Merge branch 'jn/mime-type-with-params' into maint
[git/git.git] / gitweb / gitweb.perl
1 #!/usr/bin/perl
2
3 # gitweb - simple web interface to track changes in git repositories
4 #
5 # (C) 2005-2006, Kay Sievers <kay.sievers@vrfy.org>
6 # (C) 2005, Christian Gierke
7 #
8 # This program is licensed under the GPLv2
9
10 use 5.008;
11 use strict;
12 use warnings;
13 use CGI qw(:standard :escapeHTML -nosticky);
14 use CGI::Util qw(unescape);
15 use CGI::Carp qw(fatalsToBrowser set_message);
16 use Encode;
17 use Fcntl ':mode';
18 use File::Find qw();
19 use File::Basename qw(basename);
20 use Time::HiRes qw(gettimeofday tv_interval);
21 binmode STDOUT, ':utf8';
22
23 our $t0 = [ gettimeofday() ];
24 our $number_of_git_cmds = 0;
25
26 BEGIN {
27 CGI->compile() if $ENV{'MOD_PERL'};
28 }
29
30 our $version = "++GIT_VERSION++";
31
32 our ($my_url, $my_uri, $base_url, $path_info, $home_link);
33 sub evaluate_uri {
34 our $cgi;
35
36 our $my_url = $cgi->url();
37 our $my_uri = $cgi->url(-absolute => 1);
38
39 # Base URL for relative URLs in gitweb ($logo, $favicon, ...),
40 # needed and used only for URLs with nonempty PATH_INFO
41 our $base_url = $my_url;
42
43 # When the script is used as DirectoryIndex, the URL does not contain the name
44 # of the script file itself, and $cgi->url() fails to strip PATH_INFO, so we
45 # have to do it ourselves. We make $path_info global because it's also used
46 # later on.
47 #
48 # Another issue with the script being the DirectoryIndex is that the resulting
49 # $my_url data is not the full script URL: this is good, because we want
50 # generated links to keep implying the script name if it wasn't explicitly
51 # indicated in the URL we're handling, but it means that $my_url cannot be used
52 # as base URL.
53 # Therefore, if we needed to strip PATH_INFO, then we know that we have
54 # to build the base URL ourselves:
55 our $path_info = $ENV{"PATH_INFO"};
56 if ($path_info) {
57 if ($my_url =~ s,\Q$path_info\E$,, &&
58 $my_uri =~ s,\Q$path_info\E$,, &&
59 defined $ENV{'SCRIPT_NAME'}) {
60 $base_url = $cgi->url(-base => 1) . $ENV{'SCRIPT_NAME'};
61 }
62 }
63
64 # target of the home link on top of all pages
65 our $home_link = $my_uri || "/";
66 }
67
68 # core git executable to use
69 # this can just be "git" if your webserver has a sensible PATH
70 our $GIT = "++GIT_BINDIR++/git";
71
72 # absolute fs-path which will be prepended to the project path
73 #our $projectroot = "/pub/scm";
74 our $projectroot = "++GITWEB_PROJECTROOT++";
75
76 # fs traversing limit for getting project list
77 # the number is relative to the projectroot
78 our $project_maxdepth = "++GITWEB_PROJECT_MAXDEPTH++";
79
80 # string of the home link on top of all pages
81 our $home_link_str = "++GITWEB_HOME_LINK_STR++";
82
83 # name of your site or organization to appear in page titles
84 # replace this with something more descriptive for clearer bookmarks
85 our $site_name = "++GITWEB_SITENAME++"
86 || ($ENV{'SERVER_NAME'} || "Untitled") . " Git";
87
88 # filename of html text to include at top of each page
89 our $site_header = "++GITWEB_SITE_HEADER++";
90 # html text to include at home page
91 our $home_text = "++GITWEB_HOMETEXT++";
92 # filename of html text to include at bottom of each page
93 our $site_footer = "++GITWEB_SITE_FOOTER++";
94
95 # URI of stylesheets
96 our @stylesheets = ("++GITWEB_CSS++");
97 # URI of a single stylesheet, which can be overridden in GITWEB_CONFIG.
98 our $stylesheet = undef;
99 # URI of GIT logo (72x27 size)
100 our $logo = "++GITWEB_LOGO++";
101 # URI of GIT favicon, assumed to be image/png type
102 our $favicon = "++GITWEB_FAVICON++";
103 # URI of gitweb.js (JavaScript code for gitweb)
104 our $javascript = "++GITWEB_JS++";
105
106 # URI and label (title) of GIT logo link
107 #our $logo_url = "http://www.kernel.org/pub/software/scm/git/docs/";
108 #our $logo_label = "git documentation";
109 our $logo_url = "http://git-scm.com/";
110 our $logo_label = "git homepage";
111
112 # source of projects list
113 our $projects_list = "++GITWEB_LIST++";
114
115 # the width (in characters) of the projects list "Description" column
116 our $projects_list_description_width = 25;
117
118 # group projects by category on the projects list
119 # (enabled if this variable evaluates to true)
120 our $projects_list_group_categories = 0;
121
122 # default category if none specified
123 # (leave the empty string for no category)
124 our $project_list_default_category = "";
125
126 # default order of projects list
127 # valid values are none, project, descr, owner, and age
128 our $default_projects_order = "project";
129
130 # show repository only if this file exists
131 # (only effective if this variable evaluates to true)
132 our $export_ok = "++GITWEB_EXPORT_OK++";
133
134 # show repository only if this subroutine returns true
135 # when given the path to the project, for example:
136 # sub { return -e "$_[0]/git-daemon-export-ok"; }
137 our $export_auth_hook = undef;
138
139 # only allow viewing of repositories also shown on the overview page
140 our $strict_export = "++GITWEB_STRICT_EXPORT++";
141
142 # list of git base URLs used for URL to where fetch project from,
143 # i.e. full URL is "$git_base_url/$project"
144 our @git_base_url_list = grep { $_ ne '' } ("++GITWEB_BASE_URL++");
145
146 # default blob_plain mimetype and default charset for text/plain blob
147 our $default_blob_plain_mimetype = 'text/plain';
148 our $default_text_plain_charset = undef;
149
150 # file to use for guessing MIME types before trying /etc/mime.types
151 # (relative to the current git repository)
152 our $mimetypes_file = undef;
153
154 # assume this charset if line contains non-UTF-8 characters;
155 # it should be valid encoding (see Encoding::Supported(3pm) for list),
156 # for which encoding all byte sequences are valid, for example
157 # 'iso-8859-1' aka 'latin1' (it is decoded without checking, so it
158 # could be even 'utf-8' for the old behavior)
159 our $fallback_encoding = 'latin1';
160
161 # rename detection options for git-diff and git-diff-tree
162 # - default is '-M', with the cost proportional to
163 # (number of removed files) * (number of new files).
164 # - more costly is '-C' (which implies '-M'), with the cost proportional to
165 # (number of changed files + number of removed files) * (number of new files)
166 # - even more costly is '-C', '--find-copies-harder' with cost
167 # (number of files in the original tree) * (number of new files)
168 # - one might want to include '-B' option, e.g. '-B', '-M'
169 our @diff_opts = ('-M'); # taken from git_commit
170
171 # Disables features that would allow repository owners to inject script into
172 # the gitweb domain.
173 our $prevent_xss = 0;
174
175 # Path to the highlight executable to use (must be the one from
176 # http://www.andre-simon.de due to assumptions about parameters and output).
177 # Useful if highlight is not installed on your webserver's PATH.
178 # [Default: highlight]
179 our $highlight_bin = "++HIGHLIGHT_BIN++";
180
181 # information about snapshot formats that gitweb is capable of serving
182 our %known_snapshot_formats = (
183 # name => {
184 # 'display' => display name,
185 # 'type' => mime type,
186 # 'suffix' => filename suffix,
187 # 'format' => --format for git-archive,
188 # 'compressor' => [compressor command and arguments]
189 # (array reference, optional)
190 # 'disabled' => boolean (optional)}
191 #
192 'tgz' => {
193 'display' => 'tar.gz',
194 'type' => 'application/x-gzip',
195 'suffix' => '.tar.gz',
196 'format' => 'tar',
197 'compressor' => ['gzip', '-n']},
198
199 'tbz2' => {
200 'display' => 'tar.bz2',
201 'type' => 'application/x-bzip2',
202 'suffix' => '.tar.bz2',
203 'format' => 'tar',
204 'compressor' => ['bzip2']},
205
206 'txz' => {
207 'display' => 'tar.xz',
208 'type' => 'application/x-xz',
209 'suffix' => '.tar.xz',
210 'format' => 'tar',
211 'compressor' => ['xz'],
212 'disabled' => 1},
213
214 'zip' => {
215 'display' => 'zip',
216 'type' => 'application/x-zip',
217 'suffix' => '.zip',
218 'format' => 'zip'},
219 );
220
221 # Aliases so we understand old gitweb.snapshot values in repository
222 # configuration.
223 our %known_snapshot_format_aliases = (
224 'gzip' => 'tgz',
225 'bzip2' => 'tbz2',
226 'xz' => 'txz',
227
228 # backward compatibility: legacy gitweb config support
229 'x-gzip' => undef, 'gz' => undef,
230 'x-bzip2' => undef, 'bz2' => undef,
231 'x-zip' => undef, '' => undef,
232 );
233
234 # Pixel sizes for icons and avatars. If the default font sizes or lineheights
235 # are changed, it may be appropriate to change these values too via
236 # $GITWEB_CONFIG.
237 our %avatar_size = (
238 'default' => 16,
239 'double' => 32
240 );
241
242 # Used to set the maximum load that we will still respond to gitweb queries.
243 # If server load exceed this value then return "503 server busy" error.
244 # If gitweb cannot determined server load, it is taken to be 0.
245 # Leave it undefined (or set to 'undef') to turn off load checking.
246 our $maxload = 300;
247
248 # configuration for 'highlight' (http://www.andre-simon.de/)
249 # match by basename
250 our %highlight_basename = (
251 #'Program' => 'py',
252 #'Library' => 'py',
253 'SConstruct' => 'py', # SCons equivalent of Makefile
254 'Makefile' => 'make',
255 );
256 # match by extension
257 our %highlight_ext = (
258 # main extensions, defining name of syntax;
259 # see files in /usr/share/highlight/langDefs/ directory
260 map { $_ => $_ }
261 qw(py c cpp rb java css php sh pl js tex bib xml awk bat ini spec tcl sql make),
262 # alternate extensions, see /etc/highlight/filetypes.conf
263 'h' => 'c',
264 map { $_ => 'sh' } qw(bash zsh ksh),
265 map { $_ => 'cpp' } qw(cxx c++ cc),
266 map { $_ => 'php' } qw(php3 php4 php5 phps),
267 map { $_ => 'pl' } qw(perl pm), # perhaps also 'cgi'
268 map { $_ => 'make'} qw(mak mk),
269 map { $_ => 'xml' } qw(xhtml html htm),
270 );
271
272 # You define site-wide feature defaults here; override them with
273 # $GITWEB_CONFIG as necessary.
274 our %feature = (
275 # feature => {
276 # 'sub' => feature-sub (subroutine),
277 # 'override' => allow-override (boolean),
278 # 'default' => [ default options...] (array reference)}
279 #
280 # if feature is overridable (it means that allow-override has true value),
281 # then feature-sub will be called with default options as parameters;
282 # return value of feature-sub indicates if to enable specified feature
283 #
284 # if there is no 'sub' key (no feature-sub), then feature cannot be
285 # overridden
286 #
287 # use gitweb_get_feature(<feature>) to retrieve the <feature> value
288 # (an array) or gitweb_check_feature(<feature>) to check if <feature>
289 # is enabled
290
291 # Enable the 'blame' blob view, showing the last commit that modified
292 # each line in the file. This can be very CPU-intensive.
293
294 # To enable system wide have in $GITWEB_CONFIG
295 # $feature{'blame'}{'default'} = [1];
296 # To have project specific config enable override in $GITWEB_CONFIG
297 # $feature{'blame'}{'override'} = 1;
298 # and in project config gitweb.blame = 0|1;
299 'blame' => {
300 'sub' => sub { feature_bool('blame', @_) },
301 'override' => 0,
302 'default' => [0]},
303
304 # Enable the 'snapshot' link, providing a compressed archive of any
305 # tree. This can potentially generate high traffic if you have large
306 # project.
307
308 # Value is a list of formats defined in %known_snapshot_formats that
309 # you wish to offer.
310 # To disable system wide have in $GITWEB_CONFIG
311 # $feature{'snapshot'}{'default'} = [];
312 # To have project specific config enable override in $GITWEB_CONFIG
313 # $feature{'snapshot'}{'override'} = 1;
314 # and in project config, a comma-separated list of formats or "none"
315 # to disable. Example: gitweb.snapshot = tbz2,zip;
316 'snapshot' => {
317 'sub' => \&feature_snapshot,
318 'override' => 0,
319 'default' => ['tgz']},
320
321 # Enable text search, which will list the commits which match author,
322 # committer or commit text to a given string. Enabled by default.
323 # Project specific override is not supported.
324 'search' => {
325 'override' => 0,
326 'default' => [1]},
327
328 # Enable grep search, which will list the files in currently selected
329 # tree containing the given string. Enabled by default. This can be
330 # potentially CPU-intensive, of course.
331 # Note that you need to have 'search' feature enabled too.
332
333 # To enable system wide have in $GITWEB_CONFIG
334 # $feature{'grep'}{'default'} = [1];
335 # To have project specific config enable override in $GITWEB_CONFIG
336 # $feature{'grep'}{'override'} = 1;
337 # and in project config gitweb.grep = 0|1;
338 'grep' => {
339 'sub' => sub { feature_bool('grep', @_) },
340 'override' => 0,
341 'default' => [1]},
342
343 # Enable the pickaxe search, which will list the commits that modified
344 # a given string in a file. This can be practical and quite faster
345 # alternative to 'blame', but still potentially CPU-intensive.
346 # Note that you need to have 'search' feature enabled too.
347
348 # To enable system wide have in $GITWEB_CONFIG
349 # $feature{'pickaxe'}{'default'} = [1];
350 # To have project specific config enable override in $GITWEB_CONFIG
351 # $feature{'pickaxe'}{'override'} = 1;
352 # and in project config gitweb.pickaxe = 0|1;
353 'pickaxe' => {
354 'sub' => sub { feature_bool('pickaxe', @_) },
355 'override' => 0,
356 'default' => [1]},
357
358 # Enable showing size of blobs in a 'tree' view, in a separate
359 # column, similar to what 'ls -l' does. This cost a bit of IO.
360
361 # To disable system wide have in $GITWEB_CONFIG
362 # $feature{'show-sizes'}{'default'} = [0];
363 # To have project specific config enable override in $GITWEB_CONFIG
364 # $feature{'show-sizes'}{'override'} = 1;
365 # and in project config gitweb.showsizes = 0|1;
366 'show-sizes' => {
367 'sub' => sub { feature_bool('showsizes', @_) },
368 'override' => 0,
369 'default' => [1]},
370
371 # Make gitweb use an alternative format of the URLs which can be
372 # more readable and natural-looking: project name is embedded
373 # directly in the path and the query string contains other
374 # auxiliary information. All gitweb installations recognize
375 # URL in either format; this configures in which formats gitweb
376 # generates links.
377
378 # To enable system wide have in $GITWEB_CONFIG
379 # $feature{'pathinfo'}{'default'} = [1];
380 # Project specific override is not supported.
381
382 # Note that you will need to change the default location of CSS,
383 # favicon, logo and possibly other files to an absolute URL. Also,
384 # if gitweb.cgi serves as your indexfile, you will need to force
385 # $my_uri to contain the script name in your $GITWEB_CONFIG.
386 'pathinfo' => {
387 'override' => 0,
388 'default' => [0]},
389
390 # Make gitweb consider projects in project root subdirectories
391 # to be forks of existing projects. Given project $projname.git,
392 # projects matching $projname/*.git will not be shown in the main
393 # projects list, instead a '+' mark will be added to $projname
394 # there and a 'forks' view will be enabled for the project, listing
395 # all the forks. If project list is taken from a file, forks have
396 # to be listed after the main project.
397
398 # To enable system wide have in $GITWEB_CONFIG
399 # $feature{'forks'}{'default'} = [1];
400 # Project specific override is not supported.
401 'forks' => {
402 'override' => 0,
403 'default' => [0]},
404
405 # Insert custom links to the action bar of all project pages.
406 # This enables you mainly to link to third-party scripts integrating
407 # into gitweb; e.g. git-browser for graphical history representation
408 # or custom web-based repository administration interface.
409
410 # The 'default' value consists of a list of triplets in the form
411 # (label, link, position) where position is the label after which
412 # to insert the link and link is a format string where %n expands
413 # to the project name, %f to the project path within the filesystem,
414 # %h to the current hash (h gitweb parameter) and %b to the current
415 # hash base (hb gitweb parameter); %% expands to %.
