?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
t4012: use 'printf' instead of 'dd' to generate a binary file
[git/gitweb.git] / git-svn.perl
blob0b074c4c63ebb6e1334cf05371cd22087fb79f63
1 #!/usr/bin/perl
2 # Copyright (C) 2006, Eric Wong <normalperson@yhbt.net>
3 # License: GPL v2 or later
4 use 5.008;
5 use warnings;
6 use strict;
7 use vars qw/ $AUTHOR $VERSION
8 $sha1 $sha1_short $_revision $_repository
9 $_q $_authors $_authors_prog %users/;
10 $AUTHOR = 'Eric Wong <normalperson@yhbt.net>';
11 $VERSION = '@@GIT_VERSION@@';
12
13 # From which subdir have we been invoked?
14 my $cmd_dir_prefix = eval {
15 command_oneline([qw/rev-parse --show-prefix/], STDERR => 0)
16 } || '';
17
18 my $git_dir_user_set = 1 if defined $ENV{GIT_DIR};
19 $ENV{GIT_DIR} ||= '.git';
20 $Git::SVN::default_repo_id = 'svn';
21 $Git::SVN::default_ref_id = $ENV{GIT_SVN_ID} || 'git-svn';
22 $Git::SVN::Ra::_log_window_size = 100;
23 $Git::SVN::_minimize_url = 'unset';
24
25 if (! exists $ENV{SVN_SSH} && exists $ENV{GIT_SSH}) {
26 $ENV{SVN_SSH} = $ENV{GIT_SSH};
27 }
28
29 if (exists $ENV{SVN_SSH} && $^O eq 'msys') {
30 $ENV{SVN_SSH} =~ s/\\/\\\\/g;
31 $ENV{SVN_SSH} =~ s/(.*)/"$1"/;
32 }
33
34 $Git::SVN::Log::TZ = $ENV{TZ};
35 $ENV{TZ} = 'UTC';
36 $| = 1; # unbuffer STDOUT
37
38 sub fatal (@) { print STDERR "@_\n"; exit 1 }
39
40 # All SVN commands do it. Otherwise we may die on SIGPIPE when the remote
41 # repository decides to close the connection which we expect to be kept alive.
42 $SIG{PIPE} = 'IGNORE';
43
44 # Given a dot separated version number, "subtract" it from
45 # the SVN::Core::VERSION; non-negaitive return means the SVN::Core
46 # is at least at the version the caller asked for.
47 sub compare_svn_version {
48 my (@ours) = split(/\./, $SVN::Core::VERSION);
49 my (@theirs) = split(/\./, $_[0]);
50 my ($i, $diff);
51
52 for ($i = 0; $i < @ours && $i < @theirs; $i++) {
53 $diff = $ours[$i] - $theirs[$i];
54 return $diff if ($diff);
55 }
56 return 1 if ($i < @ours);
57 return -1 if ($i < @theirs);
58 return 0;
59 }
60
61 sub _req_svn {
62 require SVN::Core; # use()-ing this causes segfaults for me... *shrug*
63 require SVN::Ra;
64 require SVN::Delta;
65 if (::compare_svn_version('1.1.0') < 0) {
66 fatal "Need SVN::Core 1.1.0 or better (got $SVN::Core::VERSION)";
67 }
68 }
69 my $can_compress = eval { require Compress::Zlib; 1};
70 use Carp qw/croak/;
71 use Digest::MD5;
72 use IO::File qw//;
73 use File::Basename qw/dirname basename/;
74 use File::Path qw/mkpath/;
75 use File::Spec;
76 use File::Find;
77 use Getopt::Long qw/:config gnu_getopt no_ignore_case auto_abbrev/;
78 use IPC::Open3;
79 use Git;
80 use Git::SVN::Editor qw//;
81 use Git::SVN::Fetcher qw//;
82 use Git::SVN::Ra qw//;
83 use Git::SVN::Prompt qw//;
84 use Memoize; # core since 5.8.0, Jul 2002
85
86 BEGIN {
87 # import functions from Git into our packages, en masse
88 no strict 'refs';
89 foreach (qw/command command_oneline command_noisy command_output_pipe
90 command_input_pipe command_close_pipe
91 command_bidi_pipe command_close_bidi_pipe/) {
92 for my $package ( qw(Git::SVN::Migration Git::SVN::Log Git::SVN),
93 __PACKAGE__) {
94 *{"${package}::$_"} = \&{"Git::$_"};
95 }
96 }
97 Memoize::memoize 'Git::config';
98 Memoize::memoize 'Git::config_bool';
99 }
100
101 my ($SVN);
102
103 $sha1 = qr/[a-f\d]{40}/;
104 $sha1_short = qr/[a-f\d]{4,40}/;
105 my ($_stdin, $_help, $_edit,
106 $_message, $_file, $_branch_dest,
107 $_template, $_shared,
108 $_version, $_fetch_all, $_no_rebase, $_fetch_parent,
109 $_merge, $_strategy, $_preserve_merges, $_dry_run, $_local,
110 $_prefix, $_no_checkout, $_url, $_verbose,
111 $_git_format, $_commit_url, $_tag, $_merge_info, $_interactive);
112 $Git::SVN::_follow_parent = 1;
113 $Git::SVN::Fetcher::_placeholder_filename = ".gitignore";
114 $_q ||= 0;
115 my %remote_opts = ( 'username=s' => \$Git::SVN::Prompt::_username,
116 'config-dir=s' => \$Git::SVN::Ra::config_dir,
117 'no-auth-cache' => \$Git::SVN::Prompt::_no_auth_cache,
118 'ignore-paths=s' => \$Git::SVN::Fetcher::_ignore_regex,
119 'ignore-refs=s' => \$Git::SVN::Ra::_ignore_refs_regex );
120 my %fc_opts = ( 'follow-parent|follow!' => \$Git::SVN::_follow_parent,
121 'authors-file|A=s' => \$_authors,
122 'authors-prog=s' => \$_authors_prog,
123 'repack:i' => \$Git::SVN::_repack,
124 'noMetadata' => \$Git::SVN::_no_metadata,
125 'useSvmProps' => \$Git::SVN::_use_svm_props,
126 'useSvnsyncProps' => \$Git::SVN::_use_svnsync_props,
127 'log-window-size=i' => \$Git::SVN::Ra::_log_window_size,
128 'no-checkout' => \$_no_checkout,
129 'quiet|q+' => \$_q,
130 'repack-flags|repack-args|repack-opts=s' =>
131 \$Git::SVN::_repack_flags,
132 'use-log-author' => \$Git::SVN::_use_log_author,
133 'add-author-from' => \$Git::SVN::_add_author_from,
134 'localtime' => \$Git::SVN::_localtime,
135 %remote_opts );
136
137 my ($_trunk, @_tags, @_branches, $_stdlayout);
138 my %icv;
139 my %init_opts = ( 'template=s' => \$_template, 'shared:s' => \$_shared,
140 'trunk|T=s' => \$_trunk, 'tags|t=s@' => \@_tags,
141 'branches|b=s@' => \@_branches, 'prefix=s' => \$_prefix,
142 'stdlayout|s' => \$_stdlayout,
143 'minimize-url|m!' => \$Git::SVN::_minimize_url,
144 'no-metadata' => sub { $icv{noMetadata} = 1 },
145 'use-svm-props' => sub { $icv{useSvmProps} = 1 },
146 'use-svnsync-props' => sub { $icv{useSvnsyncProps} = 1 },
147 'rewrite-root=s' => sub { $icv{rewriteRoot} = $_[1] },
148 'rewrite-uuid=s' => sub { $icv{rewriteUUID} = $_[1] },
149 %remote_opts );
150 my %cmt_opts = ( 'edit|e' => \$_edit,
151 'rmdir' => \$Git::SVN::Editor::_rmdir,
152 'find-copies-harder' => \$Git::SVN::Editor::_find_copies_harder,
153 'l=i' => \$Git::SVN::Editor::_rename_limit,
154 'copy-similarity|C=i'=> \$Git::SVN::Editor::_cp_similarity
155 );
156
157 my %cmd = (
158 fetch => [ \&cmd_fetch, "Download new revisions from SVN",
159 { 'revision|r=s' => \$_revision,
160 'fetch-all|all' => \$_fetch_all,
161 'parent|p' => \$_fetch_parent,
162 %fc_opts } ],
163 clone => [ \&cmd_clone, "Initialize and fetch revisions",
164 { 'revision|r=s' => \$_revision,
165 'preserve-empty-dirs' =>
166 \$Git::SVN::Fetcher::_preserve_empty_dirs,
167 'placeholder-filename=s' =>
168 \$Git::SVN::Fetcher::_placeholder_filename,
169 %fc_opts, %init_opts } ],
170 init => [ \&cmd_init, "Initialize a repo for tracking" .
171 " (requires URL argument)",
172 \%init_opts ],
173 'multi-init' => [ \&cmd_multi_init,
174 "Deprecated alias for ".
175 "'$0 init -T<trunk> -b<branches> -t<tags>'",
176 \%init_opts ],
177 dcommit => [ \&cmd_dcommit,
178 'Commit several diffs to merge with upstream',
179 { 'merge|m|M' => \$_merge,
180 'strategy|s=s' => \$_strategy,
181 'verbose|v' => \$_verbose,
182 'dry-run|n' => \$_dry_run,
183 'fetch-all|all' => \$_fetch_all,
184 'commit-url=s' => \$_commit_url,
185 'revision|r=i' => \$_revision,
186 'no-rebase' => \$_no_rebase,
187 'mergeinfo=s' => \$_merge_info,
188 'interactive|i' => \$_interactive,
189 %cmt_opts, %fc_opts } ],
190 branch => [ \&cmd_branch,
191 'Create a branch in the SVN repository',
192 { 'message|m=s' => \$_message,
193 'destination|d=s' => \$_branch_dest,
194 'dry-run|n' => \$_dry_run,
195 'tag|t' => \$_tag,
196 'username=s' => \$Git::SVN::Prompt::_username,
197 'commit-url=s' => \$_commit_url } ],
198 tag => [ sub { $_tag = 1; cmd_branch(@_) },
199 'Create a tag in the SVN repository',
200 { 'message|m=s' => \$_message,
201 'destination|d=s' => \$_branch_dest,
202 'dry-run|n' => \$_dry_run,
203 'username=s' => \$Git::SVN::Prompt::_username,
204 'commit-url=s' => \$_commit_url } ],
205 'set-tree' => [ \&cmd_set_tree,
206 "Set an SVN repository to a git tree-ish",
207 { 'stdin' => \$_stdin, %cmt_opts, %fc_opts, } ],
208 'create-ignore' => [ \&cmd_create_ignore,
209 'Create a .gitignore per svn:ignore',
210 { 'revision|r=i' => \$_revision
211 } ],
212 'mkdirs' => [ \&cmd_mkdirs ,
213 "recreate empty directories after a checkout",
214 { 'revision|r=i' => \$_revision } ],
215 'propget' => [ \&cmd_propget,
216 'Print the value of a property on a file or directory',
217 { 'revision|r=i' => \$_revision } ],
218 'proplist' => [ \&cmd_proplist,
219 'List all properties of a file or directory',
220 { 'revision|r=i' => \$_revision } ],
221 'show-ignore' => [ \&cmd_show_ignore, "Show svn:ignore listings",
222 { 'revision|r=i' => \$_revision
223 } ],
224 'show-externals' => [ \&cmd_show_externals, "Show svn:externals listings",
225 { 'revision|r=i' => \$_revision
226 } ],
227 'multi-fetch' => [ \&cmd_multi_fetch,
228 "Deprecated alias for $0 fetch --all",
229 { 'revision|r=s' => \$_revision, %fc_opts } ],
230 'migrate' => [ sub { },
231 # no-op, we automatically run this anyways,
232 'Migrate configuration/metadata/layout from
233 previous versions of git-svn',
234 { 'minimize' => \$Git::SVN::Migration::_minimize,
235 %remote_opts } ],
236 'log' => [ \&Git::SVN::Log::cmd_show_log, 'Show commit logs',
237 { 'limit=i' => \$Git::SVN::Log::limit,
238 'revision|r=s' => \$_revision,
239 'verbose|v' => \$Git::SVN::Log::verbose,
240 'incremental' => \$Git::SVN::Log::incremental,
241 'oneline' => \$Git::SVN::Log::oneline,
242 'show-commit' => \$Git::SVN::Log::show_commit,
243 'non-recursive' => \$Git::SVN::Log::non_recursive,
244 'authors-file|A=s' => \$_authors,
245 'color' => \$Git::SVN::Log::color,
246 'pager=s' => \$Git::SVN::Log::pager
247 } ],
248 'find-rev' => [ \&cmd_find_rev,
249 "Translate between SVN revision numbers and tree-ish",
250 {} ],
251 'rebase' => [ \&cmd_rebase, "Fetch and rebase your working directory",
252 { 'merge|m|M' => \$_merge,
253 'verbose|v' => \$_verbose,
254 'strategy|s=s' => \$_strategy,
255 'local|l' => \$_local,
256 'fetch-all|all' => \$_fetch_all,
257 'dry-run|n' => \$_dry_run,
258 'preserve-merges|p' => \$_preserve_merges,
259 %fc_opts } ],
260 'commit-diff' => [ \&cmd_commit_diff,
261 'Commit a diff between two trees',
262 { 'message|m=s' => \$_message,
263 'file|F=s' => \$_file,
264 'revision|r=s' => \$_revision,
265 %cmt_opts } ],
266 'info' => [ \&cmd_info,
267 "Show info about the latest SVN revision
268 on the current branch",
269 { 'url' => \$_url, } ],
270 'blame' => [ \&Git::SVN::Log::cmd_blame,
271 "Show what revision and author last modified each line of a file",
272 { 'git-format' => \$_git_format } ],
273 'reset' => [ \&cmd_reset,
274 "Undo fetches back to the specified SVN revision",
275 { 'revision|r=s' => \$_revision,
276 'parent|p' => \$_fetch_parent } ],
277 'gc' => [ \&cmd_gc,
278 "Compress unhandled.log files in .git/svn and remove " .
279 "index files in .git/svn",
280 {} ],
281 );
282
283 use Term::ReadLine;
284 package FakeTerm;
285 sub new {
286 my ($class, $reason) = @_;
287 return bless \$reason, shift;
288 }
289 sub readline {
290 my $self = shift;
291 die "Cannot use readline on FakeTerm: $$self";
292 }
293 package main;
294
295 my $term = eval {
296 $ENV{"GIT_SVN_NOTTY"}
297 ? new Term::ReadLine 'git-svn', \*STDIN, \*STDOUT
298 : new Term::ReadLine 'git-svn';
299 };
300 if ($@) {
301 $term = new FakeTerm "$@: going non-interactive";
302 }
303
304 my $cmd;
305 for (my $i = 0; $i < @ARGV; $i++) {
306 if (defined $cmd{$ARGV[$i]}) {
307 $cmd = $ARGV[$i];
308 splice @ARGV, $i, 1;
309 last;
310 } elsif ($ARGV[$i] eq 'help') {
311 $cmd = $ARGV[$i+1];
312 usage(0);
313 }
314 };
315
316 # make sure we're always running at the top-level working directory
317 unless ($cmd && $cmd =~ /(?:clone|init|multi-init)$/) {
318 unless (-d $ENV{GIT_DIR}) {
319 if ($git_dir_user_set) {
320 die "GIT_DIR=$ENV{GIT_DIR} explicitly set, ",
321 "but it is not a directory\n";
322 }
323 my $git_dir = delete $ENV{GIT_DIR};
324 my $cdup = undef;
325 git_cmd_try {
326 $cdup = command_oneline(qw/rev-parse --show-cdup/);
327 $git_dir = '.' unless ($cdup);
328 chomp $cdup if ($cdup);
329 $cdup = "." unless ($cdup && length $cdup);
330 } "Already at toplevel, but $git_dir not found\n";
331 chdir $cdup or die "Unable to chdir up to '$cdup'\n";
332 unless (-d $git_dir) {
333 die "$git_dir still not found after going to ",
334 "'$cdup'\n";
335 }
336 $ENV{GIT_DIR} = $git_dir;
337 }
338 $_repository = Git->repository(Repository => $ENV{GIT_DIR});
339 }
340
341 my %opts = %{$cmd{$cmd}->[2]} if (defined $cmd);
342
343 read_git_config(\%opts);
344 if ($cmd && ($cmd eq 'log' || $cmd eq 'blame')) {
345 Getopt::Long::Configure('pass_through');
346 }
347 my $rv = GetOptions(%opts, 'h|H' => \$_help, 'version|V' => \$_version,
348 'minimize-connections' => \$Git::SVN::Migration::_minimize,
349 'id|i=s' => \$Git::SVN::default_ref_id,
350 'svn-remote|remote|R=s' => sub {
351 $Git::SVN::no_reuse_existing = 1;
352 $Git::SVN::default_repo_id = $_[1] });
353 exit 1 if (!$rv && $cmd && $cmd ne 'log');
354
355 usage(0) if $_help;
356 version() if $_version;
357 usage(1) unless defined $cmd;
358 load_authors() if $_authors;
359 if (defined $_authors_prog) {
360 $_authors_prog = "'" . File::Spec->rel2abs($_authors_prog) . "'";
361 }
362
363 unless ($cmd =~ /^(?:clone|init|multi-init|commit-diff)$/) {
364 Git::SVN::Migration::migration_check();
365 }
366 Git::SVN::init_vars();
367 eval {
368 Git::SVN::verify_remotes_sanity();
369 $cmd{$cmd}->[0]->(@ARGV);
370 };
371 fatal $@ if $@;
372 post_fetch_checkout();
373 exit 0;
374
375 ####################### primary functions ######################
376 sub usage {
377 my $exit = shift || 0;
378 my $fd = $exit ? \*STDERR : \*STDOUT;
379 print $fd <<"";
380 git-svn - bidirectional operations between a single Subversion tree and git
381 Usage: git svn <command> [options] [arguments]\n
382
383 print $fd "Available commands:\n" unless $cmd;
384
385 foreach (sort keys %cmd) {
386 next if $cmd && $cmd ne $_;
387 next if /^multi-/; # don't show deprecated commands
388 print $fd ' ',pack('A17',$_),$cmd{$_}->[1],"\n";
389 foreach (sort keys %{$cmd{$_}->[2]}) {
390 # mixed-case options are for .git/config only
391 next if /[A-Z]/ && /^[a-z]+$/i;
392 # prints out arguments as they should be passed:
393 my $x = s#[:=]s$## ? '<arg>' : s#[:=]i$## ? '<num>' : '';
394 print $fd ' ' x 21, join(', ', map { length $_ > 1 ?
395 "--$_" : "-$_" }
396 split /\|/,$_)," $x\n";
397 }
398 }
399 print $fd <<"";
400 \nGIT_SVN_ID may be set in the environment or via the --id/-i switch to an
401 arbitrary identifier if you're tracking multiple SVN branches/repositories in
402 one git repository and want to keep them separate. See git-svn(1) for more
403 information.
404
405 exit $exit;
406 }
407
408 sub version {
409 ::_req_svn();
410 print "git-svn version $VERSION (svn $SVN::Core::VERSION)\n";
411 exit 0;
412 }
413
414 sub ask {
415 my ($prompt, %arg) = @_;
416 my $valid_re = $arg{valid_re};
417 my $default = $arg{default};
418 my $resp;
419 my $i = 0;
420
421 if ( !( defined($term->IN)
422 && defined( fileno($term->IN) )
423 && defined( $term->OUT )
424 && defined( fileno($term->OUT) ) ) ){
425 return defined($default) ? $default : undef;
426 }
427
428 while ($i++ < 10) {
429 $resp = $term->readline($prompt);
430 if (!defined $resp) { # EOF
431 print "\n";
432 return defined $default ? $default : undef;
433 }
434 if ($resp eq '' and defined $default) {
435 return $default;
436 }
437 if (!defined $valid_re or $resp =~ /$valid_re/) {
438 return $resp;
439 }
440 }
441 return undef;
442 }
443
444 sub do_git_init_db {
445 unless (-d $ENV{GIT_DIR}) {
446 my @init_db = ('init');
447 push @init_db, "--template=$_template" if defined $_template;
448 if (defined $_shared) {
449 if ($_shared =~ /[a-z]/) {
450 push @init_db, "--shared=$_shared";
451 } else {
452 push @init_db, "--shared";
453 }
454 }
455 command_noisy(@init_db);
456 $_repository = Git->repository(Repository => ".git");
457 }
458 my $set;
459 my $pfx = "svn-remote.$Git::SVN::default_repo_id";
460 foreach my $i (keys %icv) {
461 die "'$set' and '$i' cannot both be set\n" if $set;
462 next unless defined $icv{$i};
463 command_noisy('config', "$pfx.$i", $icv{$i});
464 $set = $i;
465 }
466 my $ignore_paths_regex = \$Git::SVN::Fetcher::_ignore_regex;
467 command_noisy('config', "$pfx.ignore-paths", $$ignore_paths_regex)
468 if defined $$ignore_paths_regex;
469 my $ignore_refs_regex = \$Git::SVN::Ra::_ignore_refs_regex;
470 command_noisy('config', "$pfx.ignore-refs", $$ignore_refs_regex)
471 if defined $$ignore_refs_regex;
472
473 if (defined $Git::SVN::Fetcher::_preserve_empty_dirs) {
474 my $fname = \$Git::SVN::Fetcher::_placeholder_filename;
475 command_noisy('config', "$pfx.preserve-empty-dirs", 'true');
476 command_noisy('config', "$pfx.placeholder-filename", $$fname);
477 }
478 }
479
480 sub init_subdir {
481 my $repo_path = shift or return;
482 mkpath([$repo_path]) unless -d $repo_path;
483 chdir $repo_path or die "Couldn't chdir to $repo_path: $!\n";
484 $ENV{GIT_DIR} = '.git';
485 $_repository = Git->repository(Repository => $ENV{GIT_DIR});
486 }
487
488 sub cmd_clone {
489 my ($url, $path) = @_;
490 if (!defined $path &&
491 (defined $_trunk || @_branches || @_tags ||
492 defined $_stdlayout) &&
493 $url !~ m#^[a-z\+]+://#) {
494 $path = $url;
495 }
496 $path = basename($url) if !defined $path || !length $path;
497 my $authors_absolute = $_authors ? File::Spec->rel2abs($_authors) : "";
498 cmd_init($url, $path);
499 command_oneline('config', 'svn.authorsfile', $authors_absolute)
500 if $_authors;
501 Git::SVN::fetch_all($Git::SVN::default_repo_id);
502 }
503
504 sub cmd_init {
505 if (defined $_stdlayout) {
506 $_trunk = 'trunk' if (!defined $_trunk);
507 @_tags = 'tags' if (! @_tags);
508 @_branches = 'branches' if (! @_branches);
509 }
510 if (defined $_trunk || @_branches || @_tags) {
511 return cmd_multi_init(@_);
512 }
513 my $url = shift or die "SVN repository location required ",
514 "as a command-line argument\n";
515 $url = canonicalize_url($url);
516 init_subdir(@_);
517 do_git_init_db();
518
519 if ($Git::SVN::_minimize_url eq 'unset') {
520 $Git::SVN::_minimize_url = 0;
521 }
522
523 Git::SVN->init($url);
524 }
525
526 sub cmd_fetch {
527 if (grep /^\d+=./, @_) {
528 die "'<rev>=<commit>' fetch arguments are ",
529 "no longer supported.\n";
530 }
531 my ($remote) = @_;
532 if (@_ > 1) {
533 die "Usage: $0 fetch [--all] [--parent] [svn-remote]\n";
534 }
535 $Git::SVN::no_reuse_existing = undef;
536 if ($_fetch_parent) {
537 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
538 unless ($gs) {
539 die "Unable to determine upstream SVN information from ",
540 "working tree history\n";
541 }
542 # just fetch, don't checkout.
