22ff00fc | 689 | my $bname = git_branchname($ps->{id}); |
d3968363 | 690 | |
f88961a8 | 691 | mkpath($tmp); |
d3968363 | 692 | |
f88961a8 | 693 | safe_pipe_capture($TLA,'get','-s','--no-pristine',$ps->{id},"$tmp/import"); |
d3968363 | 694 | die "Cannot get import: $!" if $?; |
f88961a8 EW |
695 | system('rsync','-aI','--delete', '--exclude',$git_dir, |
696 | '--exclude','.arch-ids','--exclude','{arch}', | |
697 | "$tmp/import/", './'); | |
d3968363 ML |
698 | die "Cannot rsync import:$!" if $?; |
699 | ||
f88961a8 | 700 | rmtree("$tmp/import"); |
d3968363 ML |
701 | die "Cannot remove tempdir: $!" if $?; |
702 | ||
703 | ||
704 | return 1; | |
705 | } | |
706 | ||
707 | sub apply_cset { | |
708 | my $ps = shift; | |
709 | ||
f88961a8 | 710 | mkpath($tmp); |
d3968363 ML |
711 | |
712 | # get the changeset | |
f88961a8 | 713 | safe_pipe_capture($TLA,'get-changeset',$ps->{id},"$tmp/changeset"); |
d3968363 ML |
714 | die "Cannot get changeset: $!" if $?; |
715 | ||
716 | # apply patches | |
717 | if (`find $tmp/changeset/patches -type f -name '*.patch'`) { | |
718 | # this can be sped up considerably by doing | |
719 | # (find | xargs cat) | patch | |
82e5a82f | 720 | # but that can get mucked up by patches |
d3968363 ML |
721 | # with missing trailing newlines or the standard |
722 | # 'missing newline' flag in the patch - possibly | |
723 | # produced with an old/buggy diff. | |
724 | # slow and safe, we invoke patch once per patchfile | |
725 | `find $tmp/changeset/patches -type f -name '*.patch' -print0 | grep -zv '{arch}' | xargs -iFILE -0 --no-run-if-empty patch -p1 --forward -iFILE`; | |
726 | die "Problem applying patches! $!" if $?; | |
727 | } | |
728 | ||
729 | # apply changed binary files | |
730 | if (my @modified = `find $tmp/changeset/patches -type f -name '*.modified'`) { | |
731 | foreach my $mod (@modified) { | |
732 | chomp $mod; | |
733 | my $orig = $mod; | |
734 | $orig =~ s/\.modified$//; # lazy | |
735 | $orig =~ s!^\Q$tmp\E/changeset/patches/!!; | |
736 | #print "rsync -p '$mod' '$orig'"; | |
f88961a8 | 737 | system('rsync','-p',$mod,"./$orig"); |
d3968363 ML |
738 | die "Problem applying binary changes! $!" if $?; |
739 | } | |
740 | } | |
741 | ||
742 | # bring in new files | |
f88961a8 EW |
743 | system('rsync','-aI','--exclude',$git_dir, |
744 | '--exclude','.arch-ids', | |
745 | '--exclude', '{arch}', | |
746 | "$tmp/changeset/new-files-archive/",'./'); | |
d3968363 ML |
747 | |
748 | # deleted files are hinted from the commitlog processing | |
749 | ||
f88961a8 | 750 | rmtree("$tmp/changeset"); |
d3968363 ML |
751 | } |
752 | ||
753 | ||
754 | # =for reference | |
6df896b5 EW |
755 | # notes: *-files/-directories keys cannot have spaces, they're always |
756 | # pika-escaped. Everything after the first newline | |
757 | # A log entry looks like: | |
d3968363 ML |
758 | # Revision: moodle-org--moodle--1.3.3--patch-15 |
759 | # Archive: arch-eduforge@catalyst.net.nz--2004 | |
760 | # Creator: Penny Leach <penny@catalyst.net.nz> | |
761 | # Date: Wed May 25 14:15:34 NZST 2005 | |
762 | # Standard-date: 2005-05-25 02:15:34 GMT | |
763 | # New-files: lang/de/.arch-ids/block_glossary_random.php.id | |
764 | # lang/de/.arch-ids/block_html.php.id | |
765 | # New-directories: lang/de/help/questionnaire | |
766 | # lang/de/help/questionnaire/.arch-ids | |
767 | # Renamed-files: .arch-ids/db_sears.sql.id db/.arch-ids/db_sears.sql.id | |