416
417 # To enable system wide have in $GITWEB_CONFIG e.g.
418 # $feature{'actions'}{'default'} = [('graphiclog',
419 # '/git-browser/by-commit.html?r=%n', 'summary')];
420 # Project specific override is not supported.
421 'actions' => {
422 'override' => 0,
423 'default' => []},
424
425 # Allow gitweb scan project content tags of project repository,
426 # and display the popular Web 2.0-ish "tag cloud" near the projects
427 # list. Note that this is something COMPLETELY different from the
428 # normal Git tags.
429
430 # gitweb by itself can show existing tags, but it does not handle
431 # tagging itself; you need to do it externally, outside gitweb.
432 # The format is described in git_get_project_ctags() subroutine.
433 # You may want to install the HTML::TagCloud Perl module to get
434 # a pretty tag cloud instead of just a list of tags.
435
436 # To enable system wide have in $GITWEB_CONFIG
437 # $feature{'ctags'}{'default'} = [1];
438 # Project specific override is not supported.
439
440 # In the future whether ctags editing is enabled might depend
441 # on the value, but using 1 should always mean no editing of ctags.
442 'ctags' => {
443 'override' => 0,
444 'default' => [0]},
445
446 # The maximum number of patches in a patchset generated in patch
447 # view. Set this to 0 or undef to disable patch view, or to a
448 # negative number to remove any limit.
449
450 # To disable system wide have in $GITWEB_CONFIG
451 # $feature{'patches'}{'default'} = [0];
452 # To have project specific config enable override in $GITWEB_CONFIG
453 # $feature{'patches'}{'override'} = 1;
454 # and in project config gitweb.patches = 0|n;
455 # where n is the maximum number of patches allowed in a patchset.
456 'patches' => {
457 'sub' => \&feature_patches,
458 'override' => 0,
459 'default' => [16]},
460
461 # Avatar support. When this feature is enabled, views such as
462 # shortlog or commit will display an avatar associated with
463 # the email of the committer(s) and/or author(s).
464
465 # Currently available providers are gravatar and picon.
466 # If an unknown provider is specified, the feature is disabled.
467
468 # Gravatar depends on Digest::MD5.
469 # Picon currently relies on the indiana.edu database.
470
471 # To enable system wide have in $GITWEB_CONFIG
472 # $feature{'avatar'}{'default'} = ['<provider>'];
473 # where <provider> is either gravatar or picon.
474 # To have project specific config enable override in $GITWEB_CONFIG
475 # $feature{'avatar'}{'override'} = 1;
476 # and in project config gitweb.avatar = <provider>;
477 'avatar' => {
478 'sub' => \&feature_avatar,
479 'override' => 0,
480 'default' => ['']},
481
482 # Enable displaying how much time and how many git commands
483 # it took to generate and display page. Disabled by default.
484 # Project specific override is not supported.
485 'timed' => {
486 'override' => 0,
487 'default' => [0]},
488
489 # Enable turning some links into links to actions which require
490 # JavaScript to run (like 'blame_incremental'). Not enabled by
491 # default. Project specific override is currently not supported.
492 'javascript-actions' => {
493 'override' => 0,
494 'default' => [0]},
495
496 # Enable and configure ability to change common timezone for dates
497 # in gitweb output via JavaScript. Enabled by default.
498 # Project specific override is not supported.
499 'javascript-timezone' => {
500 'override' => 0,
501 'default' => [
502 'local', # default timezone: 'utc', 'local', or '(-|+)HHMM' format,
503 # or undef to turn off this feature
504 'gitweb_tz', # name of cookie where to store selected timezone
505 'datetime', # CSS class used to mark up dates for manipulation
506 ]},
507
508 # Syntax highlighting support. This is based on Daniel Svensson's
509 # and Sham Chukoury's work in gitweb-xmms2.git.
510 # It requires the 'highlight' program present in $PATH,
511 # and therefore is disabled by default.
512
513 # To enable system wide have in $GITWEB_CONFIG
514 # $feature{'highlight'}{'default'} = [1];
515
516 'highlight' => {
517 'sub' => sub { feature_bool('highlight', @_) },
518 'override' => 0,
519 'default' => [0]},
520
521 # Enable displaying of remote heads in the heads list
522
523 # To enable system wide have in $GITWEB_CONFIG
524 # $feature{'remote_heads'}{'default'} = [1];
525 # To have project specific config enable override in $GITWEB_CONFIG
526 # $feature{'remote_heads'}{'override'} = 1;
527 # and in project config gitweb.remote_heads = 0|1;
528 'remote_heads' => {
529 'sub' => sub { feature_bool('remote_heads', @_) },
530 'override' => 0,
531 'default' => [0]},
532 );
533
534 sub gitweb_get_feature {
535 my ($name) = @_;
536 return unless exists $feature{$name};
537 my ($sub, $override, @defaults) = (
538 $feature{$name}{'sub'},
539 $feature{$name}{'override'},
540 @{$feature{$name}{'default'}});
541 # project specific override is possible only if we have project
542 our $git_dir; # global variable, declared later
543 if (!$override || !defined $git_dir) {
544 return @defaults;
545 }
546 if (!defined $sub) {
547 warn "feature $name is not overridable";
548 return @defaults;
549 }
550 return $sub->(@defaults);
551 }
552
553 # A wrapper to check if a given feature is enabled.
554 # With this, you can say
555 #
556 # my $bool_feat = gitweb_check_feature('bool_feat');
557 # gitweb_check_feature('bool_feat') or somecode;
558 #
559 # instead of
560 #
561 # my ($bool_feat) = gitweb_get_feature('bool_feat');
562 # (gitweb_get_feature('bool_feat'))[0] or somecode;
563 #
564 sub gitweb_check_feature {
565 return (gitweb_get_feature(@_))[0];
566 }
567
568
569 sub feature_bool {
570 my $key = shift;
571 my ($val) = git_get_project_config($key, '--bool');
572
573 if (!defined $val) {
574 return ($_[0]);
575 } elsif ($val eq 'true') {
576 return (1);
577 } elsif ($val eq 'false') {
578 return (0);
579 }
580 }
581
582 sub feature_snapshot {
583 my (@fmts) = @_;
584
585 my ($val) = git_get_project_config('snapshot');
586
587 if ($val) {
588 @fmts = ($val eq 'none' ? () : split /\s*[,\s]\s*/, $val);
589 }
590
591 return @fmts;
592 }
593
594 sub feature_patches {
595 my @val = (git_get_project_config('patches', '--int'));
596
597 if (@val) {
598 return @val;
599 }
600
601 return ($_[0]);
602 }
603
604 sub feature_avatar {
605 my @val = (git_get_project_config('avatar'));
606
607 return @val ? @val : @_;
608 }
609
610 # checking HEAD file with -e is fragile if the repository was
611 # initialized long time ago (i.e. symlink HEAD) and was pack-ref'ed
612 # and then pruned.
613 sub check_head_link {
614 my ($dir) = @_;
615 my $headfile = "$dir/HEAD";
616 return ((-e $headfile) ||
617 (-l $headfile && readlink($headfile) =~ /^refs\/heads\//));
618 }
619
620 sub check_export_ok {
621 my ($dir) = @_;
622 return (check_head_link($dir) &&
623 (!$export_ok || -e "$dir/$export_ok") &&
624 (!$export_auth_hook || $export_auth_hook->($dir)));
625 }
626
627 # process alternate names for backward compatibility
628 # filter out unsupported (unknown) snapshot formats
629 sub filter_snapshot_fmts {
630 my @fmts = @_;
631
632 @fmts = map {
633 exists $known_snapshot_format_aliases{$_} ?
634 $known_snapshot_format_aliases{$_} : $_} @fmts;
635 @fmts = grep {
636 exists $known_snapshot_formats{$_} &&
637 !$known_snapshot_formats{$_}{'disabled'}} @fmts;
638 }
639
640 # If it is set to code reference, it is code that it is to be run once per
641 # request, allowing updating configurations that change with each request,
642 # while running other code in config file only once.
643 #
644 # Otherwise, if it is false then gitweb would process config file only once;
645 # if it is true then gitweb config would be run for each request.
646 our $per_request_config = 1;
647
648 # read and parse gitweb config file given by its parameter.
649 # returns true on success, false on recoverable error, allowing
650 # to chain this subroutine, using first file that exists.
651 # dies on errors during parsing config file, as it is unrecoverable.
652 sub read_config_file {
653 my $filename = shift;
654 return unless defined $filename;
655 # die if there are errors parsing config file
656 if (-e $filename) {
657 do $filename;
658 die $@ if $@;
659 return 1;
660 }
661 return;
662 }
663
664 our ($GITWEB_CONFIG, $GITWEB_CONFIG_SYSTEM);
665 sub evaluate_gitweb_config {
666 our $GITWEB_CONFIG = $ENV{'GITWEB_CONFIG'} || "++GITWEB_CONFIG++";
667 our $GITWEB_CONFIG_SYSTEM = $ENV{'GITWEB_CONFIG_SYSTEM'} || "++GITWEB_CONFIG_SYSTEM++";
668
669 # use first config file that exists
670 read_config_file($GITWEB_CONFIG) or
671 read_config_file($GITWEB_CONFIG_SYSTEM);
672 }
673
674 # Get loadavg of system, to compare against $maxload.
675 # Currently it requires '/proc/loadavg' present to get loadavg;
676 # if it is not present it returns 0, which means no load checking.
677 sub get_loadavg {
678 if( -e '/proc/loadavg' ){
679 open my $fd, '<', '/proc/loadavg'
680 or return 0;
681 my @load = split(/\s+/, scalar <$fd>);
682 close $fd;
683
684 # The first three columns measure CPU and IO utilization of the last one,
685 # five, and 10 minute periods. The fourth column shows the number of
686 # currently running processes and the total number of processes in the m/n
687 # format. The last column displays the last process ID used.
688 return $load[0] || 0;
689 }
690 # additional checks for load average should go here for things that don't export
691 # /proc/loadavg
692
693 return 0;
694 }
695
696 # version of the core git binary
697 our $git_version;
698 sub evaluate_git_version {
699 our $git_version = qx("$GIT" --version) =~ m/git version (.*)$/ ? $1 : "unknown";
700 $number_of_git_cmds++;
701 }
702
703 sub check_loadavg {
704 if (defined $maxload && get_loadavg() > $maxload) {
705 die_error(503, "The load average on the server is too high");
706 }
707 }
708
709 # ======================================================================
710 # input validation and dispatch
711
712 # input parameters can be collected from a variety of sources (presently, CGI
713 # and PATH_INFO), so we define an %input_params hash that collects them all
714 # together during validation: this allows subsequent uses (e.g. href()) to be
715 # agnostic of the parameter origin
716
717 our %input_params = ();
718
719 # input parameters are stored with the long parameter name as key. This will
720 # also be used in the href subroutine to convert parameters to their CGI
721 # equivalent, and since the href() usage is the most frequent one, we store
722 # the name -> CGI key mapping here, instead of the reverse.
723 #
724 # XXX: Warning: If you touch this, check the search form for updating,
725 # too.
726
727 our @cgi_param_mapping = (
728 project => "p",
729 action => "a",
730 file_name => "f",
731 file_parent => "fp",
732 hash => "h",
733 hash_parent => "hp",
734 hash_base => "hb",
735 hash_parent_base => "hpb",
736 page => "pg",
737 order => "o",
738 searchtext => "s",
739 searchtype => "st",
740 snapshot_format => "sf",
741 extra_options => "opt",
742 search_use_regexp => "sr",
743 ctag => "by_tag",
744 # this must be last entry (for manipulation from JavaScript)
745 javascript => "js"
746 );
747 our %cgi_param_mapping = @cgi_param_mapping;
748
749 # we will also need to know the possible actions, for validation
750 our %actions = (
751 "blame" => \&git_blame,
752 "blame_incremental" => \&git_blame_incremental,
753 "blame_data" => \&git_blame_data,
754 "blobdiff" => \&git_blobdiff,
755 "blobdiff_plain" => \&git_blobdiff_plain,
756 "blob" => \&git_blob,
757 "blob_plain" => \&git_blob_plain,
758 "commitdiff" => \&git_commitdiff,
759 "commitdiff_plain" => \&git_commitdiff_plain,
760 "commit" => \&git_commit,
761 "forks" => \&git_forks,
762 "heads" => \&git_heads,
763 "history" => \&git_history,
764 "log" => \&git_log,
765 "patch" => \&git_patch,
766 "patches" => \&git_patches,
767 "remotes" => \&git_remotes,
768 "rss" => \&git_rss,
769 "atom" => \&git_atom,
770 "search" => \&git_search,
771 "search_help" => \&git_search_help,
772 "shortlog" => \&git_shortlog,
773 "summary" => \&git_summary,
774 "tag" => \&git_tag,
775 "tags" => \&git_tags,
776 "tree" => \&git_tree,
777 "snapshot" => \&git_snapshot,
778 "object" => \&git_object,
779 # those below don't need $project
780 "opml" => \&git_opml,
781 "project_list" => \&git_project_list,
782 "project_index" => \&git_project_index,
783 );
784
785 # finally, we have the hash of allowed extra_options for the commands that
786 # allow them
787 our %allowed_options = (
788 "--no-merges" => [ qw(rss atom log shortlog history) ],
789 );
790
791 # fill %input_params with the CGI parameters. All values except for 'opt'
792 # should be single values, but opt can be an array. We should probably
793 # build an array of parameters that can be multi-valued, but since for the time
794 # being it's only this one, we just single it out
795 sub evaluate_query_params {
796 our $cgi;
797
798 while (my ($name, $symbol) = each %cgi_param_mapping) {
799 if ($symbol eq 'opt') {
800 $input_params{$name} = [ $cgi->param($symbol) ];
801 } else {
802 $input_params{$name} = $cgi->param($symbol);
803 }
804 }
805 }
806
807 # now read PATH_INFO and update the parameter list for missing parameters
808 sub evaluate_path_info {
809 return if defined $input_params{'project'};
810 return if !$path_info;
811 $path_info =~ s,^/+,,;
812 return if !$path_info;
813
814 # find which part of PATH_INFO is project
815 my $project = $path_info;
816 $project =~ s,/+$,,;
817 while ($project && !check_head_link("$projectroot/$project")) {
818 $project =~ s,/*[^/]*$,,;
819 }
820 return unless $project;
821 $input_params{'project'} = $project;
822
823 # do not change any parameters if an action is given using the query string
824 return if $input_params{'action'};
825 $path_info =~ s,^\Q$project\E/*,,;
826
827 # next, check if we have an action
828 my $action = $path_info;
829 $action =~ s,/.*$,,;
830 if (exists $actions{$action}) {
831 $path_info =~ s,^$action/*,,;
832 $input_params{'action'} = $action;
833 }
834
835 # list of actions that want hash_base instead of hash, but can have no
836 # pathname (f) parameter
837 my @wants_base = (
838 'tree',
839 'history',
840 );
841
842 # we want to catch, among others
843 # [$hash_parent_base[:$file_parent]..]$hash_parent[:$file_name]
844 my ($parentrefname, $parentpathname, $refname, $pathname) =
845 ($path_info =~ /^(?:(.+?)(?::(.+))?\.\.)?([^:]+?)?(?::(.+))?$/);
846
847 # first, analyze the 'current' part
848 if (defined $pathname) {
849 # we got "branch:filename" or "branch:dir/"
850 # we could use git_get_type(branch:pathname), but:
851 # - it needs $git_dir
852 # - it does a git() call
853 # - the convention of terminating directories with a slash
854 # makes it superfluous
855 # - embedding the action in the PATH_INFO would make it even
856 # more superfluous
857 $pathname =~ s,^/+,,;
858 if (!$pathname || substr($pathname, -1) eq "/") {
859 $input_params{'action'} ||= "tree";
860 $pathname =~ s,/$,,;
861 } else {
862 # the default action depends on whether we had parent info
863 # or not
864 if ($parentrefname) {
865 $input_params{'action'} ||= "blobdiff_plain";
866 } else {
867 $input_params{'action'} ||= "blob_plain";
868 }
869 }
870 $input_params{'hash_base'} ||= $refname;
871 $input_params{'file_name'} ||= $pathname;
872 } elsif (defined $refname) {
873 # we got "branch". In this case we have to choose if we have to
874 # set hash or hash_base.
875 #
876 # Most of the actions without a pathname only want hash to be
877 # set, except for the ones specified in @wants_base that want
878 # hash_base instead. It should also be noted that hand-crafted
879 # links having 'history' as an action and no pathname or hash
880 # set will fail, but that happens regardless of PATH_INFO.
