Raw
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 require v5.26;
11 use strict;
12 use warnings;
13 # handle ACL in file access tests
14 use filetest 'access';
15 use CGI qw(:standard :escapeHTML -nosticky);
16 use CGI::Util qw(unescape);
17 use CGI::Carp qw(fatalsToBrowser set_message);
18 use Encode;
19 use Fcntl ':mode';
20 use File::Find qw();
21 use File::Basename qw(basename);
22 use Time::HiRes qw(gettimeofday tv_interval);
23 use Digest::MD5 qw(md5_hex);
24
25 binmode STDOUT, ':utf8';
26
27 if (!defined($CGI::VERSION) || $CGI::VERSION < 4.08) {
28 eval 'sub CGI::multi_param { CGI::param(@_) }'
29 }
30
31 our $t0 = [ gettimeofday() ];
32 our $number_of_git_cmds = 0;
33
34 BEGIN {
35 CGI->compile() if $ENV{'MOD_PERL'};
36 }
37
38 our $version = "@GIT_VERSION@";
39
40 our ($my_url, $my_uri, $base_url, $path_info, $home_link);
41 sub evaluate_uri {
42 our $cgi;
43
44 our $my_url = $cgi->url();
45 our $my_uri = $cgi->url(-absolute => 1);
46
47 # Base URL for relative URLs in gitweb ($logo, $favicon, ...),
48 # needed and used only for URLs with nonempty PATH_INFO
49 our $base_url = $my_url;
50
51 # When the script is used as DirectoryIndex, the URL does not contain the name
52 # of the script file itself, and $cgi->url() fails to strip PATH_INFO, so we
53 # have to do it ourselves. We make $path_info global because it's also used
54 # later on.
55 #
56 # Another issue with the script being the DirectoryIndex is that the resulting
57 # $my_url data is not the full script URL: this is good, because we want
58 # generated links to keep implying the script name if it wasn't explicitly
59 # indicated in the URL we're handling, but it means that $my_url cannot be used
60 # as base URL.
61 # Therefore, if we needed to strip PATH_INFO, then we know that we have
62 # to build the base URL ourselves:
63 our $path_info = decode_utf8($ENV{"PATH_INFO"});
64 if ($path_info) {
65 # $path_info has already been URL-decoded by the web server, but
66 # $my_url and $my_uri have not. URL-decode them so we can properly
67 # strip $path_info.
68 $my_url = unescape($my_url);
69 $my_uri = unescape($my_uri);
70 if ($my_url =~ s,\Q$path_info\E$,, &&
71 $my_uri =~ s,\Q$path_info\E$,, &&
72 defined $ENV{'SCRIPT_NAME'}) {
73 $base_url = $cgi->url(-base => 1) . $ENV{'SCRIPT_NAME'};
74 }
75 }
76
77 # target of the home link on top of all pages
78 our $home_link = $my_uri || "/";
79 }
80
81 # core git executable to use
82 # this can just be "git" if your webserver has a sensible PATH
83 our $GIT = "@GIT_BINDIR@/git";
84
85 # absolute fs-path which will be prepended to the project path
86 #our $projectroot = "/pub/scm";
87 our $projectroot = "@GITWEB_PROJECTROOT@";
88
89 # fs traversing limit for getting project list
90 # the number is relative to the projectroot
91 our $project_maxdepth = @GITWEB_PROJECT_MAXDEPTH@;
92
93 # string of the home link on top of all pages
94 our $home_link_str = "@GITWEB_HOME_LINK_STR@";
95
96 # extra breadcrumbs preceding the home link
97 our @extra_breadcrumbs = ();
98
99 # name of your site or organization to appear in page titles
100 # replace this with something more descriptive for clearer bookmarks
101 our $site_name = "@GITWEB_SITENAME@"
102 || ($ENV{'SERVER_NAME'} || "Untitled") . " Git";
103
104 # html snippet to include in the <head> section of each page
105 our $site_html_head_string = "@GITWEB_SITE_HTML_HEAD_STRING@";
106 # filename of html text to include at top of each page
107 our $site_header = "@GITWEB_SITE_HEADER@";
108 # html text to include at home page
109 our $home_text = "@GITWEB_HOMETEXT@";
110 # filename of html text to include at bottom of each page
111 our $site_footer = "@GITWEB_SITE_FOOTER@";
112
113 # URI of stylesheets
114 our @stylesheets = ("@GITWEB_CSS@");
115 # URI of a single stylesheet, which can be overridden in GITWEB_CONFIG.
116 our $stylesheet = undef;
117 # URI of GIT logo (72x27 size)
118 our $logo = "@GITWEB_LOGO@";
119 # URI of GIT favicon, assumed to be image/png type
120 our $favicon = "@GITWEB_FAVICON@";
121 # URI of gitweb.js (JavaScript code for gitweb)
122 our $javascript = "@GITWEB_JS@";
123
124 # URI and label (title) of GIT logo link
125 #our $logo_url = "https://www.kernel.org/pub/software/scm/git/docs/";
126 #our $logo_label = "git documentation";
127 our $logo_url = "https://git-scm.com/";
128 our $logo_label = "git homepage";
129
130 # source of projects list
131 our $projects_list = "@GITWEB_LIST@";
132
133 # the width (in characters) of the projects list "Description" column
134 our $projects_list_description_width = 25;
135
136 # group projects by category on the projects list
137 # (enabled if this variable evaluates to true)
138 our $projects_list_group_categories = 0;
139
140 # default category if none specified
141 # (leave the empty string for no category)
142 our $project_list_default_category = "";
143
144 # default order of projects list
145 # valid values are none, project, descr, owner, and age
146 our $default_projects_order = "project";
147
148 # show repository only if this file exists
149 # (only effective if this variable evaluates to true)
150 our $export_ok = "@GITWEB_EXPORT_OK@";
151
152 # don't generate age column on the projects list page
153 our $omit_age_column = 0;
154
155 # don't generate information about owners of repositories
156 our $omit_owner=0;
157
158 # show repository only if this subroutine returns true
159 # when given the path to the project, for example:
160 # sub { return -e "$_[0]/git-daemon-export-ok"; }
161 our $export_auth_hook = undef;
162
163 # only allow viewing of repositories also shown on the overview page
164 our $strict_export = "@GITWEB_STRICT_EXPORT@";
165
166 # list of git base URLs used for URL to where fetch project from,
167 # i.e. full URL is "$git_base_url/$project"
168 our @git_base_url_list = grep { $_ ne '' } ("@GITWEB_BASE_URL@");
169
170 # default blob_plain mimetype and default charset for text/plain blob
171 our $default_blob_plain_mimetype = 'text/plain';
172 our $default_text_plain_charset = undef;
173
174 # file to use for guessing MIME types before trying /etc/mime.types
175 # (relative to the current git repository)
176 our $mimetypes_file = undef;
177
178 # assume this charset if line contains non-UTF-8 characters;
179 # it should be valid encoding (see Encoding::Supported(3pm) for list),
180 # for which encoding all byte sequences are valid, for example
181 # 'iso-8859-1' aka 'latin1' (it is decoded without checking, so it
182 # could be even 'utf-8' for the old behavior)
183 our $fallback_encoding = 'latin1';
184
185 # rename detection options for git-diff and git-diff-tree
186 # - default is '-M', with the cost proportional to
187 # (number of removed files) * (number of new files).
188 # - more costly is '-C' (which implies '-M'), with the cost proportional to
189 # (number of changed files + number of removed files) * (number of new files)
190 # - even more costly is '-C', '--find-copies-harder' with cost
191 # (number of files in the original tree) * (number of new files)
192 # - one might want to include '-B' option, e.g. '-B', '-M'
193 our @diff_opts = ('-M'); # taken from git_commit
194
195 # Disables features that would allow repository owners to inject script into
196 # the gitweb domain.
197 our $prevent_xss = 0;
198
199 # Path to the highlight executable to use (must be the one from
200 # http://andre-simon.de/zip/download.php due to assumptions about parameters and output).
201 # Useful if highlight is not installed on your webserver's PATH.
202 # [Default: highlight]
203 our $highlight_bin = "@HIGHLIGHT_BIN@";
204
205 # information about snapshot formats that gitweb is capable of serving
206 our %known_snapshot_formats = (
207 # name => {
208 # 'display' => display name,
209 # 'type' => mime type,
210 # 'suffix' => filename suffix,
211 # 'format' => --format for git-archive,
212 # 'compressor' => [compressor command and arguments]
213 # (array reference, optional)
214 # 'disabled' => boolean (optional)}
215 #
216 'tgz' => {
217 'display' => 'tar.gz',
218 'type' => 'application/x-gzip',
219 'suffix' => '.tar.gz',
220 'format' => 'tar',
221 'compressor' => ['gzip', '-n']},
222
223 'tbz2' => {
224 'display' => 'tar.bz2',
225 'type' => 'application/x-bzip2',
226 'suffix' => '.tar.bz2',
227 'format' => 'tar',
228 'compressor' => ['bzip2']},
229
230 'txz' => {
231 'display' => 'tar.xz',
232 'type' => 'application/x-xz',
233 'suffix' => '.tar.xz',
234 'format' => 'tar',
235 'compressor' => ['xz'],
236 'disabled' => 1},
237
238 'zip' => {
239 'display' => 'zip',
240 'type' => 'application/x-zip',
241 'suffix' => '.zip',
242 'format' => 'zip'},
243 );
244
245 # Aliases so we understand old gitweb.snapshot values in repository
246 # configuration.
247 our %known_snapshot_format_aliases = (
248 'gzip' => 'tgz',
249 'bzip2' => 'tbz2',
250 'xz' => 'txz',
251
252 # backward compatibility: legacy gitweb config support
253 'x-gzip' => undef, 'gz' => undef,
254 'x-bzip2' => undef, 'bz2' => undef,
255 'x-zip' => undef, '' => undef,
256 );
257
258 # Pixel sizes for icons and avatars. If the default font sizes or lineheights
259 # are changed, it may be appropriate to change these values too via
260 # $GITWEB_CONFIG.
261 our %avatar_size = (
262 'default' => 16,
263 'double' => 32
264 );
265
266 # Used to set the maximum load that we will still respond to gitweb queries.
267 # If server load exceed this value then return "503 server busy" error.
268 # If gitweb cannot determined server load, it is taken to be 0.
269 # Leave it undefined (or set to 'undef') to turn off load checking.
270 our $maxload = 300;
271
272 # configuration for 'highlight' (http://andre-simon.de/doku/highlight/en/highlight.php)
273 # match by basename
274 our %highlight_basename = (
275 #'Program' => 'py',
276 #'Library' => 'py',
277 'SConstruct' => 'py', # SCons equivalent of Makefile
278 'Makefile' => 'make',
279 );
280 # match by extension
281 our %highlight_ext = (
282 # main extensions, defining name of syntax;
283 # see files in /usr/share/highlight/langDefs/ directory
284 (map { $_ => $_ } qw(py rb java css js tex bib xml awk bat ini spec tcl sql)),
285 # alternate extensions, see /etc/highlight/filetypes.conf
286 (map { $_ => 'c' } qw(c h)),
287 (map { $_ => 'sh' } qw(sh bash zsh ksh)),
288 (map { $_ => 'cpp' } qw(cpp cxx c++ cc)),
289 (map { $_ => 'php' } qw(php php3 php4 php5 phps)),
290 (map { $_ => 'pl' } qw(pl perl pm)), # perhaps also 'cgi'
291 (map { $_ => 'make'} qw(make mak mk)),
292 (map { $_ => 'xml' } qw(xml xhtml html htm)),
293 );
294
295 # You define site-wide feature defaults here; override them with
296 # $GITWEB_CONFIG as necessary.
297 our %feature = (
298 # feature => {
299 # 'sub' => feature-sub (subroutine),
300 # 'override' => allow-override (boolean),
301 # 'default' => [ default options...] (array reference)}
302 #
303 # if feature is overridable (it means that allow-override has true value),
304 # then feature-sub will be called with default options as parameters;
305 # return value of feature-sub indicates if to enable specified feature
306 #
307 # if there is no 'sub' key (no feature-sub), then feature cannot be
308 # overridden
309 #
310 # use gitweb_get_feature(<feature>) to retrieve the <feature> value
311 # (an array) or gitweb_check_feature(<feature>) to check if <feature>
312 # is enabled
313
314 # Enable the 'blame' blob view, showing the last commit that modified
315 # each line in the file. This can be very CPU-intensive.
316
317 # To enable system wide have in $GITWEB_CONFIG
318 # $feature{'blame'}{'default'} = [1];
319 # To have project specific config enable override in $GITWEB_CONFIG
320 # $feature{'blame'}{'override'} = 1;
321 # and in project config gitweb.blame = 0|1;
322 'blame' => {
323 'sub' => sub { feature_bool('blame', @_) },
324 'override' => 0,
325 'default' => [0]},
326
327 # Enable the 'snapshot' link, providing a compressed archive of any
328 # tree. This can potentially generate high traffic if you have large
329 # project.
330
331 # Value is a list of formats defined in %known_snapshot_formats that
332 # you wish to offer.
333 # To disable system wide have in $GITWEB_CONFIG
334 # $feature{'snapshot'}{'default'} = [];
335 # To have project specific config enable override in $GITWEB_CONFIG
336 # $feature{'snapshot'}{'override'} = 1;
337 # and in project config, a comma-separated list of formats or "none"
338 # to disable. Example: gitweb.snapshot = tbz2,zip;
339 'snapshot' => {
340 'sub' => \&feature_snapshot,
341 'override' => 0,
342 'default' => ['tgz']},
343
344 # Enable text search, which will list the commits which match author,
345 # committer or commit text to a given string. Enabled by default.
346 # Project specific override is not supported.
347 #
348 # Note that this controls all search features, which means that if
349 # it is disabled, then 'grep' and 'pickaxe' search would also be
350 # disabled.
351 'search' => {
352 'override' => 0,
353 'default' => [1]},
354
355 # Enable grep search, which will list the files in currently selected
356 # tree containing the given string. Enabled by default. This can be
357 # potentially CPU-intensive, of course.
358 # Note that you need to have 'search' feature enabled too.
359
360 # To enable system wide have in $GITWEB_CONFIG
361 # $feature{'grep'}{'default'} = [1];
362 # To have project specific config enable override in $GITWEB_CONFIG
363 # $feature{'grep'}{'override'} = 1;
364 # and in project config gitweb.grep = 0|1;
365 'grep' => {
366 'sub' => sub { feature_bool('grep', @_) },
367 'override' => 0,
368 'default' => [1]},
369
370 # Enable the pickaxe search, which will list the commits that modified
371 # a given string in a file. This can be practical and quite faster
372 # alternative to 'blame', but still potentially CPU-intensive.
373 # Note that you need to have 'search' feature enabled too.
374
375 # To enable system wide have in $GITWEB_CONFIG
376 # $feature{'pickaxe'}{'default'} = [1];
377 # To have project specific config enable override in $GITWEB_CONFIG
378 # $feature{'pickaxe'}{'override'} = 1;
379 # and in project config gitweb.pickaxe = 0|1;
380 'pickaxe' => {
381 'sub' => sub { feature_bool('pickaxe', @_) },
382 'override' => 0,
383 'default' => [1]},
384
385 # Enable showing size of blobs in a 'tree' view, in a separate
386 # column, similar to what 'ls -l' does. This cost a bit of IO.
387
388 # To disable system wide have in $GITWEB_CONFIG
389 # $feature{'show-sizes'}{'default'} = [0];
390 # To have project specific config enable override in $GITWEB_CONFIG
391 # $feature{'show-sizes'}{'override'} = 1;
392 # and in project config gitweb.showsizes = 0|1;
393 'show-sizes' => {
394 'sub' => sub { feature_bool('showsizes', @_) },
395 'override' => 0,
396 'default' => [1]},
397
398 # Make gitweb use an alternative format of the URLs which can be
399 # more readable and natural-looking: project name is embedded
400 # directly in the path and the query string contains other
401 # auxiliary information. All gitweb installations recognize
402 # URL in either format; this configures in which formats gitweb
403 # generates links.
404
405 # To enable system wide have in $GITWEB_CONFIG
406 # $feature{'pathinfo'}{'default'} = [1];
407 # Project specific override is not supported.
408
409 # Note that you will need to change the default location of CSS,
410 # favicon, logo and possibly other files to an absolute URL. Also,
411 # if gitweb.cgi serves as your indexfile, you will need to force
412 # $my_uri to contain the script name in your $GITWEB_CONFIG.
413 'pathinfo' => {
414 'override' => 0,
415 'default' => [0]},
416
417 # Make gitweb consider projects in project root subdirectories
418 # to be forks of existing projects. Given project $projname.git,
419 # projects matching $projname/*.git will not be shown in the main
420 # projects list, instead a '+' mark will be added to $projname
421 # there and a 'forks' view will be enabled for the project, listing
422 # all the forks. If project list is taken from a file, forks have
423 # to be listed after the main project.
424
425 # To enable system wide have in $GITWEB_CONFIG
426 # $feature{'forks'}{'default'} = [1];
427 # Project specific override is not supported.
428 'forks' => {
429 'override' => 0,
430 'default' => [0]},
431
432 # Insert custom links to the action bar of all project pages.
433 # This enables you mainly to link to third-party scripts integrating
434 # into gitweb; e.g. git-browser for graphical history representation
435 # or custom web-based repository administration interface.
436
437 # The 'default' value consists of a list of triplets in the form
438 # (label, link, position) where position is the label after which
439 # to insert the link and link is a format string where %n expands
440 # to the project name, %f to the project path within the filesystem,
441 # %h to the current hash (h gitweb parameter) and %b to the current
442 # hash base (hb gitweb parameter); %% expands to %.
443
444 # To enable system wide have in $GITWEB_CONFIG e.g.
445 # $feature{'actions'}{'default'} = [('graphiclog',
446 # '/git-browser/by-commit.html?r=%n', 'summary')];
447 # Project specific override is not supported.
448 'actions' => {
449 'override' => 0,
450 'default' => []},
451
452 # Allow gitweb scan project content tags of project repository,
453 # and display the popular Web 2.0-ish "tag cloud" near the projects
454 # list. Note that this is something COMPLETELY different from the
455 # normal Git tags.
456
457 # gitweb by itself can show existing tags, but it does not handle
458 # tagging itself; you need to do it externally, outside gitweb.
459 # The format is described in git_get_project_ctags() subroutine.
460 # You may want to install the HTML::TagCloud Perl module to get
461 # a pretty tag cloud instead of just a list of tags.
462
463 # To enable system wide have in $GITWEB_CONFIG
464 # $feature{'ctags'}{'default'} = [1];
465 # Project specific override is not supported.
466
467 # In the future whether ctags editing is enabled might depend
468 # on the value, but using 1 should always mean no editing of ctags.
469 'ctags' => {
470 'override' => 0,
471 'default' => [0]},
472
473 # The maximum number of patches in a patchset generated in patch
474 # view. Set this to 0 or undef to disable patch view, or to a
475 # negative number to remove any limit.
476
477 # To disable system wide have in $GITWEB_CONFIG
478 # $feature{'patches'}{'default'} = [0];
479 # To have project specific config enable override in $GITWEB_CONFIG
480 # $feature{'patches'}{'override'} = 1;
481 # and in project config gitweb.patches = 0|n;
482 # where n is the maximum number of patches allowed in a patchset.
483 'patches' => {
484 'sub' => \&feature_patches,
485 'override' => 0,
486 'default' => [16]},
487
488 # Avatar support. When this feature is enabled, views such as
489 # shortlog or commit will display an avatar associated with
490 # the email of the committer(s) and/or author(s).
491
492 # Currently available providers are gravatar and picon.
493 # If an unknown provider is specified, the feature is disabled.
494
495 # Picon currently relies on the indiana.edu database.
496
497 # To enable system wide have in $GITWEB_CONFIG
498 # $feature{'avatar'}{'default'} = ['<provider>'];
499 # where <provider> is either gravatar or picon.
500 # To have project specific config enable override in $GITWEB_CONFIG
501 # $feature{'avatar'}{'override'} = 1;
502 # and in project config gitweb.avatar = <provider>;
503 'avatar' => {
504 'sub' => \&feature_avatar,
505 'override' => 0,
506 'default' => ['']},
507
508 # Enable displaying how much time and how many git commands
509 # it took to generate and display page. Disabled by default.
510 # Project specific override is not supported.
511 'timed' => {
512 'override' => 0,
513 'default' => [0]},
514
515 # Enable turning some links into links to actions which require
516 # JavaScript to run (like 'blame_incremental'). Not enabled by
517 # default. Project specific override is currently not supported.
518 'javascript-actions' => {
519 'override' => 0,
520 'default' => [0]},
521
522 # Enable and configure ability to change common timezone for dates
523 # in gitweb output via JavaScript. Enabled by default.
524 # Project specific override is not supported.
525 'javascript-timezone' => {
526 'override' => 0,
527 'default' => [
528 'local', # default timezone: 'utc', 'local', or '(-|+)HHMM' format,
529 # or undef to turn off this feature
530 'gitweb_tz', # name of cookie where to store selected timezone
531 'datetime', # CSS class used to mark up dates for manipulation
532 ]},
533
534 # Syntax highlighting support. This is based on Daniel Svensson's
535 # and Sham Chukoury's work in gitweb-xmms2.git.
536 # It requires the 'highlight' program present in $PATH,
537 # and therefore is disabled by default.