768 | # db_sears.sql db/db_sears.sql | |
769 | # Removed-files: lang/be/docs/.arch-ids/release.html.id | |
770 | # lang/be/docs/.arch-ids/releaseold.html.id | |
771 | # Modified-files: admin/cron.php admin/delete.php | |
772 | # admin/editor.html backup/lib.php backup/restore.php | |
773 | # New-patches: arch-eduforge@catalyst.net.nz--2004/moodle-org--moodle--1.3.3--patch-15 | |
774 | # Summary: Updating to latest from MOODLE_14_STABLE (1.4.5+) | |
6df896b5 | 775 | # summary can be multiline with a leading space just like the above fields |
d3968363 ML |
776 | # Keywords: |
777 | # | |
778 | # Updating yadda tadda tadda madda | |
779 | sub parselog { | |
6df896b5 EW |
780 | my ($ps, $log) = @_; |
781 | my $key = undef; | |
782 | ||
783 | # headers we want that contain filenames: | |
784 | my %want_headers = ( | |
785 | new_files => 1, | |
786 | modified_files => 1, | |
787 | renamed_files => 1, | |
788 | renamed_directories => 1, | |
789 | removed_files => 1, | |
790 | removed_directories => 1, | |
791 | ); | |
d3968363 | 792 | |
6df896b5 EW |
793 | chomp (@$log); |
794 | while ($_ = shift @$log) { | |
795 | if (/^Continuation-of:\s*(.*)/) { | |
796 | $ps->{tag} = $1; | |
797 | $key = undef; | |
798 | } elsif (/^Summary:\s*(.*)$/ ) { | |
a94f457e PB |
799 | # summary can be multiline as long as it has a leading space. |
800 | # we squeeze it onto a single line, though. | |
6df896b5 EW |
801 | $ps->{summary} = [ $1 ]; |
802 | $key = 'summary'; | |
803 | } elsif (/^Creator: (.*)\s*<([^\>]+)>/) { | |
804 | $ps->{author} = $1; | |
805 | $ps->{email} = $2; | |
806 | $key = undef; | |
807 | # any *-files or *-directories can be read here: | |
808 | } elsif (/^([A-Z][a-z\-]+):\s*(.*)$/) { | |
809 | my $val = $2; | |
810 | $key = lc $1; | |
811 | $key =~ tr/-/_/; # too lazy to quote :P | |
812 | if ($want_headers{$key}) { | |
813 | push @{$ps->{$key}}, split(/\s+/, $val); | |
814 | } else { | |
815 | $key = undef; | |
816 | } | |
817 | } elsif (/^$/) { | |
818 | last; # remainder of @$log that didn't get shifted off is message | |
819 | } elsif ($key) { | |
820 | if (/^\s+(.*)$/) { | |
821 | if ($key eq 'summary') { | |
822 | push @{$ps->{$key}}, $1; | |
823 | } else { # files/directories: | |
824 | push @{$ps->{$key}}, split(/\s+/, $1); | |
825 | } | |
826 | } else { | |
827 | $key = undef; | |
828 | } | |
829 | } | |
d3968363 | 830 | } |
6df896b5 | 831 | |
a94f457e PB |
832 | # drop leading empty lines from the log message |
833 | while (@$log && $log->[0] eq '') { | |
834 | shift @$log; | |
835 | } | |
836 | if (exists $ps->{summary} && @{$ps->{summary}}) { | |
837 | $ps->{summary} = join(' ', @{$ps->{summary}}); | |
838 | } | |
839 | elsif (@$log == 0) { | |
840 | $ps->{summary} = 'empty commit message'; | |
841 | } else { | |
842 | $ps->{summary} = $log->[0] . '...'; | |
843 | } | |
6df896b5 | 844 | $ps->{message} = join("\n",@$log); |
d3968363 | 845 | |
6df896b5 EW |
846 | # skip Arch control files, unescape pika-escaped files |
847 | foreach my $k (keys %want_headers) { | |
848 | next unless (defined $ps->{$k}); | |
6e33101a | 849 | my @tmp = (); |
6df896b5 EW |
850 | foreach my $t (@{$ps->{$k}}) { |
851 | next unless length ($t); | |
852 | next if $t =~ m!\{arch\}/!; | |
853 | next if $t =~ m!\.arch-ids/!; | |
854 | # should we skip this? | |
855 | next if $t =~ m!\.arch-inventory$!; | |
f84f9d38 ML |