881 if (defined $parentrefname) {
882 # if there is parent let the default be 'shortlog' action
883 # (for http://git.example.com/repo.git/A..B links); if there
884 # is no parent, dispatch will detect type of object and set
885 # action appropriately if required (if action is not set)
886 $input_params{'action'} ||= "shortlog";
887 }
888 if ($input_params{'action'} &&
889 grep { $_ eq $input_params{'action'} } @wants_base) {
890 $input_params{'hash_base'} ||= $refname;
891 } else {
892 $input_params{'hash'} ||= $refname;
893 }
894 }
895
896 # next, handle the 'parent' part, if present
897 if (defined $parentrefname) {
898 # a missing pathspec defaults to the 'current' filename, allowing e.g.
899 # someproject/blobdiff/oldrev..newrev:/filename
900 if ($parentpathname) {
901 $parentpathname =~ s,^/+,,;
902 $parentpathname =~ s,/$,,;
903 $input_params{'file_parent'} ||= $parentpathname;
904 } else {
905 $input_params{'file_parent'} ||= $input_params{'file_name'};
906 }
907 # we assume that hash_parent_base is wanted if a path was specified,
908 # or if the action wants hash_base instead of hash
909 if (defined $input_params{'file_parent'} ||
910 grep { $_ eq $input_params{'action'} } @wants_base) {
911 $input_params{'hash_parent_base'} ||= $parentrefname;
912 } else {
913 $input_params{'hash_parent'} ||= $parentrefname;
914 }
915 }
916
917 # for the snapshot action, we allow URLs in the form
918 # $project/snapshot/$hash.ext
919 # where .ext determines the snapshot and gets removed from the
920 # passed $refname to provide the $hash.
921 #
922 # To be able to tell that $refname includes the format extension, we
923 # require the following two conditions to be satisfied:
924 # - the hash input parameter MUST have been set from the $refname part
925 # of the URL (i.e. they must be equal)
926 # - the snapshot format MUST NOT have been defined already (e.g. from
927 # CGI parameter sf)
928 # It's also useless to try any matching unless $refname has a dot,
929 # so we check for that too
930 if (defined $input_params{'action'} &&
931 $input_params{'action'} eq 'snapshot' &&
932 defined $refname && index($refname, '.') != -1 &&
933 $refname eq $input_params{'hash'} &&
934 !defined $input_params{'snapshot_format'}) {
935 # We loop over the known snapshot formats, checking for
936 # extensions. Allowed extensions are both the defined suffix
937 # (which includes the initial dot already) and the snapshot
938 # format key itself, with a prepended dot
939 while (my ($fmt, $opt) = each %known_snapshot_formats) {
940 my $hash = $refname;
941 unless ($hash =~ s/(\Q$opt->{'suffix'}\E|\Q.$fmt\E)$//) {
942 next;
943 }
944 my $sfx = $1;
945 # a valid suffix was found, so set the snapshot format
946 # and reset the hash parameter
947 $input_params{'snapshot_format'} = $fmt;
948 $input_params{'hash'} = $hash;
949 # we also set the format suffix to the one requested
950 # in the URL: this way a request for e.g. .tgz returns
951 # a .tgz instead of a .tar.gz
952 $known_snapshot_formats{$fmt}{'suffix'} = $sfx;
953 last;
954 }
955 }
956 }
957
958 our ($action, $project, $file_name, $file_parent, $hash, $hash_parent, $hash_base,
959 $hash_parent_base, @extra_options, $page, $searchtype, $search_use_regexp,
960 $searchtext, $search_regexp);
961 sub evaluate_and_validate_params {
962 our $action = $input_params{'action'};
963 if (defined $action) {
964 if (!validate_action($action)) {
965 die_error(400, "Invalid action parameter");
966 }
967 }
968
969 # parameters which are pathnames
970 our $project = $input_params{'project'};
971 if (defined $project) {
972 if (!validate_project($project)) {
973 undef $project;
974 die_error(404, "No such project");
975 }
976 }
977
978 our $file_name = $input_params{'file_name'};
979 if (defined $file_name) {
980 if (!validate_pathname($file_name)) {
981 die_error(400, "Invalid file parameter");
982 }
983 }
984
985 our $file_parent = $input_params{'file_parent'};
986 if (defined $file_parent) {
987 if (!validate_pathname($file_parent)) {
988 die_error(400, "Invalid file parent parameter");
989 }
990 }
991
992 # parameters which are refnames
993 our $hash = $input_params{'hash'};
994 if (defined $hash) {
995 if (!validate_refname($hash)) {
996 die_error(400, "Invalid hash parameter");
997 }
998 }
999
1000 our $hash_parent = $input_params{'hash_parent'};
1001 if (defined $hash_parent) {
1002 if (!validate_refname($hash_parent)) {
1003 die_error(400, "Invalid hash parent parameter");
1004 }
1005 }
1006
1007 our $hash_base = $input_params{'hash_base'};
1008 if (defined $hash_base) {
1009 if (!validate_refname($hash_base)) {
1010 die_error(400, "Invalid hash base parameter");
1011 }
1012 }
1013
1014 our @extra_options = @{$input_params{'extra_options'}};
1015 # @extra_options is always defined, since it can only be (currently) set from
1016 # CGI, and $cgi->param() returns the empty array in array context if the param
1017 # is not set
1018 foreach my $opt (@extra_options) {
1019 if (not exists $allowed_options{$opt}) {
1020 die_error(400, "Invalid option parameter");
1021 }
1022 if (not grep(/^$action$/, @{$allowed_options{$opt}})) {
1023 die_error(400, "Invalid option parameter for this action");
1024 }
1025 }
1026
1027 our $hash_parent_base = $input_params{'hash_parent_base'};
1028 if (defined $hash_parent_base) {
1029 if (!validate_refname($hash_parent_base)) {
1030 die_error(400, "Invalid hash parent base parameter");
1031 }
1032 }
1033
1034 # other parameters
1035 our $page = $input_params{'page'};
1036 if (defined $page) {
1037 if ($page =~ m/[^0-9]/) {
1038 die_error(400, "Invalid page parameter");
1039 }
1040 }
1041
1042 our $searchtype = $input_params{'searchtype'};
1043 if (defined $searchtype) {
1044 if ($searchtype =~ m/[^a-z]/) {
1045 die_error(400, "Invalid searchtype parameter");
1046 }
1047 }
1048
1049 our $search_use_regexp = $input_params{'search_use_regexp'};
1050
1051 our $searchtext = $input_params{'searchtext'};
1052 our $search_regexp;
1053 if (defined $searchtext) {
1054 if (length($searchtext) < 2) {
1055 die_error(403, "At least two characters are required for search parameter");
1056 }
1057 $search_regexp = $search_use_regexp ? $searchtext : quotemeta $searchtext;
1058 }
1059 }
1060
1061 # path to the current git repository
1062 our $git_dir;
1063 sub evaluate_git_dir {
1064 our $git_dir = "$projectroot/$project" if $project;
1065 }
1066
1067 our (@snapshot_fmts, $git_avatar);
1068 sub configure_gitweb_features {
1069 # list of supported snapshot formats
1070 our @snapshot_fmts = gitweb_get_feature('snapshot');
1071 @snapshot_fmts = filter_snapshot_fmts(@snapshot_fmts);
1072
1073 # check that the avatar feature is set to a known provider name,
1074 # and for each provider check if the dependencies are satisfied.
1075 # if the provider name is invalid or the dependencies are not met,
1076 # reset $git_avatar to the empty string.
1077 our ($git_avatar) = gitweb_get_feature('avatar');
1078 if ($git_avatar eq 'gravatar') {
1079 $git_avatar = '' unless (eval { require Digest::MD5; 1; });
1080 } elsif ($git_avatar eq 'picon') {
1081 # no dependencies
1082 } else {
1083 $git_avatar = '';
1084 }
1085 }
1086
1087 # custom error handler: 'die <message>' is Internal Server Error
1088 sub handle_errors_html {
1089 my $msg = shift; # it is already HTML escaped
1090
1091 # to avoid infinite loop where error occurs in die_error,
1092 # change handler to default handler, disabling handle_errors_html
1093 set_message("Error occured when inside die_error:\n$msg");
1094
1095 # you cannot jump out of die_error when called as error handler;
1096 # the subroutine set via CGI::Carp::set_message is called _after_
1097 # HTTP headers are already written, so it cannot write them itself
1098 die_error(undef, undef, $msg, -error_handler => 1, -no_http_header => 1);
1099 }
1100 set_message(\&handle_errors_html);
1101
1102 # dispatch
1103 sub dispatch {
1104 if (!defined $action) {
1105 if (defined $hash) {
1106 $action = git_get_type($hash);
1107 } elsif (defined $hash_base && defined $file_name) {
1108 $action = git_get_type("$hash_base:$file_name");
1109 } elsif (defined $project) {
1110 $action = 'summary';
1111 } else {
1112 $action = 'project_list';
1113 }
1114 }
1115 if (!defined($actions{$action})) {
1116 die_error(400, "Unknown action");
1117 }
1118 if ($action !~ m/^(?:opml|project_list|project_index)$/ &&
1119 !$project) {
1120 die_error(400, "Project needed");
1121 }
1122 $actions{$action}->();
1123 }
1124
1125 sub reset_timer {
1126 our $t0 = [ gettimeofday() ]
1127 if defined $t0;
1128 our $number_of_git_cmds = 0;
1129 }
1130
1131 our $first_request = 1;
1132 sub run_request {
1133 reset_timer();
1134
1135 evaluate_uri();
1136 if ($first_request) {
1137 evaluate_gitweb_config();
1138 evaluate_git_version();
1139 }
1140 if ($per_request_config) {
1141 if (ref($per_request_config) eq 'CODE') {
1142 $per_request_config->();
1143 } elsif (!$first_request) {
1144 evaluate_gitweb_config();
1145 }
1146 }
1147 check_loadavg();
1148
1149 # $projectroot and $projects_list might be set in gitweb config file
1150 $projects_list ||= $projectroot;
1151
1152 evaluate_query_params();
1153 evaluate_path_info();
1154 evaluate_and_validate_params();
1155 evaluate_git_dir();
1156
1157 configure_gitweb_features();
1158
1159 dispatch();
1160 }
1161
1162 our $is_last_request = sub { 1 };
1163 our ($pre_dispatch_hook, $post_dispatch_hook, $pre_listen_hook);
1164 our $CGI = 'CGI';
1165 our $cgi;
1166 sub configure_as_fcgi {
1167 require CGI::Fast;
1168 our $CGI = 'CGI::Fast';
1169
1170 my $request_number = 0;
1171 # let each child service 100 requests
1172 our $is_last_request = sub { ++$request_number > 100 };
1173 }
1174 sub evaluate_argv {
1175 my $script_name = $ENV{'SCRIPT_NAME'} || $ENV{'SCRIPT_FILENAME'} || __FILE__;
1176 configure_as_fcgi()
1177 if $script_name =~ /\.fcgi$/;
1178
1179 return unless (@ARGV);
1180
1181 require Getopt::Long;
1182 Getopt::Long::GetOptions(
1183 'fastcgi|fcgi|f' => \&configure_as_fcgi,
1184 'nproc|n=i' => sub {
1185 my ($arg, $val) = @_;
1186 return unless eval { require FCGI::ProcManager; 1; };
1187 my $proc_manager = FCGI::ProcManager->new({
1188 n_processes => $val,
1189 });
1190 our $pre_listen_hook = sub { $proc_manager->pm_manage() };
1191 our $pre_dispatch_hook = sub { $proc_manager->pm_pre_dispatch() };
1192 our $post_dispatch_hook = sub { $proc_manager->pm_post_dispatch() };
1193 },
1194 );
1195 }
1196
1197 sub run {
1198 evaluate_argv();
1199
1200 $first_request = 1;
1201 $pre_listen_hook->()
1202 if $pre_listen_hook;
1203
1204 REQUEST:
1205 while ($cgi = $CGI->new()) {
1206 $pre_dispatch_hook->()
1207 if $pre_dispatch_hook;
1208
1209 run_request();
1210
1211 $post_dispatch_hook->()
1212 if $post_dispatch_hook;
1213 $first_request = 0;
1214
1215 last REQUEST if ($is_last_request->());
1216 }
1217
1218 DONE_GITWEB:
1219 1;
1220 }
1221
1222 run();
1223
1224 if (defined caller) {
1225 # wrapped in a subroutine processing requests,
1226 # e.g. mod_perl with ModPerl::Registry, or PSGI with Plack::App::WrapCGI
1227 return;
1228 } else {
1229 # pure CGI script, serving single request
1230 exit;
1231 }
1232
1233 ## ======================================================================
1234 ## action links
1235
1236 # possible values of extra options
1237 # -full => 0|1 - use absolute/full URL ($my_uri/$my_url as base)
1238 # -replay => 1 - start from a current view (replay with modifications)
1239 # -path_info => 0|1 - don't use/use path_info URL (if possible)
1240 # -anchor => ANCHOR - add #ANCHOR to end of URL, implies -replay if used alone
1241 sub href {
1242 my %params = @_;
1243 # default is to use -absolute url() i.e. $my_uri
1244 my $href = $params{-full} ? $my_url : $my_uri;
1245
1246 # implicit -replay, must be first of implicit params
1247 $params{-replay} = 1 if (keys %params == 1 && $params{-anchor});
1248
1249 $params{'project'} = $project unless exists $params{'project'};
1250
1251 if ($params{-replay}) {
1252 while (my ($name, $symbol) = each %cgi_param_mapping) {
1253 if (!exists $params{$name}) {
1254 $params{$name} = $input_params{$name};
1255 }
1256 }
1257 }
1258
1259 my $use_pathinfo = gitweb_check_feature('pathinfo');
1260 if (defined $params{'project'} &&
1261 (exists $params{-path_info} ? $params{-path_info} : $use_pathinfo)) {
1262 # try to put as many parameters as possible in PATH_INFO:
1263 # - project name
1264 # - action
1265 # - hash_parent or hash_parent_base:/file_parent
1266 # - hash or hash_base:/filename
1267 # - the snapshot_format as an appropriate suffix
1268
1269 # When the script is the root DirectoryIndex for the domain,
1270 # $href here would be something like http://gitweb.example.com/
1271 # Thus, we strip any trailing / from $href, to spare us double
1272 # slashes in the final URL
1273 $href =~ s,/$,,;
1274
1275 # Then add the project name, if present
1276 $href .= "/".esc_path_info($params{'project'});
1277 delete $params{'project'};
1278
1279 # since we destructively absorb parameters, we keep this
1280 # boolean that remembers if we're handling a snapshot
1281 my $is_snapshot = $params{'action'} eq 'snapshot';
1282
1283 # Summary just uses the project path URL, any other action is
1284 # added to the URL
1285 if (defined $params{'action'}) {
1286 $href .= "/".esc_path_info($params{'action'})
1287 unless $params{'action'} eq 'summary';
1288 delete $params{'action'};
1289 }
1290
1291 # Next, we put hash_parent_base:/file_parent..hash_base:/file_name,
1292 # stripping nonexistent or useless pieces
1293 $href .= "/" if ($params{'hash_base'} || $params{'hash_parent_base'}
1294 || $params{'hash_parent'} || $params{'hash'});
1295 if (defined $params{'hash_base'}) {
1296 if (defined $params{'hash_parent_base'}) {
1297 $href .= esc_path_info($params{'hash_parent_base'});
1298 # skip the file_parent if it's the same as the file_name
1299 if (defined $params{'file_parent'}) {
1300 if (defined $params{'file_name'} && $params{'file_parent'} eq $params{'file_name'}) {
1301 delete $params{'file_parent'};
1302 } elsif ($params{'file_parent'} !~ /\.\./) {
1303 $href .= ":/".esc_path_info($params{'file_parent'});
1304 delete $params{'file_parent'};
1305 }
1306 }
1307 $href .= "..";
1308 delete $params{'hash_parent'};
1309 delete $params{'hash_parent_base'};
1310 } elsif (defined $params{'hash_parent'}) {
1311 $href .= esc_path_info($params{'hash_parent'}). "..";
1312 delete $params{'hash_parent'};
1313 }
1314
1315 $href .= esc_path_info($params{'hash_base'});
1316 if (defined $params{'file_name'} && $params{'file_name'} !~ /\.\./) {
1317 $href .= ":/".esc_path_info($params{'file_name'});
1318 delete $params{'file_name'};
1319 }
1320 delete $params{'hash'};
1321 delete $params{'hash_base'};
1322 } elsif (defined $params{'hash'}) {
1323 $href .= esc_path_info($params{'hash'});
1324 delete $params{'hash'};
1325 }
1326
1327 # If the action was a snapshot, we can absorb the
1328 # snapshot_format parameter too
1329 if ($is_snapshot) {
1330 my $fmt = $params{'snapshot_format'};
1331 # snapshot_format should always be defined when href()
1332 # is called, but just in case some code forgets, we
1333 # fall back to the default
1334 $fmt ||= $snapshot_fmts[0];
1335 $href .= $known_snapshot_formats{$fmt}{'suffix'};
1336 delete $params{'snapshot_format'};
1337 }
1338 }
1339
1340 # now encode the parameters explicitly
1341 my @result = ();
1342 for (my $i = 0; $i < @cgi_param_mapping; $i += 2) {
1343 my ($name, $symbol) = ($cgi_param_mapping[$i], $cgi_param_mapping[$i+1]);
1344 if (defined $params{$name}) {
1345 if (ref($params{$name}) eq "ARRAY") {
1346 foreach my $par (@{$params{$name}}) {
1347 push @result, $symbol . "=" . esc_param($par);
1348 }
1349 } else {
1350 push @result, $symbol . "=" . esc_param($params{$name});
1351 }
1352 }
1353 }
1354 $href .= "?" . join(';', @result) if scalar @result;
1355
1356 # final transformation: trailing spaces must be escaped (URI-encoded)
1357 $href =~ s/(\s+)$/CGI::escape($1)/e;
1358
1359 if ($params{-anchor}) {
1360 $href .= "#".esc_param($params{-anchor});
1361 }
1362
1363 return $href;
1364 }
1365
1366
1367 ## ======================================================================
1368 ## validation, quoting/unquoting and escaping
1369
1370 sub validate_action {
1371 my $input = shift || return undef;
1372 return undef unless exists $actions{$input};
1373 return $input;
1374 }
1375
1376 sub validate_project {
1377 my $input = shift || return undef;
1378 if (!validate_pathname($input) ||
1379 !(-d "$projectroot/$input") ||
1380 !check_export_ok("$projectroot/$input") ||
1381 ($strict_export && !project_in_list($input))) {
1382 return undef;
1383 } else {
1384 return $input;
1385 }
1386 }
1387
1388 sub validate_pathname {
1389 my $input = shift || return undef;
1390
1391 # no '.' or '..' as elements of path, i.e. no '.' nor '..'