538
539 # To enable system wide have in $GITWEB_CONFIG
540 # $feature{'highlight'}{'default'} = [1];
541
542 'highlight' => {
543 'sub' => sub { feature_bool('highlight', @_) },
544 'override' => 0,
545 'default' => [0]},
546
547 # Enable displaying of remote heads in the heads list
548
549 # To enable system wide have in $GITWEB_CONFIG
550 # $feature{'remote_heads'}{'default'} = [1];
551 # To have project specific config enable override in $GITWEB_CONFIG
552 # $feature{'remote_heads'}{'override'} = 1;
553 # and in project config gitweb.remoteheads = 0|1;
554 'remote_heads' => {
555 'sub' => sub { feature_bool('remote_heads', @_) },
556 'override' => 0,
557 'default' => [0]},
558
559 # Enable showing branches under other refs in addition to heads
560
561 # To set system wide extra branch refs have in $GITWEB_CONFIG
562 # $feature{'extra-branch-refs'}{'default'} = ['dirs', 'of', 'choice'];
563 # To have project specific config enable override in $GITWEB_CONFIG
564 # $feature{'extra-branch-refs'}{'override'} = 1;
565 # and in project config gitweb.extrabranchrefs = dirs of choice
566 # Every directory is separated with whitespace.
567
568 'extra-branch-refs' => {
569 'sub' => \&feature_extra_branch_refs,
570 'override' => 0,
571 'default' => []},
572
573 # Redact e-mail addresses.
574
575 # To enable system wide have in $GITWEB_CONFIG
576 # $feature{'email-privacy'}{'default'} = [1];
577 'email-privacy' => {
578 'sub' => sub { feature_bool('email-privacy', @_) },
579 'override' => 1,
580 'default' => [0]},
581 );
582
583 sub gitweb_get_feature {
584 my ($name) = @_;
585 return unless exists $feature{$name};
586 my ($sub, $override, @defaults) = (
587 $feature{$name}{'sub'},
588 $feature{$name}{'override'},
589 @{$feature{$name}{'default'}});
590 # project specific override is possible only if we have project
591 our $git_dir; # global variable, declared later
592 if (!$override || !defined $git_dir) {
593 return @defaults;
594 }
595 if (!defined $sub) {
596 warn "feature $name is not overridable";
597 return @defaults;
598 }
599 return $sub->(@defaults);
600 }
601
602 # A wrapper to check if a given feature is enabled.
603 # With this, you can say
604 #
605 # my $bool_feat = gitweb_check_feature('bool_feat');
606 # gitweb_check_feature('bool_feat') or somecode;
607 #
608 # instead of
609 #
610 # my ($bool_feat) = gitweb_get_feature('bool_feat');
611 # (gitweb_get_feature('bool_feat'))[0] or somecode;
612 #
613 sub gitweb_check_feature {
614 return (gitweb_get_feature(@_))[0];
615 }
616
617
618 sub feature_bool {
619 my $key = shift;
620 my ($val) = git_get_project_config($key, '--bool');
621
622 if (!defined $val) {
623 return ($_[0]);
624 } elsif ($val eq 'true') {
625 return (1);
626 } elsif ($val eq 'false') {
627 return (0);
628 }
629 }
630
631 sub feature_snapshot {
632 my (@fmts) = @_;
633
634 my ($val) = git_get_project_config('snapshot');
635
636 if ($val) {
637 @fmts = ($val eq 'none' ? () : split /\s*[,\s]\s*/, $val);
638 }
639
640 return @fmts;
641 }
642
643 sub feature_patches {
644 my @val = (git_get_project_config('patches', '--int'));
645
646 if (@val) {
647 return @val;
648 }
649
650 return ($_[0]);
651 }
652
653 sub feature_avatar {
654 my @val = (git_get_project_config('avatar'));
655
656 return @val ? @val : @_;
657 }
658
659 sub feature_extra_branch_refs {
660 my (@branch_refs) = @_;
661 my $values = git_get_project_config('extrabranchrefs');
662
663 if ($values) {
664 $values = config_to_multi ($values);
665 @branch_refs = ();
666 foreach my $value (@{$values}) {
667 push @branch_refs, split /\s+/, $value;
668 }
669 }
670
671 return @branch_refs;
672 }
673
674 # checking HEAD file with -e is fragile if the repository was
675 # initialized long time ago (i.e. symlink HEAD) and was pack-ref'ed
676 # and then pruned.
677 sub check_head_link {
678 my ($dir) = @_;
679 my $headfile = "$dir/HEAD";
680 return ((-e $headfile) ||
681 (-l $headfile && readlink($headfile) =~ /^refs\/heads\//));
682 }
683
684 sub check_export_ok {
685 my ($dir) = @_;
686 return (check_head_link($dir) &&
687 (!$export_ok || -e "$dir/$export_ok") &&
688 (!$export_auth_hook || $export_auth_hook->($dir)));
689 }
690
691 # process alternate names for backward compatibility
692 # filter out unsupported (unknown) snapshot formats
693 sub filter_snapshot_fmts {
694 my @fmts = @_;
695
696 @fmts = map {
697 exists $known_snapshot_format_aliases{$_} ?
698 $known_snapshot_format_aliases{$_} : $_} @fmts;
699 @fmts = grep {
700 exists $known_snapshot_formats{$_} &&
701 !$known_snapshot_formats{$_}{'disabled'}} @fmts;
702 }
703
704 sub filter_and_validate_refs {
705 my @refs = @_;
706 my %unique_refs = ();
707
708 foreach my $ref (@refs) {
709 die_error(500, "Invalid ref '$ref' in 'extra-branch-refs' feature") unless (is_valid_ref_format($ref));
710 # 'heads' are added implicitly in get_branch_refs().
711 $unique_refs{$ref} = 1 if ($ref ne 'heads');
712 }
713 return sort keys %unique_refs;
714 }
715
716 # If it is set to code reference, it is code that it is to be run once per
717 # request, allowing updating configurations that change with each request,
718 # while running other code in config file only once.
719 #
720 # Otherwise, if it is false then gitweb would process config file only once;
721 # if it is true then gitweb config would be run for each request.
722 our $per_request_config = 1;
723
724 # read and parse gitweb config file given by its parameter.
725 # returns true on success, false on recoverable error, allowing
726 # to chain this subroutine, using first file that exists.
727 # dies on errors during parsing config file, as it is unrecoverable.
728 sub read_config_file {
729 my $filename = shift;
730 return unless defined $filename;
731 if (-e $filename) {
732 do $filename;
733 # die if there is a problem accessing the file
734 die $! if $!;
735 # die if there are errors parsing config file
736 die $@ if $@;
737 return 1;
738 }
739 return;
740 }
741
742 our ($GITWEB_CONFIG, $GITWEB_CONFIG_SYSTEM, $GITWEB_CONFIG_COMMON);
743 sub evaluate_gitweb_config {
744 our $GITWEB_CONFIG = $ENV{'GITWEB_CONFIG'} || "@GITWEB_CONFIG@";
745 our $GITWEB_CONFIG_SYSTEM = $ENV{'GITWEB_CONFIG_SYSTEM'} || "@GITWEB_CONFIG_SYSTEM@";
746 our $GITWEB_CONFIG_COMMON = $ENV{'GITWEB_CONFIG_COMMON'} || "@GITWEB_CONFIG_COMMON@";
747
748 # Protect against duplications of file names, to not read config twice.
749 # Only one of $GITWEB_CONFIG and $GITWEB_CONFIG_SYSTEM is used, so
750 # there possibility of duplication of filename there doesn't matter.
751 $GITWEB_CONFIG = "" if ($GITWEB_CONFIG eq $GITWEB_CONFIG_COMMON);
752 $GITWEB_CONFIG_SYSTEM = "" if ($GITWEB_CONFIG_SYSTEM eq $GITWEB_CONFIG_COMMON);
753
754 # Common system-wide settings for convenience.
755 # Those settings can be overridden by GITWEB_CONFIG or GITWEB_CONFIG_SYSTEM.
756 read_config_file($GITWEB_CONFIG_COMMON);
757
758 # Use first config file that exists. This means use the per-instance
759 # GITWEB_CONFIG if exists, otherwise use GITWEB_SYSTEM_CONFIG.
760 read_config_file($GITWEB_CONFIG) and return;
761 read_config_file($GITWEB_CONFIG_SYSTEM);
762 }
763
764 # Get loadavg of system, to compare against $maxload.
765 # Currently it requires '/proc/loadavg' present to get loadavg;
766 # if it is not present it returns 0, which means no load checking.
767 sub get_loadavg {
768 if( -e '/proc/loadavg' ){
769 open my $fd, '<', '/proc/loadavg'
770 or return 0;
771 my @load = split(/\s+/, scalar <$fd>);
772 close $fd;
773
774 # The first three columns measure CPU and IO utilization of the last one,
775 # five, and 10 minute periods. The fourth column shows the number of
776 # currently running processes and the total number of processes in the m/n
777 # format. The last column displays the last process ID used.
778 return $load[0] || 0;
779 }
780 # additional checks for load average should go here for things that don't export
781 # /proc/loadavg
782
783 return 0;
784 }
785
786 # version of the core git binary
787 our $git_version;
788 sub evaluate_git_version {
789 our $git_version = qx("$GIT" --version) =~ m/git version (.*)$/ ? $1 : "unknown";
790 $number_of_git_cmds++;
791 }
792
793 sub check_loadavg {
794 if (defined $maxload && get_loadavg() > $maxload) {
795 die_error(503, "The load average on the server is too high");
796 }
797 }
798
799 # ======================================================================
800 # input validation and dispatch
801
802 # Various hash size-related values.
803 my $sha1_len = 40;
804 my $sha256_extra_len = 24;
805 my $sha256_len = $sha1_len + $sha256_extra_len;
806
807 # A regex matching $len hex characters. $len may be a range (e.g. 7,64).
808 sub oid_nlen_regex {
809 my $len = shift;
810 my $hchr = qr/[0-9a-fA-F]/;
811 return qr/(?:(?:$hchr){$len})/;
812 }
813
814 # A regex matching two sets of $nlen hex characters, prefixed by the literal
815 # string $prefix and with the literal string $infix between them.
816 sub oid_nlen_prefix_infix_regex {
817 my $nlen = shift;
818 my $prefix = shift;
819 my $infix = shift;
820
821 my $rx = oid_nlen_regex($nlen);
822
823 return qr/^\Q$prefix\E$rx\Q$infix\E$rx$/;
824 }
825
826 # A regex matching a valid object ID.
827 our $oid_regex;
828 {
829 my $x = oid_nlen_regex($sha1_len);
830 my $y = oid_nlen_regex($sha256_extra_len);
831 $oid_regex = qr/(?:$x(?:$y)?)/;
832 }
833
834 # input parameters can be collected from a variety of sources (presently, CGI
835 # and PATH_INFO), so we define an %input_params hash that collects them all
836 # together during validation: this allows subsequent uses (e.g. href()) to be
837 # agnostic of the parameter origin
838
839 our %input_params = ();
840
841 # input parameters are stored with the long parameter name as key. This will
842 # also be used in the href subroutine to convert parameters to their CGI
843 # equivalent, and since the href() usage is the most frequent one, we store
844 # the name -> CGI key mapping here, instead of the reverse.
845 #
846 # XXX: Warning: If you touch this, check the search form for updating,
847 # too.
848
849 our @cgi_param_mapping = (
850 project => "p",
851 action => "a",
852 file_name => "f",
853 file_parent => "fp",
854 hash => "h",
855 hash_parent => "hp",
856 hash_base => "hb",
857 hash_parent_base => "hpb",
858 page => "pg",
859 order => "o",
860 searchtext => "s",
861 searchtype => "st",
862 snapshot_format => "sf",
863 extra_options => "opt",
864 search_use_regexp => "sr",
865 ctag => "by_tag",
866 diff_style => "ds",
867 project_filter => "pf",
868 # this must be last entry (for manipulation from JavaScript)
869 javascript => "js"
870 );
871 our %cgi_param_mapping = @cgi_param_mapping;
872
873 # we will also need to know the possible actions, for validation
874 our %actions = (
875 "blame" => \&git_blame,
876 "blame_incremental" => \&git_blame_incremental,
877 "blame_data" => \&git_blame_data,
878 "blobdiff" => \&git_blobdiff,
879 "blobdiff_plain" => \&git_blobdiff_plain,
880 "blob" => \&git_blob,
881 "blob_plain" => \&git_blob_plain,
882 "commitdiff" => \&git_commitdiff,
883 "commitdiff_plain" => \&git_commitdiff_plain,
884 "commit" => \&git_commit,
885 "forks" => \&git_forks,
886 "heads" => \&git_heads,
887 "history" => \&git_history,
888 "log" => \&git_log,
889 "patch" => \&git_patch,
890 "patches" => \&git_patches,
891 "remotes" => \&git_remotes,
892 "rss" => \&git_rss,
893 "atom" => \&git_atom,
894 "search" => \&git_search,
895 "search_help" => \&git_search_help,
896 "shortlog" => \&git_shortlog,
897 "summary" => \&git_summary,
898 "tag" => \&git_tag,
899 "tags" => \&git_tags,
900 "tree" => \&git_tree,
901 "snapshot" => \&git_snapshot,
902 "object" => \&git_object,
903 # those below don't need $project
904 "opml" => \&git_opml,
905 "project_list" => \&git_project_list,
906 "project_index" => \&git_project_index,
907 );
908
909 # finally, we have the hash of allowed extra_options for the commands that
910 # allow them
911 our %allowed_options = (
912 "--no-merges" => [ qw(rss atom log shortlog history) ],
913 );
914
915 # fill %input_params with the CGI parameters. All values except for 'opt'
916 # should be single values, but opt can be an array. We should probably
917 # build an array of parameters that can be multi-valued, but since for the time
918 # being it's only this one, we just single it out
919 sub evaluate_query_params {
920 our $cgi;
921
922 while (my ($name, $symbol) = each %cgi_param_mapping) {
923 if ($symbol eq 'opt') {
924 $input_params{$name} = [ map { decode_utf8($_) } $cgi->multi_param($symbol) ];
925 } else {
926 $input_params{$name} = decode_utf8($cgi->param($symbol));
927 }
928 }
929 }
930
931 # now read PATH_INFO and update the parameter list for missing parameters
932 sub evaluate_path_info {
933 return if defined $input_params{'project'};
934 return if !$path_info;
935 $path_info =~ s,^/+,,;
936 return if !$path_info;
937
938 # find which part of PATH_INFO is project
939 my $project = $path_info;
940 $project =~ s,/+$,,;
941 while ($project && !check_head_link("$projectroot/$project")) {
942 $project =~ s,/*[^/]*$,,;
943 }
944 return unless $project;
945 $input_params{'project'} = $project;
946
947 # do not change any parameters if an action is given using the query string
948 return if $input_params{'action'};
949 $path_info =~ s,^\Q$project\E/*,,;
950
951 # next, check if we have an action
952 my $action = $path_info;
953 $action =~ s,/.*$,,;
954 if (exists $actions{$action}) {
955 $path_info =~ s,^$action/*,,;
956 $input_params{'action'} = $action;
957 }
958
959 # list of actions that want hash_base instead of hash, but can have no
960 # pathname (f) parameter
961 my @wants_base = (
962 'tree',
963 'history',
964 );
965
966 # we want to catch, among others
967 # [$hash_parent_base[:$file_parent]..]$hash_parent[:$file_name]
968 my ($parentrefname, $parentpathname, $refname, $pathname) =
969 ($path_info =~ /^(?:(.+?)(?::(.+))?\.\.)?([^:]+?)?(?::(.+))?$/);
970
971 # first, analyze the 'current' part
972 if (defined $pathname) {
973 # we got "branch:filename" or "branch:dir/"
974 # we could use git_get_type(branch:pathname), but:
975 # - it needs $git_dir
976 # - it does a git() call
977 # - the convention of terminating directories with a slash
978 # makes it superfluous
979 # - embedding the action in the PATH_INFO would make it even
980 # more superfluous
981 $pathname =~ s,^/+,,;
982 if (!$pathname || substr($pathname, -1) eq "/") {
983 $input_params{'action'} ||= "tree";
984 $pathname =~ s,/$,,;
985 } else {
986 # the default action depends on whether we had parent info
987 # or not
988 if ($parentrefname) {
989 $input_params{'action'} ||= "blobdiff_plain";
990 } else {
991 $input_params{'action'} ||= "blob_plain";
992 }
993 }
994 $input_params{'hash_base'} ||= $refname;
995 $input_params{'file_name'} ||= $pathname;
996 } elsif (defined $refname) {
997 # we got "branch". In this case we have to choose if we have to
998 # set hash or hash_base.
999 #
1000 # Most of the actions without a pathname only want hash to be
1001 # set, except for the ones specified in @wants_base that want
1002 # hash_base instead. It should also be noted that hand-crafted
1003 # links having 'history' as an action and no pathname or hash
1004 # set will fail, but that happens regardless of PATH_INFO.
1005 if (defined $parentrefname) {
1006 # if there is parent let the default be 'shortlog' action
1007 # (for http://git.example.com/repo.git/A..B links); if there
1008 # is no parent, dispatch will detect type of object and set
1009 # action appropriately if required (if action is not set)
1010 $input_params{'action'} ||= "shortlog";
1011 }
1012 if ($input_params{'action'} &&
1013 grep { $_ eq $input_params{'action'} } @wants_base) {
1014 $input_params{'hash_base'} ||= $refname;
1015 } else {
1016 $input_params{'hash'} ||= $refname;
1017 }
1018 }
1019
1020 # next, handle the 'parent' part, if present
1021 if (defined $parentrefname) {
1022 # a missing pathspec defaults to the 'current' filename, allowing e.g.
1023 # someproject/blobdiff/oldrev..newrev:/filename
1024 if ($parentpathname) {
1025 $parentpathname =~ s,^/+,,;
1026 $parentpathname =~ s,/$,,;
1027 $input_params{'file_parent'} ||= $parentpathname;
1028 } else {
1029 $input_params{'file_parent'} ||= $input_params{'file_name'};
1030 }
1031 # we assume that hash_parent_base is wanted if a path was specified,
1032 # or if the action wants hash_base instead of hash
1033 if (defined $input_params{'file_parent'} ||
1034 grep { $_ eq $input_params{'action'} } @wants_base) {
1035 $input_params{'hash_parent_base'} ||= $parentrefname;
1036 } else {
1037 $input_params{'hash_parent'} ||= $parentrefname;
1038 }
1039 }
1040
1041 # for the snapshot action, we allow URLs in the form
1042 # $project/snapshot/$hash.ext
1043 # where .ext determines the snapshot and gets removed from the
1044 # passed $refname to provide the $hash.