856 | # tla cat-archive-log will give us filenames with spaces as file\(sp)name - why? |
857 | # we can assume that any filename with \ indicates some pika escaping that we want to get rid of. | |
6df896b5 | 858 | if ($t =~ /\\/ ){ |
f88961a8 | 859 | $t = (safe_pipe_capture($TLA,'escape','--unescaped',$t))[0]; |
f84f9d38 | 860 | } |
6df896b5 | 861 | push @tmp, $t; |
d3968363 | 862 | } |
6e33101a | 863 | $ps->{$k} = \@tmp; |
d3968363 | 864 | } |
d3968363 ML |
865 | } |
866 | ||
867 | # write/read a tag | |
868 | sub tag { | |
869 | my ($tag, $commit) = @_; | |
a7fb51d3 | 870 | |
fee3365f ML |
871 | if ($opt_o) { |
872 | $tag =~ s|/|--|g; | |
873 | } else { | |
d9cb5399 PB |
874 | my $patchname = $tag; |
875 | $patchname =~ s/.*--//; | |
876 | $tag = git_branchname ($tag) . '--' . $patchname; | |
fee3365f | 877 | } |
d3968363 ML |
878 | |
879 | if ($commit) { | |
a7fb51d3 | 880 | open(C,">","$git_dir/refs/tags/$tag") |
d3968363 ML |
881 | or die "Cannot create tag $tag: $!\n"; |
882 | print C "$commit\n" | |
883 | or die "Cannot write tag $tag: $!\n"; | |
884 | close(C) | |
885 | or die "Cannot write tag $tag: $!\n"; | |
a7fb51d3 | 886 | print " * Created tag '$tag' on '$commit'\n" if $opt_v; |
d3968363 | 887 | } else { # read |
a7fb51d3 | 888 | open(C,"<","$git_dir/refs/tags/$tag") |
d3968363 ML |
889 | or die "Cannot read tag $tag: $!\n"; |
890 | $commit = <C>; | |
891 | chomp $commit; | |
892 | die "Error reading tag $tag: $!\n" unless length $commit == 40; | |
893 | close(C) | |
894 | or die "Cannot read tag $tag: $!\n"; | |
895 | return $commit; | |
896 | } | |
897 | } | |
898 | ||
899 | # write/read a private tag | |
900 | # reads fail softly if the tag isn't there | |
901 | sub ptag { | |
902 | my ($tag, $commit) = @_; | |
a7fb51d3 EW |
903 | |
904 | # don't use subdirs for tags yet, it could screw up other porcelains | |
905 | $tag =~ s|/|,|g; | |
d3968363 | 906 | |
a7fb51d3 EW |
907 | my $tag_file = "$ptag_dir/$tag"; |
908 | my $tag_branch_dir = dirname($tag_file); | |
909 | mkpath($tag_branch_dir) unless (-d $tag_branch_dir); | |
d3968363 ML |
910 | |
911 | if ($commit) { # write | |
a7fb51d3 | 912 | open(C,">",$tag_file) |
d3968363 ML |
913 | or die "Cannot create tag $tag: $!\n"; |
914 | print C "$commit\n" | |
915 | or die "Cannot write tag $tag: $!\n"; | |
916 | close(C) | |
917 | or die "Cannot write tag $tag: $!\n"; | |
b779d5f0 ML |
918 | $rptags{$commit} = $tag |
919 | unless $tag =~ m/--base-0$/; | |
d3968363 ML |
920 | } else { # read |
921 | # if the tag isn't there, return 0 | |
a7fb51d3 | 922 | unless ( -s $tag_file) { |
d3968363 ML |
923 | return 0; |
924 | } | |
a7fb51d3 | 925 | open(C,"<",$tag_file) |
d3968363 ML |
926 | or die "Cannot read tag $tag: $!\n"; |
927 | $commit = <C>; | |
928 | chomp $commit; | |
929 | die "Error reading tag $tag: $!\n" unless length $commit == 40; | |
930 | close(C) | |
931 | or die "Cannot read tag $tag: $!\n"; | |
b779d5f0 ML |
932 | unless (defined $rptags{$commit}) { |
933 | $rptags{$commit} = $tag; | |
934 | } | |
d3968363 ML |
935 | return $commit; |
936 | } | |
937 | } | |
b779d5f0 ML |
938 | |
939 | sub find_parents { | |
940 | # | |
941 | # Identify what branches are merging into me | |
942 | # and whether we are fully merged | |
943 | # git-merge-base <headsha> <headsha> should tell | |
944 | # me what the base of the merge should be | |
945 | # | |
946 | my $ps = shift; | |
947 | ||
948 | my %branches; # holds an arrayref per branch | |