1392 # at the beginning, at the end, and between slashes.
1393 # also this catches doubled slashes
1394 if ($input =~ m!(^|/)(|\.|\.\.)(/|$)!) {
1395 return undef;
1396 }
1397 # no null characters
1398 if ($input =~ m!\0!) {
1399 return undef;
1400 }
1401 return $input;
1402 }
1403
1404 sub validate_refname {
1405 my $input = shift || return undef;
1406
1407 # textual hashes are O.K.
1408 if ($input =~ m/^[0-9a-fA-F]{40}$/) {
1409 return $input;
1410 }
1411 # it must be correct pathname
1412 $input = validate_pathname($input)
1413 or return undef;
1414 # restrictions on ref name according to git-check-ref-format
1415 if ($input =~ m!(/\.|\.\.|[\000-\040\177 ~^:?*\[]|/$)!) {
1416 return undef;
1417 }
1418 return $input;
1419 }
1420
1421 # decode sequences of octets in utf8 into Perl's internal form,
1422 # which is utf-8 with utf8 flag set if needed. gitweb writes out
1423 # in utf-8 thanks to "binmode STDOUT, ':utf8'" at beginning
1424 sub to_utf8 {
1425 my $str = shift;
1426 return undef unless defined $str;
1427 if (utf8::valid($str)) {
1428 utf8::decode($str);
1429 return $str;
1430 } else {
1431 return decode($fallback_encoding, $str, Encode::FB_DEFAULT);
1432 }
1433 }
1434
1435 # quote unsafe chars, but keep the slash, even when it's not
1436 # correct, but quoted slashes look too horrible in bookmarks
1437 sub esc_param {
1438 my $str = shift;
1439 return undef unless defined $str;
1440 $str =~ s/([^A-Za-z0-9\-_.~()\/:@ ]+)/CGI::escape($1)/eg;
1441 $str =~ s/ /\+/g;
1442 return $str;
1443 }
1444
1445 # the quoting rules for path_info fragment are slightly different
1446 sub esc_path_info {
1447 my $str = shift;
1448 return undef unless defined $str;
1449
1450 # path_info doesn't treat '+' as space (specially), but '?' must be escaped
1451 $str =~ s/([^A-Za-z0-9\-_.~();\/;:@&= +]+)/CGI::escape($1)/eg;
1452
1453 return $str;
1454 }
1455
1456 # quote unsafe chars in whole URL, so some characters cannot be quoted
1457 sub esc_url {
1458 my $str = shift;
1459 return undef unless defined $str;
1460 $str =~ s/([^A-Za-z0-9\-_.~();\/;?:@&= ]+)/CGI::escape($1)/eg;
1461 $str =~ s/ /\+/g;
1462 return $str;
1463 }
1464
1465 # quote unsafe characters in HTML attributes
1466 sub esc_attr {
1467
1468 # for XHTML conformance escaping '"' to '&quot;' is not enough
1469 return esc_html(@_);
1470 }
1471
1472 # replace invalid utf8 character with SUBSTITUTION sequence
1473 sub esc_html {
1474 my $str = shift;
1475 my %opts = @_;
1476
1477 return undef unless defined $str;
1478
1479 $str = to_utf8($str);
1480 $str = $cgi->escapeHTML($str);
1481 if ($opts{'-nbsp'}) {
1482 $str =~ s/ /&nbsp;/g;
1483 }
1484 $str =~ s|([[:cntrl:]])|(($1 ne "\t") ? quot_cec($1) : $1)|eg;
1485 return $str;
1486 }
1487
1488 # quote control characters and escape filename to HTML
1489 sub esc_path {
1490 my $str = shift;
1491 my %opts = @_;
1492
1493 return undef unless defined $str;
1494
1495 $str = to_utf8($str);
1496 $str = $cgi->escapeHTML($str);
1497 if ($opts{'-nbsp'}) {
1498 $str =~ s/ /&nbsp;/g;
1499 }
1500 $str =~ s|([[:cntrl:]])|quot_cec($1)|eg;
1501 return $str;
1502 }
1503
1504 # Make control characters "printable", using character escape codes (CEC)
1505 sub quot_cec {
1506 my $cntrl = shift;
1507 my %opts = @_;
1508 my %es = ( # character escape codes, aka escape sequences
1509 "\t" => '\t', # tab (HT)
1510 "\n" => '\n', # line feed (LF)
1511 "\r" => '\r', # carrige return (CR)
1512 "\f" => '\f', # form feed (FF)
1513 "\b" => '\b', # backspace (BS)
1514 "\a" => '\a', # alarm (bell) (BEL)
1515 "\e" => '\e', # escape (ESC)
1516 "\013" => '\v', # vertical tab (VT)
1517 "\000" => '\0', # nul character (NUL)
1518 );
1519 my $chr = ( (exists $es{$cntrl})
1520 ? $es{$cntrl}
1521 : sprintf('\%2x', ord($cntrl)) );
1522 if ($opts{-nohtml}) {
1523 return $chr;
1524 } else {
1525 return "<span class=\"cntrl\">$chr</span>";
1526 }
1527 }
1528
1529 # Alternatively use unicode control pictures codepoints,
1530 # Unicode "printable representation" (PR)
1531 sub quot_upr {
1532 my $cntrl = shift;
1533 my %opts = @_;
1534
1535 my $chr = sprintf('&#%04d;', 0x2400+ord($cntrl));
1536 if ($opts{-nohtml}) {
1537 return $chr;
1538 } else {
1539 return "<span class=\"cntrl\">$chr</span>";
1540 }
1541 }
1542
1543 # git may return quoted and escaped filenames
1544 sub unquote {
1545 my $str = shift;
1546
1547 sub unq {
1548 my $seq = shift;
1549 my %es = ( # character escape codes, aka escape sequences
1550 't' => "\t", # tab (HT, TAB)
1551 'n' => "\n", # newline (NL)
1552 'r' => "\r", # return (CR)
1553 'f' => "\f", # form feed (FF)
1554 'b' => "\b", # backspace (BS)
1555 'a' => "\a", # alarm (bell) (BEL)
1556 'e' => "\e", # escape (ESC)
1557 'v' => "\013", # vertical tab (VT)
1558 );
1559
1560 if ($seq =~ m/^[0-7]{1,3}$/) {
1561 # octal char sequence
1562 return chr(oct($seq));
1563 } elsif (exists $es{$seq}) {
1564 # C escape sequence, aka character escape code
1565 return $es{$seq};
1566 }
1567 # quoted ordinary character
1568 return $seq;
1569 }
1570
1571 if ($str =~ m/^"(.*)"$/) {
1572 # needs unquoting
1573 $str = $1;
1574 $str =~ s/\\([^0-7]|[0-7]{1,3})/unq($1)/eg;
1575 }
1576 return $str;
1577 }
1578
1579 # escape tabs (convert tabs to spaces)
1580 sub untabify {
1581 my $line = shift;
1582
1583 while ((my $pos = index($line, "\t")) != -1) {
1584 if (my $count = (8 - ($pos % 8))) {
1585 my $spaces = ' ' x $count;
1586 $line =~ s/\t/$spaces/;
1587 }
1588 }
1589
1590 return $line;
1591 }
1592
1593 sub project_in_list {
1594 my $project = shift;
1595 my @list = git_get_projects_list();
1596 return @list && scalar(grep { $_->{'path'} eq $project } @list);
1597 }
1598
1599 ## ----------------------------------------------------------------------
1600 ## HTML aware string manipulation
1601
1602 # Try to chop given string on a word boundary between position
1603 # $len and $len+$add_len. If there is no word boundary there,
1604 # chop at $len+$add_len. Do not chop if chopped part plus ellipsis
1605 # (marking chopped part) would be longer than given string.
1606 sub chop_str {
1607 my $str = shift;
1608 my $len = shift;
1609 my $add_len = shift || 10;
1610 my $where = shift || 'right'; # 'left' | 'center' | 'right'
1611
1612 # Make sure perl knows it is utf8 encoded so we don't
1613 # cut in the middle of a utf8 multibyte char.
1614 $str = to_utf8($str);
1615
1616 # allow only $len chars, but don't cut a word if it would fit in $add_len
1617 # if it doesn't fit, cut it if it's still longer than the dots we would add
1618 # remove chopped character entities entirely
1619
1620 # when chopping in the middle, distribute $len into left and right part
1621 # return early if chopping wouldn't make string shorter
1622 if ($where eq 'center') {
1623 return $str if ($len + 5 >= length($str)); # filler is length 5
1624 $len = int($len/2);
1625 } else {
1626 return $str if ($len + 4 >= length($str)); # filler is length 4
1627 }
1628
1629 # regexps: ending and beginning with word part up to $add_len
1630 my $endre = qr/.{$len}\w{0,$add_len}/;
1631 my $begre = qr/\w{0,$add_len}.{$len}/;
1632
1633 if ($where eq 'left') {
1634 $str =~ m/^(.*?)($begre)$/;
1635 my ($lead, $body) = ($1, $2);
1636 if (length($lead) > 4) {
1637 $lead = " ...";
1638 }
1639 return "$lead$body";
1640
1641 } elsif ($where eq 'center') {
1642 $str =~ m/^($endre)(.*)$/;
1643 my ($left, $str) = ($1, $2);
1644 $str =~ m/^(.*?)($begre)$/;
1645 my ($mid, $right) = ($1, $2);
1646 if (length($mid) > 5) {
1647 $mid = " ... ";
1648 }
1649 return "$left$mid$right";
1650
1651 } else {
1652 $str =~ m/^($endre)(.*)$/;
1653 my $body = $1;
1654 my $tail = $2;
1655 if (length($tail) > 4) {
1656 $tail = "... ";
1657 }
1658 return "$body$tail";
1659 }
1660 }
1661
1662 # takes the same arguments as chop_str, but also wraps a <span> around the
1663 # result with a title attribute if it does get chopped. Additionally, the
1664 # string is HTML-escaped.
1665 sub chop_and_escape_str {
1666 my ($str) = @_;
1667
1668 my $chopped = chop_str(@_);
1669 if ($chopped eq $str) {
1670 return esc_html($chopped);
1671 } else {
1672 $str =~ s/[[:cntrl:]]/?/g;
1673 return $cgi->span({-title=>$str}, esc_html($chopped));
1674 }
1675 }
1676
1677 ## ----------------------------------------------------------------------
1678 ## functions returning short strings
1679
1680 # CSS class for given age value (in seconds)
1681 sub age_class {
1682 my $age = shift;
1683
1684 if (!defined $age) {
1685 return "noage";
1686 } elsif ($age < 60*60*2) {
1687 return "age0";
1688 } elsif ($age < 60*60*24*2) {
1689 return "age1";
1690 } else {
1691 return "age2";
1692 }
1693 }
1694
1695 # convert age in seconds to "nn units ago" string
1696 sub age_string {
1697 my $age = shift;
1698 my $age_str;
1699
1700 if ($age > 60*60*24*365*2) {
1701 $age_str = (int $age/60/60/24/365);
1702 $age_str .= " years ago";
1703 } elsif ($age > 60*60*24*(365/12)*2) {
1704 $age_str = int $age/60/60/24/(365/12);
1705 $age_str .= " months ago";
1706 } elsif ($age > 60*60*24*7*2) {
1707 $age_str = int $age/60/60/24/7;
1708 $age_str .= " weeks ago";
1709 } elsif ($age > 60*60*24*2) {
1710 $age_str = int $age/60/60/24;
1711 $age_str .= " days ago";
1712 } elsif ($age > 60*60*2) {
1713 $age_str = int $age/60/60;
1714 $age_str .= " hours ago";
1715 } elsif ($age > 60*2) {
1716 $age_str = int $age/60;
1717 $age_str .= " min ago";
1718 } elsif ($age > 2) {
1719 $age_str = int $age;
1720 $age_str .= " sec ago";
1721 } else {
1722 $age_str .= " right now";
1723 }
1724 return $age_str;
1725 }
1726
1727 use constant {
1728 S_IFINVALID => 0030000,
1729 S_IFGITLINK => 0160000,
1730 };
1731
1732 # submodule/subproject, a commit object reference
1733 sub S_ISGITLINK {
1734 my $mode = shift;
1735
1736 return (($mode & S_IFMT) == S_IFGITLINK)
1737 }
1738
1739 # convert file mode in octal to symbolic file mode string
1740 sub mode_str {
1741 my $mode = oct shift;
1742
1743 if (S_ISGITLINK($mode)) {
1744 return 'm---------';
1745 } elsif (S_ISDIR($mode & S_IFMT)) {
1746 return 'drwxr-xr-x';
1747 } elsif (S_ISLNK($mode)) {
1748 return 'lrwxrwxrwx';
1749 } elsif (S_ISREG($mode)) {
1750 # git cares only about the executable bit
1751 if ($mode & S_IXUSR) {
1752 return '-rwxr-xr-x';
1753 } else {
1754 return '-rw-r--r--';
1755 };
1756 } else {
1757 return '----------';
1758 }
1759 }
1760
1761 # convert file mode in octal to file type string
1762 sub file_type {
1763 my $mode = shift;
1764
1765 if ($mode !~ m/^[0-7]+$/) {
1766 return $mode;
1767 } else {
1768 $mode = oct $mode;
1769 }
1770
1771 if (S_ISGITLINK($mode)) {
1772 return "submodule";
1773 } elsif (S_ISDIR($mode & S_IFMT)) {
1774 return "directory";
1775 } elsif (S_ISLNK($mode)) {
1776 return "symlink";
1777 } elsif (S_ISREG($mode)) {
1778 return "file";
1779 } else {
1780 return "unknown";
1781 }
1782 }
1783
1784 # convert file mode in octal to file type description string
1785 sub file_type_long {
1786 my $mode = shift;
1787
1788 if ($mode !~ m/^[0-7]+$/) {
1789 return $mode;
1790 } else {
1791 $mode = oct $mode;
1792 }
1793
1794 if (S_ISGITLINK($mode)) {
1795 return "submodule";
1796 } elsif (S_ISDIR($mode & S_IFMT)) {
1797 return "directory";
1798 } elsif (S_ISLNK($mode)) {
1799 return "symlink";
1800 } elsif (S_ISREG($mode)) {
1801 if ($mode & S_IXUSR) {
1802 return "executable";
1803 } else {
1804 return "file";
1805 };
1806 } else {
1807 return "unknown";
1808 }
1809 }
1810
1811
1812 ## ----------------------------------------------------------------------
1813 ## functions returning short HTML fragments, or transforming HTML fragments
1814 ## which don't belong to other sections
1815
1816 # format line of commit message.