543 $_no_checkout = 'true';
544 $_fetch_all ? $gs->fetch_all : $gs->fetch;
545 } elsif ($_fetch_all) {
546 cmd_multi_fetch();
547 } else {
548 $remote ||= $Git::SVN::default_repo_id;
549 Git::SVN::fetch_all($remote, Git::SVN::read_all_remotes());
550 }
551 }
552
553 sub cmd_set_tree {
554 my (@commits) = @_;
555 if ($_stdin || !@commits) {
556 print "Reading from stdin...\n";
557 @commits = ();
558 while (<STDIN>) {
559 if (/\b($sha1_short)\b/o) {
560 unshift @commits, $1;
561 }
562 }
563 }
564 my @revs;
565 foreach my $c (@commits) {
566 my @tmp = command('rev-parse',$c);
567 if (scalar @tmp == 1) {
568 push @revs, $tmp[0];
569 } elsif (scalar @tmp > 1) {
570 push @revs, reverse(command('rev-list',@tmp));
571 } else {
572 fatal "Failed to rev-parse $c";
573 }
574 }
575 my $gs = Git::SVN->new;
576 my ($r_last, $cmt_last) = $gs->last_rev_commit;
577 $gs->fetch;
578 if (defined $gs->{last_rev} && $r_last != $gs->{last_rev}) {
579 fatal "There are new revisions that were fetched ",
580 "and need to be merged (or acknowledged) ",
581 "before committing.\nlast rev: $r_last\n",
582 " current: $gs->{last_rev}";
583 }
584 $gs->set_tree($_) foreach @revs;
585 print "Done committing ",scalar @revs," revisions to SVN\n";
586 unlink $gs->{index};
587 }
588
589 sub split_merge_info_range {
590 my ($range) = @_;
591 if ($range =~ /(\d+)-(\d+)/) {
592 return (int($1), int($2));
593 } else {
594 return (int($range), int($range));
595 }
596 }
597
598 sub combine_ranges {
599 my ($in) = @_;
600
601 my @fnums = ();
602 my @arr = split(/,/, $in);
603 for my $element (@arr) {
604 my ($start, $end) = split_merge_info_range($element);
605 push @fnums, $start;
606 }
607
608 my @sorted = @arr [ sort {
609 $fnums[$a] <=> $fnums[$b]
610 } 0..$#arr ];
611
612 my @return = ();
613 my $last = -1;
614 my $first = -1;
615 for my $element (@sorted) {
616 my ($start, $end) = split_merge_info_range($element);
617
618 if ($last == -1) {
619 $first = $start;
620 $last = $end;
621 next;
622 }
623 if ($start <= $last+1) {
624 if ($end > $last) {
625 $last = $end;
626 }
627 next;
628 }
629 if ($first == $last) {
630 push @return, "$first";
631 } else {
632 push @return, "$first-$last";
633 }
634 $first = $start;
635 $last = $end;
636 }
637
638 if ($first != -1) {
639 if ($first == $last) {
640 push @return, "$first";
641 } else {
642 push @return, "$first-$last";
643 }
644 }
645
646 return join(',', @return);
647 }
648
649 sub merge_revs_into_hash {
650 my ($hash, $minfo) = @_;
651 my @lines = split(' ', $minfo);
652
653 for my $line (@lines) {
654 my ($branchpath, $revs) = split(/:/, $line);
655
656 if (exists($hash->{$branchpath})) {
657 # Merge the two revision sets
658 my $combined = "$hash->{$branchpath},$revs";
659 $hash->{$branchpath} = combine_ranges($combined);
660 } else {
661 # Just do range combining for consolidation
662 $hash->{$branchpath} = combine_ranges($revs);
663 }
664 }
665 }
666
667 sub merge_merge_info {
668 my ($mergeinfo_one, $mergeinfo_two) = @_;
669 my %result_hash = ();
670
671 merge_revs_into_hash(\%result_hash, $mergeinfo_one);
672 merge_revs_into_hash(\%result_hash, $mergeinfo_two);
673
674 my $result = '';
675 # Sort below is for consistency's sake
676 for my $branchname (sort keys(%result_hash)) {
677 my $revlist = $result_hash{$branchname};
678 $result .= "$branchname:$revlist\n"
679 }
680 return $result;
681 }
682
683 sub populate_merge_info {
684 my ($d, $gs, $uuid, $linear_refs, $rewritten_parent) = @_;
685
686 my %parentshash;
687 read_commit_parents(\%parentshash, $d);
688 my @parents = @{$parentshash{$d}};
689 if ($#parents > 0) {
690 # Merge commit
691 my $all_parents_ok = 1;
692 my $aggregate_mergeinfo = '';
693 my $rooturl = $gs->repos_root;
694
695 if (defined($rewritten_parent)) {
696 # Replace first parent with newly-rewritten version
697 shift @parents;
698 unshift @parents, $rewritten_parent;
699 }
700
701 foreach my $parent (@parents) {
702 my ($branchurl, $svnrev, $paruuid) =
703 cmt_metadata($parent);
704
705 unless (defined($svnrev)) {
706 # Should have been caught be preflight check
707 fatal "merge commit $d has ancestor $parent, but that change "
708 ."does not have git-svn metadata!";
709 }
710 unless ($branchurl =~ /^\Q$rooturl\E(.*)/) {
711 fatal "commit $parent git-svn metadata changed mid-run!";
712 }
713 my $branchpath = $1;
714
715 my $ra = Git::SVN::Ra->new($branchurl);
716 my (undef, undef, $props) =
717 $ra->get_dir(canonicalize_path("."), $svnrev);
718 my $par_mergeinfo = $props->{'svn:mergeinfo'};
719 unless (defined $par_mergeinfo) {
720 $par_mergeinfo = '';
721 }
722 # Merge previous mergeinfo values
723 $aggregate_mergeinfo =
724 merge_merge_info($aggregate_mergeinfo,
725 $par_mergeinfo, 0);
726
727 next if $parent eq $parents[0]; # Skip first parent
728 # Add new changes being placed in tree by merge
729 my @cmd = (qw/rev-list --reverse/,
730 $parent, qw/--not/);
731 foreach my $par (@parents) {
732 unless ($par eq $parent) {
733 push @cmd, $par;
734 }
735 }
736 my @revsin = ();
737 my ($revlist, $ctx) = command_output_pipe(@cmd);
738 while (<$revlist>) {
739 my $irev = $_;
740 chomp $irev;
741 my (undef, $csvnrev, undef) =
742 cmt_metadata($irev);
743 unless (defined $csvnrev) {
744 # A child is missing SVN annotations...
745 # this might be OK, or might not be.
746 warn "W:child $irev is merged into revision "
747 ."$d but does not have git-svn metadata. "
748 ."This means git-svn cannot determine the "
749 ."svn revision numbers to place into the "
750 ."svn:mergeinfo property. You must ensure "
751 ."a branch is entirely committed to "
752 ."SVN before merging it in order for "
753 ."svn:mergeinfo population to function "
754 ."properly";
755 }
756 push @revsin, $csvnrev;
757 }
758 command_close_pipe($revlist, $ctx);
759
760 last unless $all_parents_ok;
761
762 # We now have a list of all SVN revnos which are
763 # merged by this particular parent. Integrate them.
764 next if $#revsin == -1;
765 my $newmergeinfo = "$branchpath:" . join(',', @revsin);
766 $aggregate_mergeinfo =
767 merge_merge_info($aggregate_mergeinfo,
768 $newmergeinfo, 1);
769 }
770 if ($all_parents_ok and $aggregate_mergeinfo) {
771 return $aggregate_mergeinfo;
772 }
773 }
774
775 return undef;
776 }
777
778 sub cmd_dcommit {
779 my $head = shift;
780 command_noisy(qw/update-index --refresh/);
781 git_cmd_try { command_oneline(qw/diff-index --quiet HEAD/) }
782 'Cannot dcommit with a dirty index. Commit your changes first, '
783 . "or stash them with `git stash'.\n";
784 $head ||= 'HEAD';
785
786 my $old_head;
787 if ($head ne 'HEAD') {
788 $old_head = eval {
789 command_oneline([qw/symbolic-ref -q HEAD/])
790 };
791 if ($old_head) {
792 $old_head =~ s{^refs/heads/}{};
793 } else {
794 $old_head = eval { command_oneline(qw/rev-parse HEAD/) };
795 }
796 command(['checkout', $head], STDERR => 0);
797 }
798
799 my @refs;
800 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD', \@refs);
801 unless ($gs) {
802 die "Unable to determine upstream SVN information from ",
803 "$head history.\nPerhaps the repository is empty.";
804 }
805
806 if (defined $_commit_url) {
807 $url = $_commit_url;
808 } else {
809 $url = eval { command_oneline('config', '--get',
810 "svn-remote.$gs->{repo_id}.commiturl") };
811 if (!$url) {
812 $url = $gs->full_pushurl
813 }
814 }
815
816 my $last_rev = $_revision if defined $_revision;
817 if ($url) {
818 print "Committing to $url ...\n";
819 }
820 my ($linear_refs, $parents) = linearize_history($gs, \@refs);
821 if ($_no_rebase && scalar(@$linear_refs) > 1) {
822 warn "Attempting to commit more than one change while ",
823 "--no-rebase is enabled.\n",
824 "If these changes depend on each other, re-running ",
825 "without --no-rebase may be required."
826 }
827
828 if (defined $_interactive){
829 my $ask_default = "y";
830 foreach my $d (@$linear_refs){
831 my ($fh, $ctx) = command_output_pipe(qw(show --summary), "$d");
832 while (<$fh>){
833 print $_;
834 }
835 command_close_pipe($fh, $ctx);
836 $_ = ask("Commit this patch to SVN? ([y]es (default)|[n]o|[q]uit|[a]ll): ",
837 valid_re => qr/^(?:yes|y|no|n|quit|q|all|a)/i,
838 default => $ask_default);
839 die "Commit this patch reply required" unless defined $_;
840 if (/^[nq]/i) {
841 exit(0);
842 } elsif (/^a/i) {
843 last;
844 }
845 }
846 }
847
848 my $expect_url = $url;
849
850 my $push_merge_info = eval {
851 command_oneline(qw/config --get svn.pushmergeinfo/)
852 };
853 if (not defined($push_merge_info)
854 or $push_merge_info eq "false"
855 or $push_merge_info eq "no"
856 or $push_merge_info eq "never") {
857 $push_merge_info = 0;
858 }
859
860 unless (defined($_merge_info) || ! $push_merge_info) {
861 # Preflight check of changes to ensure no issues with mergeinfo
862 # This includes check for uncommitted-to-SVN parents
863 # (other than the first parent, which we will handle),
864 # information from different SVN repos, and paths
865 # which are not underneath this repository root.
866 my $rooturl = $gs->repos_root;
867 foreach my $d (@$linear_refs) {
868 my %parentshash;
869 read_commit_parents(\%parentshash, $d);
870 my @realparents = @{$parentshash{$d}};
871 if ($#realparents > 0) {
872 # Merge commit
873 shift @realparents; # Remove/ignore first parent
874 foreach my $parent (@realparents) {
875 my ($branchurl, $svnrev, $paruuid) = cmt_metadata($parent);
876 unless (defined $paruuid) {
877 # A parent is missing SVN annotations...
878 # abort the whole operation.
879 fatal "$parent is merged into revision $d, "
880 ."but does not have git-svn metadata. "
881 ."Either dcommit the branch or use a "
882 ."local cherry-pick, FF merge, or rebase "
883 ."instead of an explicit merge commit.";
884 }
885
886 unless ($paruuid eq $uuid) {
887 # Parent has SVN metadata from different repository
888 fatal "merge parent $parent for change $d has "
889 ."git-svn uuid $paruuid, while current change "
890 ."has uuid $uuid!";
891 }
892
893 unless ($branchurl =~ /^\Q$rooturl\E(.*)/) {
894 # This branch is very strange indeed.
895 fatal "merge parent $parent for $d is on branch "
896 ."$branchurl, which is not under the "
897 ."git-svn root $rooturl!";
898 }
899 }
900 }
901 }
902 }
903
904 my $rewritten_parent;
905 Git::SVN::remove_username($expect_url);
906 if (defined($_merge_info)) {
907 $_merge_info =~ tr{ }{\n};
908 }
909 while (1) {
910 my $d = shift @$linear_refs or last;
911 unless (defined $last_rev) {
912 (undef, $last_rev, undef) = cmt_metadata("$d~1");
913 unless (defined $last_rev) {
914 fatal "Unable to extract revision information ",
915 "from commit $d~1";
916 }
917 }
918 if ($_dry_run) {
919 print "diff-tree $d~1 $d\n";
920 } else {
921 my $cmt_rev;
922
923 unless (defined($_merge_info) || ! $push_merge_info) {
924 $_merge_info = populate_merge_info($d, $gs,
925 $uuid,
926 $linear_refs,
927 $rewritten_parent);
928 }
929
930 my %ed_opts = ( r => $last_rev,
931 log => get_commit_entry($d)->{log},
932 ra => Git::SVN::Ra->new($url),
933 config => SVN::Core::config_get_config(
934 $Git::SVN::Ra::config_dir
935 ),
936 tree_a => "$d~1",
937 tree_b => $d,
938 editor_cb => sub {
939 print "Committed r$_[0]\n";
940 $cmt_rev = $_[0];
941 },
942 mergeinfo => $_merge_info,
943 svn_path => '');
944 if (!Git::SVN::Editor->new(\%ed_opts)->apply_diff) {
945 print "No changes\n$d~1 == $d\n";
946 } elsif ($parents->{$d} && @{$parents->{$d}}) {
947 $gs->{inject_parents_dcommit}->{$cmt_rev} =
948 $parents->{$d};
949 }
950 $_fetch_all ? $gs->fetch_all : $gs->fetch;
951 $last_rev = $cmt_rev;
952 next if $_no_rebase;
953
954 # we always want to rebase against the current HEAD,
955 # not any head that was passed to us
956 my @diff = command('diff-tree', $d,
957 $gs->refname, '--');
958 my @finish;
959 if (@diff) {
960 @finish = rebase_cmd();
961 print STDERR "W: $d and ", $gs->refname,
962 " differ, using @finish:\n",
963 join("\n", @diff), "\n";
964 } else {
965 print "No changes between current HEAD and ",
966 $gs->refname,
967 "\nResetting to the latest ",
968 $gs->refname, "\n";
969 @finish = qw/reset --mixed/;
970 }
971 command_noisy(@finish, $gs->refname);
972
973 $rewritten_parent = command_oneline(qw/rev-parse HEAD/);
974
975 if (@diff) {
976 @refs = ();
977 my ($url_, $rev_, $uuid_, $gs_) =
978 working_head_info('HEAD', \@refs);
979 my ($linear_refs_, $parents_) =
980 linearize_history($gs_, \@refs);
981 if (scalar(@$linear_refs) !=
982 scalar(@$linear_refs_)) {
983 fatal "# of revisions changed ",
984 "\nbefore:\n",
985 join("\n", @$linear_refs),
986 "\n\nafter:\n",
987 join("\n", @$linear_refs_), "\n",
988 'If you are attempting to commit ',
989 "merges, try running:\n\t",
990 'git rebase --interactive',
991 '--preserve-merges ',
992 $gs->refname,
993 "\nBefore dcommitting";
994 }
995 if ($url_ ne $expect_url) {
996 if ($url_ eq $gs->metadata_url) {
997 print
998 "Accepting rewritten URL:",
999 " $url_\n";
1000 } else {
1001 fatal
1002 "URL mismatch after rebase:",
1003 " $url_ != $expect_url";
1004 }
1005 }
1006 if ($uuid_ ne $uuid) {
1007 fatal "uuid mismatch after rebase: ",
1008 "$uuid_ != $uuid";
1009 }
1010 # remap parents
1011 my (%p, @l, $i);
1012 for ($i = 0; $i < scalar @$linear_refs; $i++) {
1013 my $new = $linear_refs_->[$i] or next;
1014 $p{$new} =
1015 $parents->{$linear_refs->[$i]};
1016 push @l, $new;
1017 }
1018 $parents = \%p;
1019 $linear_refs = \@l;
1020 }
1021 }
1022 }
1023
1024 if ($old_head) {
1025 my $new_head = command_oneline(qw/rev-parse HEAD/);
1026 my $new_is_symbolic = eval {
1027 command_oneline(qw/symbolic-ref -q HEAD/);
1028 };
1029 if ($new_is_symbolic) {
1030 print "dcommitted the branch ", $head, "\n";
1031 } else {
1032 print "dcommitted on a detached HEAD because you gave ",
1033 "a revision argument.\n",
1034 "The rewritten commit is: ", $new_head, "\n";
1035 }
1036 command(['checkout', $old_head], STDERR => 0);
1037 }
1038
1039 unlink $gs->{index};
1040 }
1041
1042 sub cmd_branch {
1043 my ($branch_name, $head) = @_;
1044
1045 unless (defined $branch_name && length $branch_name) {
1046 die(($_tag ? "tag" : "branch") . " name required\n");
1047 }
1048 $head ||= 'HEAD';
1049
1050 my (undef, $rev, undef, $gs) = working_head_info($head);
1051 my $src = $gs->full_pushurl;
1052
1053 my $remote = Git::SVN::read_all_remotes()->{$gs->{repo_id}};
1054 my $allglobs = $remote->{ $_tag ? 'tags' : 'branches' };
1055 my $glob;
1056 if ($#{$allglobs} == 0) {
1057 $glob = $allglobs->[0];
1058 } else {
1059 unless(defined $_branch_dest) {
1060 die "Multiple ",
1061 $_tag ? "tag" : "branch",
1062 " paths defined for Subversion repository.\n",
1063 "You must specify where you want to create the ",
1064 $_tag ? "tag" : "branch",
1065 " with the --destination argument.\n";
1066 }
1067 foreach my $g (@{$allglobs}) {
1068 my $re = Git::SVN::Editor::glob2pat($g->{path}->{left});
1069 if ($_branch_dest =~ /$re/) {
1070 $glob = $g;
1071 last;
1072 }
1073 }
1074 unless (defined $glob) {
1075 my $dest_re = qr/\b\Q$_branch_dest\E\b/;
1076 foreach my $g (@{$allglobs}) {
1077 $g->{path}->{left} =~ /$dest_re/ or next;
1078 if (defined $glob) {
1079 die "Ambiguous destination: ",
1080 $_branch_dest, "\nmatches both '",
1081 $glob->{path}->{left}, "' and '",
1082 $g->{path}->{left}, "'\n";
1083 }
1084 $glob = $g;
1085 }
1086 unless (defined $glob) {
1087 die "Unknown ",
1088 $_tag ? "tag" : "branch",
1089 " destination $_branch_dest\n";
1090 }
1091 }
1092 }
1093 my ($lft, $rgt) = @{ $glob->{path} }{qw/left right/};
1094 my $url;
1095 if (defined $_commit_url) {
1096 $url = $_commit_url;
1097 } else {
1098 $url = eval { command_oneline('config', '--get',
1099 "svn-remote.$gs->{repo_id}.commiturl") };
1100 if (!$url) {
1101 $url = $remote->{pushurl} || $remote->{url};
1102 }
1103 }
1104 my $dst = join '/', $url, $lft, $branch_name, ($rgt || ());
1105
1106 if ($dst =~ /^https:/ && $src =~ /^http:/) {
1107 $src=~s/^http:/https:/;
1108 }
1109
1110 ::_req_svn();
1111
1112 my $ctx = SVN::Client->new(
1113 auth => Git::SVN::Ra::_auth_providers(),
1114 log_msg => sub {
1115 ${ $_[0] } = defined $_message
1116 ? $_message
1117 : 'Create ' . ($_tag ? 'tag ' : 'branch ' )
1118 . $branch_name;
1119 },
1120 );
1121
1122 eval {
1123 $ctx->ls($dst, 'HEAD', 0);
1124 } and die "branch ${branch_name} already exists\n";
1125
1126 print "Copying ${src} at r${rev} to ${dst}...\n";
1127 $ctx->copy($src, $rev, $dst)
1128 unless $_dry_run;
1129
1130 $gs->fetch_all;
1131 }
1132
1133 sub cmd_find_rev {
1134 my $revision_or_hash = shift or die "SVN or git revision required ",
1135 "as a command-line argument\n";
1136 my $result;
1137 if ($revision_or_hash =~ /^r\d+$/) {
1138 my $head = shift;
1139 $head ||= 'HEAD';
1140 my @refs;
1141 my (undef, undef, $uuid, $gs) = working_head_info($head, \@refs);
1142 unless ($gs) {
1143 die "Unable to determine upstream SVN information from ",
1144 "$head history\n";
1145 }
1146 my $desired_revision = substr($revision_or_hash, 1);
1147 $result = $gs->rev_map_get($desired_revision, $uuid);
1148 } else {
1149 my (undef, $rev, undef) = cmt_metadata($revision_or_hash);
1150 $result = $rev;
1151 }
1152 print "$result\n" if $result;
1153 }
1154
1155 sub auto_create_empty_directories {
1156 my ($gs) = @_;
1157 my $var = eval { command_oneline('config', '--get', '--bool',
1158 "svn-remote.$gs->{repo_id}.automkdirs") };
1159 # By default, create empty directories by consulting the unhandled log,
1160 # but allow setting it to 'false' to skip it.
1161 return !($var && $var eq 'false');
1162 }
1163
1164 sub cmd_rebase {
1165 command_noisy(qw/update-index --refresh/);
1166 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
1167 unless ($gs) {
1168 die "Unable to determine upstream SVN information from ",
1169 "working tree history\n";
1170 }
1171 if ($_dry_run) {
1172 print "Remote Branch: " . $gs->refname . "\n";
1173 print "SVN URL: " . $url . "\n";
1174 return;
1175 }
1176 if (command(qw/diff-index HEAD --/)) {
1177 print STDERR "Cannot rebase with uncommited changes:\n";
1178 command_noisy('status');
1179 exit 1;
1180 }
1181 unless ($_local) {
1182 # rebase will checkout for us, so no need to do it explicitly
1183 $_no_checkout = 'true';
1184 $_fetch_all ? $gs->fetch_all : $gs->fetch;
1185 }
1186 command_noisy(rebase_cmd(), $gs->refname);
1187 if (auto_create_empty_directories($gs)) {
1188 $gs->mkemptydirs;
1189 }
1190 }
1191
1192 sub cmd_show_ignore {
1193 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
1194 $gs ||= Git::SVN->new;
1195 my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
1196 $gs->prop_walk($gs->{path}, $r, sub {
1197 my ($gs, $path, $props) = @_;
1198 print STDOUT "\n# $path\n";
1199 my $s = $props->{'svn:ignore'} or return;
1200 $s =~ s/[\r\n]+/\n/g;
1201 $s =~ s/^\n+//;
1202 chomp $s;
1203 $s =~ s#^#$path#gm;
1204 print STDOUT "$s\n";
1205 });
1206 }
1207
1208 sub cmd_show_externals {
1209 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
1210 $gs ||= Git::SVN->new;
1211 my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
1212 $gs->prop_walk($gs->{path}, $r, sub {
1213 my ($gs, $path, $props) = @_;
1214 print STDOUT "\n# $path\n";
1215 my $s = $props->{'svn:externals'} or return;
1216 $s =~ s/[\r\n]+/\n/g;
1217 chomp $s;
1218 $s =~ s#^#$path#gm;
1219 print STDOUT "$s\n";
1220 });
1221 }
1222
1223 sub cmd_create_ignore {
1224 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
1225 $gs ||= Git::SVN->new;
1226 my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
1227 $gs->prop_walk($gs->{path}, $r, sub {
1228 my ($gs, $path, $props) = @_;
1229 # $path is of the form /path/to/dir/
1230 $path = '.' . $path;
1231 # SVN can have attributes on empty directories,
1232 # which git won't track
1233 mkpath([$path]) unless -d $path;
1234 my $ignore = $path . '.gitignore';
1235 my $s = $props->{'svn:ignore'} or return;
1236 open(GITIGNORE, '>', $ignore)
1237 or fatal("Failed to open `$ignore' for writing: $!");
1238 $s =~ s/[\r\n]+/\n/g;
1239 $s =~ s/^\n+//;
1240 chomp $s;
1241 # Prefix all patterns so that the ignore doesn't apply
1242 # to sub-directories.
1243 $s =~ s#^#/#gm;
1244 print GITIGNORE "$s\n";
1245 close(GITIGNORE)
1246 or fatal("Failed to close `$ignore': $!");
1247 command_noisy('add', '-f', $ignore);
1248 });
1249 }
1250
1251 sub cmd_mkdirs {
1252 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
1253 $gs ||= Git::SVN->new;
1254 $gs->mkemptydirs($_revision);
1255 }
1256
1257 sub canonicalize_path {
1258 my ($path) = @_;
1259 my $dot_slash_added = 0;
1260 if (substr($path, 0, 1) ne "/") {
1261 $path = "./" . $path;
1262 $dot_slash_added = 1;
1263 }
1264 # File::Spec->canonpath doesn't collapse x/../y into y (for a
1265 # good reason), so let's do this manually.
1266 $path =~ s#/+#/#g;
1267 $path =~ s#/\.(?:/|$)#/#g;
1268 $path =~ s#/[^/]+/\.\.##g;
1269 $path =~ s#/$##g;
1270 $path =~ s#^\./## if $dot_slash_added;
1271 $path =~ s#^/##;
1272 $path =~ s#^\.$##;
1273 return $path;
1274 }
1275
1276 sub canonicalize_url {
1277 my ($url) = @_;
1278 $url =~ s#^([^:]+://[^/]*/)(.*)$#$1 . canonicalize_path($2)#e;
1279 return $url;
1280 }
1281
1282 # get_svnprops(PATH)
1283 # ------------------
1284 # Helper for cmd_propget and cmd_proplist below.
1285 sub get_svnprops {
1286 my $path = shift;
1287 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
1288 $gs ||= Git::SVN->new;
1289
1290 # prefix THE PATH by the sub-directory from which the user
1291 # invoked us.
1292 $path = $cmd_dir_prefix . $path;
1293 fatal("No such file or directory: $path") unless -e $path;
1294 my $is_dir = -d $path ? 1 : 0;
1295 $path = $gs->{path} . '/' . $path;
1296
1297 # canonicalize the path (otherwise libsvn will abort or fail to
1298 # find the file)
1299 $path = canonicalize_path($path);
1300
1301 my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
1302 my $props;
1303 if ($is_dir) {
1304 (undef, undef, $props) = $gs->ra->get_dir($path, $r);
1305 }
1306 else {
1307 (undef, $props) = $gs->ra->get_file($path, $r, undef);
1308 }
1309 return $props;
1310 }
1311
1312 # cmd_propget (PROP, PATH)
1313 # ------------------------
1314 # Print the SVN property PROP for PATH.
1315 sub cmd_propget {
1316 my ($prop, $path) = @_;
1317 $path = '.' if not defined $path;
1318 usage(1) if not defined $prop;
1319 my $props = get_svnprops($path);
1320 if (not defined $props->{$prop}) {
1321 fatal("`$path' does not have a `$prop' SVN property.");
1322 }
1323 print $props->{$prop} . "\n";
1324 }
1325
1326 # cmd_proplist (PATH)
1327 # -------------------
1328 # Print the list of SVN properties for PATH.