1045 #
1046 # To be able to tell that $refname includes the format extension, we
1047 # require the following two conditions to be satisfied:
1048 # - the hash input parameter MUST have been set from the $refname part
1049 # of the URL (i.e. they must be equal)
1050 # - the snapshot format MUST NOT have been defined already (e.g. from
1051 # CGI parameter sf)
1052 # It's also useless to try any matching unless $refname has a dot,
1053 # so we check for that too
1054 if (defined $input_params{'action'} &&
1055 $input_params{'action'} eq 'snapshot' &&
1056 defined $refname && index($refname, '.') != -1 &&
1057 $refname eq $input_params{'hash'} &&
1058 !defined $input_params{'snapshot_format'}) {
1059 # We loop over the known snapshot formats, checking for
1060 # extensions. Allowed extensions are both the defined suffix
1061 # (which includes the initial dot already) and the snapshot
1062 # format key itself, with a prepended dot
1063 while (my ($fmt, $opt) = each %known_snapshot_formats) {
1064 my $hash = $refname;
1065 unless ($hash =~ s/(\Q$opt->{'suffix'}\E|\Q.$fmt\E)$//) {
1066 next;
1067 }
1068 my $sfx = $1;
1069 # a valid suffix was found, so set the snapshot format
1070 # and reset the hash parameter
1071 $input_params{'snapshot_format'} = $fmt;
1072 $input_params{'hash'} = $hash;
1073 # we also set the format suffix to the one requested
1074 # in the URL: this way a request for e.g. .tgz returns
1075 # a .tgz instead of a .tar.gz
1076 $known_snapshot_formats{$fmt}{'suffix'} = $sfx;
1077 last;
1078 }
1079 }
1080 }
1081
1082 our ($action, $project, $file_name, $file_parent, $hash, $hash_parent, $hash_base,
1083 $hash_parent_base, @extra_options, $page, $searchtype, $search_use_regexp,
1084 $searchtext, $search_regexp, $project_filter);
1085 sub evaluate_and_validate_params {
1086 our $action = $input_params{'action'};
1087 if (defined $action) {
1088 if (!is_valid_action($action)) {
1089 die_error(400, "Invalid action parameter");
1090 }
1091 }
1092
1093 # parameters which are pathnames
1094 our $project = $input_params{'project'};
1095 if (defined $project) {
1096 if (!is_valid_project($project)) {
1097 undef $project;
1098 die_error(404, "No such project");
1099 }
1100 }
1101
1102 our $project_filter = $input_params{'project_filter'};
1103 if (defined $project_filter) {
1104 if (!is_valid_pathname($project_filter)) {
1105 die_error(404, "Invalid project_filter parameter");
1106 }
1107 }
1108
1109 our $file_name = $input_params{'file_name'};
1110 if (defined $file_name) {
1111 if (!is_valid_pathname($file_name)) {
1112 die_error(400, "Invalid file parameter");
1113 }
1114 }
1115
1116 our $file_parent = $input_params{'file_parent'};
1117 if (defined $file_parent) {
1118 if (!is_valid_pathname($file_parent)) {
1119 die_error(400, "Invalid file parent parameter");
1120 }
1121 }
1122
1123 # parameters which are refnames
1124 our $hash = $input_params{'hash'};
1125 if (defined $hash) {
1126 if (!is_valid_refname($hash)) {
1127 die_error(400, "Invalid hash parameter");
1128 }
1129 }
1130
1131 our $hash_parent = $input_params{'hash_parent'};
1132 if (defined $hash_parent) {
1133 if (!is_valid_refname($hash_parent)) {
1134 die_error(400, "Invalid hash parent parameter");
1135 }
1136 }
1137
1138 our $hash_base = $input_params{'hash_base'};
1139 if (defined $hash_base) {
1140 if (!is_valid_refname($hash_base)) {
1141 die_error(400, "Invalid hash base parameter");
1142 }
1143 }
1144
1145 our @extra_options = @{$input_params{'extra_options'}};
1146 # @extra_options is always defined, since it can only be (currently) set from
1147 # CGI, and $cgi->param() returns the empty array in array context if the param
1148 # is not set
1149 foreach my $opt (@extra_options) {
1150 if (not exists $allowed_options{$opt}) {
1151 die_error(400, "Invalid option parameter");
1152 }
1153 if (not grep(/^$action$/, @{$allowed_options{$opt}})) {
1154 die_error(400, "Invalid option parameter for this action");
1155 }
1156 }
1157
1158 our $hash_parent_base = $input_params{'hash_parent_base'};
1159 if (defined $hash_parent_base) {
1160 if (!is_valid_refname($hash_parent_base)) {
1161 die_error(400, "Invalid hash parent base parameter");
1162 }
1163 }
1164
1165 # other parameters
1166 our $page = $input_params{'page'};
1167 if (defined $page) {
1168 if ($page =~ m/[^0-9]/) {
1169 die_error(400, "Invalid page parameter");
1170 }
1171 }
1172
1173 our $searchtype = $input_params{'searchtype'};
1174 if (defined $searchtype) {
1175 if ($searchtype =~ m/[^a-z]/) {
1176 die_error(400, "Invalid searchtype parameter");
1177 }
1178 }
1179
1180 our $search_use_regexp = $input_params{'search_use_regexp'};
1181
1182 our $searchtext = $input_params{'searchtext'};
1183 our $search_regexp = undef;
1184 if (defined $searchtext) {
1185 if (length($searchtext) < 2) {
1186 die_error(403, "At least two characters are required for search parameter");
1187 }
1188 if ($search_use_regexp) {
1189 $search_regexp = $searchtext;
1190 if (!eval { qr/$search_regexp/; 1; }) {
1191 my $error = $@ =~ s/ at \S+ line \d+.*\n?//r;
1192 die_error(400, "Invalid search regexp '$search_regexp'",
1193 esc_html($error));
1194 }
1195 } else {
1196 $search_regexp = quotemeta $searchtext;
1197 }
1198 }
1199 }
1200
1201 # path to the current git repository
1202 our $git_dir;
1203 sub evaluate_git_dir {
1204 our $git_dir = "$projectroot/$project" if $project;
1205 }
1206
1207 our (@snapshot_fmts, $git_avatar, @extra_branch_refs);
1208 sub configure_gitweb_features {
1209 # list of supported snapshot formats
1210 our @snapshot_fmts = gitweb_get_feature('snapshot');
1211 @snapshot_fmts = filter_snapshot_fmts(@snapshot_fmts);
1212
1213 our ($git_avatar) = gitweb_get_feature('avatar');
1214 $git_avatar = '' unless $git_avatar =~ /^(?:gravatar|picon)$/s;
1215
1216 our @extra_branch_refs = gitweb_get_feature('extra-branch-refs');
1217 @extra_branch_refs = filter_and_validate_refs (@extra_branch_refs);
1218 }
1219
1220 sub get_branch_refs {
1221 return ('heads', @extra_branch_refs);
1222 }
1223
1224 # custom error handler: 'die <message>' is Internal Server Error
1225 sub handle_errors_html {
1226 my $msg = shift; # it is already HTML escaped
1227
1228 # to avoid infinite loop where error occurs in die_error,
1229 # change handler to default handler, disabling handle_errors_html
1230 set_message("Error occurred when inside die_error:\n$msg");
1231
1232 # you cannot jump out of die_error when called as error handler;
1233 # the subroutine set via CGI::Carp::set_message is called _after_
1234 # HTTP headers are already written, so it cannot write them itself
1235 die_error(undef, undef, $msg, -error_handler => 1, -no_http_header => 1);
1236 }
1237 set_message(\&handle_errors_html);
1238
1239 # dispatch
1240 sub dispatch {
1241 if (!defined $action) {
1242 if (defined $hash) {
1243 $action = git_get_type($hash);
1244 $action or die_error(404, "Object does not exist");
1245 } elsif (defined $hash_base && defined $file_name) {
1246 $action = git_get_type("$hash_base:$file_name");
1247 $action or die_error(404, "File or directory does not exist");
1248 } elsif (defined $project) {
1249 $action = 'summary';
1250 } else {
1251 $action = 'project_list';
1252 }
1253 }
1254 if (!defined($actions{$action})) {
1255 die_error(400, "Unknown action");
1256 }
1257 if ($action !~ m/^(?:opml|project_list|project_index)$/ &&
1258 !$project) {
1259 die_error(400, "Project needed");
1260 }
1261 $actions{$action}->();
1262 }
1263
1264 sub reset_timer {
1265 our $t0 = [ gettimeofday() ]
1266 if defined $t0;
1267 our $number_of_git_cmds = 0;
1268 }
1269
1270 our $first_request = 1;
1271 sub run_request {
1272 reset_timer();
1273
1274 evaluate_uri();
1275 if ($first_request) {
1276 evaluate_gitweb_config();
1277 evaluate_git_version();
1278 }
1279 if ($per_request_config) {
1280 if (ref($per_request_config) eq 'CODE') {
1281 $per_request_config->();
1282 } elsif (!$first_request) {
1283 evaluate_gitweb_config();
1284 }
1285 }
1286 check_loadavg();
1287
1288 # $projectroot and $projects_list might be set in gitweb config file
1289 $projects_list ||= $projectroot;
1290
1291 evaluate_query_params();
1292 evaluate_path_info();
1293 evaluate_and_validate_params();
1294 evaluate_git_dir();
1295
1296 configure_gitweb_features();
1297
1298 dispatch();
1299 }
1300
1301 our $is_last_request = sub { 1 };
1302 our ($pre_dispatch_hook, $post_dispatch_hook, $pre_listen_hook);
1303 our $CGI = 'CGI';
1304 our $cgi;
1305 our $FCGI_Stream_PRINT_raw = \&FCGI::Stream::PRINT;
1306 sub configure_as_fcgi {
1307 require CGI::Fast;
1308 our $CGI = 'CGI::Fast';
1309 # FCGI is not Unicode aware hence the UTF-8 encoding must be done manually.
1310 # However no encoding must be done within git_blob_plain() and git_snapshot()
1311 # which must still output in raw binary mode.
1312 no warnings 'redefine';
1313 my $enc = Encode::find_encoding('UTF-8');
1314 *FCGI::Stream::PRINT = sub {
1315 my @OUTPUT = @_;
1316 for (my $i = 1; $i < @_; $i++) {
1317 $OUTPUT[$i] = $enc->encode($_[$i], Encode::FB_CROAK|Encode::LEAVE_SRC);
1318 }
1319 @_ = @OUTPUT;
1320 goto $FCGI_Stream_PRINT_raw;
1321 };
1322
1323 my $request_number = 0;
1324 # let each child service 100 requests
1325 our $is_last_request = sub { ++$request_number > 100 };
1326 }
1327 sub evaluate_argv {
1328 my $script_name = $ENV{'SCRIPT_NAME'} || $ENV{'SCRIPT_FILENAME'} || __FILE__;
1329 configure_as_fcgi()
1330 if $script_name =~ /\.fcgi$/;
1331
1332 return unless (@ARGV);
1333
1334 require Getopt::Long;
1335 Getopt::Long::GetOptions(
1336 'fastcgi|fcgi|f' => \&configure_as_fcgi,
1337 'nproc|n=i' => sub {
1338 my ($arg, $val) = @_;
1339 return unless eval { require FCGI::ProcManager; 1; };
1340 my $proc_manager = FCGI::ProcManager->new({
1341 n_processes => $val,
1342 });
1343 our $pre_listen_hook = sub { $proc_manager->pm_manage() };
1344 our $pre_dispatch_hook = sub { $proc_manager->pm_pre_dispatch() };
1345 our $post_dispatch_hook = sub { $proc_manager->pm_post_dispatch() };
1346 },
1347 );
1348 }
1349
1350 sub run {
1351 evaluate_argv();
1352
1353 $first_request = 1;
1354 $pre_listen_hook->()
1355 if $pre_listen_hook;
1356
1357 REQUEST:
1358 while ($cgi = $CGI->new()) {
1359 $pre_dispatch_hook->()
1360 if $pre_dispatch_hook;
1361
1362 run_request();
1363
1364 $post_dispatch_hook->()
1365 if $post_dispatch_hook;
1366 $first_request = 0;
1367
1368 last REQUEST if ($is_last_request->());
1369 }
1370
1371 DONE_GITWEB:
1372 1;
1373 }
1374
1375 run();
1376
1377 if (defined caller) {
1378 # wrapped in a subroutine processing requests,
1379 # e.g. mod_perl with ModPerl::Registry, or PSGI with Plack::App::WrapCGI
1380 return;
1381 } else {
1382 # pure CGI script, serving single request
1383 exit;
1384 }
1385
1386 ## ======================================================================
1387 ## action links
1388
1389 # possible values of extra options
1390 # -full => 0|1 - use absolute/full URL ($my_uri/$my_url as base)
1391 # -replay => 1 - start from a current view (replay with modifications)
1392 # -path_info => 0|1 - don't use/use path_info URL (if possible)
1393 # -anchor => ANCHOR - add #ANCHOR to end of URL, implies -replay if used alone
1394 sub href {
1395 my %params = @_;
1396 # default is to use -absolute url() i.e. $my_uri
1397 my $href = $params{-full} ? $my_url : $my_uri;
1398
1399 # implicit -replay, must be first of implicit params
1400 $params{-replay} = 1 if (keys %params == 1 && $params{-anchor});
1401
1402 $params{'project'} = $project unless exists $params{'project'};
1403
1404 if ($params{-replay}) {
1405 while (my ($name, $symbol) = each %cgi_param_mapping) {
1406 if (!exists $params{$name}) {
1407 $params{$name} = $input_params{$name};
1408 }
1409 }
1410 }
1411
1412 my $use_pathinfo = gitweb_check_feature('pathinfo');
1413 if (defined $params{'project'} &&
1414 (exists $params{-path_info} ? $params{-path_info} : $use_pathinfo)) {
1415 # try to put as many parameters as possible in PATH_INFO:
1416 # - project name
1417 # - action
1418 # - hash_parent or hash_parent_base:/file_parent
1419 # - hash or hash_base:/filename
1420 # - the snapshot_format as an appropriate suffix
1421
1422 # When the script is the root DirectoryIndex for the domain,
1423 # $href here would be something like http://gitweb.example.com/
1424 # Thus, we strip any trailing / from $href, to spare us double
1425 # slashes in the final URL
1426 $href =~ s,/$,,;
1427
1428 # Then add the project name, if present
1429 $href .= "/".esc_path_info($params{'project'});
1430 delete $params{'project'};
1431
1432 # since we destructively absorb parameters, we keep this
1433 # boolean that remembers if we're handling a snapshot
1434 my $is_snapshot = $params{'action'} eq 'snapshot';
1435
1436 # Summary just uses the project path URL, any other action is
1437 # added to the URL
1438 if (defined $params{'action'}) {
1439 $href .= "/".esc_path_info($params{'action'})
1440 unless $params{'action'} eq 'summary';
1441 delete $params{'action'};
1442 }
1443
1444 # Next, we put hash_parent_base:/file_parent..hash_base:/file_name,
1445 # stripping nonexistent or useless pieces
1446 $href .= "/" if ($params{'hash_base'} || $params{'hash_parent_base'}
1447 || $params{'hash_parent'} || $params{'hash'});
1448 if (defined $params{'hash_base'}) {
1449 if (defined $params{'hash_parent_base'}) {
1450 $href .= esc_path_info($params{'hash_parent_base'});
1451 # skip the file_parent if it's the same as the file_name
1452 if (defined $params{'file_parent'}) {
1453 if (defined $params{'file_name'} && $params{'file_parent'} eq $params{'file_name'}) {
1454 delete $params{'file_parent'};
1455 } elsif ($params{'file_parent'} !~ /\.\./) {
1456 $href .= ":/".esc_path_info($params{'file_parent'});
1457 delete $params{'file_parent'};
1458 }
1459 }
1460 $href .= "..";
1461 delete $params{'hash_parent'};
1462 delete $params{'hash_parent_base'};
1463 } elsif (defined $params{'hash_parent'}) {
1464 $href .= esc_path_info($params{'hash_parent'}). "..";
1465 delete $params{'hash_parent'};
1466 }
1467
1468 $href .= esc_path_info($params{'hash_base'});
1469 if (defined $params{'file_name'} && $params{'file_name'} !~ /\.\./) {
1470 $href .= ":/".esc_path_info($params{'file_name'});
1471 delete $params{'file_name'};
1472 }
1473 delete $params{'hash'};
1474 delete $params{'hash_base'};
1475 } elsif (defined $params{'hash'}) {
1476 $href .= esc_path_info($params{'hash'});
1477 delete $params{'hash'};
1478 }
1479
1480 # If the action was a snapshot, we can absorb the
1481 # snapshot_format parameter too
1482 if ($is_snapshot) {
1483 my $fmt = $params{'snapshot_format'};
1484 # snapshot_format should always be defined when href()
1485 # is called, but just in case some code forgets, we
1486 # fall back to the default
1487 $fmt ||= $snapshot_fmts[0];
1488 $href .= $known_snapshot_formats{$fmt}{'suffix'};
1489 delete $params{'snapshot_format'};
1490 }
1491 }
1492
1493 # now encode the parameters explicitly
1494 my @result = ();
1495 for (my $i = 0; $i < @cgi_param_mapping; $i += 2) {
1496 my ($name, $symbol) = ($cgi_param_mapping[$i], $cgi_param_mapping[$i+1]);
1497 if (defined $params{$name}) {
1498 if (ref($params{$name}) eq "ARRAY") {
1499 foreach my $par (@{$params{$name}}) {
1500 push @result, $symbol . "=" . esc_param($par);
1501 }
1502 } else {
1503 push @result, $symbol . "=" . esc_param($params{$name});
1504 }
1505 }
1506 }
1507 $href .= "?" . join(';', @result) if scalar @result;
1508
1509 # final transformation: trailing spaces must be escaped (URI-encoded)
1510 $href =~ s/(\s+)$/CGI::escape($1)/e;
1511
1512 if ($params{-anchor}) {
1513 $href .= "#".esc_param($params{-anchor});
1514 }
1515
1516 return $href;
1517 }
1518
1519
1520 ## ======================================================================
1521 ## validation, quoting/unquoting and escaping
1522
1523 sub is_valid_action {
1524 my $input = shift;
1525 return undef unless exists $actions{$input};
1526 return 1;
1527 }
1528
1529 sub is_valid_project {
1530 my $input = shift;
1531
1532 return unless defined $input;
1533 if (!is_valid_pathname($input) ||
1534 !(-d "$projectroot/$input") ||
1535 !check_export_ok("$projectroot/$input") ||
1536 ($strict_export && !project_in_list($input))) {
1537 return undef;
1538 } else {
1539 return 1;
1540 }
1541 }
1542
1543 sub is_valid_pathname {
1544 my $input = shift;
1545
1546 return undef unless defined $input;
1547 # no '.' or '..' as elements of path, i.e. no '.' or '..'
1548 # at the beginning, at the end, and between slashes.
1549 # also this catches doubled slashes
1550 if ($input =~ m!(^|/)(|\.|\.\.)(/|$)!) {
1551 return undef;
1552 }
1553 # no null characters
1554 if ($input =~ m!\0!) {
1555 return undef;
1556 }
1557 return 1;
1558 }
1559
1560 sub is_valid_ref_format {
1561 my $input = shift;
1562
1563 return undef unless defined $input;
1564 # restrictions on ref name according to git-check-ref-format
1565 if ($input =~ m!(/\.|\.\.|[\000-\040\177 ~^:?*\[]|/$)!) {
1566 return undef;
1567 }
1568 return 1;
1569 }
1570
1571 sub is_valid_refname {
1572 my $input = shift;
1573
1574 return undef unless defined $input;
1575 # textual hashes are O.K.
1576 if ($input =~ m/^$oid_regex$/) {
1577 return 1;
1578 }
1579 # it must be correct pathname
1580 is_valid_pathname($input) or return undef;
1581 # check git-check-ref-format restrictions
1582 is_valid_ref_format($input) or return undef;
1583 return 1;
1584 }
1585
1586 # decode sequences of octets in utf8 into Perl's internal form,
1587 # which is utf-8 with utf8 flag set if needed. gitweb writes out
1588 # in utf-8 thanks to "binmode STDOUT, ':utf8'" at beginning
1589 sub to_utf8 {
1590 my $str = shift;
1591 return undef unless defined $str;
1592
1593 if (utf8::is_utf8($str) || utf8::decode($str)) {
1594 return $str;
1595 } else {
1596 return decode($fallback_encoding, $str, Encode::FB_DEFAULT);
1597 }
1598 }
1599
1600 # quote unsafe chars, but keep the slash, even when it's not
1601 # correct, but quoted slashes look too horrible in bookmarks
1602 sub esc_param {
1603 my $str = shift;
1604 return undef unless defined $str;
1605 $str =~ s/([^A-Za-z0-9\-_.~()\/:@ ]+)/CGI::escape($1)/eg;
1606 $str =~ s/ /\+/g;
1607 return $str;
1608 }
1609
1610 # the quoting rules for path_info fragment are slightly different
1611 sub esc_path_info {
1612 my $str = shift;
1613 return undef unless defined $str;
1614
1615 # path_info doesn't treat '+' as space (specially), but '?' must be escaped
1616 $str =~ s/([^A-Za-z0-9\-_.~();\/;:@&= +]+)/CGI::escape($1)/eg;
1617
1618 return $str;
1619 }
1620
1621 # quote unsafe chars in whole URL, so some characters cannot be quoted
1622 sub esc_url {
1623 my $str = shift;
1624 return undef unless defined $str;
1625 $str =~ s/([^A-Za-z0-9\-_.~();\/;?:@&= ]+)/CGI::escape($1)/eg;
1626 $str =~ s/ /\+/g;
1627 return $str;
1628 }
1629
1630 # quote unsafe characters in HTML attributes
1631 sub esc_attr {
1632
1633 # for XHTML conformance escaping '"' to '&quot;' is not enough
1634 return esc_html(@_);
1635 }
1636
1637 # replace invalid utf8 character with SUBSTITUTION sequence
1638 sub esc_html {
1639 my $str = shift;
1640 my %opts = @_;
1641
1642 return undef unless defined $str;
1643
1644 $str = to_utf8($str);
1645 $str = $cgi->escapeHTML($str);
1646 if ($opts{'-nbsp'}) {
1647 $str =~ s/ /&nbsp;/g;
1648 }
1649 $str =~ s|([[:cntrl:]])|(($1 ne "\t") ? quot_cec($1) : $1)|eg;
1650 return $str;
1651 }
1652
1653 # quote control characters and escape filename to HTML
1654 sub esc_path {
1655 my $str = shift;
1656 my %opts = @_;
1657
1658 return undef unless defined $str;
1659
1660 $str = to_utf8($str);
1661 $str = $cgi->escapeHTML($str);
1662 if ($opts{'-nbsp'}) {
1663 $str =~ s/ /&nbsp;/g;
1664 }
1665 $str =~ s|([[:cntrl:]])|quot_cec($1)|eg;
1666 return $str;
1667 }
1668
1669 # Sanitize for use in XHTML + application/xml+xhtml (valid XML 1.0)
1670 sub sanitize {
1671 my $str = shift;
1672
1673 return undef unless defined $str;
1674
1675 $str = to_utf8($str);
1676 $str =~ s|([[:cntrl:]])|(index("\t\n\r", $1) != -1 ? $1 : quot_cec($1))|eg;
1677 return $str;
1678 }
1679
1680 # Make control characters "printable", using character escape codes (CEC)
1681 sub quot_cec {
1682 my $cntrl = shift;
1683 my %opts = @_;
1684 my %es = ( # character escape codes, aka escape sequences
1685 "\t" => '\t', # tab (HT)
1686 "\n" => '\n', # line feed (LF)
1687 "\r" => '\r', # carriage return (CR)
1688 "\f" => '\f', # form feed (FF)
1689 "\b" => '\b', # backspace (BS)
1690 "\a" => '\a', # alarm (bell) (BEL)
1691 "\e" => '\e', # escape (ESC)
1692 "\013" => '\v', # vertical tab (VT)
1693 "\000" => '\0', # nul character (NUL)
1694 );
1695 my $chr = ( (exists $es{$cntrl})
1696 ? $es{$cntrl}
1697 : sprintf('\%2x', ord($cntrl)) );
1698 if ($opts{-nohtml}) {
1699 return $chr;
1700 } else {
1701 return "<span class=\"cntrl\">$chr</span>";
1702 }
1703 }
1704
1705 # Alternatively use unicode control pictures codepoints,
1706 # Unicode "printable representation" (PR)
1707 sub quot_upr {
1708 my $cntrl = shift;
1709 my %opts = @_;
1710
1711 my $chr = sprintf('&#%04d;', 0x2400+ord($cntrl));
1712 if ($opts{-nohtml}) {
1713 return $chr;
1714 } else {
1715 return "<span class=\"cntrl\">$chr</span>";
1716 }
1717 }
1718
1719 # git may return quoted and escaped filenames
1720 sub unquote {
1721 my $str = shift;
1722
1723 sub unq {
1724 my $seq = shift;
1725 my %es = ( # character escape codes, aka escape sequences
1726 't' => "\t", # tab (HT, TAB)
1727 'n' => "\n", # newline (NL)
1728 'r' => "\r", # return (CR)
1729 'f' => "\f", # form feed (FF)
1730 'b' => "\b", # backspace (BS)
1731 'a' => "\a", # alarm (bell) (BEL)
1732 'e' => "\e", # escape (ESC)
1733 'v' => "\013", # vertical tab (VT)
1734 );
1735
1736 if ($seq =~ m/^[0-7]{1,3}$/) {
1737 # octal char sequence
1738 return chr(oct($seq));
1739 } elsif (exists $es{$seq}) {
1740 # C escape sequence, aka character escape code
1741 return $es{$seq};
1742 }
1743 # quoted ordinary character
1744 return $seq;
1745 }
1746
1747 if ($str =~ m/^"(.*)"$/) {
1748 # needs unquoting
1749 $str = $1;
1750 $str =~ s/\\([^0-7]|[0-7]{1,3})/unq($1)/eg;
1751 }
1752 return $str;
1753 }
1754
1755 # escape tabs (convert tabs to spaces)
1756 sub untabify {
1757 my $line = shift;
1758
1759 while ((my $pos = index($line, "\t")) != -1) {
1760 if (my $count = (8 - ($pos % 8))) {
1761 my $spaces = ' ' x $count;
1762 $line =~ s/\t/$spaces/;
1763 }
1764 }
1765
1766 return $line;
1767 }
1768
1769 sub project_in_list {
1770 my $project = shift;
1771 my @list = git_get_projects_list();
1772 return @list && scalar(grep { $_->{'path'} eq $project } @list);
1773 }
1774
1775 ## ----------------------------------------------------------------------
1776 ## HTML aware string manipulation
1777
1778 # Try to chop given string on a word boundary between position
1779 # $len and $len+$add_len. If there is no word boundary there,
1780 # chop at $len+$add_len. Do not chop if chopped part plus ellipsis
1781 # (marking chopped part) would be longer than given string.