1817 sub format_log_line_html {
1818 my $line = shift;
1819
1820 $line = esc_html($line, -nbsp=>1);
1821 $line =~ s{\b([0-9a-fA-F]{8,40})\b}{
1822 $cgi->a({-href => href(action=>"object", hash=>$1),
1823 -class => "text"}, $1);
1824 }eg;
1825
1826 return $line;
1827 }
1828
1829 # format marker of refs pointing to given object
1830
1831 # the destination action is chosen based on object type and current context:
1832 # - for annotated tags, we choose the tag view unless it's the current view
1833 # already, in which case we go to shortlog view
1834 # - for other refs, we keep the current view if we're in history, shortlog or
1835 # log view, and select shortlog otherwise
1836 sub format_ref_marker {
1837 my ($refs, $id) = @_;
1838 my $markers = '';
1839
1840 if (defined $refs->{$id}) {
1841 foreach my $ref (@{$refs->{$id}}) {
1842 # this code exploits the fact that non-lightweight tags are the
1843 # only indirect objects, and that they are the only objects for which
1844 # we want to use tag instead of shortlog as action
1845 my ($type, $name) = qw();
1846 my $indirect = ($ref =~ s/\^\{\}$//);
1847 # e.g. tags/v2.6.11 or heads/next
1848 if ($ref =~ m!^(.*?)s?/(.*)$!) {
1849 $type = $1;
1850 $name = $2;
1851 } else {
1852 $type = "ref";
1853 $name = $ref;
1854 }
1855
1856 my $class = $type;
1857 $class .= " indirect" if $indirect;
1858
1859 my $dest_action = "shortlog";
1860
1861 if ($indirect) {
1862 $dest_action = "tag" unless $action eq "tag";
1863 } elsif ($action =~ /^(history|(short)?log)$/) {
1864 $dest_action = $action;
1865 }
1866
1867 my $dest = "";
1868 $dest .= "refs/" unless $ref =~ m!^refs/!;
1869 $dest .= $ref;
1870
1871 my $link = $cgi->a({
1872 -href => href(
1873 action=>$dest_action,
1874 hash=>$dest
1875 )}, $name);
1876
1877 $markers .= " <span class=\"".esc_attr($class)."\" title=\"".esc_attr($ref)."\">" .
1878 $link . "</span>";
1879 }
1880 }
1881
1882 if ($markers) {
1883 return ' <span class="refs">'. $markers . '</span>';
1884 } else {
1885 return "";
1886 }
1887 }
1888
1889 # format, perhaps shortened and with markers, title line
1890 sub format_subject_html {
1891 my ($long, $short, $href, $extra) = @_;
1892 $extra = '' unless defined($extra);
1893
1894 if (length($short) < length($long)) {
1895 $long =~ s/[[:cntrl:]]/?/g;
1896 return $cgi->a({-href => $href, -class => "list subject",
1897 -title => to_utf8($long)},
1898 esc_html($short)) . $extra;
1899 } else {
1900 return $cgi->a({-href => $href, -class => "list subject"},
1901 esc_html($long)) . $extra;
1902 }
1903 }
1904
1905 # Rather than recomputing the url for an email multiple times, we cache it
1906 # after the first hit. This gives a visible benefit in views where the avatar
1907 # for the same email is used repeatedly (e.g. shortlog).
1908 # The cache is shared by all avatar engines (currently gravatar only), which
1909 # are free to use it as preferred. Since only one avatar engine is used for any
1910 # given page, there's no risk for cache conflicts.
1911 our %avatar_cache = ();
1912
1913 # Compute the picon url for a given email, by using the picon search service over at
1914 # http://www.cs.indiana.edu/picons/search.html
1915 sub picon_url {
1916 my $email = lc shift;
1917 if (!$avatar_cache{$email}) {
1918 my ($user, $domain) = split('@', $email);
1919 $avatar_cache{$email} =
1920 "http://www.cs.indiana.edu/cgi-pub/kinzler/piconsearch.cgi/" .
1921 "$domain/$user/" .
1922 "users+domains+unknown/up/single";
1923 }
1924 return $avatar_cache{$email};
1925 }
1926
1927 # Compute the gravatar url for a given email, if it's not in the cache already.
1928 # Gravatar stores only the part of the URL before the size, since that's the
1929 # one computationally more expensive. This also allows reuse of the cache for
1930 # different sizes (for this particular engine).
1931 sub gravatar_url {
1932 my $email = lc shift;
1933 my $size = shift;
1934 $avatar_cache{$email} ||=
1935 "http://www.gravatar.com/avatar/" .
1936 Digest::MD5::md5_hex($email) . "?s=";
1937 return $avatar_cache{$email} . $size;
1938 }
1939
1940 # Insert an avatar for the given $email at the given $size if the feature
1941 # is enabled.
1942 sub git_get_avatar {
1943 my ($email, %opts) = @_;
1944 my $pre_white = ($opts{-pad_before} ? "&nbsp;" : "");
1945 my $post_white = ($opts{-pad_after} ? "&nbsp;" : "");
1946 $opts{-size} ||= 'default';
1947 my $size = $avatar_size{$opts{-size}} || $avatar_size{'default'};
1948 my $url = "";
1949 if ($git_avatar eq 'gravatar') {
1950 $url = gravatar_url($email, $size);
1951 } elsif ($git_avatar eq 'picon') {
1952 $url = picon_url($email);
1953 }
1954 # Other providers can be added by extending the if chain, defining $url
1955 # as needed. If no variant puts something in $url, we assume avatars
1956 # are completely disabled/unavailable.
1957 if ($url) {
1958 return $pre_white .
1959 "<img width=\"$size\" " .
1960 "class=\"avatar\" " .
1961 "src=\"".esc_url($url)."\" " .
1962 "alt=\"\" " .
1963 "/>" . $post_white;
1964 } else {
1965 return "";
1966 }
1967 }
1968
1969 sub format_search_author {
1970 my ($author, $searchtype, $displaytext) = @_;
1971 my $have_search = gitweb_check_feature('search');
1972
1973 if ($have_search) {
1974 my $performed = "";
1975 if ($searchtype eq 'author') {
1976 $performed = "authored";
1977 } elsif ($searchtype eq 'committer') {
1978 $performed = "committed";
1979 }
1980
1981 return $cgi->a({-href => href(action=>"search", hash=>$hash,
1982 searchtext=>$author,
1983 searchtype=>$searchtype), class=>"list",
1984 title=>"Search for commits $performed by $author"},
1985 $displaytext);
1986
1987 } else {
1988 return $displaytext;
1989 }
1990 }
1991
1992 # format the author name of the given commit with the given tag
1993 # the author name is chopped and escaped according to the other
1994 # optional parameters (see chop_str).
1995 sub format_author_html {
1996 my $tag = shift;
1997 my $co = shift;
1998 my $author = chop_and_escape_str($co->{'author_name'}, @_);
1999 return "<$tag class=\"author\">" .
2000 format_search_author($co->{'author_name'}, "author",
2001 git_get_avatar($co->{'author_email'}, -pad_after => 1) .
2002 $author) .
2003 "</$tag>";
2004 }
2005
2006 # format git diff header line, i.e. "diff --(git|combined|cc) ..."
2007 sub format_git_diff_header_line {
2008 my $line = shift;
2009 my $diffinfo = shift;
2010 my ($from, $to) = @_;
2011
2012 if ($diffinfo->{'nparents'}) {
2013 # combined diff
2014 $line =~ s!^(diff (.*?) )"?.*$!$1!;
2015 if ($to->{'href'}) {
2016 $line .= $cgi->a({-href => $to->{'href'}, -class => "path"},
2017 esc_path($to->{'file'}));
2018 } else { # file was deleted (no href)
2019 $line .= esc_path($to->{'file'});
2020 }
2021 } else {
2022 # "ordinary" diff
2023 $line =~ s!^(diff (.*?) )"?a/.*$!$1!;
2024 if ($from->{'href'}) {
2025 $line .= $cgi->a({-href => $from->{'href'}, -class => "path"},
2026 'a/' . esc_path($from->{'file'}));
2027 } else { # file was added (no href)
2028 $line .= 'a/' . esc_path($from->{'file'});
2029 }
2030 $line .= ' ';
2031 if ($to->{'href'}) {
2032 $line .= $cgi->a({-href => $to->{'href'}, -class => "path"},
2033 'b/' . esc_path($to->{'file'}));
2034 } else { # file was deleted
2035 $line .= 'b/' . esc_path($to->{'file'});
2036 }
2037 }
2038
2039 return "<div class=\"diff header\">$line</div>\n";
2040 }
2041
2042 # format extended diff header line, before patch itself
2043 sub format_extended_diff_header_line {
2044 my $line = shift;
2045 my $diffinfo = shift;
2046 my ($from, $to) = @_;
2047
2048 # match <path>
2049 if ($line =~ s!^((copy|rename) from ).*$!$1! && $from->{'href'}) {
2050 $line .= $cgi->a({-href=>$from->{'href'}, -class=>"path"},
2051 esc_path($from->{'file'}));
2052 }
2053 if ($line =~ s!^((copy|rename) to ).*$!$1! && $to->{'href'}) {
2054 $line .= $cgi->a({-href=>$to->{'href'}, -class=>"path"},
2055 esc_path($to->{'file'}));
2056 }
2057 # match single <mode>
2058 if ($line =~ m/\s(\d{6})$/) {
2059 $line .= '<span class="info"> (' .
2060 file_type_long($1) .
2061 ')</span>';
2062 }
2063 # match <hash>
2064 if ($line =~ m/^index [0-9a-fA-F]{40},[0-9a-fA-F]{40}/) {
2065 # can match only for combined diff
2066 $line = 'index ';
2067 for (my $i = 0; $i < $diffinfo->{'nparents'}; $i++) {
2068 if ($from->{'href'}[$i]) {
2069 $line .= $cgi->a({-href=>$from->{'href'}[$i],
2070 -class=>"hash"},
2071 substr($diffinfo->{'from_id'}[$i],0,7));
2072 } else {
2073 $line .= '0' x 7;
2074 }
2075 # separator
2076 $line .= ',' if ($i < $diffinfo->{'nparents'} - 1);
2077 }
2078 $line .= '..';
2079 if ($to->{'href'}) {
2080 $line .= $cgi->a({-href=>$to->{'href'}, -class=>"hash"},
2081 substr($diffinfo->{'to_id'},0,7));
2082 } else {
2083 $line .= '0' x 7;
2084 }
2085
2086 } elsif ($line =~ m/^index [0-9a-fA-F]{40}..[0-9a-fA-F]{40}/) {
2087 # can match only for ordinary diff
2088 my ($from_link, $to_link);
2089 if ($from->{'href'}) {
2090 $from_link = $cgi->a({-href=>$from->{'href'}, -class=>"hash"},
2091 substr($diffinfo->{'from_id'},0,7));
2092 } else {
2093 $from_link = '0' x 7;
2094 }
2095 if ($to->{'href'}) {
2096 $to_link = $cgi->a({-href=>$to->{'href'}, -class=>"hash"},
2097 substr($diffinfo->{'to_id'},0,7));
2098 } else {
2099 $to_link = '0' x 7;
2100 }
2101 my ($from_id, $to_id) = ($diffinfo->{'from_id'}, $diffinfo->{'to_id'});
2102 $line =~ s!$from_id\.\.$to_id!$from_link..$to_link!;
2103 }
2104
2105 return $line . "<br/>\n";
2106 }
2107
2108 # format from-file/to-file diff header
2109 sub format_diff_from_to_header {
2110 my ($from_line, $to_line, $diffinfo, $from, $to, @parents) = @_;
2111 my $line;
2112 my $result = '';
2113
2114 $line = $from_line;
2115 #assert($line =~ m/^---/) if DEBUG;
2116 # no extra formatting for "^--- /dev/null"
2117 if (! $diffinfo->{'nparents'}) {
2118 # ordinary (single parent) diff
2119 if ($line =~ m!^--- "?a/!) {
2120 if ($from->{'href'}) {
2121 $line = '--- a/' .
2122 $cgi->a({-href=>$from->{'href'}, -class=>"path"},
2123 esc_path($from->{'file'}));
2124 } else {
2125 $line = '--- a/' .
2126 esc_path($from->{'file'});
2127 }
2128 }
2129 $result .= qq!<div class="diff from_file">$line</div>\n!;
2130
2131 } else {
2132 # combined diff (merge commit)
2133 for (my $i = 0; $i < $diffinfo->{'nparents'}; $i++) {
2134 if ($from->{'href'}[$i]) {
2135 $line = '--- ' .
2136 $cgi->a({-href=>href(action=>"blobdiff",
2137 hash_parent=>$diffinfo->{'from_id'}[$i],
2138 hash_parent_base=>$parents[$i],
2139 file_parent=>$from->{'file'}[$i],
2140 hash=>$diffinfo->{'to_id'},
2141 hash_base=>$hash,
2142 file_name=>$to->{'file'}),
2143 -class=>"path",
2144 -title=>"diff" . ($i+1)},
2145 $i+1) .
2146 '/' .
2147 $cgi->a({-href=>$from->{'href'}[$i], -class=>"path"},
2148 esc_path($from->{'file'}[$i]));
2149 } else {
2150 $line = '--- /dev/null';
2151 }
2152 $result .= qq!<div class="diff from_file">$line</div>\n!;
2153 }
2154 }
2155
2156 $line = $to_line;
2157 #assert($line =~ m/^\+\+\+/) if DEBUG;
2158 # no extra formatting for "^+++ /dev/null"
2159 if ($line =~ m!^\+\+\+ "?b/!) {
2160 if ($to->{'href'}) {
2161 $line = '+++ b/' .
2162 $cgi->a({-href=>$to->{'href'}, -class=>"path"},
2163 esc_path($to->{'file'}));
2164 } else {
2165 $line = '+++ b/' .
2166 esc_path($to->{'file'});
2167 }
2168 }
2169 $result .= qq!<div class="diff to_file">$line</div>\n!;
2170
2171 return $result;
2172 }
2173
2174 # create note for patch simplified by combined diff
2175 sub format_diff_cc_simplified {
2176 my ($diffinfo, @parents) = @_;
2177 my $result = '';
2178
2179 $result .= "<div class=\"diff header\">" .
2180 "diff --cc ";
2181 if (!is_deleted($diffinfo)) {
2182 $result .= $cgi->a({-href => href(action=>"blob",
2183 hash_base=>$hash,
2184 hash=>$diffinfo->{'to_id'},
2185 file_name=>$diffinfo->{'to_file'}),
2186 -class => "path"},
2187 esc_path($diffinfo->{'to_file'}));
2188 } else {
2189 $result .= esc_path($diffinfo->{'to_file'});
2190 }
2191 $result .= "</div>\n" . # class="diff header"
2192 "<div class=\"diff nodifferences\">" .
2193 "Simple merge" .
2194 "</div>\n"; # class="diff nodifferences"
2195
2196 return $result;
2197 }
2198
2199 # format patch (diff) line (not to be used for diff headers)
2200 sub format_diff_line {
2201 my $line = shift;
2202 my ($from, $to) = @_;
2203 my $diff_class = "";
2204
2205 chomp $line;
2206
2207 if ($from && $to && ref($from->{'href'}) eq "ARRAY") {
2208 # combined diff
2209 my $prefix = substr($line, 0, scalar @{$from->{'href'}});
2210 if ($line =~ m/^\@{3}/) {
2211 $diff_class = " chunk_header";
2212 } elsif ($line =~ m/^\\/) {
2213 $diff_class = " incomplete";
2214 } elsif ($prefix =~ tr/+/+/) {
2215 $diff_class = " add";
2216 } elsif ($prefix =~ tr/-/-/) {
2217 $diff_class = " rem";
2218 }
2219 } else {
2220 # assume ordinary diff
2221 my $char = substr($line, 0, 1);
2222 if ($char eq '+') {
2223 $diff_class = " add";
2224 } elsif ($char eq '-') {
2225 $diff_class = " rem";
2226 } elsif ($char eq '@') {
2227 $diff_class = " chunk_header";
2228 } elsif ($char eq "\\") {
2229 $diff_class = " incomplete";
2230 }
2231 }
2232 $line = untabify($line);
2233 if ($from && $to && $line =~ m/^\@{2} /) {
2234 my ($from_text, $from_start, $from_lines, $to_text, $to_start, $to_lines, $section) =
2235 $line =~ m/^\@{2} (-(\d+)(?:,(\d+))?) (\+(\d+)(?:,(\d+))?) \@{2}(.*)$/;
2236
2237 $from_lines = 0 unless defined $from_lines;
2238 $to_lines = 0 unless defined $to_lines;
2239
2240 if ($from->{'href'}) {
2241 $from_text = $cgi->a({-href=>"$from->{'href'}#l$from_start",
2242 -class=>"list"}, $from_text);
2243 }
2244 if ($to->{'href'}) {
2245 $to_text = $cgi->a({-href=>"$to->{'href'}#l$to_start",
2246 -class=>"list"}, $to_text);
2247 }
2248 $line = "<span class=\"chunk_info\">@@ $from_text $to_text @@</span>" .