1329 sub cmd_proplist {
1330 my $path = shift;
1331 $path = '.' if not defined $path;
1332 my $props = get_svnprops($path);
1333 print "Properties on '$path':\n";
1334 foreach (sort keys %{$props}) {
1335 print " $_\n";
1336 }
1337 }
1338
1339 sub cmd_multi_init {
1340 my $url = shift;
1341 unless (defined $_trunk || @_branches || @_tags) {
1342 usage(1);
1343 }
1344
1345 $_prefix = '' unless defined $_prefix;
1346 if (defined $url) {
1347 $url = canonicalize_url($url);
1348 init_subdir(@_);
1349 }
1350 do_git_init_db();
1351 if (defined $_trunk) {
1352 $_trunk =~ s#^/+##;
1353 my $trunk_ref = 'refs/remotes/' . $_prefix . 'trunk';
1354 # try both old-style and new-style lookups:
1355 my $gs_trunk = eval { Git::SVN->new($trunk_ref) };
1356 unless ($gs_trunk) {
1357 my ($trunk_url, $trunk_path) =
1358 complete_svn_url($url, $_trunk);
1359 $gs_trunk = Git::SVN->init($trunk_url, $trunk_path,
1360 undef, $trunk_ref);
1361 }
1362 }
1363 return unless @_branches || @_tags;
1364 my $ra = $url ? Git::SVN::Ra->new($url) : undef;
1365 foreach my $path (@_branches) {
1366 complete_url_ls_init($ra, $path, '--branches/-b', $_prefix);
1367 }
1368 foreach my $path (@_tags) {
1369 complete_url_ls_init($ra, $path, '--tags/-t', $_prefix.'tags/');
1370 }
1371 }
1372
1373 sub cmd_multi_fetch {
1374 $Git::SVN::no_reuse_existing = undef;
1375 my $remotes = Git::SVN::read_all_remotes();
1376 foreach my $repo_id (sort keys %$remotes) {
1377 if ($remotes->{$repo_id}->{url}) {
1378 Git::SVN::fetch_all($repo_id, $remotes);
1379 }
1380 }
1381 }
1382
1383 # this command is special because it requires no metadata
1384 sub cmd_commit_diff {
1385 my ($ta, $tb, $url) = @_;
1386 my $usage = "Usage: $0 commit-diff -r<revision> ".
1387 "<tree-ish> <tree-ish> [<URL>]";
1388 fatal($usage) if (!defined $ta || !defined $tb);
1389 my $svn_path = '';
1390 if (!defined $url) {
1391 my $gs = eval { Git::SVN->new };
1392 if (!$gs) {
1393 fatal("Needed URL or usable git-svn --id in ",
1394 "the command-line\n", $usage);
1395 }
1396 $url = $gs->{url};
1397 $svn_path = $gs->{path};
1398 }
1399 unless (defined $_revision) {
1400 fatal("-r|--revision is a required argument\n", $usage);
1401 }
1402 if (defined $_message && defined $_file) {
1403 fatal("Both --message/-m and --file/-F specified ",
1404 "for the commit message.\n",
1405 "I have no idea what you mean");
1406 }
1407 if (defined $_file) {
1408 $_message = file_to_s($_file);
1409 } else {
1410 $_message ||= get_commit_entry($tb)->{log};
1411 }
1412 my $ra ||= Git::SVN::Ra->new($url);
1413 my $r = $_revision;
1414 if ($r eq 'HEAD') {
1415 $r = $ra->get_latest_revnum;
1416 } elsif ($r !~ /^\d+$/) {
1417 die "revision argument: $r not understood by git-svn\n";
1418 }
1419 my %ed_opts = ( r => $r,
1420 log => $_message,
1421 ra => $ra,
1422 tree_a => $ta,
1423 tree_b => $tb,
1424 editor_cb => sub { print "Committed r$_[0]\n" },
1425 svn_path => $svn_path );
1426 if (!Git::SVN::Editor->new(\%ed_opts)->apply_diff) {
1427 print "No changes\n$ta == $tb\n";
1428 }
1429 }
1430
1431 sub escape_uri_only {
1432 my ($uri) = @_;
1433 my @tmp;
1434 foreach (split m{/}, $uri) {
1435 s/([^~\w.%+-]|%(?![a-fA-F0-9]{2}))/sprintf("%%%02X",ord($1))/eg;
1436 push @tmp, $_;
1437 }
1438 join('/', @tmp);
1439 }
1440
1441 sub escape_url {
1442 my ($url) = @_;
1443 if ($url =~ m#^([^:]+)://([^/]*)(.*)$#) {
1444 my ($scheme, $domain, $uri) = ($1, $2, escape_uri_only($3));
1445 $url = "$scheme://$domain$uri";
1446 }
1447 $url;
1448 }
1449
1450 sub cmd_info {
1451 my $path = canonicalize_path(defined($_[0]) ? $_[0] : ".");
1452 my $fullpath = canonicalize_path($cmd_dir_prefix . $path);
1453 if (exists $_[1]) {
1454 die "Too many arguments specified\n";
1455 }
1456
1457 my ($file_type, $diff_status) = find_file_type_and_diff_status($path);
1458
1459 if (!$file_type && !$diff_status) {
1460 print STDERR "svn: '$path' is not under version control\n";
1461 exit 1;
1462 }
1463
1464 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
1465 unless ($gs) {
1466 die "Unable to determine upstream SVN information from ",
1467 "working tree history\n";
1468 }
1469
1470 # canonicalize_path() will return "" to make libsvn 1.5.x happy,
1471 $path = "." if $path eq "";
1472
1473 my $full_url = $url . ($fullpath eq "" ? "" : "/$fullpath");
1474
1475 if ($_url) {
1476 print escape_url($full_url), "\n";
1477 return;
1478 }
1479
1480 my $result = "Path: $path\n";
1481 $result .= "Name: " . basename($path) . "\n" if $file_type ne "dir";
1482 $result .= "URL: " . escape_url($full_url) . "\n";
1483
1484 eval {
1485 my $repos_root = $gs->repos_root;
1486 Git::SVN::remove_username($repos_root);
1487 $result .= "Repository Root: " . escape_url($repos_root) . "\n";
1488 };
1489 if ($@) {
1490 $result .= "Repository Root: (offline)\n";
1491 }
1492 ::_req_svn();
1493 $result .= "Repository UUID: $uuid\n" unless $diff_status eq "A" &&
1494 (::compare_svn_version('1.5.4') <= 0 || $file_type ne "dir");
1495 $result .= "Revision: " . ($diff_status eq "A" ? 0 : $rev) . "\n";
1496
1497 $result .= "Node Kind: " .
1498 ($file_type eq "dir" ? "directory" : "file") . "\n";
1499
1500 my $schedule = $diff_status eq "A"
1501 ? "add"
1502 : ($diff_status eq "D" ? "delete" : "normal");
1503 $result .= "Schedule: $schedule\n";
1504
1505 if ($diff_status eq "A") {
1506 print $result, "\n";
1507 return;
1508 }
1509
1510 my ($lc_author, $lc_rev, $lc_date_utc);
1511 my @args = Git::SVN::Log::git_svn_log_cmd($rev, $rev, "--", $fullpath);
1512 my $log = command_output_pipe(@args);
1513 my $esc_color = qr/(?:\033\[(?:(?:\d+;)*\d*)?m)*/;
1514 while (<$log>) {
1515 if (/^${esc_color}author (.+) <[^>]+> (\d+) ([\-\+]?\d+)$/o) {
1516 $lc_author = $1;
1517 $lc_date_utc = Git::SVN::Log::parse_git_date($2, $3);
1518 } elsif (/^${esc_color} (git-svn-id:.+)$/o) {
1519 (undef, $lc_rev, undef) = ::extract_metadata($1);
1520 }
1521 }
1522 close $log;
1523
1524 Git::SVN::Log::set_local_timezone();
1525
1526 $result .= "Last Changed Author: $lc_author\n";
1527 $result .= "Last Changed Rev: $lc_rev\n";
1528 $result .= "Last Changed Date: " .
1529 Git::SVN::Log::format_svn_date($lc_date_utc) . "\n";
1530
1531 if ($file_type ne "dir") {
1532 my $text_last_updated_date =
1533 ($diff_status eq "D" ? $lc_date_utc : (stat $path)[9]);
1534 $result .=
1535 "Text Last Updated: " .
1536 Git::SVN::Log::format_svn_date($text_last_updated_date) .
1537 "\n";
1538 my $checksum;
1539 if ($diff_status eq "D") {
1540 my ($fh, $ctx) =
1541 command_output_pipe(qw(cat-file blob), "HEAD:$path");
1542 if ($file_type eq "link") {
1543 my $file_name = <$fh>;
1544 $checksum = md5sum("link $file_name");
1545 } else {
1546 $checksum = md5sum($fh);
1547 }
1548 command_close_pipe($fh, $ctx);
1549 } elsif ($file_type eq "link") {
1550 my $file_name =
1551 command(qw(cat-file blob), "HEAD:$path");
1552 $checksum =
1553 md5sum("link " . $file_name);
1554 } else {
1555 open FILE, "<", $path or die $!;
1556 $checksum = md5sum(\*FILE);
1557 close FILE or die $!;
1558 }
1559 $result .= "Checksum: " . $checksum . "\n";
1560 }
1561
1562 print $result, "\n";
1563 }
1564
1565 sub cmd_reset {
1566 my $target = shift || $_revision or die "SVN revision required\n";
1567 $target = $1 if $target =~ /^r(\d+)$/;
1568 $target =~ /^\d+$/ or die "Numeric SVN revision expected\n";
1569 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
1570 unless ($gs) {
1571 die "Unable to determine upstream SVN information from ".
1572 "history\n";
1573 }
1574 my ($r, $c) = $gs->find_rev_before($target, not $_fetch_parent);
1575 die "Cannot find SVN revision $target\n" unless defined($c);
1576 $gs->rev_map_set($r, $c, 'reset', $uuid);
1577 print "r$r = $c ($gs->{ref_id})\n";
1578 }
1579
1580 sub cmd_gc {
1581 if (!$can_compress) {
1582 warn "Compress::Zlib could not be found; unhandled.log " .
1583 "files will not be compressed.\n";
1584 }
1585 find({ wanted => \&gc_directory, no_chdir => 1}, "$ENV{GIT_DIR}/svn");
1586 }
1587
1588 ########################### utility functions #########################
1589
1590 sub rebase_cmd {
1591 my @cmd = qw/rebase/;
1592 push @cmd, '-v' if $_verbose;
1593 push @cmd, qw/--merge/ if $_merge;
1594 push @cmd, "--strategy=$_strategy" if $_strategy;
1595 push @cmd, "--preserve-merges" if $_preserve_merges;
1596 @cmd;
1597 }
1598
1599 sub post_fetch_checkout {
1600 return if $_no_checkout;
1601 my $gs = $Git::SVN::_head or return;
1602 return if verify_ref('refs/heads/master^0');
1603
1604 # look for "trunk" ref if it exists
1605 my $remote = Git::SVN::read_all_remotes()->{$gs->{repo_id}};
1606 my $fetch = $remote->{fetch};
1607 if ($fetch) {
1608 foreach my $p (keys %$fetch) {
1609 basename($fetch->{$p}) eq 'trunk' or next;
1610 $gs = Git::SVN->new($fetch->{$p}, $gs->{repo_id}, $p);
1611 last;
1612 }
1613 }
1614
1615 my $valid_head = verify_ref('HEAD^0');
1616 command_noisy(qw(update-ref refs/heads/master), $gs->refname);
1617 return if ($valid_head || !verify_ref('HEAD^0'));
1618
1619 return if $ENV{GIT_DIR} !~ m#^(?:.*/)?\.git$#;
1620 my $index = $ENV{GIT_INDEX_FILE} || "$ENV{GIT_DIR}/index";
1621 return if -f $index;
1622
1623 return if command_oneline(qw/rev-parse --is-inside-work-tree/) eq 'false';
1624 return if command_oneline(qw/rev-parse --is-inside-git-dir/) eq 'true';
1625 command_noisy(qw/read-tree -m -u -v HEAD HEAD/);
1626 print STDERR "Checked out HEAD:\n ",
1627 $gs->full_url, " r", $gs->last_rev, "\n";
1628 if (auto_create_empty_directories($gs)) {
1629 $gs->mkemptydirs($gs->last_rev);
1630 }
1631 }
1632
1633 sub complete_svn_url {
1634 my ($url, $path) = @_;
1635 $path =~ s#/+$##;
1636 if ($path !~ m#^[a-z\+]+://#) {
1637 if (!defined $url || $url !~ m#^[a-z\+]+://#) {
1638 fatal("E: '$path' is not a complete URL ",
1639 "and a separate URL is not specified");
1640 }
1641 return ($url, $path);
1642 }
1643 return ($path, '');
1644 }
1645
1646 sub complete_url_ls_init {
1647 my ($ra, $repo_path, $switch, $pfx) = @_;
1648 unless ($repo_path) {
1649 print STDERR "W: $switch not specified\n";
1650 return;
1651 }
1652 $repo_path =~ s#/+$##;
1653 if ($repo_path =~ m#^[a-z\+]+://#) {
1654 $ra = Git::SVN::Ra->new($repo_path);
1655 $repo_path = '';
1656 } else {
1657 $repo_path =~ s#^/+##;
1658 unless ($ra) {
1659 fatal("E: '$repo_path' is not a complete URL ",
1660 "and a separate URL is not specified");
1661 }
1662 }
1663 my $url = $ra->{url};
1664 my $gs = Git::SVN->init($url, undef, undef, undef, 1);
1665 my $k = "svn-remote.$gs->{repo_id}.url";
1666 my $orig_url = eval { command_oneline(qw/config --get/, $k) };
1667 if ($orig_url && ($orig_url ne $gs->{url})) {
1668 die "$k already set: $orig_url\n",
1669 "wanted to set to: $gs->{url}\n";
1670 }
1671 command_oneline('config', $k, $gs->{url}) unless $orig_url;
1672 my $remote_path = "$gs->{path}/$repo_path";
1673 $remote_path =~ s{%([0-9A-F]{2})}{chr hex($1)}ieg;
1674 $remote_path =~ s#/+#/#g;
1675 $remote_path =~ s#^/##g;
1676 $remote_path .= "/*" if $remote_path !~ /\*/;
1677 my ($n) = ($switch =~ /^--(\w+)/);
1678 if (length $pfx && $pfx !~ m#/$#) {
1679 die "--prefix='$pfx' must have a trailing slash '/'\n";
1680 }
1681 command_noisy('config',
1682 '--add',
1683 "svn-remote.$gs->{repo_id}.$n",
1684 "$remote_path:refs/remotes/$pfx*" .
1685 ('/*' x (($remote_path =~ tr/*/*/) - 1)) );
1686 }
1687
1688 sub verify_ref {
1689 my ($ref) = @_;
1690 eval { command_oneline([ 'rev-parse', '--verify', $ref ],
1691 { STDERR => 0 }); };
1692 }
1693
1694 sub get_tree_from_treeish {
1695 my ($treeish) = @_;
1696 # $treeish can be a symbolic ref, too:
1697 my $type = command_oneline(qw/cat-file -t/, $treeish);
1698 my $expected;
1699 while ($type eq 'tag') {
1700 ($treeish, $type) = command(qw/cat-file tag/, $treeish);
1701 }
1702 if ($type eq 'commit') {
1703 $expected = (grep /^tree /, command(qw/cat-file commit/,
1704 $treeish))[0];
1705 ($expected) = ($expected =~ /^tree ($sha1)$/o);
1706 die "Unable to get tree from $treeish\n" unless $expected;
1707 } elsif ($type eq 'tree') {
1708 $expected = $treeish;
1709 } else {
1710 die "$treeish is a $type, expected tree, tag or commit\n";
1711 }
1712 return $expected;
1713 }
1714
1715 sub get_commit_entry {
1716 my ($treeish) = shift;
1717 my %log_entry = ( log => '', tree => get_tree_from_treeish($treeish) );
1718 my $commit_editmsg = "$ENV{GIT_DIR}/COMMIT_EDITMSG";
1719 my $commit_msg = "$ENV{GIT_DIR}/COMMIT_MSG";
1720 open my $log_fh, '>', $commit_editmsg or croak $!;
1721
1722 my $type = command_oneline(qw/cat-file -t/, $treeish);
1723 if ($type eq 'commit' || $type eq 'tag') {
1724 my ($msg_fh, $ctx) = command_output_pipe('cat-file',
1725 $type, $treeish);
1726 my $in_msg = 0;
1727 my $author;
1728 my $saw_from = 0;
1729 my $msgbuf = "";
1730 while (<$msg_fh>) {
1731 if (!$in_msg) {
1732 $in_msg = 1 if (/^\s*$/);
1733 $author = $1 if (/^author (.*>)/);
1734 } elsif (/^git-svn-id: /) {
1735 # skip this for now, we regenerate the
1736 # correct one on re-fetch anyways
1737 # TODO: set *:merge properties or like...
1738 } else {
1739 if (/^From:/ || /^Signed-off-by:/) {
1740 $saw_from = 1;
1741 }
1742 $msgbuf .= $_;
1743 }
1744 }
1745 $msgbuf =~ s/\s+$//s;
1746 if ($Git::SVN::_add_author_from && defined($author)
1747 && !$saw_from) {
1748 $msgbuf .= "\n\nFrom: $author";
1749 }
1750 print $log_fh $msgbuf or croak $!;
1751 command_close_pipe($msg_fh, $ctx);
1752 }
1753 close $log_fh or croak $!;
1754
1755 if ($_edit || ($type eq 'tree')) {
1756 chomp(my $editor = command_oneline(qw(var GIT_EDITOR)));
1757 system('sh', '-c', $editor.' "$@"', $editor, $commit_editmsg);
1758 }
1759 rename $commit_editmsg, $commit_msg or croak $!;
1760 {
1761 require Encode;
1762 # SVN requires messages to be UTF-8 when entering the repo
1763 local $/;
1764 open $log_fh, '<', $commit_msg or croak $!;
1765 binmode $log_fh;
1766 chomp($log_entry{log} = <$log_fh>);
1767
1768 my $enc = Git::config('i18n.commitencoding') || 'UTF-8';
1769 my $msg = $log_entry{log};
1770
1771 eval { $msg = Encode::decode($enc, $msg, 1) };
1772 if ($@) {
1773 die "Could not decode as $enc:\n", $msg,
1774 "\nPerhaps you need to set i18n.commitencoding\n";
1775 }
1776
1777 eval { $msg = Encode::encode('UTF-8', $msg, 1) };
1778 die "Could not encode as UTF-8:\n$msg\n" if $@;
1779
1780 $log_entry{log} = $msg;
1781
1782 close $log_fh or croak $!;
1783 }
1784 unlink $commit_msg;
1785 \%log_entry;
1786 }
1787
1788 sub s_to_file {
1789 my ($str, $file, $mode) = @_;
1790 open my $fd,'>',$file or croak $!;
1791 print $fd $str,"\n" or croak $!;
1792 close $fd or croak $!;
1793 chmod ($mode &~ umask, $file) if (defined $mode);
1794 }
1795
1796 sub file_to_s {
1797 my $file = shift;
1798 open my $fd,'<',$file or croak "$!: file: $file\n";
1799 local $/;
1800 my $ret = <$fd>;
1801 close $fd or croak $!;
1802 $ret =~ s/\s*$//s;
1803 return $ret;
1804 }
1805
1806 # '<svn username> = real-name <email address>' mapping based on git-svnimport:
1807 sub load_authors {
1808 open my $authors, '<', $_authors or die "Can't open $_authors $!\n";
1809 my $log = $cmd eq 'log';
1810 while (<$authors>) {
1811 chomp;
1812 next unless /^(.+?|\(no author\))\s*=\s*(.+?)\s*<(.+)>\s*$/;
1813 my ($user, $name, $email) = ($1, $2, $3);
1814 if ($log) {
1815 $Git::SVN::Log::rusers{"$name <$email>"} = $user;
1816 } else {
1817 $users{$user} = [$name, $email];
1818 }
1819 }
1820 close $authors or croak $!;
1821 }
1822
1823 # convert GetOpt::Long specs for use by git-config
1824 sub read_git_config {
1825 my $opts = shift;
1826 my @config_only;
1827 foreach my $o (keys %$opts) {
1828 # if we have mixedCase and a long option-only, then
1829 # it's a config-only variable that we don't need for
1830 # the command-line.