1782 sub chop_str {
1783 my $str = shift;
1784 my $len = shift;
1785 my $add_len = shift || 10;
1786 my $where = shift || 'right'; # 'left' | 'center' | 'right'
1787
1788 # Make sure perl knows it is utf8 encoded so we don't
1789 # cut in the middle of a utf8 multibyte char.
1790 $str = to_utf8($str);
1791
1792 # allow only $len chars, but don't cut a word if it would fit in $add_len
1793 # if it doesn't fit, cut it if it's still longer than the dots we would add
1794 # remove chopped character entities entirely
1795
1796 # when chopping in the middle, distribute $len into left and right part
1797 # return early if chopping wouldn't make string shorter
1798 if ($where eq 'center') {
1799 return $str if ($len + 5 >= length($str)); # filler is length 5
1800 $len = int($len/2);
1801 } else {
1802 return $str if ($len + 4 >= length($str)); # filler is length 4
1803 }
1804
1805 # regexps: ending and beginning with word part up to $add_len
1806 my $endre = qr/.{$len}\w{0,$add_len}/;
1807 my $begre = qr/\w{0,$add_len}.{$len}/;
1808
1809 if ($where eq 'left') {
1810 $str =~ m/^(.*?)($begre)$/;
1811 my ($lead, $body) = ($1, $2);
1812 if (length($lead) > 4) {
1813 $lead = " ...";
1814 }
1815 return "$lead$body";
1816
1817 } elsif ($where eq 'center') {
1818 $str =~ m/^($endre)(.*)$/;
1819 my ($left, $str) = ($1, $2);
1820 $str =~ m/^(.*?)($begre)$/;
1821 my ($mid, $right) = ($1, $2);
1822 if (length($mid) > 5) {
1823 $mid = " ... ";
1824 }
1825 return "$left$mid$right";
1826
1827 } else {
1828 $str =~ m/^($endre)(.*)$/;
1829 my $body = $1;
1830 my $tail = $2;
1831 if (length($tail) > 4) {
1832 $tail = "... ";
1833 }
1834 return "$body$tail";
1835 }
1836 }
1837
1838 # takes the same arguments as chop_str, but also wraps a <span> around the
1839 # result with a title attribute if it does get chopped. Additionally, the
1840 # string is HTML-escaped.
1841 sub chop_and_escape_str {
1842 my ($str) = @_;
1843
1844 my $chopped = chop_str(@_);
1845 $str = to_utf8($str);
1846 if ($chopped eq $str) {
1847 return esc_html($chopped);
1848 } else {
1849 $str =~ s/[[:cntrl:]]/?/g;
1850 return $cgi->span({-title=>$str}, esc_html($chopped));
1851 }
1852 }
1853
1854 # Highlight selected fragments of string, using given CSS class,
1855 # and escape HTML. It is assumed that fragments do not overlap.
1856 # Regions are passed as list of pairs (array references).
1857 #
1858 # Example: esc_html_hl_regions("foobar", "mark", [ 0, 3 ]) returns
1859 # '<span class="mark">foo</span>bar'
1860 sub esc_html_hl_regions {
1861 my ($str, $css_class, @sel) = @_;
1862 my %opts = grep { ref($_) ne 'ARRAY' } @sel;
1863 @sel = grep { ref($_) eq 'ARRAY' } @sel;
1864 return esc_html($str, %opts) unless @sel;
1865
1866 my $out = '';
1867 my $pos = 0;
1868
1869 for my $s (@sel) {
1870 my ($begin, $end) = @$s;
1871
1872 # Don't create empty <span> elements.
1873 next if $end <= $begin;
1874
1875 my $escaped = esc_html(substr($str, $begin, $end - $begin),
1876 %opts);
1877
1878 $out .= esc_html(substr($str, $pos, $begin - $pos), %opts)
1879 if ($begin - $pos > 0);
1880 $out .= $cgi->span({-class => $css_class}, $escaped);
1881
1882 $pos = $end;
1883 }
1884 $out .= esc_html(substr($str, $pos), %opts)
1885 if ($pos < length($str));
1886
1887 return $out;
1888 }
1889
1890 # return positions of beginning and end of each match
1891 sub matchpos_list {
1892 my ($str, $regexp) = @_;
1893 return unless (defined $str && defined $regexp);
1894
1895 my @matches;
1896 while ($str =~ /$regexp/g) {
1897 push @matches, [$-[0], $+[0]];
1898 }
1899 return @matches;
1900 }
1901
1902 # highlight match (if any), and escape HTML
1903 sub esc_html_match_hl {
1904 my ($str, $regexp) = @_;
1905 return esc_html($str) unless defined $regexp;
1906
1907 my @matches = matchpos_list($str, $regexp);
1908 return esc_html($str) unless @matches;
1909
1910 return esc_html_hl_regions($str, 'match', @matches);
1911 }
1912
1913
1914 # highlight match (if any) of shortened string, and escape HTML
1915 sub esc_html_match_hl_chopped {
1916 my ($str, $chopped, $regexp) = @_;
1917 return esc_html_match_hl($str, $regexp) unless defined $chopped;
1918
1919 my @matches = matchpos_list($str, $regexp);
1920 return esc_html($chopped) unless @matches;
1921
1922 # filter matches so that we mark chopped string
1923 my $tail = "... "; # see chop_str
1924 unless ($chopped =~ s/\Q$tail\E$//) {
1925 $tail = '';
1926 }
1927 my $chop_len = length($chopped);
1928 my $tail_len = length($tail);
1929 my @filtered;
1930
1931 for my $m (@matches) {
1932 if ($m->[0] > $chop_len) {
1933 push @filtered, [ $chop_len, $chop_len + $tail_len ] if ($tail_len > 0);
1934 last;
1935 } elsif ($m->[1] > $chop_len) {
1936 push @filtered, [ $m->[0], $chop_len + $tail_len ];
1937 last;
1938 }
1939 push @filtered, $m;
1940 }
1941
1942 return esc_html_hl_regions($chopped . $tail, 'match', @filtered);
1943 }
1944
1945 ## ----------------------------------------------------------------------
1946 ## functions returning short strings
1947
1948 # CSS class for given age value (in seconds)
1949 sub age_class {
1950 my $age = shift;
1951
1952 if (!defined $age) {
1953 return "noage";
1954 } elsif ($age < 60*60*2) {
1955 return "age0";
1956 } elsif ($age < 60*60*24*2) {
1957 return "age1";
1958 } else {
1959 return "age2";
1960 }
1961 }
1962
1963 # convert age in seconds to "nn units ago" string
1964 sub age_string {
1965 my $age = shift;
1966 my $age_str;
1967
1968 if ($age > 60*60*24*365*2) {
1969 $age_str = (int $age/60/60/24/365);
1970 $age_str .= " years ago";
1971 } elsif ($age > 60*60*24*(365/12)*2) {
1972 $age_str = int $age/60/60/24/(365/12);
1973 $age_str .= " months ago";
1974 } elsif ($age > 60*60*24*7*2) {
1975 $age_str = int $age/60/60/24/7;
1976 $age_str .= " weeks ago";
1977 } elsif ($age > 60*60*24*2) {
1978 $age_str = int $age/60/60/24;
1979 $age_str .= " days ago";
1980 } elsif ($age > 60*60*2) {
1981 $age_str = int $age/60/60;
1982 $age_str .= " hours ago";
1983 } elsif ($age > 60*2) {
1984 $age_str = int $age/60;
1985 $age_str .= " min ago";
1986 } elsif ($age > 2) {
1987 $age_str = int $age;
1988 $age_str .= " sec ago";
1989 } else {
1990 $age_str .= " right now";
1991 }
1992 return $age_str;
1993 }
1994
1995 use constant {
1996 S_IFINVALID => 0030000,
1997 S_IFGITLINK => 0160000,
1998 };
1999
2000 # submodule/subproject, a commit object reference
2001 sub S_ISGITLINK {
2002 my $mode = shift;
2003
2004 return (($mode & S_IFMT) == S_IFGITLINK)
2005 }
2006
2007 # convert file mode in octal to symbolic file mode string
2008 sub mode_str {
2009 my $mode = oct shift;
2010
2011 if (S_ISGITLINK($mode)) {
2012 return 'm---------';
2013 } elsif (S_ISDIR($mode & S_IFMT)) {
2014 return 'drwxr-xr-x';
2015 } elsif (S_ISLNK($mode)) {
2016 return 'lrwxrwxrwx';
2017 } elsif (S_ISREG($mode)) {
2018 # git cares only about the executable bit
2019 if ($mode & S_IXUSR) {
2020 return '-rwxr-xr-x';
2021 } else {
2022 return '-rw-r--r--';
2023 };
2024 } else {
2025 return '----------';
2026 }
2027 }
2028
2029 # convert file mode in octal to file type string
2030 sub file_type {
2031 my $mode = shift;
2032
2033 if ($mode !~ m/^[0-7]+$/) {
2034 return $mode;
2035 } else {
2036 $mode = oct $mode;
2037 }
2038
2039 if (S_ISGITLINK($mode)) {
2040 return "submodule";
2041 } elsif (S_ISDIR($mode & S_IFMT)) {
2042 return "directory";
2043 } elsif (S_ISLNK($mode)) {
2044 return "symlink";
2045 } elsif (S_ISREG($mode)) {
2046 return "file";
2047 } else {
2048 return "unknown";
2049 }
2050 }
2051
2052 # convert file mode in octal to file type description string
2053 sub file_type_long {
2054 my $mode = shift;
2055
2056 if ($mode !~ m/^[0-7]+$/) {
2057 return $mode;
2058 } else {
2059 $mode = oct $mode;
2060 }
2061
2062 if (S_ISGITLINK($mode)) {
2063 return "submodule";
2064 } elsif (S_ISDIR($mode & S_IFMT)) {
2065 return "directory";
2066 } elsif (S_ISLNK($mode)) {
2067 return "symlink";
2068 } elsif (S_ISREG($mode)) {
2069 if ($mode & S_IXUSR) {
2070 return "executable";
2071 } else {
2072 return "file";
2073 };
2074 } else {
2075 return "unknown";
2076 }
2077 }
2078
2079
2080 ## ----------------------------------------------------------------------
2081 ## functions returning short HTML fragments, or transforming HTML fragments
2082 ## which don't belong to other sections
2083
2084 # format line of commit message.
2085 sub format_log_line_html {
2086 my $line = shift;
2087
2088 # Potentially abbreviated OID.
2089 my $regex = oid_nlen_regex("7,64");
2090
2091 $line = esc_html($line, -nbsp=>1);
2092 $line =~ s{
2093 \b
2094 (
2095 # The output of "git describe", e.g. v2.10.0-297-gf6727b0
2096 # or hadoop-20160921-113441-20-g094fb7d
2097 (?<!-) # see check_tag_ref(). Tags can't start with -
2098 [A-Za-z0-9.-]+
2099 (?!\.) # refs can't end with ".", see check_refname_format()
2100 -g$regex
2101 |
2102 # Just a normal looking Git SHA1
2103 $regex
2104 )
2105 \b
2106 }{
2107 $cgi->a({-href => href(action=>"object", hash=>$1),
2108 -class => "text"}, $1);
2109 }egx;
2110
2111 return $line;
2112 }
2113
2114 # format marker of refs pointing to given object
2115
2116 # the destination action is chosen based on object type and current context:
2117 # - for annotated tags, we choose the tag view unless it's the current view
2118 # already, in which case we go to shortlog view
2119 # - for other refs, we keep the current view if we're in history, shortlog or
2120 # log view, and select shortlog otherwise
2121 sub format_ref_marker {
2122 my ($refs, $id) = @_;
2123 my $markers = '';
2124
2125 if (defined $refs->{$id}) {
2126 foreach my $ref (@{$refs->{$id}}) {
2127 # this code exploits the fact that non-lightweight tags are the
2128 # only indirect objects, and that they are the only objects for which
2129 # we want to use tag instead of shortlog as action
2130 my ($type, $name) = qw();
2131 my $indirect = ($ref =~ s/\^\{\}$//);
2132 # e.g. tags/v2.6.11 or heads/next
2133 if ($ref =~ m!^(.*?)s?/(.*)$!) {
2134 $type = $1;
2135 $name = $2;
2136 } else {
2137 $type = "ref";
2138 $name = $ref;
2139 }
2140
2141 my $class = $type;
2142 $class .= " indirect" if $indirect;
2143
2144 my $dest_action = "shortlog";
2145
2146 if ($indirect) {
2147 $dest_action = "tag" unless $action eq "tag";
2148 } elsif ($action =~ /^(history|(short)?log)$/) {
2149 $dest_action = $action;
2150 }
2151
2152 my $dest = "";
2153 $dest .= "refs/" unless $ref =~ m!^refs/!;
2154 $dest .= $ref;
2155
2156 my $link = $cgi->a({
2157 -href => href(
2158 action=>$dest_action,
2159 hash=>$dest
2160 )}, esc_html($name));
2161
2162 $markers .= " <span class=\"".esc_attr($class)."\" title=\"".esc_attr($ref)."\">" .
2163 $link . "</span>";
2164 }
2165 }
2166
2167 if ($markers) {
2168 return ' <span class="refs">'. $markers . '</span>';
2169 } else {
2170 return "";
2171 }
2172 }
2173
2174 # format, perhaps shortened and with markers, title line
2175 sub format_subject_html {
2176 my ($long, $short, $href, $extra) = @_;
2177 $extra = '' unless defined($extra);
2178
2179 if (length($short) < length($long)) {
2180 $long =~ s/[[:cntrl:]]/?/g;
2181 return $cgi->a({-href => $href, -class => "list subject",
2182 -title => to_utf8($long)},
2183 esc_html($short)) . $extra;
2184 } else {
2185 return $cgi->a({-href => $href, -class => "list subject"},
2186 esc_html($long)) . $extra;
2187 }
2188 }
2189
2190 # Rather than recomputing the url for an email multiple times, we cache it
2191 # after the first hit. This gives a visible benefit in views where the avatar
2192 # for the same email is used repeatedly (e.g. shortlog).
2193 # The cache is shared by all avatar engines (currently gravatar only), which
2194 # are free to use it as preferred. Since only one avatar engine is used for any
2195 # given page, there's no risk for cache conflicts.
2196 our %avatar_cache = ();
2197
2198 # Compute the picon url for a given email, by using the picon search service over at
2199 # http://www.cs.indiana.edu/picons/search.html
2200 sub picon_url {
2201 my $email = lc shift;
2202 if (!$avatar_cache{$email}) {
2203 my ($user, $domain) = split('@', $email);
2204 $avatar_cache{$email} =
2205 "//www.cs.indiana.edu/cgi-pub/kinzler/piconsearch.cgi/" .
2206 "$domain/$user/" .
2207 "users+domains+unknown/up/single";
2208 }
2209 return $avatar_cache{$email};
2210 }
2211
2212 # Compute the gravatar url for a given email, if it's not in the cache already.
2213 # Gravatar stores only the part of the URL before the size, since that's the
2214 # one computationally more expensive. This also allows reuse of the cache for
2215 # different sizes (for this particular engine).
2216 sub gravatar_url {
2217 my $email = lc shift;
2218 my $size = shift;
2219 $avatar_cache{$email} ||=
2220 "//www.gravatar.com/avatar/" .
2221 md5_hex($email) . "?s=";
2222 return $avatar_cache{$email} . $size;
2223 }
2224
2225 # Insert an avatar for the given $email at the given $size if the feature
2226 # is enabled.
2227 sub git_get_avatar {
2228 my ($email, %opts) = @_;
2229 my $pre_white = ($opts{-pad_before} ? "&nbsp;" : "");
2230 my $post_white = ($opts{-pad_after} ? "&nbsp;" : "");
2231 $opts{-size} ||= 'default';
2232 my $size = $avatar_size{$opts{-size}} || $avatar_size{'default'};
2233 my $url = "";
2234 if ($git_avatar eq 'gravatar') {
2235 $url = gravatar_url($email, $size);
2236 } elsif ($git_avatar eq 'picon') {
2237 $url = picon_url($email);
2238 }
2239 # Other providers can be added by extending the if chain, defining $url
2240 # as needed. If no variant puts something in $url, we assume avatars
2241 # are completely disabled/unavailable.
2242 if ($url) {
2243 return $pre_white .
2244 "<img width=\"$size\" " .
2245 "class=\"avatar\" " .
2246 "src=\"".esc_url($url)."\" " .
2247 "alt=\"\" " .
2248 "/>" . $post_white;
2249 } else {
2250 return "";
2251 }
2252 }
2253
2254 sub format_search_author {
2255 my ($author, $searchtype, $displaytext) = @_;
2256 my $have_search = gitweb_check_feature('search');
2257
2258 if ($have_search) {
2259 my $performed = "";
2260 if ($searchtype eq 'author') {
2261 $performed = "authored";
2262 } elsif ($searchtype eq 'committer') {
2263 $performed = "committed";
2264 }
2265
2266 return $cgi->a({-href => href(action=>"search", hash=>$hash,
2267 searchtext=>$author,
2268 searchtype=>$searchtype), class=>"list",
2269 title=>"Search for commits $performed by $author"},
2270 $displaytext);
2271
2272 } else {
2273 return $displaytext;
2274 }
2275 }
2276
2277 # format the author name of the given commit with the given tag
2278 # the author name is chopped and escaped according to the other
2279 # optional parameters (see chop_str).
2280 sub format_author_html {
2281 my $tag = shift;
2282 my $co = shift;
2283 my $author = chop_and_escape_str($co->{'author_name'}, @_);
2284 return "<$tag class=\"author\">" .
2285 format_search_author($co->{'author_name'}, "author",
2286 git_get_avatar($co->{'author_email'}, -pad_after => 1) .
2287 $author) .
2288 "</$tag>";
2289 }
2290
2291 # format git diff header line, i.e. "diff --(git|combined|cc) ..."
2292 sub format_git_diff_header_line {
2293 my $line = shift;
2294 my $diffinfo = shift;
2295 my ($from, $to) = @_;
2296
2297 if ($diffinfo->{'nparents'}) {
2298 # combined diff
2299 $line =~ s!^(diff (.*?) )"?.*$!$1!;
2300 if ($to->{'href'}) {
2301 $line .= $cgi->a({-href => $to->{'href'}, -class => "path"},
2302 esc_path($to->{'file'}));
2303 } else { # file was deleted (no href)
2304 $line .= esc_path($to->{'file'});
2305 }
2306 } else {
2307 # "ordinary" diff
2308 $line =~ s!^(diff (.*?) )"?a/.*$!$1!;
2309 if ($from->{'href'}) {
2310 $line .= $cgi->a({-href => $from->{'href'}, -class => "path"},
2311 'a/' . esc_path($from->{'file'}));
2312 } else { # file was added (no href)
2313 $line .= 'a/' . esc_path($from->{'file'});
2314 }
2315 $line .= ' ';
2316 if ($to->{'href'}) {
2317 $line .= $cgi->a({-href => $to->{'href'}, -class => "path"},
2318 'b/' . esc_path($to->{'file'}));
2319 } else { # file was deleted
2320 $line .= 'b/' . esc_path($to->{'file'});
2321 }
2322 }
2323
2324 return "<div class=\"diff header\">$line</div>\n";
2325 }
2326
2327 # format extended diff header line, before patch itself
2328 sub format_extended_diff_header_line {
2329 my $line = shift;
2330 my $diffinfo = shift;
2331 my ($from, $to) = @_;
2332
2333 # match <path>
2334 if ($line =~ s!^((copy|rename) from ).*$!$1! && $from->{'href'}) {
2335 $line .= $cgi->a({-href=>$from->{'href'}, -class=>"path"},
2336 esc_path($from->{'file'}));
2337 }
2338 if ($line =~ s!^((copy|rename) to ).*$!$1! && $to->{'href'}) {
2339 $line .= $cgi->a({-href=>$to->{'href'}, -class=>"path"},
2340 esc_path($to->{'file'}));
2341 }
2342
2343 # Temporarily remove a trailing <mode> so an index line ends with its
2344 # object IDs and can be shortened below.