2249 "<span class=\"section\">" . esc_html($section, -nbsp=>1) . "</span>";
2250 return "<div class=\"diff$diff_class\">$line</div>\n";
2251 } elsif ($from && $to && $line =~ m/^\@{3}/) {
2252 my ($prefix, $ranges, $section) = $line =~ m/^(\@+) (.*?) \@+(.*)$/;
2253 my (@from_text, @from_start, @from_nlines, $to_text, $to_start, $to_nlines);
2254
2255 @from_text = split(' ', $ranges);
2256 for (my $i = 0; $i < @from_text; ++$i) {
2257 ($from_start[$i], $from_nlines[$i]) =
2258 (split(',', substr($from_text[$i], 1)), 0);
2259 }
2260
2261 $to_text = pop @from_text;
2262 $to_start = pop @from_start;
2263 $to_nlines = pop @from_nlines;
2264
2265 $line = "<span class=\"chunk_info\">$prefix ";
2266 for (my $i = 0; $i < @from_text; ++$i) {
2267 if ($from->{'href'}[$i]) {
2268 $line .= $cgi->a({-href=>"$from->{'href'}[$i]#l$from_start[$i]",
2269 -class=>"list"}, $from_text[$i]);
2270 } else {
2271 $line .= $from_text[$i];
2272 }
2273 $line .= " ";
2274 }
2275 if ($to->{'href'}) {
2276 $line .= $cgi->a({-href=>"$to->{'href'}#l$to_start",
2277 -class=>"list"}, $to_text);
2278 } else {
2279 $line .= $to_text;
2280 }
2281 $line .= " $prefix</span>" .
2282 "<span class=\"section\">" . esc_html($section, -nbsp=>1) . "</span>";
2283 return "<div class=\"diff$diff_class\">$line</div>\n";
2284 }
2285 return "<div class=\"diff$diff_class\">" . esc_html($line, -nbsp=>1) . "</div>\n";
2286 }
2287
2288 # Generates undef or something like "_snapshot_" or "snapshot (_tbz2_ _zip_)",
2289 # linked. Pass the hash of the tree/commit to snapshot.
2290 sub format_snapshot_links {
2291 my ($hash) = @_;
2292 my $num_fmts = @snapshot_fmts;
2293 if ($num_fmts > 1) {
2294 # A parenthesized list of links bearing format names.
2295 # e.g. "snapshot (_tar.gz_ _zip_)"
2296 return "snapshot (" . join(' ', map
2297 $cgi->a({
2298 -href => href(
2299 action=>"snapshot",
2300 hash=>$hash,
2301 snapshot_format=>$_
2302 )
2303 }, $known_snapshot_formats{$_}{'display'})
2304 , @snapshot_fmts) . ")";
2305 } elsif ($num_fmts == 1) {
2306 # A single "snapshot" link whose tooltip bears the format name.
2307 # i.e. "_snapshot_"
2308 my ($fmt) = @snapshot_fmts;
2309 return
2310 $cgi->a({
2311 -href => href(
2312 action=>"snapshot",
2313 hash=>$hash,
2314 snapshot_format=>$fmt
2315 ),
2316 -title => "in format: $known_snapshot_formats{$fmt}{'display'}"
2317 }, "snapshot");
2318 } else { # $num_fmts == 0
2319 return undef;
2320 }
2321 }
2322
2323 ## ......................................................................
2324 ## functions returning values to be passed, perhaps after some
2325 ## transformation, to other functions; e.g. returning arguments to href()
2326
2327 # returns hash to be passed to href to generate gitweb URL
2328 # in -title key it returns description of link
2329 sub get_feed_info {
2330 my $format = shift || 'Atom';
2331 my %res = (action => lc($format));
2332
2333 # feed links are possible only for project views
2334 return unless (defined $project);
2335 # some views should link to OPML, or to generic project feed,
2336 # or don't have specific feed yet (so they should use generic)
2337 return if ($action =~ /^(?:tags|heads|forks|tag|search)$/x);
2338
2339 my $branch;
2340 # branches refs uses 'refs/heads/' prefix (fullname) to differentiate
2341 # from tag links; this also makes possible to detect branch links
2342 if ((defined $hash_base && $hash_base =~ m!^refs/heads/(.*)$!) ||
2343 (defined $hash && $hash =~ m!^refs/heads/(.*)$!)) {
2344 $branch = $1;
2345 }
2346 # find log type for feed description (title)
2347 my $type = 'log';
2348 if (defined $file_name) {
2349 $type = "history of $file_name";
2350 $type .= "/" if ($action eq 'tree');
2351 $type .= " on '$branch'" if (defined $branch);
2352 } else {
2353 $type = "log of $branch" if (defined $branch);
2354 }
2355
2356 $res{-title} = $type;
2357 $res{'hash'} = (defined $branch ? "refs/heads/$branch" : undef);
2358 $res{'file_name'} = $file_name;
2359
2360 return %res;
2361 }
2362
2363 ## ----------------------------------------------------------------------
2364 ## git utility subroutines, invoking git commands
2365
2366 # returns path to the core git executable and the --git-dir parameter as list
2367 sub git_cmd {
2368 $number_of_git_cmds++;
2369 return $GIT, '--git-dir='.$git_dir;
2370 }
2371
2372 # quote the given arguments for passing them to the shell
2373 # quote_command("command", "arg 1", "arg with ' and ! characters")
2374 # => "'command' 'arg 1' 'arg with '\'' and '\!' characters'"
2375 # Try to avoid using this function wherever possible.
2376 sub quote_command {
2377 return join(' ',
2378 map { my $a = $_; $a =~ s/(['!])/'\\$1'/g; "'$a'" } @_ );
2379 }
2380
2381 # get HEAD ref of given project as hash
2382 sub git_get_head_hash {
2383 return git_get_full_hash(shift, 'HEAD');
2384 }
2385
2386 sub git_get_full_hash {
2387 return git_get_hash(@_);
2388 }
2389
2390 sub git_get_short_hash {
2391 return git_get_hash(@_, '--short=7');
2392 }
2393
2394 sub git_get_hash {
2395 my ($project, $hash, @options) = @_;
2396 my $o_git_dir = $git_dir;
2397 my $retval = undef;
2398 $git_dir = "$projectroot/$project";
2399 if (open my $fd, '-|', git_cmd(), 'rev-parse',
2400 '--verify', '-q', @options, $hash) {
2401 $retval = <$fd>;
2402 chomp $retval if defined $retval;
2403 close $fd;
2404 }
2405 if (defined $o_git_dir) {
2406 $git_dir = $o_git_dir;
2407 }
2408 return $retval;
2409 }
2410
2411 # get type of given object
2412 sub git_get_type {
2413 my $hash = shift;
2414
2415 open my $fd, "-|", git_cmd(), "cat-file", '-t', $hash or return;
2416 my $type = <$fd>;
2417 close $fd or return;
2418 chomp $type;
2419 return $type;
2420 }
2421
2422 # repository configuration
2423 our $config_file = '';
2424 our %config;
2425
2426 # store multiple values for single key as anonymous array reference
2427 # single values stored directly in the hash, not as [ <value> ]
2428 sub hash_set_multi {
2429 my ($hash, $key, $value) = @_;
2430
2431 if (!exists $hash->{$key}) {
2432 $hash->{$key} = $value;
2433 } elsif (!ref $hash->{$key}) {
2434 $hash->{$key} = [ $hash->{$key}, $value ];
2435 } else {
2436 push @{$hash->{$key}}, $value;
2437 }
2438 }
2439
2440 # return hash of git project configuration
2441 # optionally limited to some section, e.g. 'gitweb'
2442 sub git_parse_project_config {
2443 my $section_regexp = shift;
2444 my %config;
2445
2446 local $/ = "\0";
2447
2448 open my $fh, "-|", git_cmd(), "config", '-z', '-l',
2449 or return;
2450
2451 while (my $keyval = <$fh>) {
2452 chomp $keyval;
2453 my ($key, $value) = split(/\n/, $keyval, 2);
2454
2455 hash_set_multi(\%config, $key, $value)
2456 if (!defined $section_regexp || $key =~ /^(?:$section_regexp)\./o);
2457 }
2458 close $fh;
2459
2460 return %config;
2461 }
2462
2463 # convert config value to boolean: 'true' or 'false'
2464 # no value, number > 0, 'true' and 'yes' values are true
2465 # rest of values are treated as false (never as error)
2466 sub config_to_bool {
2467 my $val = shift;
2468
2469 return 1 if !defined $val; # section.key
2470
2471 # strip leading and trailing whitespace
2472 $val =~ s/^\s+//;
2473 $val =~ s/\s+$//;
2474
2475 return (($val =~ /^\d+$/ && $val) || # section.key = 1
2476 ($val =~ /^(?:true|yes)$/i)); # section.key = true
2477 }
2478
2479 # convert config value to simple decimal number
2480 # an optional value suffix of 'k', 'm', or 'g' will cause the value
2481 # to be multiplied by 1024, 1048576, or 1073741824
2482 sub config_to_int {
2483 my $val = shift;
2484
2485 # strip leading and trailing whitespace
2486 $val =~ s/^\s+//;
2487 $val =~ s/\s+$//;
2488
2489 if (my ($num, $unit) = ($val =~ /^([0-9]*)([kmg])$/i)) {
2490 $unit = lc($unit);
2491 # unknown unit is treated as 1
2492 return $num * ($unit eq 'g' ? 1073741824 :
2493 $unit eq 'm' ? 1048576 :
2494 $unit eq 'k' ? 1024 : 1);
2495 }
2496 return $val;
2497 }
2498
2499 # convert config value to array reference, if needed
2500 sub config_to_multi {
2501 my $val = shift;
2502
2503 return ref($val) ? $val : (defined($val) ? [ $val ] : []);
2504 }
2505
2506 sub git_get_project_config {
2507 my ($key, $type) = @_;
2508
2509 return unless defined $git_dir;
2510
2511 # key sanity check
2512 return unless ($key);
2513 $key =~ s/^gitweb\.//;
2514 return if ($key =~ m/\W/);
2515
2516 # type sanity check
2517 if (defined $type) {
2518 $type =~ s/^--//;
2519 $type = undef
2520 unless ($type eq 'bool' || $type eq 'int');
2521 }
2522
2523 # get config
2524 if (!defined $config_file ||
2525 $config_file ne "$git_dir/config") {
2526 %config = git_parse_project_config('gitweb');
2527 $config_file = "$git_dir/config";
2528 }
2529
2530 # check if config variable (key) exists
2531 return unless exists $config{"gitweb.$key"};
2532
2533 # ensure given type
2534 if (!defined $type) {
2535 return $config{"gitweb.$key"};
2536 } elsif ($type eq 'bool') {
2537 # backward compatibility: 'git config --bool' returns true/false
2538 return config_to_bool($config{"gitweb.$key"}) ? 'true' : 'false';
2539 } elsif ($type eq 'int') {
2540 return config_to_int($config{"gitweb.$key"});
2541 }
2542 return $config{"gitweb.$key"};
2543 }
2544
2545 # get hash of given path at given ref
2546 sub git_get_hash_by_path {
2547 my $base = shift;
2548 my $path = shift || return undef;
2549 my $type = shift;
2550
2551 $path =~ s,/+$,,;
2552
2553 open my $fd, "-|", git_cmd(), "ls-tree", $base, "--", $path
2554 or die_error(500, "Open git-ls-tree failed");
2555 my $line = <$fd>;
2556 close $fd or return undef;
2557
2558 if (!defined $line) {
2559 # there is no tree or hash given by $path at $base
2560 return undef;
2561 }
2562
2563 #'100644 blob 0fa3f3a66fb6a137f6ec2c19351ed4d807070ffa panic.c'
2564 $line =~ m/^([0-9]+) (.+) ([0-9a-fA-F]{40})\t/;
2565 if (defined $type && $type ne $2) {
2566 # type doesn't match
2567 return undef;
2568 }
2569 return $3;
2570 }
2571
2572 # get path of entry with given hash at given tree-ish (ref)
2573 # used to get 'from' filename for combined diff (merge commit) for renames
2574 sub git_get_path_by_hash {
2575 my $base = shift || return;
2576 my $hash = shift || return;
2577
2578 local $/ = "\0";
2579
2580 open my $fd, "-|", git_cmd(), "ls-tree", '-r', '-t', '-z', $base
2581 or return undef;
2582 while (my $line = <$fd>) {
2583 chomp $line;
2584
2585 #'040000 tree 595596a6a9117ddba9fe379b6b012b558bac8423 gitweb'
2586 #'100644 blob e02e90f0429be0d2a69b76571101f20b8f75530f gitweb/README'
2587 if ($line =~ m/(?:[0-9]+) (?:.+) $hash\t(.+)$/) {
2588 close $fd;
2589 return $1;
2590 }
2591 }
2592 close $fd;
2593 return undef;
2594 }
2595
2596 ## ......................................................................
2597 ## git utility functions, directly accessing git repository
2598
2599 # get the value of config variable either from file named as the variable
2600 # itself in the repository ($GIT_DIR/$name file), or from gitweb.$name
2601 # configuration variable in the repository config file.
2602 sub git_get_file_or_project_config {
2603 my ($path, $name) = @_;
2604
2605 $git_dir = "$projectroot/$path";
2606 open my $fd, '<', "$git_dir/$name"
2607 or return git_get_project_config($name);
2608 my $conf = <$fd>;
2609 close $fd;
2610 if (defined $conf) {
2611 chomp $conf;
2612 }
2613 return $conf;
2614 }
2615
2616 sub git_get_project_description {
2617 my $path = shift;
2618 return git_get_file_or_project_config($path, 'description');
2619 }
2620
2621 sub git_get_project_category {
2622 my $path = shift;
2623 return git_get_file_or_project_config($path, 'category');
2624 }
2625
2626
2627 # supported formats:
2628 # * $GIT_DIR/ctags/<tagname> file (in 'ctags' subdirectory)
2629 # - if its contents is a number, use it as tag weight,
2630 # - otherwise add a tag with weight 1
2631 # * $GIT_DIR/ctags file, each line is a tag (with weight 1)
2632 # the same value multiple times increases tag weight
2633 # * `gitweb.ctag' multi-valued repo config variable
2634 sub git_get_project_ctags {
2635 my $project = shift;
2636 my $ctags = {};
2637
2638 $git_dir = "$projectroot/$project";
2639 if (opendir my $dh, "$git_dir/ctags") {
2640 my @files = grep { -f $_ } map { "$git_dir/ctags/$_" } readdir($dh);
2641 foreach my $tagfile (@files) {
2642 open my $ct, '<', $tagfile
2643 or next;
2644 my $val = <$ct>;
2645 chomp $val if $val;
2646 close $ct;
2647
2648 (my $ctag = $tagfile) =~ s#.*/##;
2649 if ($val =~ /^\d+$/) {
2650 $ctags->{$ctag} = $val;
2651 } else {
2652 $ctags->{$ctag} = 1;
2653 }
2654 }
2655 closedir $dh;
2656
2657 } elsif (open my $fh, '<', "$git_dir/ctags") {
2658 while (my $line = <$fh>) {
2659 chomp $line;
2660 $ctags->{$line}++ if $line;
2661 }
2662 close $fh;
2663
2664 } else {
2665 my $taglist = config_to_multi(git_get_project_config('ctag'));
2666 foreach my $tag (@$taglist) {
2667 $ctags->{$tag}++;
2668 }
2669 }
2670
2671 return $ctags;
2672 }
2673
2674 # return hash, where keys are content tags ('ctags'),
2675 # and values are sum of weights of given tag in every project
2676 sub git_gather_all_ctags {
2677 my $projects = shift;
2678 my $ctags = {};
2679
2680 foreach my $p (@$projects) {
2681 foreach my $ct (keys %{$p->{'ctags'}}) {
2682 $ctags->{$ct} += $p->{'ctags'}->{$ct};
2683 }
2684 }
2685
2686 return $ctags;
2687 }
2688
2689 sub git_populate_project_tagcloud {
2690 my $ctags = shift;
2691
2692 # First, merge different-cased tags; tags vote on casing
2693 my %ctags_lc;
2694 foreach (keys %$ctags) {
2695 $ctags_lc{lc $_}->{count} += $ctags->{$_};
2696 if (not $ctags_lc{lc $_}->{topcount}
2697 or $ctags_lc{lc $_}->{topcount} < $ctags->{$_}) {
2698 $ctags_lc{lc $_}->{topcount} = $ctags->{$_};
2699 $ctags_lc{lc $_}->{topname} = $_;
2700 }
2701 }
2702
2703 my $cloud;
2704 my $matched = $cgi->param('by_tag');
2705 if (eval { require HTML::TagCloud; 1; }) {
2706 $cloud = HTML::TagCloud->new;
2707 foreach my $ctag (sort keys %ctags_lc) {
2708 # Pad the title with spaces so that the cloud looks
2709 # less crammed.