1831 push @config_only, $o if ($o =~ /[A-Z]/ && $o =~ /^[a-z]+$/i);
1832 my $v = $opts->{$o};
1833 my ($key) = ($o =~ /^([a-zA-Z\-]+)/);
1834 $key =~ s/-//g;
1835 my $arg = 'git config';
1836 $arg .= ' --int' if ($o =~ /[:=]i$/);
1837 $arg .= ' --bool' if ($o !~ /[:=][sfi]$/);
1838 if (ref $v eq 'ARRAY') {
1839 chomp(my @tmp = `$arg --get-all svn.$key`);
1840 @$v = @tmp if @tmp;
1841 } else {
1842 chomp(my $tmp = `$arg --get svn.$key`);
1843 if ($tmp && !($arg =~ / --bool/ && $tmp eq 'false')) {
1844 $$v = $tmp;
1845 }
1846 }
1847 }
1848 delete @$opts{@config_only} if @config_only;
1849 }
1850
1851 sub extract_metadata {
1852 my $id = shift or return (undef, undef, undef);
1853 my ($url, $rev, $uuid) = ($id =~ /^\s*git-svn-id:\s+(.*)\@(\d+)
1854 \s([a-f\d\-]+)$/ix);
1855 if (!defined $rev || !$uuid || !$url) {
1856 # some of the original repositories I made had
1857 # identifiers like this:
1858 ($rev, $uuid) = ($id =~/^\s*git-svn-id:\s(\d+)\@([a-f\d\-]+)/i);
1859 }
1860 return ($url, $rev, $uuid);
1861 }
1862
1863 sub cmt_metadata {
1864 return extract_metadata((grep(/^git-svn-id: /,
1865 command(qw/cat-file commit/, shift)))[-1]);
1866 }
1867
1868 sub cmt_sha2rev_batch {
1869 my %s2r;
1870 my ($pid, $in, $out, $ctx) = command_bidi_pipe(qw/cat-file --batch/);
1871 my $list = shift;
1872
1873 foreach my $sha (@{$list}) {
1874 my $first = 1;
1875 my $size = 0;
1876 print $out $sha, "\n";
1877
1878 while (my $line = <$in>) {
1879 if ($first && $line =~ /^[[:xdigit:]]{40}\smissing$/) {
1880 last;
1881 } elsif ($first &&
1882 $line =~ /^[[:xdigit:]]{40}\scommit\s(\d+)$/) {
1883 $first = 0;
1884 $size = $1;
1885 next;
1886 } elsif ($line =~ /^(git-svn-id: )/) {
1887 my (undef, $rev, undef) =
1888 extract_metadata($line);
1889 $s2r{$sha} = $rev;
1890 }
1891
1892 $size -= length($line);
1893 last if ($size == 0);
1894 }
1895 }
1896
1897 command_close_bidi_pipe($pid, $in, $out, $ctx);
1898
1899 return \%s2r;
1900 }
1901
1902 sub working_head_info {
1903 my ($head, $refs) = @_;
1904 my @args = qw/rev-list --first-parent --pretty=medium/;
1905 my ($fh, $ctx) = command_output_pipe(@args, $head);
1906 my $hash;
1907 my %max;
1908 while (<$fh>) {
1909 if ( m{^commit ($::sha1)$} ) {
1910 unshift @$refs, $hash if $hash and $refs;
1911 $hash = $1;
1912 next;
1913 }
1914 next unless s{^\s*(git-svn-id:)}{$1};
1915 my ($url, $rev, $uuid) = extract_metadata($_);
1916 if (defined $url && defined $rev) {
1917 next if $max{$url} and $max{$url} < $rev;
1918 if (my $gs = Git::SVN->find_by_url($url)) {
1919 my $c = $gs->rev_map_get($rev, $uuid);
1920 if ($c && $c eq $hash) {
1921 close $fh; # break the pipe
1922 return ($url, $rev, $uuid, $gs);
1923 } else {
1924 $max{$url} ||= $gs->rev_map_max;
1925 }
1926 }
1927 }
1928 }
1929 command_close_pipe($fh, $ctx);
1930 (undef, undef, undef, undef);
1931 }
1932
1933 sub read_commit_parents {
1934 my ($parents, $c) = @_;
1935 chomp(my $p = command_oneline(qw/rev-list --parents -1/, $c));
1936 $p =~ s/^($c)\s*// or die "rev-list --parents -1 $c failed!\n";
1937 @{$parents->{$c}} = split(/ /, $p);
1938 }
1939
1940 sub linearize_history {
1941 my ($gs, $refs) = @_;
1942 my %parents;
1943 foreach my $c (@$refs) {
1944 read_commit_parents(\%parents, $c);
1945 }
1946
1947 my @linear_refs;
1948 my %skip = ();
1949 my $last_svn_commit = $gs->last_commit;
1950 foreach my $c (reverse @$refs) {
1951 next if $c eq $last_svn_commit;
1952 last if $skip{$c};
1953
1954 unshift @linear_refs, $c;
1955 $skip{$c} = 1;
1956
1957 # we only want the first parent to diff against for linear
1958 # history, we save the rest to inject when we finalize the
1959 # svn commit
1960 my $fp_a = verify_ref("$c~1");
1961 my $fp_b = shift @{$parents{$c}} if $parents{$c};
1962 if (!$fp_a || !$fp_b) {
1963 die "Commit $c\n",
1964 "has no parent commit, and therefore ",
1965 "nothing to diff against.\n",
1966 "You should be working from a repository ",
1967 "originally created by git-svn\n";
1968 }
1969 if ($fp_a ne $fp_b) {
1970 die "$c~1 = $fp_a, however parsing commit $c ",
1971 "revealed that:\n$c~1 = $fp_b\nBUG!\n";
1972 }
1973
1974 foreach my $p (@{$parents{$c}}) {
1975 $skip{$p} = 1;
1976 }
1977 }
1978 (\@linear_refs, \%parents);
1979 }
1980
1981 sub find_file_type_and_diff_status {
1982 my ($path) = @_;
1983 return ('dir', '') if $path eq '';
1984
1985 my $diff_output =
1986 command_oneline(qw(diff --cached --name-status --), $path) || "";
1987 my $diff_status = (split(' ', $diff_output))[0] || "";
1988
1989 my $ls_tree = command_oneline(qw(ls-tree HEAD), $path) || "";
1990
1991 return (undef, undef) if !$diff_status && !$ls_tree;
1992
1993 if ($diff_status eq "A") {
1994 return ("link", $diff_status) if -l $path;
1995 return ("dir", $diff_status) if -d $path;
1996 return ("file", $diff_status);
1997 }
1998
1999 my $mode = (split(' ', $ls_tree))[0] || "";
2000
2001 return ("link", $diff_status) if $mode eq "120000";
2002 return ("dir", $diff_status) if $mode eq "040000";
2003 return ("file", $diff_status);
2004 }
2005
2006 sub md5sum {
2007 my $arg = shift;
2008 my $ref = ref $arg;
2009 my $md5 = Digest::MD5->new();
2010 if ($ref eq 'GLOB' || $ref eq 'IO::File' || $ref eq 'File::Temp') {
2011 $md5->addfile($arg) or croak $!;
2012 } elsif ($ref eq 'SCALAR') {
2013 $md5->add($$arg) or croak $!;
2014 } elsif (!$ref) {
2015 $md5->add($arg) or croak $!;
2016 } else {
2017 ::fatal "Can't provide MD5 hash for unknown ref type: '", $ref, "'";
2018 }
2019 return $md5->hexdigest();
2020 }
2021
2022 sub gc_directory {
2023 if ($can_compress && -f $_ && basename($_) eq "unhandled.log") {
2024 my $out_filename = $_ . ".gz";
2025 open my $in_fh, "<", $_ or die "Unable to open $_: $!\n";
2026 binmode $in_fh;
2027 my $gz = Compress::Zlib::gzopen($out_filename, "ab") or
2028 die "Unable to open $out_filename: $!\n";
2029
2030 my $res;
2031 while ($res = sysread($in_fh, my $str, 1024)) {
2032 $gz->gzwrite($str) or
2033 die "Unable to write: ".$gz->gzerror()."!\n";
2034 }
2035 unlink $_ or die "unlink $File::Find::name: $!\n";
2036 } elsif (-f $_ && basename($_) eq "index") {
2037 unlink $_ or die "unlink $_: $!\n";
2038 }
2039 }
2040
2041 package Git::SVN;
2042 use strict;
2043 use warnings;
2044 use Fcntl qw/:DEFAULT :seek/;
2045 use constant rev_map_fmt => 'NH40';
2046 use vars qw/$default_repo_id $default_ref_id $_no_metadata $_follow_parent
2047 $_repack $_repack_flags $_use_svm_props $_head
2048 $_use_svnsync_props $no_reuse_existing $_minimize_url
2049 $_use_log_author $_add_author_from $_localtime/;
2050 use Carp qw/croak/;
2051 use File::Path qw/mkpath/;
2052 use File::Copy qw/copy/;
2053 use IPC::Open3;
2054 use Time::Local;
2055 use Memoize; # core since 5.8.0, Jul 2002
2056 use Memoize::Storable;
2057 use POSIX qw(:signal_h);
2058 my $can_use_yaml;
2059 BEGIN {
2060 $can_use_yaml = eval { require Git::SVN::Memoize::YAML; 1};
2061 }
2062
2063 my ($_gc_nr, $_gc_period);
2064
2065 # properties that we do not log:
2066 my %SKIP_PROP;
2067 BEGIN {
2068 %SKIP_PROP = map { $_ => 1 } qw/svn:wc:ra_dav:version-url
2069 svn:special svn:executable
2070 svn:entry:committed-rev
2071 svn:entry:last-author
2072 svn:entry:uuid
2073 svn:entry:committed-date/;
2074
2075 # some options are read globally, but can be overridden locally
2076 # per [svn-remote "..."] section. Command-line options will *NOT*
2077 # override options set in an [svn-remote "..."] section
2078 no strict 'refs';
2079 for my $option (qw/follow_parent no_metadata use_svm_props
2080 use_svnsync_props/) {
2081 my $key = $option;
2082 $key =~ tr/_//d;
2083 my $prop = "-$option";
2084 *$option = sub {
2085 my ($self) = @_;
2086 return $self->{$prop} if exists $self->{$prop};
2087 my $k = "svn-remote.$self->{repo_id}.$key";
2088 eval { command_oneline(qw/config --get/, $k) };
2089 if ($@) {
2090 $self->{$prop} = ${"Git::SVN::_$option"};
2091 } else {
2092 my $v = command_oneline(qw/config --bool/,$k);
2093 $self->{$prop} = $v eq 'false' ? 0 : 1;
2094 }
2095 return $self->{$prop};
2096 }
2097 }
2098 }
2099
2100
2101 my (%LOCKFILES, %INDEX_FILES);
2102 END {
2103 unlink keys %LOCKFILES if %LOCKFILES;
2104 unlink keys %INDEX_FILES if %INDEX_FILES;
2105 }
2106
2107 sub resolve_local_globs {
2108 my ($url, $fetch, $glob_spec) = @_;
2109 return unless defined $glob_spec;
2110 my $ref = $glob_spec->{ref};
2111 my $path = $glob_spec->{path};
2112 foreach (command(qw#for-each-ref --format=%(refname) refs/#)) {
2113 next unless m#^$ref->{regex}$#;
2114 my $p = $1;
2115 my $pathname = desanitize_refname($path->full_path($p));
2116 my $refname = desanitize_refname($ref->full_path($p));
2117 if (my $existing = $fetch->{$pathname}) {
2118 if ($existing ne $refname) {
2119 die "Refspec conflict:\n",
2120 "existing: $existing\n",
2121 " globbed: $refname\n";
2122 }
2123 my $u = (::cmt_metadata("$refname"))[0];
2124 $u =~ s!^\Q$url\E(/|$)!! or die
2125 "$refname: '$url' not found in '$u'\n";
2126 if ($pathname ne $u) {
2127 warn "W: Refspec glob conflict ",
2128 "(ref: $refname):\n",
2129 "expected path: $pathname\n",
2130 " real path: $u\n",
2131 "Continuing ahead with $u\n";
2132 next;
2133 }
2134 } else {
2135 $fetch->{$pathname} = $refname;
2136 }
2137 }
2138 }
2139
2140 sub parse_revision_argument {
2141 my ($base, $head) = @_;
2142 if (!defined $::_revision || $::_revision eq 'BASE:HEAD') {
2143 return ($base, $head);
2144 }
2145 return ($1, $2) if ($::_revision =~ /^(\d+):(\d+)$/);
2146 return ($::_revision, $::_revision) if ($::_revision =~ /^\d+$/);
2147 return ($head, $head) if ($::_revision eq 'HEAD');
2148 return ($base, $1) if ($::_revision =~ /^BASE:(\d+)$/);
2149 return ($1, $head) if ($::_revision =~ /^(\d+):HEAD$/);
2150 die "revision argument: $::_revision not understood by git-svn\n";
2151 }
2152
2153 sub fetch_all {
2154 my ($repo_id, $remotes) = @_;
2155 if (ref $repo_id) {
2156 my $gs = $repo_id;
2157 $repo_id = undef;
2158 $repo_id = $gs->{repo_id};
2159 }
2160 $remotes ||= read_all_remotes();
2161 my $remote = $remotes->{$repo_id} or
2162 die "[svn-remote \"$repo_id\"] unknown\n";
2163 my $fetch = $remote->{fetch};
2164 my $url = $remote->{url} or die "svn-remote.$repo_id.url not defined\n";
2165 my (@gs, @globs);
2166 my $ra = Git::SVN::Ra->new($url);
2167 my $uuid = $ra->get_uuid;
2168 my $head = $ra->get_latest_revnum;
2169
2170 # ignore errors, $head revision may not even exist anymore
2171 eval { $ra->get_log("", $head, 0, 1, 0, 1, sub { $head = $_[1] }) };
2172 warn "W: $@\n" if $@;
2173
2174 my $base = defined $fetch ? $head : 0;
2175
2176 # read the max revs for wildcard expansion (branches/*, tags/*)
2177 foreach my $t (qw/branches tags/) {
2178 defined $remote->{$t} or next;
2179 push @globs, @{$remote->{$t}};
2180
2181 my $max_rev = eval { tmp_config(qw/--int --get/,
2182 "svn-remote.$repo_id.${t}-maxRev") };
2183 if (defined $max_rev && ($max_rev < $base)) {
2184 $base = $max_rev;
2185 } elsif (!defined $max_rev) {
2186 $base = 0;
2187 }
2188 }
2189
2190 if ($fetch) {
2191 foreach my $p (sort keys %$fetch) {
2192 my $gs = Git::SVN->new($fetch->{$p}, $repo_id, $p);
2193 my $lr = $gs->rev_map_max;
2194 if (defined $lr) {
2195 $base = $lr if ($lr < $base);
2196 }
2197 push @gs, $gs;
2198 }
2199 }
2200
2201 ($base, $head) = parse_revision_argument($base, $head);
2202 $ra->gs_fetch_loop_common($base, $head, \@gs, \@globs);
2203 }
2204
2205 sub read_all_remotes {
2206 my $r = {};
2207 my $use_svm_props = eval { command_oneline(qw/config --bool
2208 svn.useSvmProps/) };
2209 $use_svm_props = $use_svm_props eq 'true' if $use_svm_props;
2210 my $svn_refspec = qr{\s*(.*?)\s*:\s*(.+?)\s*};
2211 foreach (grep { s/^svn-remote\.// } command(qw/config -l/)) {
2212 if (m!^(.+)\.fetch=$svn_refspec$!) {
2213 my ($remote, $local_ref, $remote_ref) = ($1, $2, $3);
2214 die("svn-remote.$remote: remote ref '$remote_ref' "
2215 . "must start with 'refs/'\n")
2216 unless $remote_ref =~ m{^refs/};
2217 $local_ref = uri_decode($local_ref);
2218 $r->{$remote}->{fetch}->{$local_ref} = $remote_ref;
2219 $r->{$remote}->{svm} = {} if $use_svm_props;
2220 } elsif (m!^(.+)\.usesvmprops=\s*(.*)\s*$!) {
2221 $r->{$1}->{svm} = {};
2222 } elsif (m!^(.+)\.url=\s*(.*)\s*$!) {
2223 $r->{$1}->{url} = $2;
2224 } elsif (m!^(.+)\.pushurl=\s*(.*)\s*$!) {
2225 $r->{$1}->{pushurl} = $2;
2226 } elsif (m!^(.+)\.ignore-refs=\s*(.*)\s*$!) {
2227 $r->{$1}->{ignore_refs_regex} = $2;
2228 } elsif (m!^(.+)\.(branches|tags)=$svn_refspec$!) {
2229 my ($remote, $t, $local_ref, $remote_ref) =
2230 ($1, $2, $3, $4);
2231 die("svn-remote.$remote: remote ref '$remote_ref' ($t) "
2232 . "must start with 'refs/'\n")
2233 unless $remote_ref =~ m{^refs/};
2234 $local_ref = uri_decode($local_ref);
2235 my $rs = {
2236 t => $t,
2237 remote => $remote,
2238 path => Git::SVN::GlobSpec->new($local_ref, 1),
2239 ref => Git::SVN::GlobSpec->new($remote_ref, 0) };
2240 if (length($rs->{ref}->{right}) != 0) {
2241 die "The '*' glob character must be the last ",
2242 "character of '$remote_ref'\n";
2243 }
2244 push @{ $r->{$remote}->{$t} }, $rs;
2245 }
2246 }
2247
2248 map {
2249 if (defined $r->{$_}->{svm}) {
2250 my $svm;
2251 eval {
2252 my $section = "svn-remote.$_";
2253 $svm = {
2254 source => tmp_config('--get',
2255 "$section.svm-source"),
2256 replace => tmp_config('--get',
2257 "$section.svm-replace"),
2258 }
2259 };
2260 $r->{$_}->{svm} = $svm;
2261 }
2262 } keys %$r;
2263
2264 foreach my $remote (keys %$r) {
2265 foreach ( grep { defined $_ }
2266 map { $r->{$remote}->{$_} } qw(branches tags) ) {
2267 foreach my $rs ( @$_ ) {
2268 $rs->{ignore_refs_regex} =
2269 $r->{$remote}->{ignore_refs_regex};
2270 }
2271 }
2272 }
2273
2274 $r;
2275 }
2276
2277 sub init_vars {
2278 $_gc_nr = $_gc_period = 1000;
2279 if (defined $_repack || defined $_repack_flags) {
2280 warn "Repack options are obsolete; they have no effect.\n";
2281 }
2282 }
2283
2284 sub verify_remotes_sanity {
2285 return unless -d $ENV{GIT_DIR};
2286 my %seen;
2287 foreach (command(qw/config -l/)) {
2288 if (m!^svn-remote\.(?:.+)\.fetch=.*:refs/remotes/(\S+)\s*$!) {
2289 if ($seen{$1}) {
2290 die "Remote ref refs/remote/$1 is tracked by",
2291 "\n \"$_\"\nand\n \"$seen{$1}\"\n",
2292 "Please resolve this ambiguity in ",
2293 "your git configuration file before ",
2294 "continuing\n";
2295 }
2296 $seen{$1} = $_;
2297 }
2298 }
2299 }
2300
2301 sub find_existing_remote {
2302 my ($url, $remotes) = @_;
2303 return undef if $no_reuse_existing;
2304 my $existing;
2305 foreach my $repo_id (keys %$remotes) {
2306 my $u = $remotes->{$repo_id}->{url} or next;
2307 next if $u ne $url;
2308 $existing = $repo_id;
2309 last;
2310 }
2311 $existing;
2312 }
2313
2314 sub init_remote_config {
2315 my ($self, $url, $no_write) = @_;
2316 $url =~ s!/+$!!; # strip trailing slash
2317 my $r = read_all_remotes();
2318 my $existing = find_existing_remote($url, $r);
2319 if ($existing) {
2320 unless ($no_write) {
2321 print STDERR "Using existing ",
2322 "[svn-remote \"$existing\"]\n";
2323 }
2324 $self->{repo_id} = $existing;
2325 } elsif ($_minimize_url) {
2326 my $min_url = Git::SVN::Ra->new($url)->minimize_url;
2327 $existing = find_existing_remote($min_url, $r);
2328 if ($existing) {
2329 unless ($no_write) {
2330 print STDERR "Using existing ",
2331 "[svn-remote \"$existing\"]\n";
2332 }
2333 $self->{repo_id} = $existing;
2334 }
2335 if ($min_url ne $url) {
2336 unless ($no_write) {
2337 print STDERR "Using higher level of URL: ",
2338 "$url => $min_url\n";
2339 }
2340 my $old_path = $self->{path};
2341 $self->{path} = $url;
2342 $self->{path} =~ s!^\Q$min_url\E(/|$)!!;
2343 if (length $old_path) {
2344 $self->{path} .= "/$old_path";
2345 }
2346 $url = $min_url;
2347 }
2348 }
2349 my $orig_url;
2350 if (!$existing) {
2351 # verify that we aren't overwriting anything:
2352 $orig_url = eval {
2353 command_oneline('config', '--get',
2354 "svn-remote.$self->{repo_id}.url")
2355 };
2356 if ($orig_url && ($orig_url ne $url)) {
2357 die "svn-remote.$self->{repo_id}.url already set: ",
2358 "$orig_url\nwanted to set to: $url\n";
2359 }
2360 }
2361 my ($xrepo_id, $xpath) = find_ref($self->refname);
2362 if (!$no_write && defined $xpath) {
2363 die "svn-remote.$xrepo_id.fetch already set to track ",
2364 "$xpath:", $self->refname, "\n";
2365 }
2366 unless ($no_write) {
2367 command_noisy('config',
2368 "svn-remote.$self->{repo_id}.url", $url);
2369 $self->{path} =~ s{^/}{};
2370 $self->{path} =~ s{%([0-9A-F]{2})}{chr hex($1)}ieg;
2371 command_noisy('config', '--add',
2372 "svn-remote.$self->{repo_id}.fetch",
2373 "$self->{path}:".$self->refname);
2374 }
2375 $self->{url} = $url;
2376 }
2377
2378 sub find_by_url { # repos_root and, path are optional
2379 my ($class, $full_url, $repos_root, $path) = @_;
2380
2381 return undef unless defined $full_url;
2382 remove_username($full_url);
2383 remove_username($repos_root) if defined $repos_root;
2384 my $remotes = read_all_remotes();
2385 if (defined $full_url && defined $repos_root && !defined $path) {
2386 $path = $full_url;
2387 $path =~ s#^\Q$repos_root\E(?:/|$)##;
2388 }
2389 foreach my $repo_id (keys %$remotes) {
2390 my $u = $remotes->{$repo_id}->{url} or next;
2391 remove_username($u);
2392 next if defined $repos_root && $repos_root ne $u;
2393
2394 my $fetch = $remotes->{$repo_id}->{fetch} || {};
2395 foreach my $t (qw/branches tags/) {
2396 foreach my $globspec (@{$remotes->{$repo_id}->{$t}}) {
2397 resolve_local_globs($u, $fetch, $globspec);
2398 }
2399 }
2400 my $p = $path;
2401 my $rwr = rewrite_root({repo_id => $repo_id});
2402 my $svm = $remotes->{$repo_id}->{svm}
2403 if defined $remotes->{$repo_id}->{svm};
2404 unless (defined $p) {
2405 $p = $full_url;
2406 my $z = $u;
2407 my $prefix = '';
2408 if ($rwr) {
2409 $z = $rwr;
2410 remove_username($z);
2411 } elsif (defined $svm) {
2412 $z = $svm->{source};
2413 $prefix = $svm->{replace};
2414 $prefix =~ s#^\Q$u\E(?:/|$)##;
2415 $prefix =~ s#/$##;
2416 }
2417 $p =~ s#^\Q$z\E(?:/|$)#$prefix# or next;
2418 }
2419 foreach my $f (keys %$fetch) {
2420 next if $f ne $p;
2421 return Git::SVN->new($fetch->{$f}, $repo_id, $f);
2422 }
2423 }
2424 undef;
2425 }
2426
2427 sub init {
2428 my ($class, $url, $path, $repo_id, $ref_id, $no_write) = @_;
2429 my $self = _new($class, $repo_id, $ref_id, $path);
2430 if (defined $url) {
2431 $self->init_remote_config($url, $no_write);
2432 }
2433 $self;
2434 }
2435
2436 sub find_ref {
2437 my ($ref_id) = @_;
2438 foreach (command(qw/config -l/)) {
2439 next unless m!^svn-remote\.(.+)\.fetch=
2440 \s*(.*?)\s*:\s*(.+?)\s*$!x;
2441 my ($repo_id, $path, $ref) = ($1, $2, $3);
2442 if ($ref eq $ref_id) {
2443 $path = '' if ($path =~ m#^\./?#);
2444 return ($repo_id, $path);
2445 }
2446 }
2447 (undef, undef, undef);
2448 }
2449
2450 sub new {
2451 my ($class, $ref_id, $repo_id, $path) = @_;
2452 if (defined $ref_id && !defined $repo_id && !defined $path) {
2453 ($repo_id, $path) = find_ref($ref_id);
2454 if (!defined $repo_id) {
2455 die "Could not find a \"svn-remote.*.fetch\" key ",
2456 "in the repository configuration matching: ",
2457 "$ref_id\n";
2458 }
2459 }
2460 my $self = _new($class, $repo_id, $ref_id, $path);
2461 if (!defined $self->{path} || !length $self->{path}) {
2462 my $fetch = command_oneline('config', '--get',
2463 "svn-remote.$repo_id.fetch",
2464 ":$ref_id\$") or
2465 die "Failed to read \"svn-remote.$repo_id.fetch\" ",
2466 "\":$ref_id\$\" in config\n";
2467 ($self->{path}, undef) = split(/\s*:\s*/, $fetch);
2468 }
2469 $self->{path} =~ s{/+}{/}g;
2470 $self->{path} =~ s{\A/}{};
2471 $self->{path} =~ s{/\z}{};
2472 $self->{url} = command_oneline('config', '--get',
2473 "svn-remote.$repo_id.url") or
2474 die "Failed to read \"svn-remote.$repo_id.url\" in config\n";
2475 $self->{pushurl} = eval { command_oneline('config', '--get',
2476 "svn-remote.$repo_id.pushurl") };
2477 $self->rebuild;
2478 $self;
2479 }
2480
2481 sub refname {
2482 my ($refname) = $_[0]->{ref_id} ;
2483
2484 # It cannot end with a slash /, we'll throw up on this because
2485 # SVN can't have directories with a slash in their name, either:
2486 if ($refname =~ m{/$}) {
2487 die "ref: '$refname' ends with a trailing slash, this is ",
2488 "not permitted by git nor Subversion\n";
2489 }
2490
2491 # It cannot have ASCII control character space, tilde ~, caret ^,
2492 # colon :, question-mark ?, asterisk *, space, or open bracket [
2493 # anywhere.
2494 #
2495 # Additionally, % must be escaped because it is used for escaping
2496 # and we want our escaped refname to be reversible
2497 $refname =~ s{([ \%~\^:\?\*\[\t])}{uc sprintf('%%%02x',ord($1))}eg;
2498
2499 # no slash-separated component can begin with a dot .
2500 # /.* becomes /%2E*
2501 $refname =~ s{/\.}{/%2E}g;
2502
2503 # It cannot have two consecutive dots .. anywhere
2504 # .. becomes %2E%2E
2505 $refname =~ s{\.\.}{%2E%2E}g;
2506
2507 # trailing dots and .lock are not allowed
2508 # .$ becomes %2E and .lock becomes %2Elock
2509 $refname =~ s{\.(?=$|lock$)}{%2E};
2510
2511 # the sequence @{ is used to access the reflog
2512 # @{ becomes %40{
2513 $refname =~ s{\@\{}{%40\{}g;
2514
2515 return $refname;
2516 }
2517
2518 sub desanitize_refname {
2519 my ($refname) = @_;
2520 $refname =~ s{%(?:([0-9A-F]{2}))}{chr hex($1)}eg;
2521 return $refname;
2522 }
2523
2524 sub svm_uuid {
2525 my ($self) = @_;
2526 return $self->{svm}->{uuid} if $self->svm;
2527 $self->ra;
2528 unless ($self->{svm}) {
2529 die "SVM UUID not cached, and reading remotely failed\n";
2530 }
2531 $self->{svm}->{uuid};
2532 }
2533
2534 sub svm {
2535 my ($self) = @_;
2536 return $self->{svm} if $self->{svm};
2537 my $svm;
2538 # see if we have it in our config, first:
2539 eval {
2540 my $section = "svn-remote.$self->{repo_id}";
2541 $svm = {
2542 source => tmp_config('--get', "$section.svm-source"),
2543 uuid => tmp_config('--get', "$section.svm-uuid"),
2544 replace => tmp_config('--get', "$section.svm-replace"),
2545 }
2546 };
2547 if ($svm && $svm->{source} && $svm->{uuid} && $svm->{replace}) {
2548 $self->{svm} = $svm;
2549 }
2550 $self->{svm};
2551 }
2552
2553 sub _set_svm_vars {
2554 my ($self, $ra) = @_;
2555 return $ra if $self->svm;
2556
2557 my @err = ( "useSvmProps set, but failed to read SVM properties\n",
2558 "(svm:source, svm:uuid) ",
2559 "from the following URLs:\n" );
2560 sub read_svm_props {
2561 my ($self, $ra, $path, $r) = @_;
2562 my $props = ($ra->get_dir($path, $r))[2];
2563 my $src = $props->{'svm:source'};
2564 my $uuid = $props->{'svm:uuid'};
2565 return undef if (!$src || !$uuid);
2566
2567 chomp($src, $uuid);
2568
2569 $uuid =~ m{^[0-9a-f\-]{30,}$}i
2570 or die "doesn't look right - svm:uuid is '$uuid'\n";
2571
2572 # the '!' is used to mark the repos_root!/relative/path
2573 $src =~ s{/?!/?}{/};
2574 $src =~ s{/+$}{}; # no trailing slashes please
2575 # username is of no interest
2576 $src =~ s{(^[a-z\+]*://)[^/@]*@}{$1};
2577
2578 my $replace = $ra->{url};
2579 $replace .= "/$path" if length $path;
2580
2581 my $section = "svn-remote.$self->{repo_id}";
2582 tmp_config("$section.svm-source", $src);
2583 tmp_config("$section.svm-replace", $replace);
2584 tmp_config("$section.svm-uuid", $uuid);
2585 $self->{svm} = {
2586 source => $src,
2587 uuid => $uuid,
2588 replace => $replace
2589 };
2590 }
2591
2592 my $r = $ra->get_latest_revnum;
2593 my $path = $self->{path};
2594 my %tried;
2595 while (length $path) {
2596 unless ($tried{"$self->{url}/$path"}) {
2597 return $ra if $self->read_svm_props($ra, $path, $r);
2598 $tried{"$self->{url}/$path"} = 1;
2599 }
2600 $path =~ s#/?[^/]+$##;
2601 }
2602 die "Path: '$path' should be ''\n" if $path ne '';
2603 return $ra if $self->read_svm_props($ra, $path, $r);
2604 $tried{"$self->{url}/$path"} = 1;
2605
2606 if ($ra->{repos_root} eq $self->{url}) {
2607 die @err, (map { " $_\n" } keys %tried), "\n";
2608 }
2609
2610 # nope, make sure we're connected to the repository root:
2611 my $ok;
2612 my @tried_b;
2613 $path = $ra->{svn_path};
2614 $ra = Git::SVN::Ra->new($ra->{repos_root});
2615 while (length $path) {
2616 unless ($tried{"$ra->{url}/$path"}) {
2617 $ok = $self->read_svm_props($ra, $path, $r);
2618 last if $ok;
2619 $tried{"$ra->{url}/$path"} = 1;
2620 }
2621 $path =~ s#/?[^/]+$##;
2622 }
2623 die "Path: '$path' should be ''\n" if $path ne '';
2624 $ok ||= $self->read_svm_props($ra, $path, $r);
2625 $tried{"$ra->{url}/$path"} = 1;
2626 if (!$ok) {
2627 die @err, (map { " $_\n" } keys %tried), "\n";
2628 }
2629 Git::SVN::Ra->new($self->{url});
2630 }
2631
2632 sub svnsync {
2633 my ($self) = @_;
2634 return $self->{svnsync} if $self->{svnsync};
2635
2636 if ($self->no_metadata) {
2637 die "Can't have both 'noMetadata' and ",
2638 "'useSvnsyncProps' options set!\n";
2639 }
2640 if ($self->rewrite_root) {
2641 die "Can't have both 'useSvnsyncProps' and 'rewriteRoot' ",
2642 "options set!\n";
2643 }
2644 if ($self->rewrite_uuid) {
2645 die "Can't have both 'useSvnsyncProps' and 'rewriteUUID' ",
2646 "options set!\n";
2647 }
2648
2649 my $svnsync;
2650 # see if we have it in our config, first:
2651 eval {
2652 my $section = "svn-remote.$self->{repo_id}";
2653
2654 my $url = tmp_config('--get', "$section.svnsync-url");
2655 ($url) = ($url =~ m{^([a-z\+]+://\S+)$}) or
2656 die "doesn't look right - svn:sync-from-url is '$url'\n";
2657
2658 my $uuid = tmp_config('--get', "$section.svnsync-uuid");
2659 ($uuid) = ($uuid =~ m{^([0-9a-f\-]{30,})$}i) or
2660 die "doesn't look right - svn:sync-from-uuid is '$uuid'\n";
2661
2662 $svnsync = { url => $url, uuid => $uuid }
2663 };
2664 if ($svnsync && $svnsync->{url} && $svnsync->{uuid}) {
2665 return $self->{svnsync} = $svnsync;
2666 }
2667
2668 my $err = "useSvnsyncProps set, but failed to read " .