2345 my $mode;
2346 if ($line =~ s/\s(\d{6})$//) {
2347 $mode = $1;
2348 }
2349
2350 # match <hash>
2351 if ($line =~ oid_nlen_prefix_infix_regex($sha1_len, "index ", ",") |
2352 $line =~ oid_nlen_prefix_infix_regex($sha256_len, "index ", ",")) {
2353 # can match only for combined diff
2354 $line = 'index ';
2355 for (my $i = 0; $i < $diffinfo->{'nparents'}; $i++) {
2356 if ($from->{'href'}[$i]) {
2357 $line .= $cgi->a({-href=>$from->{'href'}[$i],
2358 -class=>"hash"},
2359 substr($diffinfo->{'from_id'}[$i],0,7));
2360 } else {
2361 $line .= '0' x 7;
2362 }
2363 # separator
2364 $line .= ',' if ($i < $diffinfo->{'nparents'} - 1);
2365 }
2366 $line .= '..';
2367 if ($to->{'href'}) {
2368 $line .= $cgi->a({-href=>$to->{'href'}, -class=>"hash"},
2369 substr($diffinfo->{'to_id'},0,7));
2370 } else {
2371 $line .= '0' x 7;
2372 }
2373
2374 } elsif ($line =~ oid_nlen_prefix_infix_regex($sha1_len, "index ", "..") |
2375 $line =~ oid_nlen_prefix_infix_regex($sha256_len, "index ", "..")) {
2376 # can match only for ordinary diff
2377 my ($from_link, $to_link);
2378 if ($from->{'href'}) {
2379 $from_link = $cgi->a({-href=>$from->{'href'}, -class=>"hash"},
2380 substr($diffinfo->{'from_id'},0,7));
2381 } else {
2382 $from_link = '0' x 7;
2383 }
2384 if ($to->{'href'}) {
2385 $to_link = $cgi->a({-href=>$to->{'href'}, -class=>"hash"},
2386 substr($diffinfo->{'to_id'},0,7));
2387 } else {
2388 $to_link = '0' x 7;
2389 }
2390 my ($from_id, $to_id) = ($diffinfo->{'from_id'}, $diffinfo->{'to_id'});
2391 $line =~ s!$from_id\.\.$to_id!$from_link..$to_link!;
2392 }
2393 if (defined $mode) {
2394 $line .= " $mode" .
2395 '<span class="info"> (' .
2396 file_type_long($mode) .
2397 ')</span>';
2398 }
2399
2400 return $line . "<br/>\n";
2401 }
2402
2403 # format from-file/to-file diff header
2404 sub format_diff_from_to_header {
2405 my ($from_line, $to_line, $diffinfo, $from, $to, @parents) = @_;
2406 my $line;
2407 my $result = '';
2408
2409 $line = $from_line;
2410 #assert($line =~ m/^---/) if DEBUG;
2411 # no extra formatting for "^--- /dev/null"
2412 if (! $diffinfo->{'nparents'}) {
2413 # ordinary (single parent) diff
2414 if ($line =~ m!^--- "?a/!) {
2415 if ($from->{'href'}) {
2416 $line = '--- a/' .
2417 $cgi->a({-href=>$from->{'href'}, -class=>"path"},
2418 esc_path($from->{'file'}));
2419 } else {
2420 $line = '--- a/' .
2421 esc_path($from->{'file'});
2422 }
2423 }
2424 $result .= qq!<div class="diff from_file">$line</div>\n!;
2425
2426 } else {
2427 # combined diff (merge commit)
2428 for (my $i = 0; $i < $diffinfo->{'nparents'}; $i++) {
2429 if ($from->{'href'}[$i]) {
2430 $line = '--- ' .
2431 $cgi->a({-href=>href(action=>"blobdiff",
2432 hash_parent=>$diffinfo->{'from_id'}[$i],
2433 hash_parent_base=>$parents[$i],
2434 file_parent=>$from->{'file'}[$i],
2435 hash=>$diffinfo->{'to_id'},
2436 hash_base=>$hash,
2437 file_name=>$to->{'file'}),
2438 -class=>"path",
2439 -title=>"diff" . ($i+1)},
2440 $i+1) .
2441 '/' .
2442 $cgi->a({-href=>$from->{'href'}[$i], -class=>"path"},
2443 esc_path($from->{'file'}[$i]));
2444 } else {
2445 $line = '--- /dev/null';
2446 }
2447 $result .= qq!<div class="diff from_file">$line</div>\n!;
2448 }
2449 }
2450
2451 $line = $to_line;
2452 #assert($line =~ m/^\+\+\+/) if DEBUG;
2453 # no extra formatting for "^+++ /dev/null"
2454 if ($line =~ m!^\+\+\+ "?b/!) {
2455 if ($to->{'href'}) {
2456 $line = '+++ b/' .
2457 $cgi->a({-href=>$to->{'href'}, -class=>"path"},
2458 esc_path($to->{'file'}));
2459 } else {
2460 $line = '+++ b/' .
2461 esc_path($to->{'file'});
2462 }
2463 }
2464 $result .= qq!<div class="diff to_file">$line</div>\n!;
2465
2466 return $result;
2467 }
2468
2469 # create note for patch simplified by combined diff
2470 sub format_diff_cc_simplified {
2471 my ($diffinfo, @parents) = @_;
2472 my $result = '';
2473
2474 $result .= "<div class=\"diff header\">" .
2475 "diff --cc ";
2476 if (!is_deleted($diffinfo)) {
2477 $result .= $cgi->a({-href => href(action=>"blob",
2478 hash_base=>$hash,
2479 hash=>$diffinfo->{'to_id'},
2480 file_name=>$diffinfo->{'to_file'}),
2481 -class => "path"},
2482 esc_path($diffinfo->{'to_file'}));
2483 } else {
2484 $result .= esc_path($diffinfo->{'to_file'});
2485 }
2486 $result .= "</div>\n" . # class="diff header"
2487 "<div class=\"diff nodifferences\">" .
2488 "Simple merge" .
2489 "</div>\n"; # class="diff nodifferences"
2490
2491 return $result;
2492 }
2493
2494 sub diff_line_class {
2495 my ($line, $from, $to) = @_;
2496
2497 # ordinary diff
2498 my $num_sign = 1;
2499 # combined diff
2500 if ($from && $to && ref($from->{'href'}) eq "ARRAY") {
2501 $num_sign = scalar @{$from->{'href'}};
2502 }
2503
2504 my @diff_line_classifier = (
2505 { regexp => qr/^\@\@{$num_sign} /, class => "chunk_header"},
2506 { regexp => qr/^\\/, class => "incomplete" },
2507 { regexp => qr/^ {$num_sign}/, class => "ctx" },
2508 # classifier for context must come before classifier add/rem,
2509 # or we would have to use more complicated regexp, for example
2510 # qr/(?= {0,$m}\+)[+ ]{$num_sign}/, where $m = $num_sign - 1;
2511 { regexp => qr/^[+ ]{$num_sign}/, class => "add" },
2512 { regexp => qr/^[- ]{$num_sign}/, class => "rem" },
2513 );
2514 for my $clsfy (@diff_line_classifier) {
2515 return $clsfy->{'class'}
2516 if ($line =~ $clsfy->{'regexp'});
2517 }
2518
2519 # fallback
2520 return "";
2521 }
2522
2523 # assumes that $from and $to are defined and correctly filled,
2524 # and that $line holds a line of chunk header for unified diff
2525 sub format_unidiff_chunk_header {
2526 my ($line, $from, $to) = @_;
2527
2528 my ($from_text, $from_start, $from_lines, $to_text, $to_start, $to_lines, $section) =
2529 $line =~ m/^\@{2} (-(\d+)(?:,(\d+))?) (\+(\d+)(?:,(\d+))?) \@{2}(.*)$/;
2530
2531 $from_lines = 0 unless defined $from_lines;
2532 $to_lines = 0 unless defined $to_lines;
2533
2534 if ($from->{'href'}) {
2535 $from_text = $cgi->a({-href=>"$from->{'href'}#l$from_start",
2536 -class=>"list"}, $from_text);
2537 }
2538 if ($to->{'href'}) {
2539 $to_text = $cgi->a({-href=>"$to->{'href'}#l$to_start",
2540 -class=>"list"}, $to_text);
2541 }
2542 $line = "<span class=\"chunk_info\">@@ $from_text $to_text @@</span>" .
2543 "<span class=\"section\">" . esc_html($section, -nbsp=>1) . "</span>";
2544 return $line;
2545 }
2546
2547 # assumes that $from and $to are defined and correctly filled,
2548 # and that $line holds a line of chunk header for combined diff
2549 sub format_cc_diff_chunk_header {
2550 my ($line, $from, $to) = @_;
2551
2552 my ($prefix, $ranges, $section) = $line =~ m/^(\@+) (.*?) \@+(.*)$/;
2553 my (@from_text, @from_start, @from_nlines, $to_text, $to_start, $to_nlines);
2554
2555 @from_text = split(' ', $ranges);
2556 for (my $i = 0; $i < @from_text; ++$i) {
2557 ($from_start[$i], $from_nlines[$i]) =
2558 (split(',', substr($from_text[$i], 1)), 0);
2559 }
2560
2561 $to_text = pop @from_text;
2562 $to_start = pop @from_start;
2563 $to_nlines = pop @from_nlines;
2564
2565 $line = "<span class=\"chunk_info\">$prefix ";
2566 for (my $i = 0; $i < @from_text; ++$i) {
2567 if ($from->{'href'}[$i]) {
2568 $line .= $cgi->a({-href=>"$from->{'href'}[$i]#l$from_start[$i]",
2569 -class=>"list"}, $from_text[$i]);
2570 } else {
2571 $line .= $from_text[$i];
2572 }
2573 $line .= " ";
2574 }
2575 if ($to->{'href'}) {
2576 $line .= $cgi->a({-href=>"$to->{'href'}#l$to_start",
2577 -class=>"list"}, $to_text);
2578 } else {
2579 $line .= $to_text;
2580 }
2581 $line .= " $prefix</span>" .
2582 "<span class=\"section\">" . esc_html($section, -nbsp=>1) . "</span>";
2583 return $line;
2584 }
2585
2586 # process patch (diff) line (not to be used for diff headers),
2587 # returning HTML-formatted (but not wrapped) line.
2588 # If the line is passed as a reference, it is treated as HTML and not
2589 # esc_html()'ed.
2590 sub format_diff_line {
2591 my ($line, $diff_class, $from, $to) = @_;
2592
2593 if (ref($line)) {
2594 $line = $$line;
2595 } else {
2596 chomp $line;
2597 $line = untabify($line);
2598
2599 if ($from && $to && $line =~ m/^\@{2} /) {
2600 $line = format_unidiff_chunk_header($line, $from, $to);
2601 } elsif ($from && $to && $line =~ m/^\@{3}/) {
2602 $line = format_cc_diff_chunk_header($line, $from, $to);
2603 } else {
2604 $line = esc_html($line, -nbsp=>1);
2605 }
2606 }
2607
2608 my $diff_classes = "diff";
2609 $diff_classes .= " $diff_class" if ($diff_class);
2610 $line = "<div class=\"$diff_classes\">$line</div>\n";
2611
2612 return $line;
2613 }
2614
2615 # Generates undef or something like "_snapshot_" or "snapshot (_tbz2_ _zip_)",
2616 # linked. Pass the hash of the tree/commit to snapshot.
2617 sub format_snapshot_links {
2618 my ($hash) = @_;
2619 my $num_fmts = @snapshot_fmts;
2620 if ($num_fmts > 1) {
2621 # A parenthesized list of links bearing format names.
2622 # e.g. "snapshot (_tar.gz_ _zip_)"
2623 return "snapshot (" . join(' ', map
2624 $cgi->a({
2625 -href => href(
2626 action=>"snapshot",
2627 hash=>$hash,
2628 snapshot_format=>$_
2629 )
2630 }, $known_snapshot_formats{$_}{'display'})
2631 , @snapshot_fmts) . ")";
2632 } elsif ($num_fmts == 1) {
2633 # A single "snapshot" link whose tooltip bears the format name.
2634 # i.e. "_snapshot_"
2635 my ($fmt) = @snapshot_fmts;
2636 return
2637 $cgi->a({
2638 -href => href(
2639 action=>"snapshot",
2640 hash=>$hash,
2641 snapshot_format=>$fmt
2642 ),
2643 -title => "in format: $known_snapshot_formats{$fmt}{'display'}"
2644 }, "snapshot");
2645 } else { # $num_fmts == 0
2646 return undef;
2647 }
2648 }
2649
2650 ## ......................................................................
2651 ## functions returning values to be passed, perhaps after some
2652 ## transformation, to other functions; e.g. returning arguments to href()
2653
2654 # returns hash to be passed to href to generate gitweb URL
2655 # in -title key it returns description of link
2656 sub get_feed_info {
2657 my $format = shift || 'Atom';
2658 my %res = (action => lc($format));
2659 my $matched_ref = 0;
2660
2661 # feed links are possible only for project views
2662 return unless (defined $project);
2663 # some views should link to OPML, or to generic project feed,
2664 # or don't have specific feed yet (so they should use generic)
2665 return if (!$action || $action =~ /^(?:tags|heads|forks|tag|search)$/x);
2666
2667 my $branch = undef;
2668 # branches refs uses 'refs/' + $get_branch_refs()[x] + '/' prefix
2669 # (fullname) to differentiate from tag links; this also makes
2670 # possible to detect branch links
2671 for my $ref (get_branch_refs()) {
2672 if ((defined $hash_base && $hash_base =~ m!^refs/\Q$ref\E/(.*)$!) ||
2673 (defined $hash && $hash =~ m!^refs/\Q$ref\E/(.*)$!)) {
2674 $branch = $1;
2675 $matched_ref = $ref;
2676 last;
2677 }
2678 }
2679 # find log type for feed description (title)
2680 my $type = 'log';
2681 if (defined $file_name) {
2682 $type = "history of $file_name";
2683 $type .= "/" if ($action eq 'tree');
2684 $type .= " on '$branch'" if (defined $branch);
2685 } else {
2686 $type = "log of $branch" if (defined $branch);
2687 }
2688
2689 $res{-title} = $type;
2690 $res{'hash'} = (defined $branch ? "refs/$matched_ref/$branch" : undef);
2691 $res{'file_name'} = $file_name;
2692
2693 return %res;
2694 }
2695
2696 ## ----------------------------------------------------------------------
2697 ## git utility subroutines, invoking git commands
2698
2699 # returns path to the core git executable and the --git-dir parameter as list
2700 sub git_cmd {
2701 $number_of_git_cmds++;
2702 return $GIT, '--git-dir='.$git_dir;
2703 }
2704
2705 # quote the given arguments for passing them to the shell
2706 # quote_command("command", "arg 1", "arg with ' and ! characters")
2707 # => "'command' 'arg 1' 'arg with '\'' and '\!' characters'"
2708 # Try to avoid using this function wherever possible.
2709 sub quote_command {
2710 return join(' ',
2711 map { my $a = $_ =~ s/(['!])/'\\$1'/gr; "'$a'" } @_ );
2712 }
2713
2714 # get HEAD ref of given project as hash
2715 sub git_get_head_hash {
2716 return git_get_full_hash(shift, 'HEAD');
2717 }
2718
2719 sub git_get_full_hash {
2720 return git_get_hash(@_);
2721 }
2722
2723 sub git_get_short_hash {
2724 return git_get_hash(@_, '--short=7');
2725 }
2726
2727 sub git_get_hash {
2728 my ($project, $hash, @options) = @_;
2729 my $o_git_dir = $git_dir;
2730 my $retval = undef;
2731 $git_dir = "$projectroot/$project";
2732 if (open my $fd, '-|', git_cmd(), 'rev-parse',
2733 '--verify', '-q', @options, $hash) {
2734 $retval = <$fd>;
2735 chomp $retval if defined $retval;
2736 close $fd;
2737 }
2738 if (defined $o_git_dir) {
2739 $git_dir = $o_git_dir;
2740 }
2741 return $retval;
2742 }
2743
2744 # get type of given object
2745 sub git_get_type {
2746 my $hash = shift;
2747
2748 open my $fd, "-|", git_cmd(), "cat-file", '-t', $hash or return;
2749 my $type = <$fd>;
2750 close $fd or return;
2751 chomp $type;
2752 return $type;
2753 }
2754
2755 # repository configuration
2756 our $config_file = '';
2757 our %config;
2758
2759 # store multiple values for single key as anonymous array reference
2760 # single values stored directly in the hash, not as [ <value> ]
2761 sub hash_set_multi {
2762 my ($hash, $key, $value) = @_;
2763
2764 if (!exists $hash->{$key}) {
2765 $hash->{$key} = $value;
2766 } elsif (!ref $hash->{$key}) {
2767 $hash->{$key} = [ $hash->{$key}, $value ];
2768 } else {
2769 push @{$hash->{$key}}, $value;
2770 }
2771 }
2772
2773 # return hash of git project configuration
2774 # optionally limited to some section, e.g. 'gitweb'
2775 sub git_parse_project_config {
2776 my $section_regexp = shift;
2777 my %config;
2778
2779 local $/ = "\0";
2780
2781 open my $fh, "-|", git_cmd(), "config", '-z', '-l',
2782 or return;
2783
2784 while (my $keyval = <$fh>) {
2785 chomp $keyval;
2786 my ($key, $value) = split(/\n/, $keyval, 2);
2787
2788 hash_set_multi(\%config, $key, $value)
2789 if (!defined $section_regexp || $key =~ /^(?:$section_regexp)\./o);
2790 }
2791 close $fh;
2792
2793 return %config;
2794 }
2795
2796 # convert config value to boolean: 'true' or 'false'
2797 # no value, number > 0, 'true' and 'yes' values are true
2798 # rest of values are treated as false (never as error)
2799 sub config_to_bool {
2800 my $val = shift;
2801
2802 return 1 if !defined $val; # section.key
2803
2804 # strip leading and trailing whitespace
2805 $val =~ s/^\s+//;
2806 $val =~ s/\s+$//;
2807
2808 return (($val =~ /^\d+$/ && $val) || # section.key = 1
2809 ($val =~ /^(?:true|yes)$/i)); # section.key = true
2810 }
2811
2812 # convert config value to simple decimal number
2813 # an optional value suffix of 'k', 'm', or 'g' will cause the value
2814 # to be multiplied by 1024, 1048576, or 1073741824
2815 sub config_to_int {
2816 my $val = shift;
2817
2818 # strip leading and trailing whitespace
2819 $val =~ s/^\s+//;
2820 $val =~ s/\s+$//;
2821
2822 if (my ($num, $unit) = ($val =~ /^([0-9]*)([kmg])$/i)) {
2823 $unit = lc($unit);
2824 # unknown unit is treated as 1
2825 return $num * ($unit eq 'g' ? 1073741824 :
2826 $unit eq 'm' ? 1048576 :
2827 $unit eq 'k' ? 1024 : 1);
2828 }
2829 return $val;
2830 }
2831
2832 # convert config value to array reference, if needed
2833 sub config_to_multi {
2834 my $val = shift;
2835
2836 return ref($val) ? $val : (defined($val) ? [ $val ] : []);
2837 }
2838
2839 sub git_get_project_config {
2840 my ($key, $type) = @_;
2841
2842 return unless defined $git_dir;
2843
2844 # key sanity check
2845 return unless ($key);
2846 # only subsection, if exists, is case sensitive,
2847 # and not lowercased by 'git config -z -l'
2848 if (my ($hi, $mi, $lo) = ($key =~ /^([^.]*)\.(.*)\.([^.]*)$/)) {
2849 $lo =~ s/_//g;
2850 $key = join(".", lc($hi), $mi, lc($lo));
2851 return if ($lo =~ /\W/ || $hi =~ /\W/);
2852 } else {
2853 $key = lc($key);
2854 $key =~ s/_//g;
2855 return if ($key =~ /\W/);
2856 }
2857 $key =~ s/^gitweb\.//;
2858
2859 # type sanity check
2860 if (defined $type) {
2861 $type =~ s/^--//;
2862 $type = undef
2863 unless ($type eq 'bool' || $type eq 'int');
2864 }
2865
2866 # get config
2867 if (!defined $config_file ||
2868 $config_file ne "$git_dir/config") {
2869 %config = git_parse_project_config('gitweb');
2870 $config_file = "$git_dir/config";
2871 }
2872
2873 # check if config variable (key) exists
2874 return unless exists $config{"gitweb.$key"};
2875
2876 # ensure given type
2877 if (!defined $type) {
2878 return $config{"gitweb.$key"};
2879 } elsif ($type eq 'bool') {
2880 # backward compatibility: 'git config --bool' returns true/false
2881 return config_to_bool($config{"gitweb.$key"}) ? 'true' : 'false';
2882 } elsif ($type eq 'int') {
2883 return config_to_int($config{"gitweb.$key"});
2884 }
2885 return $config{"gitweb.$key"};
2886 }
2887
2888 # get hash of given path at given ref
2889 sub git_get_hash_by_path {
2890 my $base = shift;
2891 my $path = shift || return undef;
2892 my $type = shift;
2893
2894 $path =~ s,/+$,,;
2895
2896 open my $fd, "-|", git_cmd(), "ls-tree", $base, "--", $path
2897 or die_error(500, "Open git-ls-tree failed");
2898 my $line = <$fd>;
2899 close $fd or return undef;
2900
2901 if (!defined $line) {
2902 # there is no tree or hash given by $path at $base
2903 return undef;
2904 }
2905
2906 #'100644 blob 0fa3f3a66fb6a137f6ec2c19351ed4d807070ffa panic.c'
2907 $line =~ m/^([0-9]+) (.+) ($oid_regex)\t/;
2908 if (defined $type && $type ne $2) {
2909 # type doesn't match
2910 return undef;
2911 }
2912 return $3;
2913 }
2914
2915 # get path of entry with given hash at given tree-ish (ref)
2916 # used to get 'from' filename for combined diff (merge commit) for renames
2917 sub git_get_path_by_hash {
2918 my $base = shift || return;
2919 my $hash = shift || return;
2920
2921 local $/ = "\0";
2922
2923 open my $fd, "-|", git_cmd(), "ls-tree", '-r', '-t', '-z', $base
2924 or return undef;
2925 while (my $line = <$fd>) {
2926 chomp $line;
2927
2928 #'040000 tree 595596a6a9117ddba9fe379b6b012b558bac8423 gitweb'
2929 #'100644 blob e02e90f0429be0d2a69b76571101f20b8f75530f gitweb/README'
2930 if ($line =~ m/(?:[0-9]+) (?:.+) $hash\t(.+)$/) {
2931 close $fd;
2932 return $1;
2933 }
2934 }
2935 close $fd;
2936 return undef;
2937 }
2938
2939 ## ......................................................................