949 | # the arrayref contains a list of | |
950 | # merged patches between the base | |
951 | # of the merge and the current head | |
952 | ||
953 | my @parents; # parents found for this commit | |
954 | ||
955 | # simple loop to split the merges | |
956 | # per branch | |
957 | foreach my $merge (@{$ps->{merges}}) { | |
22ff00fc | 958 | my $branch = git_branchname($merge); |
b779d5f0 ML |
959 | unless (defined $branches{$branch} ){ |
960 | $branches{$branch} = []; | |
961 | } | |
962 | push @{$branches{$branch}}, $merge; | |
963 | } | |
964 | ||
965 | # | |
966 | # foreach branch find a merge base and walk it to the | |
967 | # head where we are, collecting the merged patchsets that | |
968 | # Arch has recorded. Keep that in @have | |
969 | # Compare that with the commits on the other branch | |
970 | # between merge-base and the tip of the branch (@need) | |
971 | # and see if we have a series of consecutive patches | |
972 | # starting from the merge base. The tip of the series | |
973 | # of consecutive patches merged is our new parent for | |
974 | # that branch. | |
975 | # | |
976 | foreach my $branch (keys %branches) { | |
37f15d50 ML |
977 | |
978 | # check that we actually know about the branch | |
979 | next unless -e "$git_dir/refs/heads/$branch"; | |
980 | ||
b779d5f0 | 981 | my $mergebase = `git-merge-base $branch $ps->{branch}`; |
9b626e75 EW |
982 | if ($?) { |
983 | # Don't die here, Arch supports one-way cherry-picking | |
984 | # between branches with no common base (or any relationship | |
985 | # at all beforehand) | |
986 | warn "Cannot find merge base for $branch and $ps->{branch}"; | |
987 | next; | |
988 | } | |
b779d5f0 ML |
989 | chomp $mergebase; |
990 | ||
991 | # now walk up to the mergepoint collecting what patches we have | |
992 | my $branchtip = git_rev_parse($ps->{branch}); | |
765ac8ec | 993 | my @ancestors = `git-rev-list --topo-order $branchtip ^$mergebase`; |
b779d5f0 ML |
994 | my %have; # collected merges this branch has |
995 | foreach my $merge (@{$ps->{merges}}) { | |
996 | $have{$merge} = 1; | |
997 | } | |
998 | my %ancestorshave; | |
999 | foreach my $par (@ancestors) { | |
1000 | $par = commitid2pset($par); | |
1001 | if (defined $par->{merges}) { | |
1002 | foreach my $merge (@{$par->{merges}}) { | |
1003 | $ancestorshave{$merge}=1; | |
1004 | } | |
1005 | } | |
1006 | } | |
1007 | # print "++++ Merges in $ps->{id} are....\n"; | |
1008 | # my @have = sort keys %have; print Dumper(\@have); | |
1009 | ||
1010 | # merge what we have with what ancestors have | |
1011 | %have = (%have, %ancestorshave); | |
1012 | ||
1013 | # see what the remote branch has - these are the merges we | |
1014 | # will want to have in a consecutive series from the mergebase | |
1015 | my $otherbranchtip = git_rev_parse($branch); | |
765ac8ec | 1016 | my @needraw = `git-rev-list --topo-order $otherbranchtip ^$mergebase`; |
b779d5f0 ML |
1017 | my @need; |
1018 | foreach my $needps (@needraw) { # get the psets | |
1019 | $needps = commitid2pset($needps); | |
1020 | # git-rev-list will also | |
1021 | # list commits merged in via earlier | |
1022 | # merges. we are only interested in commits | |
1023 | # from the branch we're looking at | |
1024 | if ($branch eq $needps->{branch}) { | |
1025 | push @need, $needps->{id}; | |
1026 | } | |
1027 | } | |
1028 | ||
1029 | # print "++++ Merges from $branch we want are....\n"; | |