2669 "svnsync property: svn:sync-from-";
2670 my $rp = $self->ra->rev_proplist(0);
2671
2672 my $url = $rp->{'svn:sync-from-url'} or die $err . "url\n";
2673 ($url) = ($url =~ m{^([a-z\+]+://\S+)$}) or
2674 die "doesn't look right - svn:sync-from-url is '$url'\n";
2675
2676 my $uuid = $rp->{'svn:sync-from-uuid'} or die $err . "uuid\n";
2677 ($uuid) = ($uuid =~ m{^([0-9a-f\-]{30,})$}i) or
2678 die "doesn't look right - svn:sync-from-uuid is '$uuid'\n";
2679
2680 my $section = "svn-remote.$self->{repo_id}";
2681 tmp_config('--add', "$section.svnsync-uuid", $uuid);
2682 tmp_config('--add', "$section.svnsync-url", $url);
2683 return $self->{svnsync} = { url => $url, uuid => $uuid };
2684 }
2685
2686 # this allows us to memoize our SVN::Ra UUID locally and avoid a
2687 # remote lookup (useful for 'git svn log').
2688 sub ra_uuid {
2689 my ($self) = @_;
2690 unless ($self->{ra_uuid}) {
2691 my $key = "svn-remote.$self->{repo_id}.uuid";
2692 my $uuid = eval { tmp_config('--get', $key) };
2693 if (!$@ && $uuid && $uuid =~ /^([a-f\d\-]{30,})$/i) {
2694 $self->{ra_uuid} = $uuid;
2695 } else {
2696 die "ra_uuid called without URL\n" unless $self->{url};
2697 $self->{ra_uuid} = $self->ra->get_uuid;
2698 tmp_config('--add', $key, $self->{ra_uuid});
2699 }
2700 }
2701 $self->{ra_uuid};
2702 }
2703
2704 sub _set_repos_root {
2705 my ($self, $repos_root) = @_;
2706 my $k = "svn-remote.$self->{repo_id}.reposRoot";
2707 $repos_root ||= $self->ra->{repos_root};
2708 tmp_config($k, $repos_root);
2709 $repos_root;
2710 }
2711
2712 sub repos_root {
2713 my ($self) = @_;
2714 my $k = "svn-remote.$self->{repo_id}.reposRoot";
2715 eval { tmp_config('--get', $k) } || $self->_set_repos_root;
2716 }
2717
2718 sub ra {
2719 my ($self) = shift;
2720 my $ra = Git::SVN::Ra->new($self->{url});
2721 $self->_set_repos_root($ra->{repos_root});
2722 if ($self->use_svm_props && !$self->{svm}) {
2723 if ($self->no_metadata) {
2724 die "Can't have both 'noMetadata' and ",
2725 "'useSvmProps' options set!\n";
2726 } elsif ($self->use_svnsync_props) {
2727 die "Can't have both 'useSvnsyncProps' and ",
2728 "'useSvmProps' options set!\n";
2729 }
2730 $ra = $self->_set_svm_vars($ra);
2731 $self->{-want_revprops} = 1;
2732 }
2733 $ra;
2734 }
2735
2736 # prop_walk(PATH, REV, SUB)
2737 # -------------------------
2738 # Recursively traverse PATH at revision REV and invoke SUB for each
2739 # directory that contains a SVN property. SUB will be invoked as
2740 # follows: &SUB(gs, path, props); where `gs' is this instance of
2741 # Git::SVN, `path' the path to the directory where the properties
2742 # `props' were found. The `path' will be relative to point of checkout,
2743 # that is, if url://repo/trunk is the current Git branch, and that
2744 # directory contains a sub-directory `d', SUB will be invoked with `/d/'
2745 # as `path' (note the trailing `/').
2746 sub prop_walk {
2747 my ($self, $path, $rev, $sub) = @_;
2748
2749 $path =~ s#^/##;
2750 my ($dirent, undef, $props) = $self->ra->get_dir($path, $rev);
2751 $path =~ s#^/*#/#g;
2752 my $p = $path;
2753 # Strip the irrelevant part of the path.
2754 $p =~ s#^/+\Q$self->{path}\E(/|$)#/#;
2755 # Ensure the path is terminated by a `/'.
2756 $p =~ s#/*$#/#;
2757
2758 # The properties contain all the internal SVN stuff nobody
2759 # (usually) cares about.
2760 my $interesting_props = 0;
2761 foreach (keys %{$props}) {
2762 # If it doesn't start with `svn:', it must be a
2763 # user-defined property.
2764 ++$interesting_props and next if $_ !~ /^svn:/;
2765 # FIXME: Fragile, if SVN adds new public properties,
2766 # this needs to be updated.
2767 ++$interesting_props if /^svn:(?:ignore|keywords|executable
2768 |eol-style|mime-type
2769 |externals|needs-lock)$/x;
2770 }
2771 &$sub($self, $p, $props) if $interesting_props;
2772
2773 foreach (sort keys %$dirent) {
2774 next if $dirent->{$_}->{kind} != $SVN::Node::dir;
2775 $self->prop_walk($self->{path} . $p . $_, $rev, $sub);
2776 }
2777 }
2778
2779 sub last_rev { ($_[0]->last_rev_commit)[0] }
2780 sub last_commit { ($_[0]->last_rev_commit)[1] }
2781
2782 # returns the newest SVN revision number and newest commit SHA1
2783 sub last_rev_commit {
2784 my ($self) = @_;
2785 if (defined $self->{last_rev} && defined $self->{last_commit}) {
2786 return ($self->{last_rev}, $self->{last_commit});
2787 }
2788 my $c = ::verify_ref($self->refname.'^0');
2789 if ($c && !$self->use_svm_props && !$self->no_metadata) {
2790 my $rev = (::cmt_metadata($c))[1];
2791 if (defined $rev) {
2792 ($self->{last_rev}, $self->{last_commit}) = ($rev, $c);
2793 return ($rev, $c);
2794 }
2795 }
2796 my $map_path = $self->map_path;
2797 unless (-e $map_path) {
2798 ($self->{last_rev}, $self->{last_commit}) = (undef, undef);
2799 return (undef, undef);
2800 }
2801 my ($rev, $commit) = $self->rev_map_max(1);
2802 ($self->{last_rev}, $self->{last_commit}) = ($rev, $commit);
2803 return ($rev, $commit);
2804 }
2805
2806 sub get_fetch_range {
2807 my ($self, $min, $max) = @_;
2808 $max ||= $self->ra->get_latest_revnum;
2809 $min ||= $self->rev_map_max;
2810 (++$min, $max);
2811 }
2812
2813 sub tmp_config {
2814 my (@args) = @_;
2815 my $old_def_config = "$ENV{GIT_DIR}/svn/config";
2816 my $config = "$ENV{GIT_DIR}/svn/.metadata";
2817 if (! -f $config && -f $old_def_config) {
2818 rename $old_def_config, $config or
2819 die "Failed rename $old_def_config => $config: $!\n";
2820 }
2821 my $old_config = $ENV{GIT_CONFIG};
2822 $ENV{GIT_CONFIG} = $config;
2823 $@ = undef;
2824 my @ret = eval {
2825 unless (-f $config) {
2826 mkfile($config);
2827 open my $fh, '>', $config or
2828 die "Can't open $config: $!\n";
2829 print $fh "; This file is used internally by ",
2830 "git-svn\n" or die
2831 "Couldn't write to $config: $!\n";
2832 print $fh "; You should not have to edit it\n" or
2833 die "Couldn't write to $config: $!\n";
2834 close $fh or die "Couldn't close $config: $!\n";
2835 }
2836 command('config', @args);
2837 };
2838 my $err = $@;
2839 if (defined $old_config) {
2840 $ENV{GIT_CONFIG} = $old_config;
2841 } else {
2842 delete $ENV{GIT_CONFIG};
2843 }
2844 die $err if $err;
2845 wantarray ? @ret : $ret[0];
2846 }
2847
2848 sub tmp_index_do {
2849 my ($self, $sub) = @_;
2850 my $old_index = $ENV{GIT_INDEX_FILE};
2851 $ENV{GIT_INDEX_FILE} = $self->{index};
2852 $@ = undef;
2853 my @ret = eval {
2854 my ($dir, $base) = ($self->{index} =~ m#^(.*?)/?([^/]+)$#);
2855 mkpath([$dir]) unless -d $dir;
2856 &$sub;
2857 };
2858 my $err = $@;
2859 if (defined $old_index) {
2860 $ENV{GIT_INDEX_FILE} = $old_index;
2861 } else {
2862 delete $ENV{GIT_INDEX_FILE};
2863 }
2864 die $err if $err;
2865 wantarray ? @ret : $ret[0];
2866 }
2867
2868 sub assert_index_clean {
2869 my ($self, $treeish) = @_;
2870
2871 $self->tmp_index_do(sub {
2872 command_noisy('read-tree', $treeish) unless -e $self->{index};
2873 my $x = command_oneline('write-tree');
2874 my ($y) = (command(qw/cat-file commit/, $treeish) =~
2875 /^tree ($::sha1)/mo);
2876 return if $y eq $x;
2877
2878 warn "Index mismatch: $y != $x\nrereading $treeish\n";
2879 unlink $self->{index} or die "unlink $self->{index}: $!\n";
2880 command_noisy('read-tree', $treeish);
2881 $x = command_oneline('write-tree');
2882 if ($y ne $x) {
2883 ::fatal "trees ($treeish) $y != $x\n",
2884 "Something is seriously wrong...";
2885 }
2886 });
2887 }
2888
2889 sub get_commit_parents {
2890 my ($self, $log_entry) = @_;
2891 my (%seen, @ret, @tmp);
2892 # legacy support for 'set-tree'; this is only used by set_tree_cb:
2893 if (my $ip = $self->{inject_parents}) {
2894 if (my $commit = delete $ip->{$log_entry->{revision}}) {
2895 push @tmp, $commit;
2896 }
2897 }
2898 if (my $cur = ::verify_ref($self->refname.'^0')) {
2899 push @tmp, $cur;
2900 }
2901 if (my $ipd = $self->{inject_parents_dcommit}) {
2902 if (my $commit = delete $ipd->{$log_entry->{revision}}) {
2903 push @tmp, @$commit;
2904 }
2905 }
2906 push @tmp, $_ foreach (@{$log_entry->{parents}}, @tmp);
2907 while (my $p = shift @tmp) {
2908 next if $seen{$p};
2909 $seen{$p} = 1;
2910 push @ret, $p;
2911 }
2912 @ret;
2913 }
2914
2915 sub rewrite_root {
2916 my ($self) = @_;
2917 return $self->{-rewrite_root} if exists $self->{-rewrite_root};
2918 my $k = "svn-remote.$self->{repo_id}.rewriteRoot";
2919 my $rwr = eval { command_oneline(qw/config --get/, $k) };
2920 if ($rwr) {
2921 $rwr =~ s#/+$##;
2922 if ($rwr !~ m#^[a-z\+]+://#) {
2923 die "$rwr is not a valid URL (key: $k)\n";
2924 }
2925 }
2926 $self->{-rewrite_root} = $rwr;
2927 }
2928
2929 sub rewrite_uuid {
2930 my ($self) = @_;
2931 return $self->{-rewrite_uuid} if exists $self->{-rewrite_uuid};
2932 my $k = "svn-remote.$self->{repo_id}.rewriteUUID";
2933 my $rwid = eval { command_oneline(qw/config --get/, $k) };
2934 if ($rwid) {
2935 $rwid =~ s#/+$##;
2936 if ($rwid !~ m#^[a-f0-9]{8}-(?:[a-f0-9]{4}-){3}[a-f0-9]{12}$#) {
2937 die "$rwid is not a valid UUID (key: $k)\n";
2938 }
2939 }
2940 $self->{-rewrite_uuid} = $rwid;
2941 }
2942
2943 sub metadata_url {
2944 my ($self) = @_;
2945 ($self->rewrite_root || $self->{url}) .
2946 (length $self->{path} ? '/' . $self->{path} : '');
2947 }
2948
2949 sub full_url {
2950 my ($self) = @_;
2951 $self->{url} . (length $self->{path} ? '/' . $self->{path} : '');
2952 }
2953
2954 sub full_pushurl {
2955 my ($self) = @_;
2956 if ($self->{pushurl}) {
2957 return $self->{pushurl} . (length $self->{path} ? '/' .
2958 $self->{path} : '');
2959 } else {
2960 return $self->full_url;
2961 }
2962 }
2963
2964 sub set_commit_header_env {
2965 my ($log_entry) = @_;
2966 my %env;
2967 foreach my $ned (qw/NAME EMAIL DATE/) {
2968 foreach my $ac (qw/AUTHOR COMMITTER/) {
2969 $env{"GIT_${ac}_${ned}"} = $ENV{"GIT_${ac}_${ned}"};
2970 }
2971 }
2972
2973 $ENV{GIT_AUTHOR_NAME} = $log_entry->{name};
2974 $ENV{GIT_AUTHOR_EMAIL} = $log_entry->{email};
2975 $ENV{GIT_AUTHOR_DATE} = $ENV{GIT_COMMITTER_DATE} = $log_entry->{date};
2976
2977 $ENV{GIT_COMMITTER_NAME} = (defined $log_entry->{commit_name})
2978 ? $log_entry->{commit_name}
2979 : $log_entry->{name};
2980 $ENV{GIT_COMMITTER_EMAIL} = (defined $log_entry->{commit_email})
2981 ? $log_entry->{commit_email}
2982 : $log_entry->{email};
2983 \%env;
2984 }
2985
2986 sub restore_commit_header_env {
2987 my ($env) = @_;
2988 foreach my $ned (qw/NAME EMAIL DATE/) {
2989 foreach my $ac (qw/AUTHOR COMMITTER/) {
2990 my $k = "GIT_${ac}_${ned}";
2991 if (defined $env->{$k}) {
2992 $ENV{$k} = $env->{$k};
2993 } else {
2994 delete $ENV{$k};
2995 }
2996 }
2997 }
2998 }
2999
3000 sub gc {
3001 command_noisy('gc', '--auto');
3002 };
3003
3004 sub do_git_commit {
3005 my ($self, $log_entry) = @_;
3006 my $lr = $self->last_rev;
3007 if (defined $lr && $lr >= $log_entry->{revision}) {
3008 die "Last fetched revision of ", $self->refname,
3009 " was r$lr, but we are about to fetch: ",
3010 "r$log_entry->{revision}!\n";
3011 }
3012 if (my $c = $self->rev_map_get($log_entry->{revision})) {
3013 croak "$log_entry->{revision} = $c already exists! ",
3014 "Why are we refetching it?\n";
3015 }
3016 my $old_env = set_commit_header_env($log_entry);
3017 my $tree = $log_entry->{tree};
3018 if (!defined $tree) {
3019 $tree = $self->tmp_index_do(sub {
3020 command_oneline('write-tree') });
3021 }
3022 die "Tree is not a valid sha1: $tree\n" if $tree !~ /^$::sha1$/o;
3023
3024 my @exec = ('git', 'commit-tree', $tree);
3025 foreach ($self->get_commit_parents($log_entry)) {
3026 push @exec, '-p', $_;
3027 }
3028 defined(my $pid = open3(my $msg_fh, my $out_fh, '>&STDERR', @exec))
3029 or croak $!;
3030 binmode $msg_fh;
3031
3032 # we always get UTF-8 from SVN, but we may want our commits in
3033 # a different encoding.
3034 if (my $enc = Git::config('i18n.commitencoding')) {
3035 require Encode;
3036 Encode::from_to($log_entry->{log}, 'UTF-8', $enc);
3037 }
3038 print $msg_fh $log_entry->{log} or croak $!;
3039 restore_commit_header_env($old_env);
3040 unless ($self->no_metadata) {
3041 print $msg_fh "\ngit-svn-id: $log_entry->{metadata}\n"
3042 or croak $!;
3043 }
3044 $msg_fh->flush == 0 or croak $!;
3045 close $msg_fh or croak $!;
3046 chomp(my $commit = do { local $/; <$out_fh> });
3047 close $out_fh or croak $!;
3048 waitpid $pid, 0;
3049 croak $? if $?;
3050 if ($commit !~ /^$::sha1$/o) {
3051 die "Failed to commit, invalid sha1: $commit\n";
3052 }
3053
3054 $self->rev_map_set($log_entry->{revision}, $commit, 1);
3055
3056 $self->{last_rev} = $log_entry->{revision};
3057 $self->{last_commit} = $commit;
3058 print "r$log_entry->{revision}" unless $::_q > 1;
3059 if (defined $log_entry->{svm_revision}) {
3060 print " (\@$log_entry->{svm_revision})" unless $::_q > 1;
3061 $self->rev_map_set($log_entry->{svm_revision}, $commit,
3062 0, $self->svm_uuid);
3063 }
3064 print " = $commit ($self->{ref_id})\n" unless $::_q > 1;
3065 if (--$_gc_nr == 0) {
3066 $_gc_nr = $_gc_period;
3067 gc();
3068 }
3069 return $commit;
3070 }
3071
3072 sub match_paths {
3073 my ($self, $paths, $r) = @_;
3074 return 1 if $self->{path} eq '';
3075 if (my $path = $paths->{"/$self->{path}"}) {
3076 return ($path->{action} eq 'D') ? 0 : 1;
3077 }
3078 $self->{path_regex} ||= qr/^\/\Q$self->{path}\E\//;
3079 if (grep /$self->{path_regex}/, keys %$paths) {
3080 return 1;
3081 }
3082 my $c = '';
3083 foreach (split m#/#, $self->{path}) {
3084 $c .= "/$_";
3085 next unless ($paths->{$c} &&
3086 ($paths->{$c}->{action} =~ /^[AR]$/));
3087 if ($self->ra->check_path($self->{path}, $r) ==
3088 $SVN::Node::dir) {
3089 return 1;
3090 }
3091 }
3092 return 0;
3093 }
3094
3095 sub find_parent_branch {
3096 my ($self, $paths, $rev) = @_;
3097 return undef unless $self->follow_parent;
3098 unless (defined $paths) {
3099 my $err_handler = $SVN::Error::handler;
3100 $SVN::Error::handler = \&Git::SVN::Ra::skip_unknown_revs;
3101 $self->ra->get_log([$self->{path}], $rev, $rev, 0, 1, 1,
3102 sub { $paths = $_[0] });
3103 $SVN::Error::handler = $err_handler;
3104 }
3105 return undef unless defined $paths;
3106
3107 # look for a parent from another branch:
3108 my @b_path_components = split m#/#, $self->{path};
3109 my @a_path_components;
3110 my $i;
3111 while (@b_path_components) {
3112 $i = $paths->{'/'.join('/', @b_path_components)};
3113 last if $i && defined $i->{copyfrom_path};
3114 unshift(@a_path_components, pop(@b_path_components));
3115 }
3116 return undef unless defined $i && defined $i->{copyfrom_path};
3117 my $branch_from = $i->{copyfrom_path};
3118 if (@a_path_components) {
3119 print STDERR "branch_from: $branch_from => ";
3120 $branch_from .= '/'.join('/', @a_path_components);
3121 print STDERR $branch_from, "\n";
3122 }
3123 my $r = $i->{copyfrom_rev};
3124 my $repos_root = $self->ra->{repos_root};
3125 my $url = $self->ra->{url};
3126 my $new_url = $url . $branch_from;
3127 print STDERR "Found possible branch point: ",
3128 "$new_url => ", $self->full_url, ", $r\n"
3129 unless $::_q > 1;
3130 $branch_from =~ s#^/##;
3131 my $gs = $self->other_gs($new_url, $url,
3132 $branch_from, $r, $self->{ref_id});
3133 my ($r0, $parent) = $gs->find_rev_before($r, 1);
3134 {
3135 my ($base, $head);
3136 if (!defined $r0 || !defined $parent) {
3137 ($base, $head) = parse_revision_argument(0, $r);
3138 } else {
3139 if ($r0 < $r) {
3140 $gs->ra->get_log([$gs->{path}], $r0 + 1, $r, 1,
3141 0, 1, sub { $base = $_[1] - 1 });
3142 }
3143 }
3144 if (defined $base && $base <= $r) {
3145 $gs->fetch($base, $r);
3146 }
3147 ($r0, $parent) = $gs->find_rev_before($r, 1);
3148 }
3149 if (defined $r0 && defined $parent) {
3150 print STDERR "Found branch parent: ($self->{ref_id}) $parent\n"
3151 unless $::_q > 1;
3152 my $ed;
3153 if ($self->ra->can_do_switch) {
3154 $self->assert_index_clean($parent);
3155 print STDERR "Following parent with do_switch\n"
3156 unless $::_q > 1;
3157 # do_switch works with svn/trunk >= r22312, but that
3158 # is not included with SVN 1.4.3 (the latest version
3159 # at the moment), so we can't rely on it
3160 $self->{last_rev} = $r0;
3161 $self->{last_commit} = $parent;
3162 $ed = Git::SVN::Fetcher->new($self, $gs->{path});
3163 $gs->ra->gs_do_switch($r0, $rev, $gs,
3164 $self->full_url, $ed)
3165 or die "SVN connection failed somewhere...\n";
3166 } elsif ($self->ra->trees_match($new_url, $r0,
3167 $self->full_url, $rev)) {
3168 print STDERR "Trees match:\n",
3169 " $new_url\@$r0\n",
3170 " ${\$self->full_url}\@$rev\n",
3171 "Following parent with no changes\n"
3172 unless $::_q > 1;
3173 $self->tmp_index_do(sub {
3174 command_noisy('read-tree', $parent);
3175 });
3176 $self->{last_commit} = $parent;
3177 } else {
3178 print STDERR "Following parent with do_update\n"
3179 unless $::_q > 1;
3180 $ed = Git::SVN::Fetcher->new($self);
3181 $self->ra->gs_do_update($rev, $rev, $self, $ed)
3182 or die "SVN connection failed somewhere...\n";
3183 }
3184 print STDERR "Successfully followed parent\n" unless $::_q > 1;
3185 return $self->make_log_entry($rev, [$parent], $ed);
3186 }
3187 return undef;
3188 }
3189
3190 sub do_fetch {
3191 my ($self, $paths, $rev) = @_;
3192 my $ed;
3193 my ($last_rev, @parents);
3194 if (my $lc = $self->last_commit) {
3195 # we can have a branch that was deleted, then re-added
3196 # under the same name but copied from another path, in
3197 # which case we'll have multiple parents (we don't
3198 # want to break the original ref, nor lose copypath info):
3199 if (my $log_entry = $self->find_parent_branch($paths, $rev)) {
3200 push @{$log_entry->{parents}}, $lc;
3201 return $log_entry;
3202 }
3203 $ed = Git::SVN::Fetcher->new($self);
3204 $last_rev = $self->{last_rev};
3205 $ed->{c} = $lc;
3206 @parents = ($lc);
3207 } else {
3208 $last_rev = $rev;
3209 if (my $log_entry = $self->find_parent_branch($paths, $rev)) {
3210 return $log_entry;
3211 }
3212 $ed = Git::SVN::Fetcher->new($self);
3213 }
3214 unless ($self->ra->gs_do_update($last_rev, $rev, $self, $ed)) {
3215 die "SVN connection failed somewhere...\n";
3216 }
3217 $self->make_log_entry($rev, \@parents, $ed);
3218 }
3219
3220 sub mkemptydirs {
3221 my ($self, $r) = @_;
3222
3223 sub scan {
3224 my ($r, $empty_dirs, $line) = @_;
3225 if (defined $r && $line =~ /^r(\d+)$/) {
3226 return 0 if $1 > $r;
3227 } elsif ($line =~ /^ \+empty_dir: (.+)$/) {
3228 $empty_dirs->{$1} = 1;
3229 } elsif ($line =~ /^ \-empty_dir: (.+)$/) {
3230 my @d = grep {m[^\Q$1\E(/|$)]} (keys %$empty_dirs);
3231 delete @$empty_dirs{@d};
3232 }
3233 1; # continue
3234 };
3235
3236 my %empty_dirs = ();
3237 my $gz_file = "$self->{dir}/unhandled.log.gz";
3238 if (-f $gz_file) {
3239 if (!$can_compress) {
3240 warn "Compress::Zlib could not be found; ",
3241 "empty directories in $gz_file will not be read\n";
3242 } else {
3243 my $gz = Compress::Zlib::gzopen($gz_file, "rb") or
3244 die "Unable to open $gz_file: $!\n";
3245 my $line;
3246 while ($gz->gzreadline($line) > 0) {
3247 scan($r, \%empty_dirs, $line) or last;
3248 }
3249 $gz->gzclose;
3250 }
3251 }
3252
3253 if (open my $fh, '<', "$self->{dir}/unhandled.log") {
3254 binmode $fh or croak "binmode: $!";
3255 while (<$fh>) {
3256 scan($r, \%empty_dirs, $_) or last;
3257 }
3258 close $fh;
3259 }
3260
3261 my $strip = qr/\A\Q$self->{path}\E(?:\/|$)/;
3262 foreach my $d (sort keys %empty_dirs) {
3263 $d = uri_decode($d);
3264 $d =~ s/$strip//;
3265 next unless length($d);
3266 next if -d $d;
3267 if (-e $d) {
3268 warn "$d exists but is not a directory\n";
3269 } else {
3270 print "creating empty directory: $d\n";
3271 mkpath([$d]);
3272 }
3273 }
3274 }
3275
3276 sub get_untracked {
3277 my ($self, $ed) = @_;
3278 my @out;
3279 my $h = $ed->{empty};
3280 foreach (sort keys %$h) {
3281 my $act = $h->{$_} ? '+empty_dir' : '-empty_dir';
3282 push @out, " $act: " . uri_encode($_);
3283 warn "W: $act: $_\n";
3284 }
3285 foreach my $t (qw/dir_prop file_prop/) {
3286 $h = $ed->{$t} or next;
3287 foreach my $path (sort keys %$h) {
3288 my $ppath = $path eq '' ? '.' : $path;
3289 foreach my $prop (sort keys %{$h->{$path}}) {
3290 next if $SKIP_PROP{$prop};
3291 my $v = $h->{$path}->{$prop};
3292 my $t_ppath_prop = "$t: " .