2940 ## git utility functions, directly accessing git repository
2941
2942 # get the value of config variable either from file named as the variable
2943 # itself in the repository ($GIT_DIR/$name file), or from gitweb.$name
2944 # configuration variable in the repository config file.
2945 sub git_get_file_or_project_config {
2946 my ($path, $name) = @_;
2947
2948 $git_dir = "$projectroot/$path";
2949 open my $fd, '<', "$git_dir/$name"
2950 or return git_get_project_config($name);
2951 my $conf = <$fd>;
2952 close $fd;
2953 if (defined $conf) {
2954 chomp $conf;
2955 }
2956 return $conf;
2957 }
2958
2959 sub git_get_project_description {
2960 my $path = shift;
2961 return git_get_file_or_project_config($path, 'description');
2962 }
2963
2964 sub git_get_project_category {
2965 my $path = shift;
2966 return git_get_file_or_project_config($path, 'category');
2967 }
2968
2969
2970 # supported formats:
2971 # * $GIT_DIR/ctags/<tagname> file (in 'ctags' subdirectory)
2972 # - if its contents is a number, use it as tag weight,
2973 # - otherwise add a tag with weight 1
2974 # * $GIT_DIR/ctags file, each line is a tag (with weight 1)
2975 # the same value multiple times increases tag weight
2976 # * `gitweb.ctag' multi-valued repo config variable
2977 sub git_get_project_ctags {
2978 my $project = shift;
2979 my $ctags = {};
2980
2981 $git_dir = "$projectroot/$project";
2982 if (opendir my $dh, "$git_dir/ctags") {
2983 my @files = grep { -f $_ } map { "$git_dir/ctags/$_" } readdir($dh);
2984 foreach my $tagfile (@files) {
2985 open my $ct, '<', $tagfile
2986 or next;
2987 my $val = <$ct>;
2988 chomp $val if $val;
2989 close $ct;
2990
2991 (my $ctag = $tagfile) =~ s#.*/##;
2992 if ($val =~ /^\d+$/) {
2993 $ctags->{$ctag} = $val;
2994 } else {
2995 $ctags->{$ctag} = 1;
2996 }
2997 }
2998 closedir $dh;
2999
3000 } elsif (open my $fh, '<', "$git_dir/ctags") {
3001 while (my $line = <$fh>) {
3002 chomp $line;
3003 $ctags->{$line}++ if $line;
3004 }
3005 close $fh;
3006
3007 } else {
3008 my $taglist = config_to_multi(git_get_project_config('ctag'));
3009 foreach my $tag (@$taglist) {
3010 $ctags->{$tag}++;
3011 }
3012 }
3013
3014 return $ctags;
3015 }
3016
3017 # return hash, where keys are content tags ('ctags'),
3018 # and values are sum of weights of given tag in every project
3019 sub git_gather_all_ctags {
3020 my $projects = shift;
3021 my $ctags = {};
3022
3023 foreach my $p (@$projects) {
3024 foreach my $ct (keys %{$p->{'ctags'}}) {
3025 $ctags->{$ct} += $p->{'ctags'}->{$ct};
3026 }
3027 }
3028
3029 return $ctags;
3030 }
3031
3032 sub git_populate_project_tagcloud {
3033 my $ctags = shift;
3034
3035 # First, merge different-cased tags; tags vote on casing
3036 my %ctags_lc;
3037 foreach (keys %$ctags) {
3038 $ctags_lc{lc $_}->{count} += $ctags->{$_};
3039 if (not $ctags_lc{lc $_}->{topcount}
3040 or $ctags_lc{lc $_}->{topcount} < $ctags->{$_}) {
3041 $ctags_lc{lc $_}->{topcount} = $ctags->{$_};
3042 $ctags_lc{lc $_}->{topname} = $_;
3043 }
3044 }
3045
3046 my $cloud;
3047 my $matched = $input_params{'ctag'};
3048 if (eval { require HTML::TagCloud; 1; }) {
3049 $cloud = HTML::TagCloud->new;
3050 foreach my $ctag (sort keys %ctags_lc) {
3051 # Pad the title with spaces so that the cloud looks
3052 # less crammed.
3053 my $title = esc_html($ctags_lc{$ctag}->{topname});
3054 $title =~ s/ /&nbsp;/g;
3055 $title =~ s/^/&nbsp;/g;
3056 $title =~ s/$/&nbsp;/g;
3057 if (defined $matched && $matched eq $ctag) {
3058 $title = qq(<span class="match">$title</span>);
3059 }
3060 $cloud->add($title, href(project=>undef, ctag=>$ctag),
3061 $ctags_lc{$ctag}->{count});
3062 }
3063 } else {
3064 $cloud = {};
3065 foreach my $ctag (keys %ctags_lc) {
3066 my $title = esc_html($ctags_lc{$ctag}->{topname}, -nbsp=>1);
3067 if (defined $matched && $matched eq $ctag) {
3068 $title = qq(<span class="match">$title</span>);
3069 }
3070 $cloud->{$ctag}{count} = $ctags_lc{$ctag}->{count};
3071 $cloud->{$ctag}{ctag} =
3072 $cgi->a({-href=>href(project=>undef, ctag=>$ctag)}, $title);
3073 }
3074 }
3075 return $cloud;
3076 }
3077
3078 sub git_show_project_tagcloud {
3079 my ($cloud, $count) = @_;
3080 if (ref $cloud eq 'HTML::TagCloud') {
3081 return $cloud->html_and_css($count);
3082 } else {
3083 my @tags = sort { $cloud->{$a}->{'count'} <=> $cloud->{$b}->{'count'} } keys %$cloud;
3084 return
3085 '<div id="htmltagcloud"'.($project ? '' : ' align="center"').'>' .
3086 join (', ', map {
3087 $cloud->{$_}->{'ctag'}
3088 } splice(@tags, 0, $count)) .
3089 '</div>';
3090 }
3091 }
3092
3093 sub git_get_project_url_list {
3094 my $path = shift;
3095
3096 $git_dir = "$projectroot/$path";
3097 open my $fd, '<', "$git_dir/cloneurl"
3098 or return wantarray ?
3099 @{ config_to_multi(git_get_project_config('url')) } :
3100 config_to_multi(git_get_project_config('url'));
3101 my @git_project_url_list = map { chomp; $_ } <$fd>;
3102 close $fd;
3103
3104 return wantarray ? @git_project_url_list : \@git_project_url_list;
3105 }
3106
3107 sub git_get_projects_list {
3108 my $filter = shift || '';
3109 my $paranoid = shift;
3110 my @list;
3111
3112 if (-d $projects_list) {
3113 # search in directory
3114 my $dir = $projects_list;
3115 # remove the trailing "/"
3116 $dir =~ s!/+$!!;
3117 my $pfxlen = length("$dir");
3118 my $pfxdepth = ($dir =~ tr!/!!);
3119 # when filtering, search only given subdirectory
3120 if ($filter && !$paranoid) {
3121 $dir .= "/$filter";
3122 $dir =~ s!/+$!!;
3123 }
3124
3125 File::Find::find({
3126 follow_fast => 1, # follow symbolic links
3127 follow_skip => 2, # ignore duplicates
3128 dangling_symlinks => 0, # ignore dangling symlinks, silently
3129 wanted => sub {
3130 # global variables
3131 our $project_maxdepth;
3132 our $projectroot;
3133 # skip project-list toplevel, if we get it.
3134 return if (m!^[/.]$!);
3135 # only directories can be git repositories
3136 return unless (-d $_);
3137 # need search permission
3138 return unless (-x $_);
3139 # don't traverse too deep (Find is super slow on os x)
3140 # $project_maxdepth excludes depth of $projectroot
3141 if (($File::Find::name =~ tr!/!!) - $pfxdepth > $project_maxdepth) {
3142 $File::Find::prune = 1;
3143 return;
3144 }
3145
3146 my $path = substr($File::Find::name, $pfxlen + 1);
3147 # paranoidly only filter here
3148 if ($paranoid && $filter && $path !~ m!^\Q$filter\E/!) {
3149 next;
3150 }
3151 # we check related file in $projectroot
3152 if (check_export_ok("$projectroot/$path")) {
3153 push @list, { path => $path };
3154 $File::Find::prune = 1;
3155 }
3156 },
3157 }, "$dir");
3158
3159 } elsif (-f $projects_list) {
3160 # read from file(url-encoded):
3161 # 'git%2Fgit.git Linus+Torvalds'
3162 # 'libs%2Fklibc%2Fklibc.git H.+Peter+Anvin'
3163 # 'linux%2Fhotplug%2Fudev.git Greg+Kroah-Hartman'
3164 open my $fd, '<', $projects_list or return;
3165 PROJECT:
3166 while (my $line = <$fd>) {
3167 chomp $line;
3168 my ($path, $owner) = split ' ', $line;
3169 $path = unescape($path);
3170 $owner = unescape($owner);
3171 if (!defined $path) {
3172 next;
3173 }
3174 # if $filter is rpovided, check if $path begins with $filter
3175 if ($filter && $path !~ m!^\Q$filter\E/!) {
3176 next;
3177 }
3178 if (check_export_ok("$projectroot/$path")) {
3179 my $pr = {
3180 path => $path
3181 };
3182 if ($owner) {
3183 $pr->{'owner'} = to_utf8($owner);
3184 }
3185 push @list, $pr;
3186 }
3187 }
3188 close $fd;
3189 }
3190 return @list;
3191 }
3192
3193 # written with help of Tree::Trie module (Perl Artistic License, GPL compatible)
3194 # as side effects it sets 'forks' field to list of forks for forked projects
3195 sub filter_forks_from_projects_list {
3196 my $projects = shift;
3197
3198 my %trie; # prefix tree of directories (path components)
3199 # generate trie out of those directories that might contain forks
3200 foreach my $pr (@$projects) {
3201 my $path = $pr->{'path'};
3202 $path =~ s/\.git$//; # forks of 'repo.git' are in 'repo/' directory
3203 next if ($path =~ m!/$!); # skip non-bare repositories, e.g. 'repo/.git'
3204 next unless ($path); # skip '.git' repository: tests, git-instaweb
3205 next unless (-d "$projectroot/$path"); # containing directory exists
3206 $pr->{'forks'} = []; # there can be 0 or more forks of project
3207
3208 # add to trie
3209 my @dirs = split('/', $path);
3210 # walk the trie, until either runs out of components or out of trie
3211 my $ref = \%trie;
3212 while (scalar @dirs &&
3213 exists($ref->{$dirs[0]})) {
3214 $ref = $ref->{shift @dirs};
3215 }
3216 # create rest of trie structure from rest of components
3217 foreach my $dir (@dirs) {
3218 $ref = $ref->{$dir} = {};
3219 }
3220 # create end marker, store $pr as a data
3221 $ref->{''} = $pr if (!exists $ref->{''});
3222 }
3223
3224 # filter out forks, by finding shortest prefix match for paths
3225 my @filtered;
3226 PROJECT:
3227 foreach my $pr (@$projects) {
3228 # trie lookup
3229 my $ref = \%trie;
3230 DIR:
3231 foreach my $dir (split('/', $pr->{'path'})) {
3232 if (exists $ref->{''}) {
3233 # found [shortest] prefix, is a fork - skip it
3234 push @{$ref->{''}{'forks'}}, $pr;
3235 next PROJECT;
3236 }
3237 if (!exists $ref->{$dir}) {
3238 # not in trie, cannot have prefix, not a fork
3239 push @filtered, $pr;
3240 next PROJECT;
3241 }
3242 # If the dir is there, we just walk one step down the trie.
3243 $ref = $ref->{$dir};
3244 }
3245 # we ran out of trie
3246 # (shouldn't happen: it's either no match, or end marker)
3247 push @filtered, $pr;
3248 }
3249
3250 return @filtered;
3251 }
3252
3253 # note: fill_project_list_info must be run first,
3254 # for 'descr_long' and 'ctags' to be filled
3255 sub search_projects_list {
3256 my ($projlist, %opts) = @_;
3257 my $tagfilter = $opts{'tagfilter'};
3258 my $search_re = $opts{'search_regexp'};
3259
3260 return @$projlist
3261 unless ($tagfilter || $search_re);
3262
3263 # searching projects require filling to be run before it;
3264 fill_project_list_info($projlist,
3265 $tagfilter ? 'ctags' : (),
3266 $search_re ? ('path', 'descr') : ());
3267 my @projects;
3268 PROJECT:
3269 foreach my $pr (@$projlist) {
3270
3271 if ($tagfilter) {
3272 next unless ref($pr->{'ctags'}) eq 'HASH';
3273 next unless
3274 grep { lc($_) eq lc($tagfilter) } keys %{$pr->{'ctags'}};
3275 }
3276
3277 if ($search_re) {
3278 next unless
3279 $pr->{'path'} =~ /$search_re/ ||
3280 $pr->{'descr_long'} =~ /$search_re/;
3281 }
3282
3283 push @projects, $pr;
3284 }
3285
3286 return @projects;
3287 }
3288
3289 our $gitweb_project_owner = undef;
3290 sub git_get_project_list_from_file {
3291
3292 return if (defined $gitweb_project_owner);
3293
3294 $gitweb_project_owner = {};
3295 # read from file (url-encoded):
3296 # 'git%2Fgit.git Linus+Torvalds'
3297 # 'libs%2Fklibc%2Fklibc.git H.+Peter+Anvin'
3298 # 'linux%2Fhotplug%2Fudev.git Greg+Kroah-Hartman'
3299 if (-f $projects_list) {
3300 open(my $fd, '<', $projects_list);
3301 while (my $line = <$fd>) {
3302 chomp $line;
3303 my ($pr, $ow) = split ' ', $line;
3304 $pr = unescape($pr);
3305 $ow = unescape($ow);
3306 $gitweb_project_owner->{$pr} = to_utf8($ow);
3307 }
3308 close $fd;
3309 }
3310 }
3311
3312 sub git_get_project_owner {
3313 my $project = shift;
3314 my $owner;
3315
3316 return undef unless $project;
3317 $git_dir = "$projectroot/$project";
3318
3319 if (!defined $gitweb_project_owner) {
3320 git_get_project_list_from_file();
3321 }
3322
3323 if (exists $gitweb_project_owner->{$project}) {
3324 $owner = $gitweb_project_owner->{$project};
3325 }
3326 if (!defined $owner){
3327 $owner = git_get_project_config('owner');
3328 }
3329 if (!defined $owner) {
3330 $owner = get_file_owner("$git_dir");
3331 }
3332
3333 return $owner;
3334 }
3335
3336 sub git_get_last_activity {
3337 my ($path) = @_;
3338 my $fd;
3339
3340 $git_dir = "$projectroot/$path";
3341 open($fd, "-|", git_cmd(), 'for-each-ref',
3342 '--format=%(committer)',
3343 '--sort=-committerdate',
3344 '--count=1',
3345 map { "refs/$_" } get_branch_refs ()) or return;
3346 my $most_recent = <$fd>;
3347 close $fd or return;
3348 if (defined $most_recent &&
3349 $most_recent =~ / (\d+) [-+][01]\d\d\d$/) {
3350 my $timestamp = $1;
3351 my $age = time - $timestamp;
3352 return ($age, age_string($age));
3353 }
3354 return (undef, undef);
3355 }
3356
3357 # Implementation note: when a single remote is wanted, we cannot use 'git
3358 # remote show -n' because that command always work (assuming it's a remote URL
3359 # if it's not defined), and we cannot use 'git remote show' because that would
3360 # try to make a network roundtrip. So the only way to find if that particular
3361 # remote is defined is to walk the list provided by 'git remote -v' and stop if
3362 # and when we find what we want.
3363 sub git_get_remotes_list {
3364 my $wanted = shift;
3365 my %remotes = ();
3366
3367 open my $fd, '-|' , git_cmd(), 'remote', '-v';
3368 return unless $fd;
3369 while (my $remote = <$fd>) {
3370 chomp $remote;
3371 $remote =~ s!\t(.*?)\s+\((\w+)\)$!!;
3372 next if $wanted and not $remote eq $wanted;
3373 my ($url, $key) = ($1, $2);
3374
3375 $remotes{$remote} ||= { 'heads' => () };
3376 $remotes{$remote}{$key} = $url;
3377 }
3378 close $fd or return;
3379 return wantarray ? %remotes : \%remotes;
3380 }
3381
3382 # Takes a hash of remotes as first parameter and fills it by adding the
3383 # available remote heads for each of the indicated remotes.
3384 sub fill_remote_heads {
3385 my $remotes = shift;
3386 my @heads = map { "remotes/$_" } keys %$remotes;
3387 my @remoteheads = git_get_heads_list(undef, @heads);
3388 foreach my $remote (keys %$remotes) {
3389 $remotes->{$remote}{'heads'} = [ grep {
3390 $_->{'name'} =~ s!^$remote/!!