2710 my $title = esc_html($ctags_lc{$ctag}->{topname});
2711 $title =~ s/ /&nbsp;/g;
2712 $title =~ s/^/&nbsp;/g;
2713 $title =~ s/$/&nbsp;/g;
2714 if (defined $matched && $matched eq $ctag) {
2715 $title = qq(<span class="match">$title</span>);
2716 }
2717 $cloud->add($title, href(project=>undef, ctag=>$ctag),
2718 $ctags_lc{$ctag}->{count});
2719 }
2720 } else {
2721 $cloud = {};
2722 foreach my $ctag (keys %ctags_lc) {
2723 my $title = esc_html($ctags_lc{$ctag}->{topname}, -nbsp=>1);
2724 if (defined $matched && $matched eq $ctag) {
2725 $title = qq(<span class="match">$title</span>);
2726 }
2727 $cloud->{$ctag}{count} = $ctags_lc{$ctag}->{count};
2728 $cloud->{$ctag}{ctag} =
2729 $cgi->a({-href=>href(project=>undef, ctag=>$ctag)}, $title);
2730 }
2731 }
2732 return $cloud;
2733 }
2734
2735 sub git_show_project_tagcloud {
2736 my ($cloud, $count) = @_;
2737 if (ref $cloud eq 'HTML::TagCloud') {
2738 return $cloud->html_and_css($count);
2739 } else {
2740 my @tags = sort { $cloud->{$a}->{'count'} <=> $cloud->{$b}->{'count'} } keys %$cloud;
2741 return
2742 '<div id="htmltagcloud"'.($project ? '' : ' align="center"').'>' .
2743 join (', ', map {
2744 $cloud->{$_}->{'ctag'}
2745 } splice(@tags, 0, $count)) .
2746 '</div>';
2747 }
2748 }
2749
2750 sub git_get_project_url_list {
2751 my $path = shift;
2752
2753 $git_dir = "$projectroot/$path";
2754 open my $fd, '<', "$git_dir/cloneurl"
2755 or return wantarray ?
2756 @{ config_to_multi(git_get_project_config('url')) } :
2757 config_to_multi(git_get_project_config('url'));
2758 my @git_project_url_list = map { chomp; $_ } <$fd>;
2759 close $fd;
2760
2761 return wantarray ? @git_project_url_list : \@git_project_url_list;
2762 }
2763
2764 sub git_get_projects_list {
2765 my $filter = shift || '';
2766 my @list;
2767
2768 $filter =~ s/\.git$//;
2769
2770 if (-d $projects_list) {
2771 # search in directory
2772 my $dir = $projects_list;
2773 # remove the trailing "/"
2774 $dir =~ s!/+$!!;
2775 my $pfxlen = length("$projects_list");
2776 my $pfxdepth = ($projects_list =~ tr!/!!);
2777 # when filtering, search only given subdirectory
2778 if ($filter) {
2779 $dir .= "/$filter";
2780 $dir =~ s!/+$!!;
2781 }
2782
2783 File::Find::find({
2784 follow_fast => 1, # follow symbolic links
2785 follow_skip => 2, # ignore duplicates
2786 dangling_symlinks => 0, # ignore dangling symlinks, silently
2787 wanted => sub {
2788 # global variables
2789 our $project_maxdepth;
2790 our $projectroot;
2791 # skip project-list toplevel, if we get it.
2792 return if (m!^[/.]$!);
2793 # only directories can be git repositories
2794 return unless (-d $_);
2795 # don't traverse too deep (Find is super slow on os x)
2796 # $project_maxdepth excludes depth of $projectroot
2797 if (($File::Find::name =~ tr!/!!) - $pfxdepth > $project_maxdepth) {
2798 $File::Find::prune = 1;
2799 return;
2800 }
2801
2802 my $path = substr($File::Find::name, $pfxlen + 1);
2803 # we check related file in $projectroot
2804 if (check_export_ok("$projectroot/$path")) {
2805 push @list, { path => $path };
2806 $File::Find::prune = 1;
2807 }
2808 },
2809 }, "$dir");
2810
2811 } elsif (-f $projects_list) {
2812 # read from file(url-encoded):
2813 # 'git%2Fgit.git Linus+Torvalds'
2814 # 'libs%2Fklibc%2Fklibc.git H.+Peter+Anvin'
2815 # 'linux%2Fhotplug%2Fudev.git Greg+Kroah-Hartman'
2816 open my $fd, '<', $projects_list or return;
2817 PROJECT:
2818 while (my $line = <$fd>) {
2819 chomp $line;
2820 my ($path, $owner) = split ' ', $line;
2821 $path = unescape($path);
2822 $owner = unescape($owner);
2823 if (!defined $path) {
2824 next;
2825 }
2826 # if $filter is rpovided, check if $path begins with $filter
2827 if ($filter && $path !~ m!^\Q$filter\E/!) {
2828 next;
2829 }
2830 if (check_export_ok("$projectroot/$path")) {
2831 my $pr = {
2832 path => $path,
2833 owner => to_utf8($owner),
2834 };
2835 push @list, $pr;
2836 }
2837 }
2838 close $fd;
2839 }
2840 return @list;
2841 }
2842
2843 # written with help of Tree::Trie module (Perl Artistic License, GPL compatibile)
2844 # as side effects it sets 'forks' field to list of forks for forked projects
2845 sub filter_forks_from_projects_list {
2846 my $projects = shift;
2847
2848 my %trie; # prefix tree of directories (path components)
2849 # generate trie out of those directories that might contain forks
2850 foreach my $pr (@$projects) {
2851 my $path = $pr->{'path'};
2852 $path =~ s/\.git$//; # forks of 'repo.git' are in 'repo/' directory
2853 next if ($path =~ m!/$!); # skip non-bare repositories, e.g. 'repo/.git'
2854 next unless ($path); # skip '.git' repository: tests, git-instaweb
2855 next unless (-d $path); # containing directory exists
2856 $pr->{'forks'} = []; # there can be 0 or more forks of project
2857
2858 # add to trie
2859 my @dirs = split('/', $path);
2860 # walk the trie, until either runs out of components or out of trie
2861 my $ref = \%trie;
2862 while (scalar @dirs &&
2863 exists($ref->{$dirs[0]})) {
2864 $ref = $ref->{shift @dirs};
2865 }
2866 # create rest of trie structure from rest of components
2867 foreach my $dir (@dirs) {
2868 $ref = $ref->{$dir} = {};
2869 }
2870 # create end marker, store $pr as a data
2871 $ref->{''} = $pr if (!exists $ref->{''});
2872 }
2873
2874 # filter out forks, by finding shortest prefix match for paths
2875 my @filtered;
2876 PROJECT:
2877 foreach my $pr (@$projects) {
2878 # trie lookup
2879 my $ref = \%trie;
2880 DIR:
2881 foreach my $dir (split('/', $pr->{'path'})) {
2882 if (exists $ref->{''}) {
2883 # found [shortest] prefix, is a fork - skip it
2884 push @{$ref->{''}{'forks'}}, $pr;
2885 next PROJECT;
2886 }
2887 if (!exists $ref->{$dir}) {
2888 # not in trie, cannot have prefix, not a fork
2889 push @filtered, $pr;
2890 next PROJECT;
2891 }
2892 # If the dir is there, we just walk one step down the trie.
2893 $ref = $ref->{$dir};
2894 }
2895 # we ran out of trie
2896 # (shouldn't happen: it's either no match, or end marker)
2897 push @filtered, $pr;
2898 }
2899
2900 return @filtered;
2901 }
2902
2903 # note: fill_project_list_info must be run first,
2904 # for 'descr_long' and 'ctags' to be filled
2905 sub search_projects_list {
2906 my ($projlist, %opts) = @_;
2907 my $tagfilter = $opts{'tagfilter'};
2908 my $searchtext = $opts{'searchtext'};
2909
2910 return @$projlist
2911 unless ($tagfilter || $searchtext);
2912
2913 my @projects;
2914 PROJECT:
2915 foreach my $pr (@$projlist) {
2916
2917 if ($tagfilter) {
2918 next unless ref($pr->{'ctags'}) eq 'HASH';
2919 next unless
2920 grep { lc($_) eq lc($tagfilter) } keys %{$pr->{'ctags'}};
2921 }
2922
2923 if ($searchtext) {
2924 next unless
2925 $pr->{'path'} =~ /$searchtext/ ||
2926 $pr->{'descr_long'} =~ /$searchtext/;
2927 }
2928
2929 push @projects, $pr;
2930 }
2931
2932 return @projects;
2933 }
2934
2935 our $gitweb_project_owner = undef;
2936 sub git_get_project_list_from_file {
2937
2938 return if (defined $gitweb_project_owner);
2939
2940 $gitweb_project_owner = {};
2941 # read from file (url-encoded):
2942 # 'git%2Fgit.git Linus+Torvalds'
2943 # 'libs%2Fklibc%2Fklibc.git H.+Peter+Anvin'
2944 # 'linux%2Fhotplug%2Fudev.git Greg+Kroah-Hartman'
2945 if (-f $projects_list) {
2946 open(my $fd, '<', $projects_list);
2947 while (my $line = <$fd>) {
2948 chomp $line;
2949 my ($pr, $ow) = split ' ', $line;
2950 $pr = unescape($pr);
2951 $ow = unescape($ow);
2952 $gitweb_project_owner->{$pr} = to_utf8($ow);
2953 }
2954 close $fd;
2955 }
2956 }
2957
2958 sub git_get_project_owner {
2959 my $project = shift;
2960 my $owner;
2961
2962 return undef unless $project;
2963 $git_dir = "$projectroot/$project";
2964
2965 if (!defined $gitweb_project_owner) {
2966 git_get_project_list_from_file();
2967 }
2968
2969 if (exists $gitweb_project_owner->{$project}) {
2970 $owner = $gitweb_project_owner->{$project};
2971 }
2972 if (!defined $owner){
2973 $owner = git_get_project_config('owner');
2974 }
2975 if (!defined $owner) {
2976 $owner = get_file_owner("$git_dir");
2977 }
2978
2979 return $owner;
2980 }
2981
2982 sub git_get_last_activity {
2983 my ($path) = @_;
2984 my $fd;
2985
2986 $git_dir = "$projectroot/$path";
2987 open($fd, "-|", git_cmd(), 'for-each-ref',
2988 '--format=%(committer)',
2989 '--sort=-committerdate',
2990 '--count=1',
2991 'refs/heads') or return;
2992 my $most_recent = <$fd>;
2993 close $fd or return;
2994 if (defined $most_recent &&
2995 $most_recent =~ / (\d+) [-+][01]\d\d\d$/) {
2996 my $timestamp = $1;
2997 my $age = time - $timestamp;
2998 return ($age, age_string($age));
2999 }
3000 return (undef, undef);
3001 }
3002
3003 # Implementation note: when a single remote is wanted, we cannot use 'git
3004 # remote show -n' because that command always work (assuming it's a remote URL
3005 # if it's not defined), and we cannot use 'git remote show' because that would
3006 # try to make a network roundtrip. So the only way to find if that particular
3007 # remote is defined is to walk the list provided by 'git remote -v' and stop if
3008 # and when we find what we want.
3009 sub git_get_remotes_list {
3010 my $wanted = shift;
3011 my %remotes = ();
3012
3013 open my $fd, '-|' , git_cmd(), 'remote', '-v';
3014 return unless $fd;
3015 while (my $remote = <$fd>) {
3016 chomp $remote;
3017 $remote =~ s!\t(.*?)\s+\((\w+)\)$!!;
3018 next if $wanted and not $remote eq $wanted;
3019 my ($url, $key) = ($1, $2);
3020
3021 $remotes{$remote} ||= { 'heads' => () };
3022 $remotes{$remote}{$key} = $url;
3023 }
3024 close $fd or return;
3025 return wantarray ? %remotes : \%remotes;
3026 }
3027
3028 # Takes a hash of remotes as first parameter and fills it by adding the
3029 # available remote heads for each of the indicated remotes.
3030 sub fill_remote_heads {
3031 my $remotes = shift;
3032 my @heads = map { "remotes/$_" } keys %$remotes;
3033 my @remoteheads = git_get_heads_list(undef, @heads);
3034 foreach my $remote (keys %$remotes) {
3035 $remotes->{$remote}{'heads'} = [ grep {
3036 $_->{'name'} =~ s!^$remote/!!