3293 uri_encode($ppath) . ' ' .
3294 uri_encode($prop);
3295 if (defined $v) {
3296 push @out, " +$t_ppath_prop " .
3297 uri_encode($v);
3298 } else {
3299 push @out, " -$t_ppath_prop";
3300 }
3301 }
3302 }
3303 }
3304 foreach my $t (qw/absent_file absent_directory/) {
3305 $h = $ed->{$t} or next;
3306 foreach my $parent (sort keys %$h) {
3307 foreach my $path (sort @{$h->{$parent}}) {
3308 push @out, " $t: " .
3309 uri_encode("$parent/$path");
3310 warn "W: $t: $parent/$path ",
3311 "Insufficient permissions?\n";
3312 }
3313 }
3314 }
3315 \@out;
3316 }
3317
3318 sub get_tz {
3319 # some systmes don't handle or mishandle %z, so be creative.
3320 my $t = shift || time;
3321 my $gm = timelocal(gmtime($t));
3322 my $sign = qw( + + - )[ $t <=> $gm ];
3323 return sprintf("%s%02d%02d", $sign, (gmtime(abs($t - $gm)))[2,1]);
3324 }
3325
3326 # parse_svn_date(DATE)
3327 # --------------------
3328 # Given a date (in UTC) from Subversion, return a string in the format
3329 # "<TZ Offset> <local date/time>" that Git will use.
3330 #
3331 # By default the parsed date will be in UTC; if $Git::SVN::_localtime
3332 # is true we'll convert it to the local timezone instead.
3333 sub parse_svn_date {
3334 my $date = shift || return '+0000 1970-01-01 00:00:00';
3335 my ($Y,$m,$d,$H,$M,$S) = ($date =~ /^(\d{4})\-(\d\d)\-(\d\d)T
3336 (\d\d)\:(\d\d)\:(\d\d)\.\d*Z$/x) or
3337 croak "Unable to parse date: $date\n";
3338 my $parsed_date; # Set next.
3339
3340 if ($Git::SVN::_localtime) {
3341 # Translate the Subversion datetime to an epoch time.
3342 # Begin by switching ourselves to $date's timezone, UTC.
3343 my $old_env_TZ = $ENV{TZ};
3344 $ENV{TZ} = 'UTC';
3345
3346 my $epoch_in_UTC =
3347 POSIX::strftime('%s', $S, $M, $H, $d, $m - 1, $Y - 1900);
3348
3349 # Determine our local timezone (including DST) at the
3350 # time of $epoch_in_UTC. $Git::SVN::Log::TZ stored the
3351 # value of TZ, if any, at the time we were run.
3352 if (defined $Git::SVN::Log::TZ) {
3353 $ENV{TZ} = $Git::SVN::Log::TZ;
3354 } else {
3355 delete $ENV{TZ};
3356 }
3357
3358 my $our_TZ = get_tz();
3359
3360 # This converts $epoch_in_UTC into our local timezone.
3361 my ($sec, $min, $hour, $mday, $mon, $year,
3362 $wday, $yday, $isdst) = localtime($epoch_in_UTC);
3363
3364 $parsed_date = sprintf('%s %04d-%02d-%02d %02d:%02d:%02d',
3365 $our_TZ, $year + 1900, $mon + 1,
3366 $mday, $hour, $min, $sec);
3367
3368 # Reset us to the timezone in effect when we entered
3369 # this routine.
3370 if (defined $old_env_TZ) {
3371 $ENV{TZ} = $old_env_TZ;
3372 } else {
3373 delete $ENV{TZ};
3374 }
3375 } else {
3376 $parsed_date = "+0000 $Y-$m-$d $H:$M:$S";
3377 }
3378
3379 return $parsed_date;
3380 }
3381
3382 sub other_gs {
3383 my ($self, $new_url, $url,
3384 $branch_from, $r, $old_ref_id) = @_;
3385 my $gs = Git::SVN->find_by_url($new_url, $url, $branch_from);
3386 unless ($gs) {
3387 my $ref_id = $old_ref_id;
3388 $ref_id =~ s/\@\d+-*$//;
3389 $ref_id .= "\@$r";
3390 # just grow a tail if we're not unique enough :x
3391 $ref_id .= '-' while find_ref($ref_id);
3392 my ($u, $p, $repo_id) = ($new_url, '', $ref_id);
3393 if ($u =~ s#^\Q$url\E(/|$)##) {
3394 $p = $u;
3395 $u = $url;
3396 $repo_id = $self->{repo_id};
3397 }
3398 while (1) {
3399 # It is possible to tag two different subdirectories at
3400 # the same revision. If the url for an existing ref
3401 # does not match, we must either find a ref with a
3402 # matching url or create a new ref by growing a tail.
3403 $gs = Git::SVN->init($u, $p, $repo_id, $ref_id, 1);
3404 my (undef, $max_commit) = $gs->rev_map_max(1);
3405 last if (!$max_commit);
3406 my ($url) = ::cmt_metadata($max_commit);
3407 last if ($url eq $gs->metadata_url);
3408 $ref_id .= '-';
3409 }
3410 print STDERR "Initializing parent: $ref_id\n" unless $::_q > 1;
3411 }
3412 $gs
3413 }
3414
3415 sub call_authors_prog {
3416 my ($orig_author) = @_;
3417 $orig_author = command_oneline('rev-parse', '--sq-quote', $orig_author);
3418 my $author = `$::_authors_prog $orig_author`;
3419 if ($? != 0) {
3420 die "$::_authors_prog failed with exit code $?\n"
3421 }
3422 if ($author =~ /^\s*(.+?)\s*<(.*)>\s*$/) {
3423 my ($name, $email) = ($1, $2);
3424 $email = undef if length $2 == 0;
3425 return [$name, $email];
3426 } else {
3427 die "Author: $orig_author: $::_authors_prog returned "
3428 . "invalid author format: $author\n";
3429 }
3430 }
3431
3432 sub check_author {
3433 my ($author) = @_;
3434 if (!defined $author || length $author == 0) {
3435 $author = '(no author)';
3436 }
3437 if (!defined $::users{$author}) {
3438 if (defined $::_authors_prog) {
3439 $::users{$author} = call_authors_prog($author);
3440 } elsif (defined $::_authors) {
3441 die "Author: $author not defined in $::_authors file\n";
3442 }
3443 }
3444 $author;
3445 }
3446
3447 sub find_extra_svk_parents {
3448 my ($self, $ed, $tickets, $parents) = @_;
3449 # aha! svk:merge property changed...
3450 my @tickets = split "\n", $tickets;
3451 my @known_parents;
3452 for my $ticket ( @tickets ) {
3453 my ($uuid, $path, $rev) = split /:/, $ticket;
3454 if ( $uuid eq $self->ra_uuid ) {
3455 my $url = $self->{url};
3456 my $repos_root = $url;
3457 my $branch_from = $path;
3458 $branch_from =~ s{^/}{};
3459 my $gs = $self->other_gs($repos_root."/".$branch_from,
3460 $url,
3461 $branch_from,
3462 $rev,
3463 $self->{ref_id});
3464 if ( my $commit = $gs->rev_map_get($rev, $uuid) ) {
3465 # wahey! we found it, but it might be
3466 # an old one (!)
3467 push @known_parents, [ $rev, $commit ];
3468 }
3469 }
3470 }
3471 # Ordering matters; highest-numbered commit merge tickets
3472 # first, as they may account for later merge ticket additions
3473 # or changes.
3474 @known_parents = map {$_->[1]} sort {$b->[0] <=> $a->[0]} @known_parents;
3475 for my $parent ( @known_parents ) {
3476 my @cmd = ('rev-list', $parent, map { "^$_" } @$parents );
3477 my ($msg_fh, $ctx) = command_output_pipe(@cmd);
3478 my $new;
3479 while ( <$msg_fh> ) {
3480 $new=1;last;
3481 }
3482 command_close_pipe($msg_fh, $ctx);
3483 if ( $new ) {
3484 print STDERR
3485 "Found merge parent (svk:merge ticket): $parent\n";
3486 push @$parents, $parent;
3487 }
3488 }
3489 }
3490
3491 sub lookup_svn_merge {
3492 my $uuid = shift;
3493 my $url = shift;
3494 my $merge = shift;
3495
3496 my ($source, $revs) = split ":", $merge;
3497 my $path = $source;
3498 $path =~ s{^/}{};
3499 my $gs = Git::SVN->find_by_url($url.$source, $url, $path);
3500 if ( !$gs ) {
3501 warn "Couldn't find revmap for $url$source\n";
3502 return;
3503 }
3504 my @ranges = split ",", $revs;
3505 my ($tip, $tip_commit);
3506 my @merged_commit_ranges;
3507 # find the tip
3508 for my $range ( @ranges ) {
3509 my ($bottom, $top) = split "-", $range;
3510 $top ||= $bottom;
3511 my $bottom_commit = $gs->find_rev_after( $bottom, 1, $top );
3512 my $top_commit = $gs->find_rev_before( $top, 1, $bottom );
3513
3514 unless ($top_commit and $bottom_commit) {
3515 warn "W:unknown path/rev in svn:mergeinfo "
3516 ."dirprop: $source:$range\n";
3517 next;
3518 }
3519
3520 if (scalar(command('rev-parse', "$bottom_commit^@"))) {
3521 push @merged_commit_ranges,
3522 "$bottom_commit^..$top_commit";
3523 } else {
3524 push @merged_commit_ranges, "$top_commit";
3525 }
3526
3527 if ( !defined $tip or $top > $tip ) {
3528 $tip = $top;
3529 $tip_commit = $top_commit;
3530 }
3531 }
3532 return ($tip_commit, @merged_commit_ranges);
3533 }
3534
3535 sub _rev_list {
3536 my ($msg_fh, $ctx) = command_output_pipe(
3537 "rev-list", @_,
3538 );
3539 my @rv;
3540 while ( <$msg_fh> ) {
3541 chomp;
3542 push @rv, $_;
3543 }
3544 command_close_pipe($msg_fh, $ctx);
3545 @rv;
3546 }
3547
3548 sub check_cherry_pick {
3549 my $base = shift;
3550 my $tip = shift;
3551 my $parents = shift;
3552 my @ranges = @_;
3553 my %commits = map { $_ => 1 }
3554 _rev_list("--no-merges", $tip, "--not", $base, @$parents, "--");
3555 for my $range ( @ranges ) {
3556 delete @commits{_rev_list($range, "--")};
3557 }
3558 for my $commit (keys %commits) {
3559 if (has_no_changes($commit)) {
3560 delete $commits{$commit};
3561 }
3562 }
3563 return (keys %commits);
3564 }
3565
3566 sub has_no_changes {
3567 my $commit = shift;
3568
3569 my @revs = split / /, command_oneline(
3570 qw(rev-list --parents -1 -m), $commit);
3571
3572 # Commits with no parents, e.g. the start of a partial branch,
3573 # have changes by definition.
3574 return 1 if (@revs < 2);
3575
3576 # Commits with multiple parents, e.g a merge, have no changes
3577 # by definition.
3578 return 0 if (@revs > 2);
3579
3580 return (command_oneline("rev-parse", "$commit^{tree}") eq
3581 command_oneline("rev-parse", "$commit~1^{tree}"));
3582 }
3583
3584 sub tie_for_persistent_memoization {
3585 my $hash = shift;
3586 my $path = shift;
3587
3588 if ($can_use_yaml) {
3589 tie %$hash => 'Git::SVN::Memoize::YAML', "$path.yaml";
3590 } else {
3591 tie %$hash => 'Memoize::Storable', "$path.db", 'nstore';
3592 }
3593 }
3594
3595 # The GIT_DIR environment variable is not always set until after the command
3596 # line arguments are processed, so we can't memoize in a BEGIN block.
3597 {
3598 my $memoized = 0;
3599
3600 sub memoize_svn_mergeinfo_functions {
3601 return if $memoized;
3602 $memoized = 1;
3603
3604 my $cache_path = "$ENV{GIT_DIR}/svn/.caches/";
3605 mkpath([$cache_path]) unless -d $cache_path;
3606
3607 my %lookup_svn_merge_cache;
3608 my %check_cherry_pick_cache;
3609 my %has_no_changes_cache;
3610
3611 tie_for_persistent_memoization(\%lookup_svn_merge_cache,
3612 "$cache_path/lookup_svn_merge");
3613 memoize 'lookup_svn_merge',
3614 SCALAR_CACHE => 'FAULT',
3615 LIST_CACHE => ['HASH' => \%lookup_svn_merge_cache],
3616 ;
3617
3618 tie_for_persistent_memoization(\%check_cherry_pick_cache,
3619 "$cache_path/check_cherry_pick");
3620 memoize 'check_cherry_pick',
3621 SCALAR_CACHE => 'FAULT',
3622 LIST_CACHE => ['HASH' => \%check_cherry_pick_cache],
3623 ;
3624
3625 tie_for_persistent_memoization(\%has_no_changes_cache,
3626 "$cache_path/has_no_changes");
3627 memoize 'has_no_changes',
3628 SCALAR_CACHE => ['HASH' => \%has_no_changes_cache],
3629 LIST_CACHE => 'FAULT',
3630 ;
3631 }
3632
3633 sub unmemoize_svn_mergeinfo_functions {
3634 return if not $memoized;
3635 $memoized = 0;
3636
3637 Memoize::unmemoize 'lookup_svn_merge';
3638 Memoize::unmemoize 'check_cherry_pick';
3639 Memoize::unmemoize 'has_no_changes';
3640 }
3641
3642 Memoize::memoize 'Git::SVN::repos_root';
3643 }
3644
3645 END {
3646 # Force cache writeout explicitly instead of waiting for
3647 # global destruction to avoid segfault in Storable:
3648 # http://rt.cpan.org/Public/Bug/Display.html?id=36087
3649 unmemoize_svn_mergeinfo_functions();
3650 }
3651
3652 sub parents_exclude {
3653 my $parents = shift;
3654 my @commits = @_;
3655 return unless @commits;
3656
3657 my @excluded;
3658 my $excluded;
3659 do {
3660 my @cmd = ('rev-list', "-1", @commits, "--not", @$parents );
3661 $excluded = command_oneline(@cmd);
3662 if ( $excluded ) {
3663 my @new;
3664 my $found;
3665 for my $commit ( @commits ) {
3666 if ( $commit eq $excluded ) {
3667 push @excluded, $commit;
3668 $found++;
3669 last;
3670 }
3671 else {
3672 push @new, $commit;
3673 }
3674 }
3675 die "saw commit '$excluded' in rev-list output, "
3676 ."but we didn't ask for that commit (wanted: @commits --not @$parents)"
3677 unless $found;
3678 @commits = @new;
3679 }
3680 }
3681 while ($excluded and @commits);
3682
3683 return @excluded;
3684 }
3685
3686
3687 # note: this function should only be called if the various dirprops
3688 # have actually changed
3689 sub find_extra_svn_parents {
3690 my ($self, $ed, $mergeinfo, $parents) = @_;
3691 # aha! svk:merge property changed...
3692
3693 memoize_svn_mergeinfo_functions();
3694
3695 # We first search for merged tips which are not in our
3696 # history. Then, we figure out which git revisions are in
3697 # that tip, but not this revision. If all of those revisions
3698 # are now marked as merge, we can add the tip as a parent.
3699 my @merges = split "\n", $mergeinfo;
3700 my @merge_tips;
3701 my $url = $self->{url};
3702 my $uuid = $self->ra_uuid;
3703 my %ranges;
3704 for my $merge ( @merges ) {
3705 my ($tip_commit, @ranges) =
3706 lookup_svn_merge( $uuid, $url, $merge );
3707 unless (!$tip_commit or
3708 grep { $_ eq $tip_commit } @$parents ) {
3709 push @merge_tips, $tip_commit;
3710 $ranges{$tip_commit} = \@ranges;
3711 } else {
3712 push @merge_tips, undef;
3713 }
3714 }
3715
3716 my %excluded = map { $_ => 1 }
3717 parents_exclude($parents, grep { defined } @merge_tips);
3718
3719 # check merge tips for new parents
3720 my @new_parents;
3721 for my $merge_tip ( @merge_tips ) {
3722 my $spec = shift @merges;
3723 next unless $merge_tip and $excluded{$merge_tip};
3724
3725 my $ranges = $ranges{$merge_tip};
3726
3727 # check out 'new' tips
3728 my $merge_base;
3729 eval {
3730 $merge_base = command_oneline(
3731 "merge-base",
3732 @$parents, $merge_tip,
3733 );
3734 };
3735 if ($@) {
3736 die "An error occurred during merge-base"
3737 unless $@->isa("Git::Error::Command");
3738
3739 warn "W: Cannot find common ancestor between ".
3740 "@$parents and $merge_tip. Ignoring merge info.\n";
3741 next;
3742 }
3743
3744 # double check that there are no missing non-merge commits
3745 my (@incomplete) = check_cherry_pick(
3746 $merge_base, $merge_tip,
3747 $parents,
3748 @$ranges,
3749 );
3750
3751 if ( @incomplete ) {
3752 warn "W:svn cherry-pick ignored ($spec) - missing "
3753 .@incomplete." commit(s) (eg $incomplete[0])\n";
3754 } else {
3755 warn
3756 "Found merge parent (svn:mergeinfo prop): ",
3757 $merge_tip, "\n";
3758 push @new_parents, $merge_tip;
3759 }
3760 }
3761
3762 # cater for merges which merge commits from multiple branches
3763 if ( @new_parents > 1 ) {
3764 for ( my $i = 0; $i <= $#new_parents; $i++ ) {
3765 for ( my $j = 0; $j <= $#new_parents; $j++ ) {
3766 next if $i == $j;
3767 next unless $new_parents[$i];
3768 next unless $new_parents[$j];
3769 my $revs = command_oneline(
3770 "rev-list", "-1",
3771 "$new_parents[$i]..$new_parents[$j]",
3772 );
3773 if ( !$revs ) {
3774 undef($new_parents[$j]);
3775 }
3776 }
3777 }
3778 }
3779 push @$parents, grep { defined } @new_parents;
3780 }
3781
3782 sub make_log_entry {
3783 my ($self, $rev, $parents, $ed) = @_;
3784 my $untracked = $self->get_untracked($ed);
3785
3786 my @parents = @$parents;
3787 my $ps = $ed->{path_strip} || "";
3788 for my $path ( grep { m/$ps/ } %{$ed->{dir_prop}} ) {
3789 my $props = $ed->{dir_prop}{$path};
3790 if ( $props->{"svk:merge"} ) {
3791 $self->find_extra_svk_parents
3792 ($ed, $props->{"svk:merge"}, \@parents);
3793 }
3794 if ( $props->{"svn:mergeinfo"} ) {
3795 $self->find_extra_svn_parents
3796 ($ed,
3797 $props->{"svn:mergeinfo"},
3798 \@parents);
3799 }
3800 }
3801
3802 open my $un, '>>', "$self->{dir}/unhandled.log" or croak $!;
3803 print $un "r$rev\n" or croak $!;
3804 print $un $_, "\n" foreach @$untracked;
3805 my %log_entry = ( parents => \@parents, revision => $rev,
3806 log => '');
3807
3808 my $headrev;
3809 my $logged = delete $self->{logged_rev_props};
3810 if (!$logged || $self->{-want_revprops}) {
3811 my $rp = $self->ra->rev_proplist($rev);
3812 foreach (sort keys %$rp) {
3813 my $v = $rp->{$_};
3814 if (/^svn:(author|date|log)$/) {
3815 $log_entry{$1} = $v;
3816 } elsif ($_ eq 'svm:headrev') {
3817 $headrev = $v;
3818 } else {
3819 print $un " rev_prop: ", uri_encode($_), ' ',
3820 uri_encode($v), "\n";
3821 }
3822 }
3823 } else {
3824 map { $log_entry{$_} = $logged->{$_} } keys %$logged;
3825 }
3826 close $un or croak $!;
3827
3828 $log_entry{date} = parse_svn_date($log_entry{date});
3829 $log_entry{log} .= "\n";
3830 my $author = $log_entry{author} = check_author($log_entry{author});
3831 my ($name, $email) = defined $::users{$author} ? @{$::users{$author}}
3832 : ($author, undef);
3833
3834 my ($commit_name, $commit_email) = ($name, $email);
3835 if ($_use_log_author) {
3836 my $name_field;
3837 if ($log_entry{log} =~ /From:\s+(.*\S)\s*\n/i) {
3838 $name_field = $1;
3839 } elsif ($log_entry{log} =~ /Signed-off-by:\s+(.*\S)\s*\n/i) {
3840 $name_field = $1;
3841 }
3842 if (!defined $name_field) {
3843 if (!defined $email) {
3844 $email = $name;
3845 }
3846 } elsif ($name_field =~ /(.*?)\s+<(.*)>/) {
3847 ($name, $email) = ($1, $2);
3848 } elsif ($name_field =~ /(.*)@/) {
3849 ($name, $email) = ($1, $name_field);
3850 } else {
3851 ($name, $email) = ($name_field, $name_field);
3852 }
3853 }
3854 if (defined $headrev && $self->use_svm_props) {
3855 if ($self->rewrite_root) {
3856 die "Can't have both 'useSvmProps' and 'rewriteRoot' ",
3857 "options set!\n";
3858 }
3859 if ($self->rewrite_uuid) {
3860 die "Can't have both 'useSvmProps' and 'rewriteUUID' ",
3861 "options set!\n";
3862 }
3863 my ($uuid, $r) = $headrev =~ m{^([a-f\d\-]{30,}):(\d+)$}i;
3864 # we don't want "SVM: initializing mirror for junk" ...