3391 } @remoteheads ];
3392 }
3393 }
3394
3395 sub git_get_references {
3396 my $type = shift || "";
3397 my %refs;
3398 # 5dc01c595e6c6ec9ccda4f6f69c131c0dd945f8c refs/tags/v2.6.11
3399 # c39ae07f393806ccf406ef966e9a15afc43cc36a refs/tags/v2.6.11^{}
3400 open my $fd, "-|", git_cmd(), "show-ref", "--dereference",
3401 ($type ? ("--", "refs/$type") : ()) # use -- <pattern> if $type
3402 or return;
3403
3404 while (my $line = <$fd>) {
3405 chomp $line;
3406 if ($line =~ m!^($oid_regex)\srefs/($type.*)$!) {
3407 if (defined $refs{$1}) {
3408 push @{$refs{$1}}, $2;
3409 } else {
3410 $refs{$1} = [ $2 ];
3411 }
3412 }
3413 }
3414 close $fd or return;
3415 return \%refs;
3416 }
3417
3418 sub git_get_rev_name_tags {
3419 my $hash = shift || return undef;
3420
3421 open my $fd, "-|", git_cmd(), "name-rev", "--tags", $hash
3422 or return;
3423 my $name_rev = <$fd>;
3424 close $fd;
3425
3426 if ($name_rev =~ m|^$hash tags/(.*)$|) {
3427 return $1;
3428 } else {
3429 # catches also '$hash undefined' output
3430 return undef;
3431 }
3432 }
3433
3434 ## ----------------------------------------------------------------------
3435 ## parse to hash functions
3436
3437 sub parse_date {
3438 my $epoch = shift;
3439 my $tz = shift || "-0000";
3440
3441 my %date;
3442 my @months = ("Jan", "Feb", "Mar", "Apr", "May", "Jun", "Jul", "Aug", "Sep", "Oct", "Nov", "Dec");
3443 my @days = ("Sun", "Mon", "Tue", "Wed", "Thu", "Fri", "Sat");
3444 my ($sec, $min, $hour, $mday, $mon, $year, $wday, $yday) = gmtime($epoch);
3445 $date{'hour'} = $hour;
3446 $date{'minute'} = $min;
3447 $date{'mday'} = $mday;
3448 $date{'day'} = $days[$wday];
3449 $date{'month'} = $months[$mon];
3450 $date{'rfc2822'} = sprintf "%s, %d %s %4d %02d:%02d:%02d +0000",
3451 $days[$wday], $mday, $months[$mon], 1900+$year, $hour ,$min, $sec;
3452 $date{'mday-time'} = sprintf "%d %s %02d:%02d",
3453 $mday, $months[$mon], $hour ,$min;
3454 $date{'iso-8601'} = sprintf "%04d-%02d-%02dT%02d:%02d:%02dZ",
3455 1900+$year, 1+$mon, $mday, $hour ,$min, $sec;
3456
3457 my ($tz_sign, $tz_hour, $tz_min) =
3458 ($tz =~ m/^([-+])(\d\d)(\d\d)$/);
3459 $tz_sign = ($tz_sign eq '-' ? -1 : +1);
3460 my $local = $epoch + $tz_sign*((($tz_hour*60) + $tz_min)*60);
3461 ($sec, $min, $hour, $mday, $mon, $year, $wday, $yday) = gmtime($local);
3462 $date{'hour_local'} = $hour;
3463 $date{'minute_local'} = $min;
3464 $date{'tz_local'} = $tz;
3465 $date{'iso-tz'} = sprintf("%04d-%02d-%02d %02d:%02d:%02d %s",
3466 1900+$year, $mon+1, $mday,
3467 $hour, $min, $sec, $tz);
3468 return %date;
3469 }
3470
3471 sub hide_mailaddrs_if_private {
3472 my $line = shift;
3473 return $line unless gitweb_check_feature('email-privacy');
3474 $line =~ s/<[^@>]+@[^>]+>/<redacted>/g;
3475 return $line;
3476 }
3477
3478 sub parse_tag {
3479 my $tag_id = shift;
3480 my %tag;
3481 my @comment;
3482
3483 open my $fd, "-|", git_cmd(), "cat-file", "tag", $tag_id or return;
3484 $tag{'id'} = $tag_id;
3485 while (my $line = <$fd>) {
3486 chomp $line;
3487 if ($line =~ m/^object ($oid_regex)$/) {
3488 $tag{'object'} = $1;
3489 } elsif ($line =~ m/^type (.+)$/) {
3490 $tag{'type'} = $1;
3491 } elsif ($line =~ m/^tag (.+)$/) {
3492 $tag{'name'} = $1;
3493 } elsif ($line =~ m/^tagger (.*) ([0-9]+) (.*)$/) {
3494 $tag{'author'} = hide_mailaddrs_if_private($1);
3495 $tag{'author_epoch'} = $2;
3496 $tag{'author_tz'} = $3;
3497 if ($tag{'author'} =~ m/^([^<]+) <([^>]*)>/) {
3498 $tag{'author_name'} = $1;
3499 $tag{'author_email'} = $2;
3500 } else {
3501 $tag{'author_name'} = $tag{'author'};
3502 }
3503 } elsif ($line =~ m/--BEGIN/) {
3504 push @comment, $line;
3505 last;
3506 } elsif ($line eq "") {
3507 last;
3508 }
3509 }
3510 push @comment, <$fd>;
3511 $tag{'comment'} = \@comment;
3512 close $fd or return;
3513 if (!defined $tag{'name'}) {
3514 return
3515 };
3516 return %tag
3517 }
3518
3519 sub parse_commit_text {
3520 my ($commit_text, $withparents) = @_;
3521 my @commit_lines = split '\n', $commit_text;
3522 my %co;
3523
3524 pop @commit_lines; # Remove '\0'
3525
3526 if (! @commit_lines) {
3527 return;
3528 }
3529
3530 my $header = shift @commit_lines;
3531 if ($header !~ m/^$oid_regex/) {
3532 return;
3533 }
3534 ($co{'id'}, my @parents) = split ' ', $header;
3535 while (my $line = shift @commit_lines) {
3536 last if $line eq "\n";
3537 if ($line =~ m/^tree ($oid_regex)$/) {
3538 $co{'tree'} = $1;
3539 } elsif ((!defined $withparents) && ($line =~ m/^parent ($oid_regex)$/)) {
3540 push @parents, $1;
3541 } elsif ($line =~ m/^author (.*) ([0-9]+) (.*)$/) {
3542 $co{'author'} = hide_mailaddrs_if_private(to_utf8($1));
3543 $co{'author_epoch'} = $2;
3544 $co{'author_tz'} = $3;
3545 if ($co{'author'} =~ m/^([^<]+) <([^>]*)>/) {
3546 $co{'author_name'} = $1;
3547 $co{'author_email'} = $2;
3548 } else {
3549 $co{'author_name'} = $co{'author'};
3550 }
3551 } elsif ($line =~ m/^committer (.*) ([0-9]+) (.*)$/) {
3552 $co{'committer'} = hide_mailaddrs_if_private(to_utf8($1));
3553 $co{'committer_epoch'} = $2;
3554 $co{'committer_tz'} = $3;
3555 if ($co{'committer'} =~ m/^([^<]+) <([^>]*)>/) {
3556 $co{'committer_name'} = $1;
3557 $co{'committer_email'} = $2;
3558 } else {
3559 $co{'committer_name'} = $co{'committer'};
3560 }
3561 }
3562 }
3563 if (!defined $co{'tree'}) {
3564 return;
3565 };
3566 $co{'parents'} = \@parents;
3567 $co{'parent'} = $parents[0];
3568
3569 foreach my $title (@commit_lines) {
3570 $title =~ s/^ //;
3571 if ($title ne "") {
3572 $co{'title'} = chop_str($title, 80, 5);
3573 $co{'title_short'} = chop_str($title, 50, 5);
3574 last;
3575 }
3576 }
3577 if (! defined $co{'title'} || $co{'title'} eq "") {
3578 $co{'title'} = $co{'title_short'} = '(no commit message)';
3579 }
3580 # remove added spaces, redact e-mail addresses if applicable.
3581 foreach my $line (@commit_lines) {
3582 $line =~ s/^ //;
3583 $line = hide_mailaddrs_if_private($line);
3584 }
3585 $co{'comment'} = \@commit_lines;
3586
3587 my $age = time - $co{'committer_epoch'};
3588 $co{'age'} = $age;
3589 $co{'age_string'} = age_string($age);
3590 my ($sec, $min, $hour, $mday, $mon, $year, $wday, $yday) = gmtime($co{'committer_epoch'});
3591 if ($age > 60*60*24*7*2) {
3592 $co{'age_string_date'} = sprintf "%4i-%02u-%02i", 1900 + $year, $mon+1, $mday;
3593 $co{'age_string_age'} = $co{'age_string'};
3594 } else {
3595 $co{'age_string_date'} = $co{'age_string'};
3596 $co{'age_string_age'} = sprintf "%4i-%02u-%02i", 1900 + $year, $mon+1, $mday;
3597 }
3598 return %co;
3599 }
3600
3601 sub parse_commit {
3602 my ($commit_id) = @_;
3603 my %co;
3604
3605 local $/ = "\0";
3606
3607 open my $fd, "-|", git_cmd(), "rev-list",
3608 "--parents",
3609 "--header",
3610 "--max-count=1",
3611 $commit_id,
3612 "--",
3613 or die_error(500, "Open git-rev-list failed");
3614 %co = parse_commit_text(<$fd>, 1);
3615 close $fd;
3616
3617 return %co;
3618 }
3619
3620 sub parse_commits {
3621 my ($commit_id, $maxcount, $skip, $filename, @args) = @_;
3622 my @cos;
3623
3624 $maxcount ||= 1;
3625 $skip ||= 0;
3626
3627 local $/ = "\0";
3628
3629 open my $fd, "-|", git_cmd(), "rev-list",
3630 "--header",
3631 @args,
3632 ("--max-count=" . $maxcount),
3633 ("--skip=" . $skip),
3634 @extra_options,
3635 $commit_id,
3636 "--",
3637 ($filename ? ($filename) : ())
3638 or die_error(500, "Open git-rev-list failed");
3639 while (my $line = <$fd>) {
3640 my %co = parse_commit_text($line);
3641 push @cos, \%co;
3642 }
3643 close $fd;
3644
3645 return wantarray ? @cos : \@cos;
3646 }
3647
3648 # parse line of git-diff-tree "raw" output
3649 sub parse_difftree_raw_line {
3650 my $line = shift;
3651 my %res;
3652
3653 # ':100644 100644 03b218260e99b78c6df0ed378e59ed9205ccc96d 3b93d5e7cc7f7dd4ebed13a5cc1a4ad976fc94d8 M ls-files.c'
3654 # ':100644 100644 7f9281985086971d3877aca27704f2aaf9c448ce bc190ebc71bbd923f2b728e505408f5e54bd073a M rev-tree.c'
3655 if ($line =~ m/^:([0-7]{6}) ([0-7]{6}) ($oid_regex) ($oid_regex) (.)([0-9]{0,3})\t(.*)$/) {
3656 $res{'from_mode'} = $1;
3657 $res{'to_mode'} = $2;
3658 $res{'from_id'} = $3;
3659 $res{'to_id'} = $4;
3660 $res{'status'} = $5;
3661 $res{'similarity'} = $6;
3662 if ($res{'status'} eq 'R' || $res{'status'} eq 'C') { # renamed or copied
3663 ($res{'from_file'}, $res{'to_file'}) = map { unquote($_) } split("\t", $7);
3664 } else {
3665 $res{'from_file'} = $res{'to_file'} = $res{'file'} = unquote($7);
3666 }
3667 }
3668 # '::100755 100755 100755 60e79ca1b01bc8b057abe17ddab484699a7f5fdb 94067cc5f73388f33722d52ae02f44692bc07490 94067cc5f73388f33722d52ae02f44692bc07490 MR git-gui/git-gui.sh'
3669 # combined diff (for merge commit)
3670 elsif ($line =~ s/^(::+)((?:[0-7]{6} )+)((?:$oid_regex )+)([a-zA-Z]+)\t(.*)$//) {
3671 $res{'nparents'} = length($1);
3672 $res{'from_mode'} = [ split(' ', $2) ];
3673 $res{'to_mode'} = pop @{$res{'from_mode'}};
3674 $res{'from_id'} = [ split(' ', $3) ];
3675 $res{'to_id'} = pop @{$res{'from_id'}};
3676 $res{'status'} = [ split('', $4) ];
3677 $res{'to_file'} = unquote($5);
3678 }
3679 # 'c512b523472485aef4fff9e57b229d9d243c967f'
3680 elsif ($line =~ m/^($oid_regex)$/) {
3681 $res{'commit'} = $1;
3682 }
3683
3684 return wantarray ? %res : \%res;
3685 }
3686
3687 # wrapper: return parsed line of git-diff-tree "raw" output
3688 # (the argument might be raw line, or parsed info)
3689 sub parsed_difftree_line {
3690 my $line_or_ref = shift;
3691
3692 if (ref($line_or_ref) eq "HASH") {
3693 # pre-parsed (or generated by hand)
3694 return $line_or_ref;
3695 } else {
3696 return parse_difftree_raw_line($line_or_ref);
3697 }
3698 }
3699
3700 # parse line of git-ls-tree output
3701 sub parse_ls_tree_line {
3702 my $line = shift;
3703 my %opts = @_;
3704 my %res;
3705
3706 if ($opts{'-l'}) {
3707 #'100644 blob 0fa3f3a66fb6a137f6ec2c19351ed4d807070ffa 16717 panic.c'
3708 $line =~ m/^([0-9]+) (.+) ($oid_regex) +(-|[0-9]+)\t(.+)$/s;
3709
3710 $res{'mode'} = $1;
3711 $res{'type'} = $2;
3712 $res{'hash'} = $3;
3713 $res{'size'} = $4;
3714 if ($opts{'-z'}) {
3715 $res{'name'} = $5;
3716 } else {
3717 $res{'name'} = unquote($5);
3718 }
3719 } else {
3720 #'100644 blob 0fa3f3a66fb6a137f6ec2c19351ed4d807070ffa panic.c'
3721 $line =~ m/^([0-9]+) (.+) ($oid_regex)\t(.+)$/s;
3722
3723 $res{'mode'} = $1;
3724 $res{'type'} = $2;
3725 $res{'hash'} = $3;
3726 if ($opts{'-z'}) {
3727 $res{'name'} = $4;
3728 } else {
3729 $res{'name'} = unquote($4);
3730 }
3731 }
3732
3733 return wantarray ? %res : \%res;
3734 }
3735
3736 # generates _two_ hashes, references to which are passed as 2 and 3 argument
3737 sub parse_from_to_diffinfo {
3738 my ($diffinfo, $from, $to, @parents) = @_;
3739
3740 if ($diffinfo->{'nparents'}) {
3741 # combined diff
3742 $from->{'file'} = [];
3743 $from->{'href'} = [];
3744 fill_from_file_info($diffinfo, @parents)
3745 unless exists $diffinfo->{'from_file'};
3746 for (my $i = 0; $i < $diffinfo->{'nparents'}; $i++) {
3747 $from->{'file'}[$i] =
3748 defined $diffinfo->{'from_file'}[$i] ?
3749 $diffinfo->{'from_file'}[$i] :
3750 $diffinfo->{'to_file'};
3751 if ($diffinfo->{'status'}[$i] ne "A") { # not new (added) file
3752 $from->{'href'}[$i] = href(action=>"blob",
3753 hash_base=>$parents[$i],
3754 hash=>$diffinfo->{'from_id'}[$i],
3755 file_name=>$from->{'file'}[$i]);
3756 } else {
3757 $from->{'href'}[$i] = undef;
3758 }
3759 }
3760 } else {
3761 # ordinary (not combined) diff
3762 $from->{'file'} = $diffinfo->{'from_file'};
3763 if ($diffinfo->{'status'} ne "A") { # not new (added) file
3764 $from->{'href'} = href(action=>"blob", hash_base=>$hash_parent,
3765 hash=>$diffinfo->{'from_id'},
3766 file_name=>$from->{'file'});
3767 } else {
3768 delete $from->{'href'};
3769 }
3770 }
3771
3772 $to->{'file'} = $diffinfo->{'to_file'};
3773 if (!is_deleted($diffinfo)) { # file exists in result
3774 $to->{'href'} = href(action=>"blob", hash_base=>$hash,
3775 hash=>$diffinfo->{'to_id'},
3776 file_name=>$to->{'file'});
3777 } else {
3778 delete $to->{'href'};
3779 }
3780 }
3781
3782 ## ......................................................................
3783 ## parse to array of hashes functions
3784
3785 sub git_get_heads_list {
3786 my ($limit, @classes) = @_;
3787 @classes = get_branch_refs() unless @classes;
3788 my @patterns = map { "refs/$_" } @classes;
3789 my @headslist;
3790
3791 open my $fd, '-|', git_cmd(), 'for-each-ref',
3792 ($limit ? '--count='.($limit+1) : ()),
3793 '--sort=-HEAD', '--sort=-committerdate',
3794 '--format=%(objectname) %(refname) %(subject)%00%(committer)',
3795 @patterns
3796 or return;
3797 while (my $line = <$fd>) {
3798 my %ref_item;
3799
3800 chomp $line;
3801 my ($refinfo, $committerinfo) = split(/\0/, $line);
3802 my ($hash, $name, $title) = split(' ', $refinfo, 3);
3803 my ($committer, $epoch, $tz) =
3804 ($committerinfo =~ /^(.*) ([0-9]+) (.*)$/);
3805 $ref_item{'fullname'} = $name;
3806 my $strip_refs = join '|', map { quotemeta } get_branch_refs();
3807 $name =~ s!^refs/($strip_refs|remotes)/!!;
3808 $ref_item{'name'} = $name;
3809 # for refs neither in 'heads' nor 'remotes' we want to
3810 # show their ref dir
3811 my $ref_dir = (defined $1) ? $1 : '';
3812 if ($ref_dir ne '' and $ref_dir ne 'heads' and $ref_dir ne 'remotes') {
3813 $ref_item{'name'} .= ' (' . $ref_dir . ')';
3814 }
3815
3816 $ref_item{'id'} = $hash;
3817 $ref_item{'title'} = $title || '(no commit message)';
3818 $ref_item{'epoch'} = $epoch;
3819 if ($epoch) {
3820 $ref_item{'age'} = age_string(time - $ref_item{'epoch'});
3821 } else {
3822 $ref_item{'age'} = "unknown";
3823 }
3824
3825 push @headslist, \%ref_item;
3826 }
3827 close $fd;
3828
3829 return wantarray ? @headslist : \@headslist;
3830 }
3831
3832 sub git_get_tags_list {
3833 my $limit = shift;
3834 my @tagslist;
3835
3836 open my $fd, '-|', git_cmd(), 'for-each-ref',
3837 ($limit ? '--count='.($limit+1) : ()), '--sort=-creatordate',
3838 '--format=%(objectname) %(objecttype) %(refname) '.
3839 '%(*objectname) %(*objecttype) %(subject)%00%(creator)',
3840 'refs/tags'
3841 or return;
3842 while (my $line = <$fd>) {
3843 my %ref_item;
3844
3845 chomp $line;
3846 my ($refinfo, $creatorinfo) = split(/\0/, $line);
3847 my ($id, $type, $name, $refid, $reftype, $title) = split(' ', $refinfo, 6);
3848 my ($creator, $epoch, $tz) =
3849 ($creatorinfo =~ /^(.*) ([0-9]+) (.*)$/);
3850 $ref_item{'fullname'} = $name;
3851 $name =~ s!^refs/tags/!!;
3852
3853 $ref_item{'type'} = $type;
3854 $ref_item{'id'} = $id;
3855 $ref_item{'name'} = $name;
3856 if ($type eq "tag") {
3857 $ref_item{'subject'} = $title;
3858 $ref_item{'reftype'} = $reftype;
3859 $ref_item{'refid'} = $refid;
3860 } else {
3861 $ref_item{'reftype'} = $type;
3862 $ref_item{'refid'} = $id;
3863 }
3864
3865 if ($type eq "tag" || $type eq "commit") {
3866 $ref_item{'epoch'} = $epoch;
3867 if ($epoch) {
3868 $ref_item{'age'} = age_string(time - $ref_item{'epoch'});
3869 } else {
3870 $ref_item{'age'} = "unknown";
3871 }
3872 }
3873
3874 push @tagslist, \%ref_item;
3875 }
3876 close $fd;
3877
3878 return wantarray ? @tagslist : \@tagslist;
3879 }
3880
3881 ## ----------------------------------------------------------------------
3882 ## filesystem-related functions
3883
3884 sub get_file_owner {
3885 my $path = shift;
3886
3887 my ($dev, $ino, $mode, $nlink, $st_uid, $st_gid, $rdev, $size) = stat($path);
3888 my ($name, $passwd, $uid, $gid, $quota, $comment, $gcos, $dir, $shell) = getpwuid($st_uid);
3889 if (!defined $gcos) {
3890 return undef;
3891 }
3892 my $owner = $gcos;
3893 $owner =~ s/[,;].*$//;
3894 return to_utf8($owner);
3895 }
3896
3897 # assume that file exists
3898 sub insert_file {
3899 my $filename = shift;
3900
3901 open my $fd, '<', $filename;
3902 print map { to_utf8($_) } <$fd>;
3903 close $fd;
3904 }
3905
3906 ## ......................................................................
3907 ## mimetype related functions
3908
3909 sub mimetype_guess_file {
3910 my $filename = shift;
3911 my $mimemap = shift;
3912 -r $mimemap or return undef;
3913
3914 my %mimemap;
3915 open(my $mh, '<', $mimemap) or return undef;
3916 while (<$mh>) {
3917 next if m/^#/; # skip comments
3918 my ($mimetype, @exts) = split(/\s+/);
3919 foreach my $ext (@exts) {
3920 $mimemap{$ext} = $mimetype;
3921 }
3922 }
3923 close($mh);
3924
3925 $filename =~ /\.([^.]*)$/;
3926 return $mimemap{$1};
3927 }
3928
3929 sub mimetype_guess {
3930 my $filename = shift;
3931 my $mime;
3932 $filename =~ /\./ or return undef;
3933
3934 if ($mimetypes_file) {
3935 my $file = $mimetypes_file;
3936 if ($file !~ m!^/!) { # if it is relative path
3937 # it is relative to project
3938 $file = "$projectroot/$project/$file";
3939 }
3940 $mime = mimetype_guess_file($filename, $file);
3941 }
3942 $mime ||= mimetype_guess_file($filename, '/etc/mime.types');
3943 return $mime;
3944 }
3945
3946 sub blob_mimetype {
3947 my $fd = shift;
3948 my $filename = shift;
3949
3950 if ($filename) {
3951 my $mime = mimetype_guess($filename);
3952 $mime and return $mime;
3953 }
3954
3955 # just in case
3956 return $default_blob_plain_mimetype unless $fd;
3957
3958 if (-T $fd) {
3959 return 'text/plain';
3960 } elsif (! $filename) {
3961 return 'application/octet-stream';
3962 } elsif ($filename =~ m/\.png$/i) {
3963 return 'image/png';
3964 } elsif ($filename =~ m/\.gif$/i) {
3965 return 'image/gif';
3966 } elsif ($filename =~ m/\.jpe?g$/i) {
3967 return 'image/jpeg';
3968 } else {
3969 return 'application/octet-stream';
3970 }
3971 }
3972
3973 sub blob_contenttype {
3974 my ($fd, $file_name, $type) = @_;
3975
3976 $type ||= blob_mimetype($fd, $file_name);
3977 if ($type eq 'text/plain' && defined $default_text_plain_charset) {
3978 $type .= "; charset=$default_text_plain_charset";
3979 }
3980
3981 return $type;
3982 }
3983
3984 # guess file syntax for syntax highlighting; return undef if no highlighting
3985 # the name of syntax can (in the future) depend on syntax highlighter used
3986 sub guess_file_syntax {
3987 my ($highlight, $file_name) = @_;
3988 return undef unless ($highlight && defined $file_name);
3989 my $basename = basename($file_name, '.in');
3990 return $highlight_basename{$basename}
3991 if exists $highlight_basename{$basename};
3992
3993 $basename =~ /\.([^.]*)$/;
3994 my $ext = $1 or return undef;
3995 return $highlight_ext{$ext}
3996 if exists $highlight_ext{$ext};
3997
3998 return undef;
3999 }
4000
4001 # run highlighter and return FD of its output,
4002 # or return original FD if no highlighting
4003 sub run_highlighter {
4004 my ($fd, $highlight, $syntax) = @_;
4005 return $fd unless ($highlight);
4006
4007 close $fd;
4008 my $syntax_arg = (defined $syntax) ? "--syntax $syntax" : "--force";
4009 open $fd, quote_command(git_cmd(), "cat-file", "blob", $hash)." | ".