1030 | # print Dumper(\@need); | |
1031 | ||
1032 | my $newparent; | |
1033 | while (my $needed_commit = pop @need) { | |
1034 | if ($have{$needed_commit}) { | |
1035 | $newparent = $needed_commit; | |
1036 | } else { | |
1037 | last; # break out of the while | |
1038 | } | |
1039 | } | |
1040 | if ($newparent) { | |
1041 | push @parents, $newparent; | |
1042 | } | |
1043 | ||
1044 | ||
1045 | } # end foreach branch | |
1046 | ||
1047 | # prune redundant parents | |
1048 | my %parents; | |
1049 | foreach my $p (@parents) { | |
1050 | $parents{$p} = 1; | |
1051 | } | |
1052 | foreach my $p (@parents) { | |
1053 | next unless exists $psets{$p}{merges}; | |
1054 | next unless ref $psets{$p}{merges}; | |
1055 | my @merges = @{$psets{$p}{merges}}; | |
1056 | foreach my $merge (@merges) { | |
1057 | if ($parents{$merge}) { | |
1058 | delete $parents{$merge}; | |
1059 | } | |
1060 | } | |
1061 | } | |
42f44b08 | 1062 | |
f88961a8 EW |
1063 | @parents = (); |
1064 | foreach (keys %parents) { | |
1065 | push @parents, '-p', ptag($_); | |
1066 | } | |
b779d5f0 ML |
1067 | return @parents; |
1068 | } | |
1069 | ||
1070 | sub git_rev_parse { | |
1071 | my $name = shift; | |
1072 | my $val = `git-rev-parse $name`; | |
1073 | die "Error: git-rev-parse $name" if $?; | |
1074 | chomp $val; | |
1075 | return $val; | |
1076 | } | |
1077 | ||
1078 | # resolve a SHA1 to a known patchset | |
1079 | sub commitid2pset { | |
1080 | my $commitid = shift; | |
1081 | chomp $commitid; | |
1082 | my $name = $rptags{$commitid} | |
1083 | || die "Cannot find reverse tag mapping for $commitid"; | |
a7fb51d3 | 1084 | $name =~ s|,|/|; |
b779d5f0 ML |
1085 | my $ps = $psets{$name} |
1086 | || (print Dumper(sort keys %psets)) && die "Cannot find patchset for $name"; | |
1087 | return $ps; | |
1088 | } | |
2777ef76 | 1089 | |
42f44b08 | 1090 | |
82e5a82f | 1091 | # an alternative to `command` that allows input to be passed as an array |
2777ef76 EW |
1092 | # to work around shell problems with weird characters in arguments |
1093 | sub safe_pipe_capture { | |
1094 | my @output; | |
1095 | if (my $pid = open my $child, '-|') { | |
1096 | @output = (<$child>); | |
1097 | close $child or die join(' ',@_).": $! $?"; | |
1098 | } else { | |
3e525e67 | 1099 | exec(@_) or die "$! $?"; # exec() can fail the executable can't be found |
2777ef76 EW |
1100 | } |
1101 | return wantarray ? @output : join('',@output); | |
1102 | } | |
1103 | ||
42f44b08 EW |
1104 | # `tla logs -rf -d <dir> | head -n1` or `baz tree-id <dir>` |
1105 | sub arch_tree_id { | |
1106 | my $dir = shift; | |
1107 | chomp( my $ret = (safe_pipe_capture($TLA,'logs','-rf','-d',$dir))[0] ); | |
1108 | return $ret; | |
1109 | } | |
1110 | ||
1111 | sub archive_reachable { | |
1112 | my $archive = shift; | |
1113 | return 1 if $reachable{$archive}; | |
1114 | return 0 if $unreachable{$archive}; | |
1115 | ||
1116 | if (system "$TLA whereis-archive $archive >/dev/null") { | |
1117 | if ($opt_a && (system($TLA,'register-archive', | |
1118 | "http://mirrors.sourcecontrol.net/$archive") == 0)) { | |
1119 | $reachable{$archive} = 1; | |
1120 | return 1; | |
1121 | } | |
1122 | print STDERR "Archive is unreachable: $archive\n"; | |
1123 | $unreachable{$archive} = 1; | |
1124 | return 0; | |
1125 | } else { | |
1126 | $reachable{$archive} = 1; | |
1127 | return 1; | |
1128 | } | |
1129 | } | |
2777ef76 | 1130 |