3037 } @remoteheads ];
3038 }
3039 }
3040
3041 sub git_get_references {
3042 my $type = shift || "";
3043 my %refs;
3044 # 5dc01c595e6c6ec9ccda4f6f69c131c0dd945f8c refs/tags/v2.6.11
3045 # c39ae07f393806ccf406ef966e9a15afc43cc36a refs/tags/v2.6.11^{}
3046 open my $fd, "-|", git_cmd(), "show-ref", "--dereference",
3047 ($type ? ("--", "refs/$type") : ()) # use -- <pattern> if $type
3048 or return;
3049
3050 while (my $line = <$fd>) {
3051 chomp $line;
3052 if ($line =~ m!^([0-9a-fA-F]{40})\srefs/($type.*)$!) {
3053 if (defined $refs{$1}) {
3054 push @{$refs{$1}}, $2;
3055 } else {
3056 $refs{$1} = [ $2 ];
3057 }
3058 }
3059 }
3060 close $fd or return;
3061 return \%refs;
3062 }
3063
3064 sub git_get_rev_name_tags {
3065 my $hash = shift || return undef;
3066
3067 open my $fd, "-|", git_cmd(), "name-rev", "--tags", $hash
3068 or return;
3069 my $name_rev = <$fd>;
3070 close $fd;
3071
3072 if ($name_rev =~ m|^$hash tags/(.*)$|) {
3073 return $1;
3074 } else {
3075 # catches also '$hash undefined' output
3076 return undef;
3077 }
3078 }
3079
3080 ## ----------------------------------------------------------------------
3081 ## parse to hash functions
3082
3083 sub parse_date {
3084 my $epoch = shift;
3085 my $tz = shift || "-0000";
3086
3087 my %date;
3088 my @months = ("Jan", "Feb", "Mar", "Apr", "May", "Jun", "Jul", "Aug", "Sep", "Oct", "Nov", "Dec");
3089 my @days = ("Sun", "Mon", "Tue", "Wed", "Thu", "Fri", "Sat");
3090 my ($sec, $min, $hour, $mday, $mon, $year, $wday, $yday) = gmtime($epoch);
3091 $date{'hour'} = $hour;
3092 $date{'minute'} = $min;
3093 $date{'mday'} = $mday;
3094 $date{'day'} = $days[$wday];
3095 $date{'month'} = $months[$mon];
3096 $date{'rfc2822'} = sprintf "%s, %d %s %4d %02d:%02d:%02d +0000",
3097 $days[$wday], $mday, $months[$mon], 1900+$year, $hour ,$min, $sec;
3098 $date{'mday-time'} = sprintf "%d %s %02d:%02d",
3099 $mday, $months[$mon], $hour ,$min;
3100 $date{'iso-8601'} = sprintf "%04d-%02d-%02dT%02d:%02d:%02dZ",
3101 1900+$year, 1+$mon, $mday, $hour ,$min, $sec;
3102
3103 my ($tz_sign, $tz_hour, $tz_min) =
3104 ($tz =~ m/^([-+])(\d\d)(\d\d)$/);
3105 $tz_sign = ($tz_sign eq '-' ? -1 : +1);
3106 my $local = $epoch + $tz_sign*((($tz_hour*60) + $tz_min)*60);
3107 ($sec, $min, $hour, $mday, $mon, $year, $wday, $yday) = gmtime($local);
3108 $date{'hour_local'} = $hour;
3109 $date{'minute_local'} = $min;
3110 $date{'tz_local'} = $tz;
3111 $date{'iso-tz'} = sprintf("%04d-%02d-%02d %02d:%02d:%02d %s",
3112 1900+$year, $mon+1, $mday,
3113 $hour, $min, $sec, $tz);
3114 return %date;
3115 }
3116
3117 sub parse_tag {
3118 my $tag_id = shift;
3119 my %tag;
3120 my @comment;
3121
3122 open my $fd, "-|", git_cmd(), "cat-file", "tag", $tag_id or return;
3123 $tag{'id'} = $tag_id;
3124 while (my $line = <$fd>) {
3125 chomp $line;
3126 if ($line =~ m/^object ([0-9a-fA-F]{40})$/) {
3127 $tag{'object'} = $1;
3128 } elsif ($line =~ m/^type (.+)$/) {
3129 $tag{'type'} = $1;
3130 } elsif ($line =~ m/^tag (.+)$/) {
3131 $tag{'name'} = $1;
3132 } elsif ($line =~ m/^tagger (.*) ([0-9]+) (.*)$/) {
3133 $tag{'author'} = $1;
3134 $tag{'author_epoch'} = $2;
3135 $tag{'author_tz'} = $3;
3136 if ($tag{'author'} =~ m/^([^<]+) <([^>]*)>/) {
3137 $tag{'author_name'} = $1;
3138 $tag{'author_email'} = $2;
3139 } else {
3140 $tag{'author_name'} = $tag{'author'};
3141 }
3142 } elsif ($line =~ m/--BEGIN/) {
3143 push @comment, $line;
3144 last;
3145 } elsif ($line eq "") {
3146 last;
3147 }
3148 }
3149 push @comment, <$fd>;
3150 $tag{'comment'} = \@comment;
3151 close $fd or return;
3152 if (!defined $tag{'name'}) {
3153 return
3154 };
3155 return %tag
3156 }
3157
3158 sub parse_commit_text {
3159 my ($commit_text, $withparents) = @_;
3160 my @commit_lines = split '\n', $commit_text;
3161 my %co;
3162
3163 pop @commit_lines; # Remove '\0'
3164
3165 if (! @commit_lines) {
3166 return;
3167 }
3168
3169 my $header = shift @commit_lines;
3170 if ($header !~ m/^[0-9a-fA-F]{40}/) {
3171 return;
3172 }
3173 ($co{'id'}, my @parents) = split ' ', $header;
3174 while (my $line = shift @commit_lines) {
3175 last if $line eq "\n";
3176 if ($line =~ m/^tree ([0-9a-fA-F]{40})$/) {
3177 $co{'tree'} = $1;
3178 } elsif ((!defined $withparents) && ($line =~ m/^parent ([0-9a-fA-F]{40})$/)) {
3179 push @parents, $1;
3180 } elsif ($line =~ m/^author (.*) ([0-9]+) (.*)$/) {
3181 $co{'author'} = to_utf8($1);
3182 $co{'author_epoch'} = $2;
3183 $co{'author_tz'} = $3;
3184 if ($co{'author'} =~ m/^([^<]+) <([^>]*)>/) {
3185 $co{'author_name'} = $1;
3186 $co{'author_email'} = $2;
3187 } else {
3188 $co{'author_name'} = $co{'author'};
3189 }
3190 } elsif ($line =~ m/^committer (.*) ([0-9]+) (.*)$/) {
3191 $co{'committer'} = to_utf8($1);
3192 $co{'committer_epoch'} = $2;
3193 $co{'committer_tz'} = $3;
3194 if ($co{'committer'} =~ m/^([^<]+) <([^>]*)>/) {
3195 $co{'committer_name'} = $1;
3196 $co{'committer_email'} = $2;
3197 } else {
3198 $co{'committer_name'} = $co{'committer'};
3199 }
3200 }
3201 }
3202 if (!defined $co{'tree'}) {
3203 return;
3204 };
3205 $co{'parents'} = \@parents;
3206 $co{'parent'} = $parents[0];
3207
3208 foreach my $title (@commit_lines) {
3209 $title =~ s/^ //;
3210 if ($title ne "") {
3211 $co{'title'} = chop_str($title, 80, 5);
3212 # remove leading stuff of merges to make the interesting part visible
3213 if (length($title) > 50) {
3214 $title =~ s/^Automatic //;
3215 $title =~ s/^merge (of|with) /Merge ... /i;
3216 if (length($title) > 50) {
3217 $title =~ s/(http|rsync):\/\///;
3218 }
3219 if (length($title) > 50) {
3220 $title =~ s/(master|www|rsync)\.//;
3221 }
3222 if (length($title) > 50) {
3223 $title =~ s/kernel.org:?//;
3224 }
3225 if (length($title) > 50) {
3226 $title =~ s/\/pub\/scm//;
3227 }
3228 }
3229 $co{'title_short'} = chop_str($title, 50, 5);
3230 last;
3231 }
3232 }
3233 if (! defined $co{'title'} || $co{'title'} eq "") {
3234 $co{'title'} = $co{'title_short'} = '(no commit message)';
3235 }
3236 # remove added spaces
3237 foreach my $line (@commit_lines) {
3238 $line =~ s/^ //;
3239 }
3240 $co{'comment'} = \@commit_lines;
3241
3242 my $age = time - $co{'committer_epoch'};
3243 $co{'age'} = $age;
3244 $co{'age_string'} = age_string($age);
3245 my ($sec, $min, $hour, $mday, $mon, $year, $wday, $yday) = gmtime($co{'committer_epoch'});
3246 if ($age > 60*60*24*7*2) {
3247 $co{'age_string_date'} = sprintf "%4i-%02u-%02i", 1900 + $year, $mon+1, $mday;
3248 $co{'age_string_age'} = $co{'age_string'};
3249 } else {
3250 $co{'age_string_date'} = $co{'age_string'};
3251 $co{'age_string_age'} = sprintf "%4i-%02u-%02i", 1900 + $year, $mon+1, $mday;
3252 }
3253 return %co;
3254 }
3255
3256 sub parse_commit {
3257 my ($commit_id) = @_;
3258 my %co;
3259
3260 local $/ = "\0";
3261
3262 open my $fd, "-|", git_cmd(), "rev-list",
3263 "--parents",
3264 "--header",
3265 "--max-count=1",
3266 $commit_id,
3267 "--",
3268 or die_error(500, "Open git-rev-list failed");
3269 %co = parse_commit_text(<$fd>, 1);
3270 close $fd;
3271
3272 return %co;
3273 }
3274
3275 sub parse_commits {
3276 my ($commit_id, $maxcount, $skip, $filename, @args) = @_;
3277 my @cos;
3278
3279 $maxcount ||= 1;
3280 $skip ||= 0;
3281
3282 local $/ = "\0";
3283
3284 open my $fd, "-|", git_cmd(), "rev-list",
3285 "--header",
3286 @args,
3287 ("--max-count=" . $maxcount),
3288 ("--skip=" . $skip),
3289 @extra_options,
3290 $commit_id,
3291 "--",
3292 ($filename ? ($filename) : ())
3293 or die_error(500, "Open git-rev-list failed");
3294 while (my $line = <$fd>) {
3295 my %co = parse_commit_text($line);
3296 push @cos, \%co;
3297 }
3298 close $fd;
3299
3300 return wantarray ? @cos : \@cos;
3301 }
3302
3303 # parse line of git-diff-tree "raw" output
3304 sub parse_difftree_raw_line {
3305 my $line = shift;
3306 my %res;
3307
3308 # ':100644 100644 03b218260e99b78c6df0ed378e59ed9205ccc96d 3b93d5e7cc7f7dd4ebed13a5cc1a4ad976fc94d8 M ls-files.c'
3309 # ':100644 100644 7f9281985086971d3877aca27704f2aaf9c448ce bc190ebc71bbd923f2b728e505408f5e54bd073a M rev-tree.c'
3310 if ($line =~ m/^:([0-7]{6}) ([0-7]{6}) ([0-9a-fA-F]{40}) ([0-9a-fA-F]{40}) (.)([0-9]{0,3})\t(.*)$/) {
3311 $res{'from_mode'} = $1;
3312 $res{'to_mode'} = $2;
3313 $res{'from_id'} = $3;
3314 $res{'to_id'} = $4;
3315 $res{'status'} = $5;
3316 $res{'similarity'} = $6;
3317 if ($res{'status'} eq 'R' || $res{'status'} eq 'C') { # renamed or copied
3318 ($res{'from_file'}, $res{'to_file'}) = map { unquote($_) } split("\t", $7);
3319 } else {
3320 $res{'from_file'} = $res{'to_file'} = $res{'file'} = unquote($7);
3321 }
3322 }
3323 # '::100755 100755 100755 60e79ca1b01bc8b057abe17ddab484699a7f5fdb 94067cc5f73388f33722d52ae02f44692bc07490 94067cc5f73388f33722d52ae02f44692bc07490 MR git-gui/git-gui.sh'
3324 # combined diff (for merge commit)
3325 elsif ($line =~ s/^(::+)((?:[0-7]{6} )+)((?:[0-9a-fA-F]{40} )+)([a-zA-Z]+)\t(.*)$//) {
3326 $res{'nparents'} = length($1);
3327 $res{'from_mode'} = [ split(' ', $2) ];
3328 $res{'to_mode'} = pop @{$res{'from_mode'}};
3329 $res{'from_id'} = [ split(' ', $3) ];
3330 $res{'to_id'} = pop @{$res{'from_id'}};
3331 $res{'status'} = [ split('', $4) ];
3332 $res{'to_file'} = unquote($5);
3333 }
3334 # 'c512b523472485aef4fff9e57b229d9d243c967f'
3335 elsif ($line =~ m/^([0-9a-fA-F]{40})$/) {
3336 $res{'commit'} = $1;
3337 }
3338
3339 return wantarray ? %res : \%res;
3340 }
3341
3342 # wrapper: return parsed line of git-diff-tree "raw" output
3343 # (the argument might be raw line, or parsed info)
3344 sub parsed_difftree_line {
3345 my $line_or_ref = shift;
3346
3347 if (ref($line_or_ref) eq "HASH") {
3348 # pre-parsed (or generated by hand)
3349 return $line_or_ref;
3350 } else {
3351 return parse_difftree_raw_line($line_or_ref);
3352 }
3353 }
3354
3355 # parse line of git-ls-tree output
3356 sub parse_ls_tree_line {
3357 my $line = shift;
3358 my %opts = @_;
3359 my %res;
3360
3361 if ($opts{'-l'}) {
3362 #'100644 blob 0fa3f3a66fb6a137f6ec2c19351ed4d807070ffa 16717 panic.c'
3363 $line =~ m/^([0-9]+) (.+) ([0-9a-fA-F]{40}) +(-|[0-9]+)\t(.+)$/s;
3364
3365 $res{'mode'} = $1;
3366 $res{'type'} = $2;
3367 $res{'hash'} = $3;
3368 $res{'size'} = $4;
3369 if ($opts{'-z'}) {
3370 $res{'name'} = $5;
3371 } else {
3372 $res{'name'} = unquote($5);
3373 }
3374 } else {
3375 #'100644 blob 0fa3f3a66fb6a137f6ec2c19351ed4d807070ffa panic.c'
3376 $line =~ m/^([0-9]+) (.+) ([0-9a-fA-F]{40})\t(.+)$/s;
3377
3378 $res{'mode'} = $1;
3379 $res{'type'} = $2;
3380 $res{'hash'} = $3;
3381 if ($opts{'-z'}) {
3382 $res{'name'} = $4;
3383 } else {
3384 $res{'name'} = unquote($4);
3385 }
3386 }
3387
3388 return wantarray ? %res : \%res;
3389 }
3390
3391 # generates _two_ hashes, references to which are passed as 2 and 3 argument
3392 sub parse_from_to_diffinfo {
3393 my ($diffinfo, $from, $to, @parents) = @_;
3394
3395 if ($diffinfo->{'nparents'}) {
3396 # combined diff
3397 $from->{'file'} = [];
3398 $from->{'href'} = [];
3399 fill_from_file_info($diffinfo, @parents)
3400 unless exists $diffinfo->{'from_file'};
3401 for (my $i = 0; $i < $diffinfo->{'nparents'}; $i++) {
3402 $from->{'file'}[$i] =
3403 defined $diffinfo->{'from_file'}[$i] ?
3404 $diffinfo->{'from_file'}[$i] :
3405 $diffinfo->{'to_file'};
3406 if ($diffinfo->{'status'}[$i] ne "A") { # not new (added) file
3407 $from->{'href'}[$i] = href(action=>"blob",
3408 hash_base=>$parents[$i],
3409 hash=>$diffinfo->{'from_id'}[$i],
3410 file_name=>$from->{'file'}[$i]);
3411 } else {
3412 $from->{'href'}[$i] = undef;
3413 }
3414 }
3415 } else {
3416 # ordinary (not combined) diff
3417 $from->{'file'} = $diffinfo->{'from_file'};
3418 if ($diffinfo->{'status'} ne "A") { # not new (added) file
3419 $from->{'href'} = href(action=>"blob", hash_base=>$hash_parent,
3420 hash=>$diffinfo->{'from_id'},
3421 file_name=>$from->{'file'});
3422 } else {
3423 delete $from->{'href'};
3424 }
3425 }
3426
3427 $to->{'file'} = $diffinfo->{'to_file'};
3428 if (!is_deleted($diffinfo)) { # file exists in result
3429 $to->{'href'} = href(action=>"blob", hash_base=>$hash,
3430 hash=>$diffinfo->{'to_id'},
3431 file_name=>$to->{'file'});
3432 } else {
3433 delete $to->{'href'};
3434 }
3435 }
3436
3437 ## ......................................................................
3438 ## parse to array of hashes functions
3439
3440 sub git_get_heads_list {
3441 my ($limit, @classes) = @_;
3442 @classes = ('heads') unless @classes;
3443 my @patterns = map { "refs/$_" } @classes;
3444 my @headslist;
3445
3446 open my $fd, '-|', git_cmd(), 'for-each-ref',
3447 ($limit ? '--count='.($limit+1) : ()), '--sort=-committerdate',
3448 '--format=%(objectname) %(refname) %(subject)%00%(committer)',
3449 @patterns
3450 or return;
3451 while (my $line = <$fd>) {
3452 my %ref_item;
3453
3454 chomp $line;
3455 my ($refinfo, $committerinfo) = split(/\0/, $line);
3456 my ($hash, $name, $title) = split(' ', $refinfo, 3);
3457 my ($committer, $epoch, $tz) =
3458 ($committerinfo =~ /^(.*) ([0-9]+) (.*)$/);
3459 $ref_item{'fullname'} = $name;
3460 $name =~ s!^refs/(?:head|remote)s/!!;
3461
3462 $ref_item{'name'} = $name;
3463 $ref_item{'id'} = $hash;
3464 $ref_item{'title'} = $title || '(no commit message)';
3465 $ref_item{'epoch'} = $epoch;
3466 if ($epoch) {
3467 $ref_item{'age'} = age_string(time - $ref_item{'epoch'});
3468 } else {
3469 $ref_item{'age'} = "unknown";
3470 }
3471
3472 push @headslist, \%ref_item;
3473 }
3474 close $fd;
3475
3476 return wantarray ? @headslist : \@headslist;
3477 }
3478
3479 sub git_get_tags_list {
3480 my $limit = shift;
3481 my @tagslist;
3482
3483 open my $fd, '-|', git_cmd(), 'for-each-ref',
3484 ($limit ? '--count='.($limit+1) : ()), '--sort=-creatordate',
3485 '--format=%(objectname) %(objecttype) %(refname) '.
3486 '%(*objectname) %(*objecttype) %(subject)%00%(creator)',
3487 'refs/tags'
3488 or return;
3489 while (my $line = <$fd>) {
3490 my %ref_item;
3491
3492 chomp $line;
3493 my ($refinfo, $creatorinfo) = split(/\0/, $line);
3494 my ($id, $type, $name, $refid, $reftype, $title) = split(' ', $refinfo, 6);
3495 my ($creator, $epoch, $tz) =
3496 ($creatorinfo =~ /^(.*) ([0-9]+) (.*)$/);
3497 $ref_item{'fullname'} = $name;
3498 $name =~ s!^refs/tags/!!;
3499
3500 $ref_item{'type'} = $type;
3501 $ref_item{'id'} = $id;
3502 $ref_item{'name'} = $name;
3503 if ($type eq "tag") {
3504 $ref_item{'subject'} = $title;
3505 $ref_item{'reftype'} = $reftype;
3506 $ref_item{'refid'} = $refid;
3507 } else {
3508 $ref_item{'reftype'} = $type;
3509 $ref_item{'refid'} = $id;
3510 }
3511
3512 if ($type eq "tag" || $type eq "commit") {
3513 $ref_item{'epoch'} = $epoch;
3514 if ($epoch) {
3515 $ref_item{'age'} = age_string(time - $ref_item{'epoch'});
3516 } else {
3517 $ref_item{'age'} = "unknown";
3518 }
3519 }
3520
3521 push @tagslist, \%ref_item;
3522 }
3523 close $fd;
3524
3525 return wantarray ? @tagslist : \@tagslist;
3526 }
3527
3528 ## ----------------------------------------------------------------------
3529 ## filesystem-related functions
3530
3531 sub get_file_owner {
3532 my $path = shift;
3533
3534 my ($dev, $ino, $mode, $nlink, $st_uid, $st_gid, $rdev, $size) = stat($path);
3535 my ($name, $passwd, $uid, $gid, $quota, $comment, $gcos, $dir, $shell) = getpwuid($st_uid);
3536 if (!defined $gcos) {
3537 return undef;
3538 }
3539 my $owner = $gcos;
3540 $owner =~ s/[,;].*$//;
3541 return to_utf8($owner);
3542 }
3543
3544 # assume that file exists
3545 sub insert_file {
3546 my $filename = shift;
<