3865 return undef if $r == 0;
3866 my $svm = $self->svm;
3867 if ($uuid ne $svm->{uuid}) {
3868 die "UUID mismatch on SVM path:\n",
3869 "expected: $svm->{uuid}\n",
3870 " got: $uuid\n";
3871 }
3872 my $full_url = $self->full_url;
3873 $full_url =~ s#^\Q$svm->{replace}\E(/|$)#$svm->{source}$1# or
3874 die "Failed to replace '$svm->{replace}' with ",
3875 "'$svm->{source}' in $full_url\n";
3876 # throw away username for storing in records
3877 remove_username($full_url);
3878 $log_entry{metadata} = "$full_url\@$r $uuid";
3879 $log_entry{svm_revision} = $r;
3880 $email ||= "$author\@$uuid";
3881 $commit_email ||= "$author\@$uuid";
3882 } elsif ($self->use_svnsync_props) {
3883 my $full_url = $self->svnsync->{url};
3884 $full_url .= "/$self->{path}" if length $self->{path};
3885 remove_username($full_url);
3886 my $uuid = $self->svnsync->{uuid};
3887 $log_entry{metadata} = "$full_url\@$rev $uuid";
3888 $email ||= "$author\@$uuid";
3889 $commit_email ||= "$author\@$uuid";
3890 } else {
3891 my $url = $self->metadata_url;
3892 remove_username($url);
3893 my $uuid = $self->rewrite_uuid || $self->ra->get_uuid;
3894 $log_entry{metadata} = "$url\@$rev " . $uuid;
3895 $email ||= "$author\@" . $uuid;
3896 $commit_email ||= "$author\@" . $uuid;
3897 }
3898 $log_entry{name} = $name;
3899 $log_entry{email} = $email;
3900 $log_entry{commit_name} = $commit_name;
3901 $log_entry{commit_email} = $commit_email;
3902 \%log_entry;
3903 }
3904
3905 sub fetch {
3906 my ($self, $min_rev, $max_rev, @parents) = @_;
3907 my ($last_rev, $last_commit) = $self->last_rev_commit;
3908 my ($base, $head) = $self->get_fetch_range($min_rev, $max_rev);
3909 $self->ra->gs_fetch_loop_common($base, $head, [$self]);
3910 }
3911
3912 sub set_tree_cb {
3913 my ($self, $log_entry, $tree, $rev, $date, $author) = @_;
3914 $self->{inject_parents} = { $rev => $tree };
3915 $self->fetch(undef, undef);
3916 }
3917
3918 sub set_tree {
3919 my ($self, $tree) = (shift, shift);
3920 my $log_entry = ::get_commit_entry($tree);
3921 unless ($self->{last_rev}) {
3922 ::fatal("Must have an existing revision to commit");
3923 }
3924 my %ed_opts = ( r => $self->{last_rev},
3925 log => $log_entry->{log},
3926 ra => $self->ra,
3927 tree_a => $self->{last_commit},
3928 tree_b => $tree,
3929 editor_cb => sub {
3930 $self->set_tree_cb($log_entry, $tree, @_) },
3931 svn_path => $self->{path} );
3932 if (!Git::SVN::Editor->new(\%ed_opts)->apply_diff) {
3933 print "No changes\nr$self->{last_rev} = $tree\n";
3934 }
3935 }
3936
3937 sub rebuild_from_rev_db {
3938 my ($self, $path) = @_;
3939 my $r = -1;
3940 open my $fh, '<', $path or croak "open: $!";
3941 binmode $fh or croak "binmode: $!";
3942 while (<$fh>) {
3943 length($_) == 41 or croak "inconsistent size in ($_) != 41";
3944 chomp($_);
3945 ++$r;
3946 next if $_ eq ('0' x 40);
3947 $self->rev_map_set($r, $_);
3948 print "r$r = $_\n";
3949 }
3950 close $fh or croak "close: $!";
3951 unlink $path or croak "unlink: $!";
3952 }
3953
3954 sub rebuild {
3955 my ($self) = @_;
3956 my $map_path = $self->map_path;
3957 my $partial = (-e $map_path && ! -z $map_path);
3958 return unless ::verify_ref($self->refname.'^0');
3959 if (!$partial && ($self->use_svm_props || $self->no_metadata)) {
3960 my $rev_db = $self->rev_db_path;
3961 $self->rebuild_from_rev_db($rev_db);
3962 if ($self->use_svm_props) {
3963 my $svm_rev_db = $self->rev_db_path($self->svm_uuid);
3964 $self->rebuild_from_rev_db($svm_rev_db);
3965 }
3966 $self->unlink_rev_db_symlink;
3967 return;
3968 }
3969 print "Rebuilding $map_path ...\n" if (!$partial);
3970 my ($base_rev, $head) = ($partial ? $self->rev_map_max_norebuild(1) :
3971 (undef, undef));
3972 my ($log, $ctx) =
3973 command_output_pipe(qw/rev-list --pretty=raw --reverse/,
3974 ($head ? "$head.." : "") . $self->refname,
3975 '--');
3976 my $metadata_url = $self->metadata_url;
3977 remove_username($metadata_url);
3978 my $svn_uuid = $self->rewrite_uuid || $self->ra_uuid;
3979 my $c;
3980 while (<$log>) {
3981 if ( m{^commit ($::sha1)$} ) {
3982 $c = $1;
3983 next;
3984 }
3985 next unless s{^\s*(git-svn-id:)}{$1};
3986 my ($url, $rev, $uuid) = ::extract_metadata($_);
3987 remove_username($url);
3988
3989 # ignore merges (from set-tree)
3990 next if (!defined $rev || !$uuid);
3991
3992 # if we merged or otherwise started elsewhere, this is
3993 # how we break out of it
3994 if (($uuid ne $svn_uuid) ||
3995 ($metadata_url && $url && ($url ne $metadata_url))) {
3996 next;
3997 }
3998 if ($partial && $head) {
3999 print "Partial-rebuilding $map_path ...\n";
4000 print "Currently at $base_rev = $head\n";
4001 $head = undef;
4002 }
4003
4004 $self->rev_map_set($rev, $c);
4005 print "r$rev = $c\n";
4006 }
4007 command_close_pipe($log, $ctx);
4008 print "Done rebuilding $map_path\n" if (!$partial || !$head);
4009 my $rev_db_path = $self->rev_db_path;
4010 if (-f $self->rev_db_path) {
4011 unlink $self->rev_db_path or croak "unlink: $!";
4012 }
4013 $self->unlink_rev_db_symlink;
4014 }
4015
4016 # rev_map:
4017 # Tie::File seems to be prone to offset errors if revisions get sparse,
4018 # it's not that fast, either. Tie::File is also not in Perl 5.6. So
4019 # one of my favorite modules is out :< Next up would be one of the DBM
4020 # modules, but I'm not sure which is most portable...
4021 #
4022 # This is the replacement for the rev_db format, which was too big
4023 # and inefficient for large repositories with a lot of sparse history
4024 # (mainly tags)
4025 #
4026 # The format is this:
4027 # - 24 bytes for every record,
4028 # * 4 bytes for the integer representing an SVN revision number
4029 # * 20 bytes representing the sha1 of a git commit
4030 # - No empty padding records like the old format
4031 # (except the last record, which can be overwritten)
4032 # - new records are written append-only since SVN revision numbers
4033 # increase monotonically
4034 # - lookups on SVN revision number are done via a binary search
4035 # - Piping the file to xxd -c24 is a good way of dumping it for
4036 # viewing or editing (piped back through xxd -r), should the need
4037 # ever arise.
4038 # - The last record can be padding revision with an all-zero sha1
4039 # This is used to optimize fetch performance when using multiple
4040 # "fetch" directives in .git/config
4041 #
4042 # These files are disposable unless noMetadata or useSvmProps is set
4043
4044 sub _rev_map_set {
4045 my ($fh, $rev, $commit) = @_;
4046
4047 binmode $fh or croak "binmode: $!";
4048 my $size = (stat($fh))[7];
4049 ($size % 24) == 0 or croak "inconsistent size: $size";
4050
4051 my $wr_offset = 0;
4052 if ($size > 0) {
4053 sysseek($fh, -24, SEEK_END) or croak "seek: $!";
4054 my $read = sysread($fh, my $buf, 24) or croak "read: $!";
4055 $read == 24 or croak "read only $read bytes (!= 24)";
4056 my ($last_rev, $last_commit) = unpack(rev_map_fmt, $buf);
4057 if ($last_commit eq ('0' x40)) {
4058 if ($size >= 48) {
4059 sysseek($fh, -48, SEEK_END) or croak "seek: $!";
4060 $read = sysread($fh, $buf, 24) or
4061 croak "read: $!";
4062 $read == 24 or
4063 croak "read only $read bytes (!= 24)";
4064 ($last_rev, $last_commit) =
4065 unpack(rev_map_fmt, $buf);
4066 if ($last_commit eq ('0' x40)) {
4067 croak "inconsistent .rev_map\n";
4068 }
4069 }
4070 if ($last_rev >= $rev) {
4071 croak "last_rev is higher!: $last_rev >= $rev";
4072 }
4073 $wr_offset = -24;
4074 }
4075 }
4076 sysseek($fh, $wr_offset, SEEK_END) or croak "seek: $!";
4077 syswrite($fh, pack(rev_map_fmt, $rev, $commit), 24) == 24 or
4078 croak "write: $!";
4079 }
4080
4081 sub _rev_map_reset {
4082 my ($fh, $rev, $commit) = @_;
4083 my $c = _rev_map_get($fh, $rev);
4084 $c eq $commit or die "_rev_map_reset(@_) commit $c does not match!\n";
4085 my $offset = sysseek($fh, 0, SEEK_CUR) or croak "seek: $!";
4086 truncate $fh, $offset or croak "truncate: $!";
4087 }
4088
4089 sub mkfile {
4090 my ($path) = @_;
4091 unless (-e $path) {
4092 my ($dir, $base) = ($path =~ m#^(.*?)/?([^/]+)$#);
4093 mkpath([$dir]) unless -d $dir;
4094 open my $fh, '>>', $path or die "Couldn't create $path: $!\n";
4095 close $fh or die "Couldn't close (create) $path: $!\n";
4096 }
4097 }
4098
4099 sub rev_map_set {
4100 my ($self, $rev, $commit, $update_ref, $uuid) = @_;
4101 defined $commit or die "missing arg3\n";
4102 length $commit == 40 or die "arg3 must be a full SHA1 hexsum\n";
4103 my $db = $self->map_path($uuid);
4104 my $db_lock = "$db.lock";
4105 my $sigmask;
4106 $update_ref ||= 0;
4107 if ($update_ref) {
4108 $sigmask = POSIX::SigSet->new();
4109 my $signew = POSIX::SigSet->new(SIGINT, SIGHUP, SIGTERM,
4110 SIGALRM, SIGUSR1, SIGUSR2);
4111 sigprocmask(SIG_BLOCK, $signew, $sigmask) or
4112 croak "Can't block signals: $!";
4113 }
4114 mkfile($db);
4115
4116 $LOCKFILES{$db_lock} = 1;
4117 my $sync;
4118 # both of these options make our .rev_db file very, very important
4119 # and we can't afford to lose it because rebuild() won't work
4120 if ($self->use_svm_props || $self->no_metadata) {
4121 $sync = 1;
4122 copy($db, $db_lock) or die "rev_map_set(@_): ",
4123 "Failed to copy: ",
4124 "$db => $db_lock ($!)\n";
4125 } else {
4126 rename $db, $db_lock or die "rev_map_set(@_): ",
4127 "Failed to rename: ",
4128 "$db => $db_lock ($!)\n";
4129 }
4130
4131 sysopen(my $fh, $db_lock, O_RDWR | O_CREAT)
4132 or croak "Couldn't open $db_lock: $!\n";
4133 $update_ref eq 'reset' ? _rev_map_reset($fh, $rev, $commit) :
4134 _rev_map_set($fh, $rev, $commit);
4135 if ($sync) {
4136 $fh->flush or die "Couldn't flush $db_lock: $!\n";
4137 $fh->sync or die "Couldn't sync $db_lock: $!\n";
4138 }
4139 close $fh or croak $!;
4140 if ($update_ref) {
4141 $_head = $self;
4142 my $note = "";
4143 $note = " ($update_ref)" if ($update_ref !~ /^\d*$/);
4144 command_noisy('update-ref', '-m', "r$rev$note",
4145 $self->refname, $commit);
4146 }
4147 rename $db_lock, $db or die "rev_map_set(@_): ", "Failed to rename: ",
4148 "$db_lock => $db ($!)\n";
4149 delete $LOCKFILES{$db_lock};
4150 if ($update_ref) {
4151 sigprocmask(SIG_SETMASK, $sigmask) or
4152 croak "Can't restore signal mask: $!";
4153 }
4154 }
4155
4156 # If want_commit, this will return an array of (rev, commit) where
4157 # commit _must_ be a valid commit in the archive.
4158 # Otherwise, it'll return the max revision (whether or not the
4159 # commit is valid or just a 0x40 placeholder).
4160 sub rev_map_max {
4161 my ($self, $want_commit) = @_;
4162 $self->rebuild;
4163 my ($r, $c) = $self->rev_map_max_norebuild($want_commit);
4164 $want_commit ? ($r, $c) : $r;
4165 }
4166
4167 sub rev_map_max_norebuild {
4168 my ($self, $want_commit) = @_;
4169 my $map_path = $self->map_path;
4170 stat $map_path or return $want_commit ? (0, undef) : 0;
4171 sysopen(my $fh, $map_path, O_RDONLY) or croak "open: $!";
4172 binmode $fh or croak "binmode: $!";
4173 my $size = (stat($fh))[7];
4174 ($size % 24) == 0 or croak "inconsistent size: $size";
4175
4176 if ($size == 0) {
4177 close $fh or croak "close: $!";
4178 return $want_commit ? (0, undef) : 0;
4179 }
4180
4181 sysseek($fh, -24, SEEK_END) or croak "seek: $!";
4182 sysread($fh, my $buf, 24) == 24 or croak "read: $!";
4183 my ($r, $c) = unpack(rev_map_fmt, $buf);
4184 if ($want_commit && $c eq ('0' x40)) {
4185 if ($size < 48) {
4186 return $want_commit ? (0, undef) : 0;
4187 }
4188 sysseek($fh, -48, SEEK_END) or croak "seek: $!";
4189 sysread($fh, $buf, 24) == 24 or croak "read: $!";
4190 ($r, $c) = unpack(rev_map_fmt, $buf);
4191 if ($c eq ('0'x40)) {
4192 croak "Penultimate record is all-zeroes in $map_path";
4193 }
4194 }
4195 close $fh or croak "close: $!";
4196 $want_commit ? ($r, $c) : $r;
4197 }
4198
4199 sub rev_map_get {
4200 my ($self, $rev, $uuid) = @_;
4201 my $map_path = $self->map_path($uuid);
4202 return undef unless -e $map_path;
4203
4204 sysopen(my $fh, $map_path, O_RDONLY) or croak "open: $!";
4205 my $c = _rev_map_get($fh, $rev);
4206 close($fh) or croak "close: $!";
4207 $c
4208 }
4209
4210 sub _rev_map_get {
4211 my ($fh, $rev) = @_;
4212
4213 binmode $fh or croak "binmode: $!";
4214 my $size = (stat($fh))[7];
4215 ($size % 24) == 0 or croak "inconsistent size: $size";
4216
4217 if ($size == 0) {
4218 return undef;
4219 }
4220
4221 my ($l, $u) = (0, $size - 24);
4222 my ($r, $c, $buf);
4223
4224 while ($l <= $u) {
4225 my $i = int(($l/24 + $u/24) / 2) * 24;
4226 sysseek($fh, $i, SEEK_SET) or croak "seek: $!";
4227 sysread($fh, my $buf, 24) == 24 or croak "read: $!";
4228 my ($r, $c) = unpack(rev_map_fmt, $buf);
4229
4230 if ($r < $rev) {
4231 $l = $i + 24;
4232 } elsif ($r > $rev) {
4233 $u = $i - 24;
4234 } else { # $r == $rev
4235 return $c eq ('0' x 40) ? undef : $c;
4236 }
4237 }
4238 undef;
4239 }
4240
4241 # Finds the first svn revision that exists on (if $eq_ok is true) or
4242 # before $rev for the current branch. It will not search any lower
4243 # than $min_rev. Returns the git commit hash and svn revision number
4244 # if found, else (undef, undef).
4245 sub find_rev_before {
4246 my ($self, $rev, $eq_ok, $min_rev) = @_;
4247 --$rev unless $eq_ok;
4248 $min_rev ||= 1;
4249 my $max_rev = $self->rev_map_max;
4250 $rev = $max_rev if ($rev > $max_rev);
4251 while ($rev >= $min_rev) {
4252 if (my $c = $self->rev_map_get($rev)) {
4253 return ($rev, $c);
4254 }
4255 --$rev;
4256 }
4257 return (undef, undef);
4258 }
4259
4260 # Finds the first svn revision that exists on (if $eq_ok is true) or
4261 # after $rev for the current branch. It will not search any higher
4262 # than $max_rev. Returns the git commit hash and svn revision number
4263 # if found, else (undef, undef).
4264 sub find_rev_after {
4265 my ($self, $rev, $eq_ok, $max_rev) = @_;
4266 ++$rev unless $eq_ok;
4267 $max_rev ||= $self->rev_map_max;
4268 while ($rev <= $max_rev) {
4269 if (my $c = $self->rev_map_get($rev)) {
4270 return ($rev, $c);
4271 }
4272 ++$rev;
4273 }
4274 return (undef, undef);
4275 }
4276
4277 sub _new {
4278 my ($class, $repo_id, $ref_id, $path) = @_;
4279 unless (defined $repo_id && length $repo_id) {
4280 $repo_id = $Git::SVN::default_repo_id;
4281 }
4282 unless (defined $ref_id && length $ref_id) {
4283 $_prefix = '' unless defined($_prefix);
4284 $_[2] = $ref_id =
4285 "refs/remotes/$_prefix$Git::SVN::default_ref_id";
4286 }
4287 $_[1] = $repo_id;
4288 my $dir = "$ENV{GIT_DIR}/svn/$ref_id";
4289
4290 # Older repos imported by us used $GIT_DIR/svn/foo instead of
4291 # $GIT_DIR/svn/refs/remotes/foo when tracking refs/remotes/foo
4292 if ($ref_id =~ m{^refs/remotes/(.*)}) {
4293 my $old_dir = "$ENV{GIT_DIR}/svn/$1";
4294 if (-d $old_dir && ! -d $dir) {
4295 $dir = $old_dir;
4296 }
4297 }
4298
4299 $_[3] = $path = '' unless (defined $path);
4300 mkpath([$dir]);
4301 bless {
4302 ref_id => $ref_id, dir => $dir, index => "$dir/index",
4303 path => $path, config => "$ENV{GIT_DIR}/svn/config",
4304 map_root => "$dir/.rev_map", repo_id => $repo_id }, $class;
4305 }
4306
4307 # for read-only access of old .rev_db formats
4308 sub unlink_rev_db_symlink {
4309 my ($self) = @_;
4310 my $link = $self->rev_db_path;
4311 $link =~ s/\.[\w-]+$// or croak "missing UUID at the end of $link";
4312 if (-l $link) {
4313 unlink $link or croak "unlink: $link failed!";
4314 }
4315 }
4316
4317 sub rev_db_path {
4318 my ($self, $uuid) = @_;
4319 my $db_path = $self->map_path($uuid);
4320 $db_path =~ s{/\.rev_map\.}{/\.rev_db\.}
4321 or croak "map_path: $db_path does not contain '/.rev_map.' !";
4322 $db_path;
4323 }
4324
4325 # the new replacement for .rev_db
4326 sub map_path {
4327 my ($self, $uuid) = @_;
4328 $uuid ||= $self->ra_uuid;
4329 "$self->{map_root}.$uuid";
4330 }
4331
4332 sub uri_encode {
4333 my ($f) = @_;
4334 $f =~ s#([^a-zA-Z0-9\*!\:_\./\-])#uc sprintf("%%%02x",ord($1))#eg;
4335 $f
4336 }
4337
4338 sub uri_decode {
4339 my ($f) = @_;
4340 $f =~ s#%([0-9a-fA-F]{2})#chr(hex($1))#eg;
4341 $f
4342 }
4343
4344 sub remove_username {
4345 $_[0] =~ s{^([^:]*://)[^@]+@}{$1};
4346 }
4347
4348 package Git::SVN::Log;
4349 use strict;
4350 use warnings;
4351 use POSIX qw/strftime/;
4352 use constant commit_log_separator => ('-' x 72) . "\n";
4353 use vars qw/$TZ $limit $color $pager $non_recursive $verbose $oneline
4354 %rusers $show_commit $incremental/;
4355 my $l_fmt;
4356
4357 sub cmt_showable {
4358 my ($c) = @_;
4359 return 1 if defined $c->{r};
4360
4361 # big commit message got truncated by the 16k pretty buffer in rev-list
4362 if ($c->{l} && $c->{l}->[-1] eq "...\n" &&
4363 $c->{a_raw} =~ /\@([a-f\d\-]+)>$/) {
4364 @{$c->{l}} = ();
4365 my @log = command(qw/cat-file commit/, $c->{c});
4366
4367 # shift off the headers
4368 shift @log while ($log[0] ne '');
4369 shift @log;
4370
4371 # TODO: make $c->{l} not have a trailing newline in the future
4372 @{$c->{l}} = map { "$_\n" } grep !/^git-svn-id: /, @log;
4373
4374 (undef, $c->{r}, undef) = ::extract_metadata(
4375 (grep(/^git-svn-id: /, @log))[-1]);
4376 }
4377 return defined $c->{r};
4378 }
4379
4380 sub log_use_color {
4381 return $color || Git->repository->get_colorbool('color.diff');
4382 }
4383
4384 sub git_svn_log_cmd {
4385 my ($r_min, $r_max, @args) = @_;
4386 my $head = 'HEAD';
4387 my (@files, @log_opts);
4388 foreach my $x (@args) {
4389 if ($x eq '--' || @files) {
4390 push @files, $x;
4391 } else {
4392 if (::verify_ref("$x^0")) {
4393 $head = $x;
4394 } else {
4395 push @log_opts, $x;
4396 }
4397 }
4398 }
4399
4400 my ($url, $rev, $uuid, $gs) = ::working_head_info($head);
4401 $gs ||= Git::SVN->_new;
4402 my @cmd = (qw/log --abbrev-commit --pretty=raw --default/,
4403 $gs->refname);
4404 push @cmd, '-r' unless $non_recursive;
4405 push @cmd, qw/--raw --name-status/ if $verbose;
4406 push @cmd, '--color' if log_use_color();
4407 push @cmd, @log_opts;
4408 if (defined $r_max && $r_max == $r_min) {
4409 push @cmd, '--max-count=1';
4410 if (my $c = $gs->rev_map_get($r_max)) {
4411 push @cmd, $c;
4412 }
4413 } elsif (defined $r_max) {
4414 if ($r_max < $r_min) {
4415 ($r_min, $r_max) = ($r_max, $r_min);
4416 }
4417 my (undef, $c_max) = $gs->find_rev_before($r_max, 1, $r_min);
4418 my (undef, $c_min) = $gs->find_rev_after($r_min, 1, $r_max);
4419 # If there are no commits in the range, both $c_max and $c_min
4420 # will be undefined. If there is at least 1 commit in the
4421 # range, both will be defined.
4422 return () if !defined $c_min || !defined $c_max;
4423 if ($c_min eq $c_max) {
4424 push @cmd, '--max-count=1', $c_min;
4425 } else {
4426 push @cmd, '--boundary', "$c_min..$c_max";
4427 }
4428 }
4429 return (@cmd, @files);
4430 }
4431
4432 # adapted from pager.c
4433 sub config_pager {
4434 if (! -t *STDOUT) {
4435 $ENV{GIT_PAGER_IN_USE} = 'false';
4436 $pager = undef;
4437 return;
4438 }
4439 chomp($pager = command_oneline(qw(var GIT_PAGER)));
4440 if ($pager eq 'cat') {
4441 $pager = undef;
4442 }
4443 $ENV{GIT_PAGER_IN_USE} = defined($pager);
4444 }
4445
4446 sub run_pager {
4447 return unless defined $pager;
4448 pipe my ($rfd, $wfd) or return;
4449 defined(my $pid = fork) or ::fatal "Can't fork: $!";
4450 if (!$pid) {
4451 open STDOUT, '>&', $wfd or
4452 ::fatal "Can't redirect to stdout: $!";
4453 return;
4454 }
4455 open STDIN, '<&', $rfd or ::fatal "Can't redirect stdin: $!";
4456 $ENV{LESS} ||= 'FRSX';
4457 exec $pager or ::fatal "Can't run pager: $! ($pager)";
4458 }
4459
4460 sub format_svn_date {
4461 my $t = shift || time;
4462 my $gmoff = Git::SVN::get_tz($t);
4463 return strftime("%Y-%m-%d %H:%M:%S $gmoff (%a, %d %b %Y)", localtime($t));
4464 }
4465
4466 sub parse_git_date {
4467 my ($t, $tz) = @_;
4468 # Date::Parse isn't in the standard Perl distro :(
4469 if ($tz =~ s/^\+//) {
4470 $t += tz_to_s_offset($tz);
4471 } elsif ($tz =~ s/^\-//) {
4472 $t -= tz_to_s_offset($tz);
4473 }
4474 return $t;
4475 }
4476
4477 sub set_local_timezone {
4478 if (defined $TZ) {
4479 $ENV{TZ} = $TZ;
4480 } else {
4481 delete $ENV{TZ};
4482 }
4483 }
4484
4485 sub tz_to_s_offset {
4486 my ($tz) = @_;
4487 $tz =~ s/(\d\d)$//;
4488 return ($1 * 60) + ($tz * 3600);
4489 }
4490
4491 sub get_author_info {
4492 my ($dest, $author, $t, $tz) = @_;
4493 $author =~ s/(?:^\s*|\s*$)//g;
4494 $dest->{a_raw} = $author;
4495 my $au;
4496 if ($::_authors) {
4497 $au = $rusers{$author} || undef;
4498 }
4499 if (!$au) {
4500 ($au) = ($author =~ /<([^>]+)\@[^>]+>$/);
4501 }
4502 $dest->{t} = $t;
4503 $dest->{tz} = $tz;
4504 $dest->{a} = $au;
4505 $dest->{t_utc} = parse_git_date($t, $tz);
4506 }
4507
4508 sub process_commit {
4509 my ($c, $r_min, $r_max, $defer) = @_;
4510 if (defined $r_min && defined $r_max) {
4511 if ($r_min == $c->{r} && $r_min == $r_max) {
4512 show_commit($c);
4513 return 0;
4514 }
4515 return 1 if $r_min == $r_max;
4516 if ($r_min < $r_max) {
4517 # we need to reverse the print order
4518 return 0 if (defined $limit && --$limit < 0);
4519 push @$defer, $c;
4520 return 1;
4521 }
4522 if ($r_min != $r_max) {
4523 return 1 if ($r_min < $c->{r});
4524 return 1 if ($r_max > $c->{r});
4525 }
4526 }
4527 return 0 if (defined $limit && --$limit < 0);
4528 show_commit($c);
4529 return 1;
4530 }
4531
4532 sub show_commit {
4533 my $c = shift;
4534 if ($oneline) {
4535 my $x = "\n";
4536 if (my $l = $c->{l}) {
4537 while ($l->[0] =~ /^\s*$/) { shift @$l }
4538 $x = $l->[0];
4539 }
4540 $l_fmt ||= 'A' . length($c->{r});
4541 print 'r',pack($l_fmt, $c->{r}),' | ';
4542 print "$c->{c} | " if $show_commit;
4543 print $x;
4544 } else {
4545 show_commit_normal($c);
4546 }
4547 }
4548
4549 sub show_commit_changed_paths {
4550 my ($c) = @_;
4551 return unless $c->{changed};
4552 print "Changed paths:\n", @{$c->{changed}};
4553 }
4554
4555 sub show_commit_normal {
4556 my ($c) = @_;
4557 print commit_log_separator, "r$c->{r} | ";
4558 print "$c->{c} | " if $show_commit;
4559 print "$c->{a} | ", format_svn_date($c->{t_utc}), ' | ';
4560 my $nr_line = 0;
4561
4562 if (my $l = $c->{l}) {
4563 while ($l->[$#$l] eq "\n" && $#$l > 0
4564 && $l->[($#$l - 1)] eq "\n") {
4565 pop @$l;
4566 }
4567 $nr_line = scalar @$l;
4568 if (!$nr_line) {
4569 print "1 line\n\n\n";
4570 } else {
4571 if ($nr_line == 1) {
4572 $nr_line = '1 line';
4573 } else {
4574 $nr_line .= ' lines';
4575 }
4576 print $nr_line, "\n";
4577 show_commit_changed_paths($c);
4578 print "\n";
4579 print $_ foreach @$l;
4580 }
4581 } else {
4582 print "1 line\n";
4583 show_commit_changed_paths($c);
4584 print "\n";
4585
4586 }
4587 foreach my $x (qw/raw stat diff/) {
4588 if ($c->{$x}) {
4589 print "\n";
4590 print $_ foreach @{$c->{$x}}
4591 }
4592 }
4593 }
4594
4595 sub cmd_show_log {
4596 my (@args) = @_;
4597 my ($r_min, $r_max);
4598 my $r_last = -1; # prevent dupes
4599 set_local_timezone();
4600 if (defined $::_revision) {
4601 if ($::_revision =~ /^(\d+):(\d+)$/) {
4602 ($r_min, $r_max) = ($1, $2);
4603 } elsif ($::_revision =~ /^\d+$/) {
4604 $r_min = $r_max = $::_revision;
4605 } else {
4606 ::fatal "-r$::_revision is not supported, use ",
4607 "standard 'git log' arguments instead";
4608 }
4609 }
4610
4611 config_pager();
4612 @args = git_svn_log_cmd($r_min, $r_max, @args);
4613 if (!@args) {
4614 print commit_log_separator unless $incremental || $oneline;
4615 return;
4616 }
4617 my $log = command_output_pipe(@args);
4618 run_pager();
4619 my (@k, $c, $d, $stat);
4620 my $esc_color = qr/(?:\033\[(?:(?:\d+;)*\d*)?m)*/;
4621 while (<$log>) {
4622 if (/^${esc_color}commit (?:- )?($::sha1_short)/o) {
4623 my $cmt = $1;
4624 if ($c && cmt_showable($c) && $c->{r} != $r_last) {
4625 $r_last = $c->{r};
4626 process_commit($c, $r_min, $r_max, \@k) or
4627 goto out;
4628 }
4629 $d = undef;
4630 $c = { c => $cmt };
4631 } elsif (/^${esc_color}author (.+) (\d+) ([\-\+]?\d+)$/o) {
4632 get_author_info($c, $1, $2, $3);
4633 } elsif (/^${esc_color}(?:tree|parent|committer) /o) {
4634 # ignore
4635 } elsif (/^${esc_color}:\d{6} \d{6} $::sha1_short/o) {
4636 push @{$c->{raw}}, $_;
4637 } elsif (/^${esc_color}[ACRMDT]\t/) {
4638 # we could add $SVN->{svn_path} here, but that requires
4639 # remote access at the moment (repo_path_split)...
4640 s#^(${esc_color})([ACRMDT])\t#$1 $2 #o;
4641 push @{$c->{changed}}, $_;
4642 } elsif (/^${esc_color}diff /o) {
4643 $d = 1;
4644 push @{$c->{diff}}, $_;
4645 } elsif ($d) {
4646 push @{$c->{diff}}, $_;
4647 } elsif (/^\ .+\ \|\s*\d+\ $esc_color[\+\-]*
4648 $esc_color*[\+\-]*$esc_color$/x) {
4649 $stat = 1;
4650 push @{$c->{stat}}, $_;
4651 } elsif ($stat && /^ \d+ files changed, \d+ insertions/) {
4652 push @{$c->{stat}}, $_;
4653 $stat = undef;
4654 } elsif (/^${esc_color} (git-svn-id:.+)$/o) {
4655 ($c->{url}, $c->{r}, undef) = ::extract_metadata($1);
4656 } elsif (s/^${esc_color} //o) {
4657 push @{$c->{l}}, $_;
4658 }
4659 }
4660 if ($c && defined $c->{r} && $c->{r} != $r_last) {
4661 $r_last = $c->{r};
4662 process_commit($c, $r_min, $r_max, \@k);
4663 }
4664 if (@k) {
4665 ($r_min, $r_max) = ($r_max, $r_min);
4666 process_commit($_, $r_min, $r_max) foreach reverse @k;
4667 }
4668 out:
4669 close $log;
4670 print commit_log_separator unless $incremental || $oneline;
4671 }
4672
4673 sub cmd_blame {
4674 my $path = pop;
4675
4676 config_pager();
4677 run_pager();
4678
4679 my ($fh, $ctx, $rev);
4680
4681 if ($_git_format) {
4682 ($fh, $ctx) = command_output_pipe('blame', @_, $path);
4683 while (my $line = <$fh>) {
4684 if ($line =~ /^\^?([[:xdigit:]]+)\s/) {
4685 # Uncommitted edits show up as a rev ID of
4686 # all zeros, which we can't look up with
4687 # cmt_metadata
4688 if ($1 !~ /^0+$/) {
4689 (undef, $rev, undef) =
4690 ::cmt_metadata($1);
4691 $rev = '0' if (!$rev);
4692 } else {
4693 $rev = '0';
4694 }
4695 $rev = sprintf('%-10s', $rev);
4696 $line =~ s/^\^?[[:xdigit:]]+(\s)/$rev$1/;
4697 }
4698 print $line;
4699 }
4700 } else {
4701 ($fh, $ctx) = command_output_pipe('blame', '-p', @_, 'HEAD',
4702 '--', $path);
4703 my ($sha1);
4704 my %authors;
4705 my @buffer;
4706 my %dsha; #distinct sha keys
4707
4708 while (my $line = <$fh>) {
4709 push @buffer, $line;
4710 if ($line =~ /^([[:xdigit:]]{40})\s\d+\s\d+/) {
4711 $dsha{$1} = 1;
4712 }
4713 }
4714
4715 my $s2r = ::cmt_sha2rev_batch([keys %dsha]);
4716
4717 foreach my $line (@buffer) {
4718 if ($line =~ /^([[:xdigit:]]{40})\s\d+\s\d+/) {
4719 $rev = $s2r->{$1};
4720 $rev = '0' if (!$rev)
4721 }
4722 elsif ($line =~ /^author (.*)/) {
4723 $authors{$rev} = $1;
4724 $authors{$rev} =~ s/\s/_/g;
4725 }
4726 elsif ($line =~ /^\t(.*)$/) {
4727 printf("%6s %10s %s\n", $rev, $authors{$rev}, $1);
4728 }
4729 }
4730 }
4731 command_close_pipe($fh, $ctx);
4732 }
4733
4734 package Git::SVN::Migration;
4735 # these version numbers do NOT correspond to actual version numbers
4736 # of git nor git-svn. They are just relative.
4737 #
4738 # v0 layout: .git/$id/info/url, refs/heads/$id-HEAD
4739 #
4740 # v1 layout: .git/$id/info/url, refs/remotes/$id
4741 #
4742 # v2 layout: .git/svn/$id/info/url, refs/remotes/$id
4743 #
4744 # v3 layout: .git/svn/$id, refs/remotes/$id
4745 # - info/url may remain for backwards compatibility
4746 # - this is what we migrate up to this layout automatically,
4747 # - this will be used by git svn init on single branches
4748 # v3.1 layout (auto migrated):
4749 # - .rev_db => .rev_db.$UUID, .rev_db will remain as a symlink
4750 # for backwards compatibility
4751 #
4752 # v4 layout: .git/svn/$repo_id/$id, refs/remotes/$repo_id/$id
4753 # - this is only created for newly multi-init-ed
4754 # repositories. Similar in spirit to the
4755 # --use-separate-remotes option in git-clone (now default)
4756 # - we do not automatically migrate to this (following
4757 # the example set by core git)
4758 #
4759 # v5 layout: .rev_db.$UUID => .rev_map.$UUID
4760 # - newer, more-efficient format that uses 24-bytes per record
4761 # with no filler space.
4762 # - use xxd -c24 < .rev_map.$UUID to view and debug
4763 # - This is a one-way migration, repositories updated to the
4764 # new format will not be able to use old git-svn without
4765 # rebuilding the .rev_db. Rebuilding the rev_db is not
4766 # possible if noMetadata or useSvmProps are set; but should
4767 # be no problem for users that use the (sensible) defaults.
4768 use strict;
4769 use warnings;
4770 use Carp qw/croak/;
4771 use File::Path qw/mkpath/;
4772 use File::Basename qw/dirname basename/;
4773 use vars qw/$_minimize/;
4774
4775 sub migrate_from_v0 {
4776 my $git_dir = $ENV{GIT_DIR};
4777 return undef unless -d $git_dir;
4778 my ($fh, $ctx) = command_output_pipe(qw/rev-parse --symbolic --all/);
4779 my $migrated = 0;
4780 while (<$fh>) {
4781 chomp;
4782 my ($id, $orig_ref) = ($_, $_);
4783 next unless $id =~ s#^refs/heads/(.+)-HEAD$#$1#;
4784 next unless -f "$git_dir/$id/info/url";
4785 my $new_ref = "refs/remotes/$id";
4786 if (::verify_ref("$new_ref^0")) {
4787 print STDERR "W: $orig_ref is probably an old ",
4788 "branch used by an ancient version of ",
4789 "git-svn.\n",
4790 "However, $new_ref also exists.\n",
4791 "We will not be able ",
4792 "to use this branch until this ",
4793 "ambiguity is resolved.\n";
4794 next;
4795 }
4796 print STDERR "Migrating from v0 layout...\n" if !$migrated;
4797 print STDERR "Renaming ref: $orig_ref => $new_ref\n";
4798 command_noisy('update-ref', $new_ref, $orig_ref);
4799 command_noisy('update-ref', '-d', $orig_ref, $orig_ref);
4800 $migrated++;
4801 }
4802 command_close_pipe($fh, $ctx);
4803 print STDERR "Done migrating from v0 layout...\n" if $migrated;
4804 $migrated;
4805 }
4806
4807 sub migrate_from_v1 {
4808 my $git_dir = $ENV{GIT_DIR};
4809 my $migrated = 0;
4810 return $migrated unless -d $git_dir;
4811 my $svn_dir = "$git_dir/svn";
4812
4813 # just in case somebody used 'svn' as their $id at some point...
4814 return $migrated if -d $svn_dir && ! -f "$svn_dir/info/url";
4815
4816 print STDERR "Migrating from a git-svn v1 layout...\n";
4817 mkpath([$svn_dir]);
4818 print STDERR "Data from a previous version of git-svn exists, but\n\t",
4819 "$svn_dir\n\t(required for this version ",
4820 "($::VERSION) of git-svn) does not exist.\n";
4821 my ($fh, $ctx) = command_output_pipe(qw/rev-parse --symbolic --all/);
4822 while (<$fh>) {
4823 my $x = $_;
4824 next unless $x =~ s#^refs/remotes/##;
4825 chomp $x;
4826 next unless -f "$git_dir/$x/info/url";
4827 my $u = eval { ::file_to_s("$git_dir/$x/info/url") };
4828 next unless $u;
4829 my $dn = dirname("$git_dir/svn/$x");
4830 mkpath([$dn]) unless -d $dn;
4831 if ($x eq 'svn') { # they used 'svn' as GIT_SVN_ID:
4832 mkpath(["$git_dir/svn/svn"]);
4833 print STDERR " - $git_dir/$x/info => ",
4834 "$git_dir/svn/$x/info\n";
4835 rename "$git_dir/$x/info", "$git_dir/svn/$x/info" or
4836 croak "$!: $x";
4837 # don't worry too much about these, they probably
4838 # don't exist with repos this old (save for index,
4839 # and we can easily regenerate that)
4840 foreach my $f (qw/unhandled.log index .rev_db/) {
4841 rename "$git_dir/$x/$f", "$git_dir/svn/$x/$f";
4842 }
4843 } else {
4844 print STDERR " - $git_dir/$x => $git_dir/svn/$x\n";
4845 rename "$git_dir/$x", "$git_dir/svn/$x" or
4846 croak "$!: $x";
4847 }
4848 $migrated++;
4849 }
4850 command_close_pipe($fh, $ctx);
4851 print STDERR "Done migrating from a git-svn v1 layout\n";
4852 $migrated;
4853 }
4854
4855 sub read_old_urls {
4856 my ($l_map, $pfx, $path) = @_;
4857 my @dir;
4858 foreach (<$path/*>) {
4859 if (-r "$_/info/url") {
4860 $pfx .= '/' if $pfx && $pfx !~ m!/$!;
4861 my $ref_id = $pfx . basename $_;
4862 my $url = ::file_to_s("$_/info/url");
4863 $l_map->{$ref_id} = $url;
4864 } elsif (-d $_) {
4865 push @dir, $_;
4866 }
4867 }
4868 foreach (@dir) {
4869 my $x = $_;
4870 $x =~ s!^\Q$ENV{GIT_DIR}\E/svn/!!o;
4871 read_old_urls($l_map, $x, $_);
4872 }
4873 }
4874
4875 sub migrate_from_v2 {
4876 my @cfg = command(qw/config -l/);
4877 return if grep /^svn-remote\..+\.url=/, @cfg;
4878 my %l_map;
4879 read_old_urls(\%l_map, '', "$ENV{GIT_DIR}/svn");
4880 my $migrated = 0;
4881
4882 foreach my $ref_id (sort keys %l_map) {
4883 eval { Git::SVN->init($l_map{$ref_id}, '', undef, $ref_id) };
4884 if ($@) {
4885 Git::SVN->init($l_map{$ref_id}, '', $ref_id, $ref_id);
4886 }
4887 $migrated++;
4888 }
4889 $migrated;
4890 }
4891
4892 sub minimize_connections {
4893 my $r = Git::SVN::read_all_remotes();
4894 my $new_urls = {};
4895 my $root_repos = {};
4896 foreach my $repo_id (keys %$r) {
4897 my $url = $r->{$repo_id}->{url} or next;
4898 my $fetch = $r->{$repo_id}->{fetch} or next;
4899 my $ra = Git::SVN::Ra->new($url);
4900
4901 # skip existing cases where we already connect to the root
4902 if (($ra->{url} eq $ra->{repos_root}) ||
4903 ($ra->{repos_root} eq $repo_id)) {
4904 $root_repos->{$ra->{url}} = $repo_id;
4905 next;
4906 }
4907
4908 my $root_ra = Git::SVN::Ra->new($ra->{repos_root});
4909 my $root_path = $ra->{url};
4910 $root_path =~ s#^\Q$ra->{repos_root}\E(/|$)##;
4911 foreach my $path (keys %$fetch) {
4912 my $ref_id = $fetch->{$path};
4913 my $gs = Git::SVN->new($ref_id, $repo_id, $path);
4914
4915 # make sure we can read when connecting to
4916 # a higher level of a repository
4917 my ($last_rev, undef) = $gs->last_rev_commit;
4918 if (!defined $last_rev) {
4919 $last_rev = eval {
4920 $root_ra->get_latest_revnum;
4921 };
4922 next if $@;
4923 }
4924 my $new = $root_path;
4925 $new .= length $path ? "/$path" : '';
4926 eval {
4927 $root_ra->get_log([$new], $last_rev, $last_rev,
4928 0, 0, 1, sub { });
4929 };
4930 next if $@;
4931 $new_urls->{$ra->{repos_root}}->{$new} =
4932 { ref_id => $ref_id,
4933 old_repo_id => $repo_id,
4934 old_path => $path };
4935 }
4936 }
4937
4938 my @emptied;
4939 foreach my $url (keys %$new_urls) {
4940 # see if we can re-use an existing [svn-remote "repo_id"]
4941 # instead of creating a(n ugly) new section:
4942 my $repo_id = $root_repos->{$url} || $url;
4943
4944 my $fetch = $new_urls->{$url};
4945 foreach my $path (keys %$fetch) {
4946 my $x = $fetch->{$path};
4947 Git::SVN->init($url, $path, $repo_id, $x->{ref_id});
4948 my $pfx = "svn-remote.$x->{old_repo_id}";
4949
4950 my $old_fetch = quotemeta("$x->{old_path}:".
4951 "$x->{ref_id}");
4952 command_noisy(qw/config --unset/,
4953 "$pfx.fetch", '^'. $old_fetch . '$');
4954 delete $r->{$x->{old_repo_id}}->
4955 {fetch}->{$x->{old_path}};
4956 if (!keys %{$r->{$x->{old_repo_id}}->{fetch}}) {
4957 command_noisy(qw/config --unset/,
4958 "$pfx.url");
4959 push @emptied, $x->{old_repo_id}
4960 }
4961 }
4962 }
4963 if (@emptied) {
4964 my $file = $ENV{GIT_CONFIG} || "$ENV{GIT_DIR}/config";
4965 print STDERR <<EOF;
4966 The following [svn-remote] sections in your config file ($file) are empty
4967 and can be safely removed:
4968 EOF
4969 print STDERR "[svn-remote \"$_\"]\n" foreach @emptied;
4970 }
4971 }
4972
4973 sub migration_check {
4974 migrate_from_v0();
4975 migrate_from_v1();
4976 migrate_from_v2();
4977 minimize_connections() if $_minimize;
4978 }
4979
4980 package Git::IndexInfo;
4981 use strict;
4982 use warnings;
4983 use Git qw/command_input_pipe command_close_pipe/;
4984
4985 sub new {
4986 my ($class) = @_;
4987 my ($gui, $ctx) = command_input_pipe(qw/update-index -z --index-info/);
4988 bless { gui => $gui, ctx => $ctx, nr => 0}, $class;
4989 }
4990
4991 sub remove {
4992 my ($self, $path) = @_;
4993 if (print { $self->{gui} } '0 ', 0 x 40, "\t", $path, "\0") {
4994 return ++$self->{nr};
4995 }
4996 undef;
4997 }
4998
4999 sub update {
5000 my ($self, $mode, $hash, $path) = @_;
5001 if (print { $self->{gui} } $mode, ' ', $hash, "\t", $path, "\0") {
5002 return ++$self->{nr};
5003 }
5004 undef;
5005 }
5006
5007 sub DESTROY {
5008 my ($self) = @_;
5009 command_close_pipe($self->{gui}, $self->{ctx});
5010 }
5011
5012 package Git::SVN::GlobSpec;
5013 use strict;
5014 use warnings;
5015
5016 sub new {
5017 my ($class, $glob, $pattern_ok) = @_;
5018 my $re = $glob;
5019 $re =~ s!/+$!!g; # no need for trailing slashes
5020 my (@left, @right, @patterns);
5021 my $state = "left";
5022 my $die_msg = "Only one set of wildcard directories " .
5023 "(e.g. '*' or '*/*/*') is supported: '$glob'\n";
5024 for my $part (split(m|/|, $glob)) {
5025 if ($part =~ /\*/ && $part ne "*") {
5026 die "Invalid pattern in '$glob': $part\n";
5027 } elsif ($pattern_ok && $part =~ /[{}]/ &&
5028 $part !~ /^\{[^{}]+\}/) {
5029 die "Invalid pattern in '$glob': $part\n";
5030 }
5031 if ($part eq "*") {
5032 die $die_msg if $state eq "right";
5033 $state = "pattern";
5034 push(@patterns, "[^/]*");
5035 } elsif ($pattern_ok && $part =~ /^\{(.*)\}$/) {
5036 die $die_msg if $state eq "right";
5037 $state = "pattern";
5038 my $p = quotemeta($1);
5039 $p =~ s/\\,/|/g;
5040 push(@patterns, "(?:$p)");
5041 } else {
5042 if ($state eq "left") {
5043 push(@left, $part);
5044 } else {
5045 push(@right, $part);
5046 $state = "right";
5047 }
5048 }
5049 }
5050 my $depth = @patterns;
5051 if ($depth == 0) {
5052 die "One '*' is needed in glob: '$glob'\n";
5053 }
5054 my $left = join('/', @left);
5055 my $right = join('/', @right);
5056 $re = join('/', @patterns);
5057 $re = join('\/',
5058 grep(length, quotemeta($left), "($re)", quotemeta($right)));
5059 my $left_re = qr/^\/\Q$left\E(\/|$)/;
5060 bless { left => $left, right => $right, left_regex => $left_re,
5061 regex => qr/$re/, glob => $glob, depth => $depth }, $class;
5062 }
5063
5064 sub full_path {
5065 my ($self, $path) = @_;
5066 return (length $self->{left} ? "$self->{left}/" : '') .
5067 $path . (length $self->{right} ? "/$self->{right}" : '');
5068 }
5069
5070 __END__
5071
5072 Data structures:
5073
5074
5075 $remotes = { # returned by read_all_remotes()
5076 'svn' => {
5077 # svn-remote.svn.url=https://svn.musicpd.org
5078 url => 'https://svn.musicpd.org',
5079 # svn-remote.svn.fetch=mpd/trunk:trunk
5080 fetch => {
5081 'mpd/trunk' => 'trunk',
5082 },
5083 # svn-remote.svn.tags=mpd/tags/*:tags/*
5084 tags => {
5085 path => {
5086 left => 'mpd/tags',
5087 right => '',
5088 regex => qr!mpd/tags/([^/]+)$!,
5089 glob => 'tags/*',
5090 },
5091 ref => {
5092 left => 'tags',
5093 right => '',
5094 regex => qr!tags/([^/]+)$!,
5095 glob => 'tags/*',
5096 },
5097 }
5098 }
5099 };
5100
5101 $log_entry hashref as returned by libsvn_log_entry()
5102 {
5103 log => 'whitespace-formatted log entry
5104 ', # trailing newline is preserved
5105 revision => '8', # integer
5106 date => '2004-02-24T17:01:44.108345Z', # commit date
5107 author => 'committer name'
5108 };
5109
5110
5111 # this is generated by generate_diff();
5112 @mods = array of diff-index line hashes, each element represents one line
5113 of diff-index output
5114
5115 diff-index line ($m hash)
5116 {
5117 mode_a => first column of diff-index output, no leading ':',
5118 mode_b => second column of diff-index output,
5119 sha1_b => sha1sum of the final blob,
5120 chg => change type [MCRADT],
5121 file_a => original file name of a file (iff chg is 'C' or 'R')
5122 file_b => new/current file name of a file (any chg)
5123 }
5124 ;
5125
5126 # retval of read_url_paths{,_all}();
5127 $l_map = {
5128 # repository root url
5129 'https://svn.musicpd.org' => {
5130 # repository path # GIT_SVN_ID
5131 'mpd/trunk' => 'trunk',
5132 'mpd/tags/0.11.5' => 'tags/0.11.5',
5133 },
5134 }
5135
5136 Notes:
5137 I don't trust the each() function on unless I created %hash myself
5138 because the internal iterator may not have started at base.
Gitweb development nest