4010 quote_command($^X, '-CO', '-MEncode=decode,FB_DEFAULT', '-pse',
4011 '$_ = decode($fe, $_, FB_DEFAULT) if !utf8::decode($_);',
4012 '--', "-fe=$fallback_encoding")." | ".
4013 quote_command($highlight_bin).
4014 " --replace-tabs=8 --fragment $syntax_arg |"
4015 or die_error(500, "Couldn't open file or run syntax highlighter");
4016 return $fd;
4017 }
4018
4019 ## ======================================================================
4020 ## functions printing HTML: header, footer, error page
4021
4022 sub get_page_title {
4023 my $title = to_utf8($site_name);
4024
4025 unless (defined $project) {
4026 if (defined $project_filter) {
4027 $title .= " - projects in '" . esc_path($project_filter) . "'";
4028 }
4029 return $title;
4030 }
4031 $title .= " - " . to_utf8($project);
4032
4033 return $title unless (defined $action);
4034 $title .= "/$action"; # $action is US-ASCII (7bit ASCII)
4035
4036 return $title unless (defined $file_name);
4037 $title .= " - " . esc_path($file_name);
4038 if ($action eq "tree" && $file_name !~ m|/$|) {
4039 $title .= "/";
4040 }
4041
4042 return $title;
4043 }
4044
4045 sub get_content_type_html {
4046 # require explicit support from the UA if we are to send the page as
4047 # 'application/xhtml+xml', otherwise send it as plain old 'text/html'.
4048 # we have to do this because MSIE sometimes globs '*/*', pretending to
4049 # support xhtml+xml but choking when it gets what it asked for.
4050 if (defined $cgi->http('HTTP_ACCEPT') &&
4051 $cgi->http('HTTP_ACCEPT') =~ m/(,|;|\s|^)application\/xhtml\+xml(,|;|\s|$)/ &&
4052 $cgi->Accept('application/xhtml+xml') != 0) {
4053 return 'application/xhtml+xml';
4054 } else {
4055 return 'text/html';
4056 }
4057 }
4058
4059 sub print_feed_meta {
4060 if (defined $project) {
4061 my %href_params = get_feed_info();
4062 if (!exists $href_params{'-title'}) {
4063 $href_params{'-title'} = 'log';
4064 }
4065
4066 foreach my $format (qw(RSS Atom)) {
4067 my $type = lc($format);
4068 my %link_attr = (
4069 '-rel' => 'alternate',
4070 '-title' => esc_attr("$project - $href_params{'-title'} - $format feed"),
4071 '-type' => "application/$type+xml"
4072 );
4073
4074 $href_params{'extra_options'} = undef;
4075 $href_params{'action'} = $type;
4076 $link_attr{'-href'} = esc_attr(href(%href_params));
4077 print "<link ".
4078 "rel=\"$link_attr{'-rel'}\" ".
4079 "title=\"$link_attr{'-title'}\" ".
4080 "href=\"$link_attr{'-href'}\" ".
4081 "type=\"$link_attr{'-type'}\" ".
4082 "/>\n";
4083
4084 $href_params{'extra_options'} = '--no-merges';
4085 $link_attr{'-href'} = esc_attr(href(%href_params));
4086 $link_attr{'-title'} .= ' (no merges)';
4087 print "<link ".
4088 "rel=\"$link_attr{'-rel'}\" ".
4089 "title=\"$link_attr{'-title'}\" ".
4090 "href=\"$link_attr{'-href'}\" ".
4091 "type=\"$link_attr{'-type'}\" ".
4092 "/>\n";
4093 }
4094
4095 } else {
4096 printf('<link rel="alternate" title="%s projects list" '.
4097 'href="%s" type="text/plain; charset=utf-8" />'."\n",
4098 esc_attr($site_name),
4099 esc_attr(href(project=>undef, action=>"project_index")));
4100 printf('<link rel="alternate" title="%s projects feeds" '.
4101 'href="%s" type="text/x-opml" />'."\n",
4102 esc_attr($site_name),
4103 esc_attr(href(project=>undef, action=>"opml")));
4104 }
4105 }
4106
4107 sub print_header_links {
4108 my $status = shift;
4109
4110 # print out each stylesheet that exist, providing backwards capability
4111 # for those people who defined $stylesheet in a config file
4112 if (defined $stylesheet) {
4113 print '<link rel="stylesheet" type="text/css" href="'.esc_url($stylesheet).'"/>'."\n";
4114 } else {
4115 foreach my $stylesheet (@stylesheets) {
4116 next unless $stylesheet;
4117 print '<link rel="stylesheet" type="text/css" href="'.esc_url($stylesheet).'"/>'."\n";
4118 }
4119 }
4120 print_feed_meta()
4121 if ($status eq '200 OK');
4122 if (defined $favicon) {
4123 print qq(<link rel="shortcut icon" href=").esc_url($favicon).qq(" type="image/png" />\n);
4124 }
4125 }
4126
4127 sub print_nav_breadcrumbs_path {
4128 my $dirprefix = undef;
4129 while (my $part = shift) {
4130 $dirprefix .= "/" if defined $dirprefix;
4131 $dirprefix .= $part;
4132 print $cgi->a({-href => href(project => undef,
4133 project_filter => $dirprefix,
4134 action => "project_list")},
4135 esc_html($part)) . " / ";
4136 }
4137 }
4138
4139 sub print_nav_breadcrumbs {
4140 my %opts = @_;
4141
4142 for my $crumb (@extra_breadcrumbs, [ $home_link_str => $home_link ]) {
4143 print $cgi->a({-href => esc_url($crumb->[1])}, $crumb->[0]) . " / ";
4144 }
4145 if (defined $project) {
4146 my @dirname = split '/', $project;
4147 my $projectbasename = pop @dirname;
4148 print_nav_breadcrumbs_path(@dirname);
4149 print $cgi->a({-href => href(action=>"summary")}, esc_html($projectbasename));
4150 if (defined $action) {
4151 my $action_print = $action ;
4152 if (defined $opts{-action_extra}) {
4153 $action_print = $cgi->a({-href => href(action=>$action)},
4154 $action);
4155 }
4156 print " / $action_print";
4157 }
4158 if (defined $opts{-action_extra}) {
4159 print " / $opts{-action_extra}";
4160 }
4161 print "\n";
4162 } elsif (defined $project_filter) {
4163 print_nav_breadcrumbs_path(split '/', $project_filter);
4164 }
4165 }
4166
4167 sub print_search_form {
4168 if (!defined $searchtext) {
4169 $searchtext = "";
4170 }
4171 my $search_hash;
4172 if (defined $hash_base) {
4173 $search_hash = $hash_base;
4174 } elsif (defined $hash) {
4175 $search_hash = $hash;
4176 } else {
4177 $search_hash = "HEAD";
4178 }
4179 my $action = $my_uri;
4180 my $use_pathinfo = gitweb_check_feature('pathinfo');
4181 if ($use_pathinfo) {
4182 $action .= "/".esc_url($project);
4183 }
4184 print $cgi->start_form(-method => "get", -action => $action) .
4185 "<div class=\"search\">\n" .
4186 (!$use_pathinfo &&
4187 $cgi->input({-name=>"p", -value=>$project, -type=>"hidden"}) . "\n") .
4188 $cgi->input({-name=>"a", -value=>"search", -type=>"hidden"}) . "\n" .
4189 $cgi->input({-name=>"h", -value=>$search_hash, -type=>"hidden"}) . "\n" .
4190 $cgi->popup_menu(-name => 'st', -default => 'commit',
4191 -values => ['commit', 'grep', 'author', 'committer', 'pickaxe']) .
4192 " " . $cgi->a({-href => href(action=>"search_help"),
4193 -title => "search help" }, "?") . " search:\n",
4194 $cgi->textfield(-name => "s", -value => $searchtext, -override => 1) . "\n" .
4195 "<span title=\"Extended regular expression\">" .
4196 $cgi->checkbox(-name => 'sr', -value => 1, -label => 're',
4197 -checked => $search_use_regexp) .
4198 "</span>" .
4199 "</div>" .
4200 $cgi->end_form() . "\n";
4201 }
4202
4203 sub git_header_html {
4204 my $status = shift || "200 OK";
4205 my $expires = shift;
4206 my %opts = @_;
4207
4208 my $title = get_page_title();
4209 print $cgi->header(-type=>get_content_type_html(), -charset => 'utf-8',
4210 -status=> $status, -expires => $expires)
4211 unless ($opts{'-no_http_header'});
4212 my $mod_perl_version = $ENV{'MOD_PERL'} ? " $ENV{'MOD_PERL'}" : '';
4213 print <<EOF;
4214 <?xml version="1.0" encoding="utf-8"?>
4215 <!DOCTYPE html [
4216 <!ENTITY nbsp "&#xA0;">
4217 <!ENTITY sdot "&#x22C5;">
4218 ]>
4219 <html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en-US" lang="en-US">
4220 <!-- git web interface version $version, (C) 2005-2006, Kay Sievers <kay.sievers\@vrfy.org>, Christian Gierke -->
4221 <!-- git core binaries version $git_version -->
4222 <head>
4223 <meta name="generator" content="gitweb/$version git/$git_version$mod_perl_version"/>
4224 <meta name="robots" content="index, nofollow"/>
4225 <meta name="viewport" content="width=device-width, initial-scale=1"/>
4226 <title>$title</title>
4227 EOF
4228 # the stylesheet, favicon etc urls won't work correctly with path_info
4229 # unless we set the appropriate base URL
4230 if ($ENV{'PATH_INFO'}) {
4231 print "<base href=\"".esc_url($base_url)."\" />\n";
4232 }
4233 print_header_links($status);
4234
4235 if (defined $site_html_head_string) {
4236 print to_utf8($site_html_head_string);
4237 }
4238
4239 print "</head>\n" .
4240 "<body>\n";
4241
4242 if (defined $site_header && -f $site_header) {
4243 insert_file($site_header);
4244 }
4245
4246 print "<div class=\"page_header\">\n";
4247 if (defined $logo) {
4248 print $cgi->a({-href => esc_url($logo_url),
4249 -title => $logo_label},
4250 $cgi->img({-src => esc_url($logo),
4251 -width => 72, -height => 27,
4252 -alt => "git",
4253 -class => "logo"}));
4254 }
4255 print_nav_breadcrumbs(%opts);
4256 print "</div>\n";
4257
4258 my $have_search = gitweb_check_feature('search');
4259 if (defined $project && $have_search) {
4260 print_search_form();
4261 }
4262 }
4263
4264 sub git_footer_html {
4265 my $feed_class = 'rss_logo';
4266
4267 print "<div class=\"page_footer\">\n";
4268 if (defined $project) {
4269 my $descr = git_get_project_description($project);
4270 if (defined $descr) {
4271 print "<div class=\"page_footer_text\">" . esc_html($descr) . "</div>\n";
4272 }
4273
4274 my %href_params = get_feed_info();
4275 if (!%href_params) {
4276 $feed_class .= ' generic';
4277 }
4278 $href_params{'-title'} ||= 'log';
4279
4280 foreach my $format (qw(RSS Atom)) {
4281 $href_params{'action'} = lc($format);
4282 print $cgi->a({-href => href(%href_params),
4283 -title => "$href_params{'-title'} $format feed",
4284 -class => $feed_class}, $format)."\n";
4285 }
4286
4287 } else {
4288 print $cgi->a({-href => href(project=>undef, action=>"opml",
4289 project_filter => $project_filter),
4290 -class => $feed_class}, "OPML") . " ";
4291 print $cgi->a({-href => href(project=>undef, action=>"project_index",
4292 project_filter => $project_filter),
4293 -class => $feed_class}, "TXT") . "\n";
4294 }
4295 print "</div>\n"; # class="page_footer"
4296
4297 if (defined $t0 && gitweb_check_feature('timed')) {
4298 print "<div id=\"generating_info\">\n";
4299 print 'This page took '.
4300 '<span id="generating_time" class="time_span">'.
4301 tv_interval($t0, [ gettimeofday() ]).
4302 ' seconds </span>'.
4303 ' and '.
4304 '<span id="generating_cmd">'.
4305 $number_of_git_cmds.
4306 '</span> git commands '.
4307 " to generate.\n";
4308 print "</div>\n"; # class="page_footer"
4309 }
4310
4311 if (defined $site_footer && -f $site_footer) {
4312 insert_file($site_footer);
4313 }
4314
4315 print qq!<script type="text/javascript" src="!.esc_url($javascript).qq!"></script>\n!;
4316 if (defined $action &&
4317 $action eq 'blame_incremental') {
4318 print qq!<script type="text/javascript">\n!.
4319 qq!startBlame("!. esc_attr(href(action=>"blame_data", -replay=>1)) .qq!",\n!.
4320 qq! "!. esc_attr(href()) .qq!");\n!.
4321 qq!</script>\n!;
4322 } else {
4323 my ($jstimezone, $tz_cookie, $datetime_class) =
4324 gitweb_get_feature('javascript-timezone');
4325
4326 print qq!<script type="text/javascript">\n!.
4327 qq!window.onload = function () {\n!;
4328 if (gitweb_check_feature('javascript-actions')) {
4329 print qq! fixLinks();\n!;
4330 }
4331 if ($jstimezone && $tz_cookie && $datetime_class) {
4332 print qq! var tz_cookie = { name: '$tz_cookie', expires: 14, path: '/' };\n!. # in days
4333 qq! onloadTZSetup('$jstimezone', tz_cookie, '$datetime_class');\n!;
4334 }
4335 print qq!};\n!.
4336 qq!</script>\n!;
4337 }
4338
4339 print "</body>\n" .
4340 "</html>";
4341 }
4342
4343 # die_error(<http_status_code>, <error_message>[, <detailed_html_description>])
4344 # Example: die_error(404, 'Hash not found')
4345 # By convention, use the following status codes (as defined in RFC 2616):
4346 # 400: Invalid or missing CGI parameters, or
4347 # requested object exists but has wrong type.
4348 # 403: Requested feature (like "pickaxe" or "snapshot") not enabled on
4349 # this server or project.
4350 # 404: Requested object/revision/project doesn't exist.
4351 # 500: The server isn't configured properly, or
4352 # an internal error occurred (e.g. failed assertions caused by bugs), or
4353 # an unknown error occurred (e.g. the git binary died unexpectedly).
4354 # 503: The server is currently unavailable (because it is overloaded,
4355 # or down for maintenance). Generally, this is a temporary state.
4356 sub die_error {
4357 my $status = shift || 500;
4358 my $error = esc_html(shift) || "Internal Server Error";
4359 my $extra = shift;
4360 my %opts = @_;
4361
4362 my %http_responses = (
4363 400 => '400 Bad Request',
4364 403 => '403 Forbidden',
4365 404 => '404 Not Found',
4366 500 => '500 Internal Server Error',
4367 503 => '503 Service Unavailable',
4368 );
4369 git_header_html($http_responses{$status}, undef, %opts);
4370 print <<EOF;
4371 <div class="page_body">
4372 <br /><br />
4373 $status - $error
4374 <br />
4375 EOF
4376 if (defined $extra) {
4377 print "<hr />\n" .
4378 "$extra\n";
4379 }
4380 print "</div>\n";
4381
4382 git_footer_html();
4383 goto DONE_GITWEB
4384 unless ($opts{'-error_handler'});
4385 }
4386
4387 ## ----------------------------------------------------------------------
4388 ## functions printing or outputting HTML: navigation
4389
4390 sub git_print_page_nav {
4391 my ($current, $suppress, $head, $treehead, $treebase, $extra) = @_;
4392 $extra = '' if !defined $extra; # pager or formats
4393
4394 my @navs = qw(summary shortlog log commit commitdiff tree);
4395 if ($suppress) {
4396 @navs = grep { $_ ne $suppress } @navs;
4397 }
4398
4399 my %arg = map { $_ => {action=>$_} } @navs;
4400 if (defined $head) {
4401 for (qw(commit commitdiff)) {
4402 $arg{$_}{'hash'} = $head;
4403 }
4404 if ($current =~ m/^(tree | log | shortlog | commit | commitdiff | search)$/x) {
4405 for (qw(shortlog log)) {
4406 $arg{$_}{'hash'} = $head;
4407 }
4408 }
4409 }
4410
4411 $arg{'tree'}{'hash'} = $treehead if defined $treehead;
4412 $arg{'tree'}{'hash_base'} = $treebase if defined $treebase;
4413
4414 my @actions = gitweb_get_feature('actions');
4415 my %repl = (
4416 '%' => '%',
4417 'n' => $project, # project name
4418 'f' => $git_dir, # project path within filesystem
4419 'h' => $treehead || '', # current hash ('h' parameter)
4420 'b' => $treebase || '', # hash base ('hb' parameter)
4421 );
4422 while (@actions) {
4423 my ($label, $link, $pos) = splice(@actions,0,3);
4424 # insert
4425 @navs = map { $_ eq $pos ? ($_, $label) : $_ } @navs;
4426 # munch munch
4427 $link =~ s/%([%nfhb])/$repl{$1}/g;
4428 $arg{$label}{'_href'} = $link;
4429 }
4430
4431 print "<div class=\"page_nav\">\n" .
4432 (join " | ",
4433 map { $_ eq $current ?
4434 $_ : $cgi->a({-href => ($arg{$_}{_href} ? $arg{$_}{_href} : href(%{$arg{$_}}))}, "$_")
4435 } @navs);
4436 print "<br/>\n$extra<br/>\n" .
4437 "</div>\n";
4438 }
4439
4440 # returns a submenu for the navigation of the refs views (tags, heads,
4441 # remotes) with the current view disabled and the remotes view only
4442 # available if the feature is enabled
4443 sub format_ref_views {
4444 my ($current) = @_;
4445 my @ref_views = qw{tags heads};
4446 push @ref_views, 'remotes' if gitweb_check_feature('remote_heads');
4447 return join " | ", map {
4448 $_ eq $current ? $_ :
4449 $cgi->a({-href => href(action=>$_)}, $_)
4450 } @ref_views
4451 }
4452
4453 sub format_paging_nav {
4454 my ($action, $page, $has_next_link) = @_;
4455 my $paging_nav;
4456
4457
4458 if ($page > 0) {
4459 $paging_nav .=
4460 $cgi->a({-href => href(-replay=>1, page=>undef)}, "first") .
4461 " &sdot; " .
4462 $cgi->a({-href => href(-replay=>1, page=>$page-1),
4463 -accesskey => "p", -title => "Alt-p"}, "prev");
4464 } else {
4465 $paging_nav .= "first &sdot; prev";
4466 }
4467
4468 if ($has_next_link) {
4469 $paging_nav .= " &sdot; " .
4470 $cgi->a({-href => href(-replay=>1, page=>$page+1),
4471 -accesskey => "n", -title => "Alt-n"}, "next");
4472 } else {
4473 $paging_nav .= " &sdot; next";
4474 }
4475
4476 return $paging_nav;
4477 }
4478
4479 ## ......................................................................
4480 ## functions printing or outputting HTML: div
4481
4482 sub git_print_header_div {
4483 my ($action, $title, $hash, $hash_base) = @_;
4484 my %args = ();
4485
4486 $args{'action'} = $action;
4487 $args{'hash'} = $hash if $hash;
4488 $args{'hash_base'} = $hash_base if $hash_base;
4489
4490 print "<div class=\"header\">\n" .
4491 $cgi->a({-href => href(%args), -class => "title"},
4492 $title ? $title : $action) .
4493 "\n</div>\n";
4494 }
4495
4496 sub format_repo_url {
4497 my ($name, $url) = @_;
4498 return "<tr class=\"metadata_url\"><td>$name</td><td>$url</td></tr>\n";
4499 }
4500
4501 # Group output by placing it in a DIV element and adding a header.
4502 # Options for start_div() can be provided by passing a hash reference as the
4503 # first parameter to the function.
4504 # Options to git_print_header_div() can be provided by passing an array
4505 # reference. This must follow the options to start_div if they are present.
4506 # The content can be a scalar, which is output as-is, a scalar reference, which
4507 # is output after html escaping, an IO handle passed either as *handle or
4508 # *handle{IO}, or a function reference. In the latter case all following
4509 # parameters will be taken as argument to the content function call.
4510 sub git_print_section {
4511 my ($div_args, $header_args, $content);
4512 my $arg = shift;
4513 if (ref($arg) eq 'HASH') {
4514 $div_args = $arg;
4515 $arg = shift;
4516 }
4517 if (ref($arg) eq 'ARRAY') {
4518 $header_args = $arg;
4519 $arg = shift;
4520 }
4521 $content = $arg;
4522
4523 print $cgi->start_div($div_args);
4524 git_print_header_div(@$header_args);
4525
4526 if (ref($content) eq 'CODE') {
4527 $content->(@_);
4528 } elsif (ref($content) eq 'SCALAR') {
4529 print esc_html($$content);
4530 } elsif (ref($content) eq 'GLOB' or ref($content) eq 'IO::Handle') {
4531 print <$content>;
4532 } elsif (!ref($content) && defined($content)) {
4533 print $content;
4534 }
4535
4536 print $cgi->end_div;
4537 }
4538
4539 sub format_timestamp_html {
4540 my $date = shift;
4541 my $strtime = $date->{'rfc2822'};
4542
4543 my (undef, undef, $datetime_class) =
4544 gitweb_get_feature('javascript-timezone');
4545 if ($datetime_class) {
4546 $strtime = qq!<span class="$datetime_class">$strtime</span>!;
4547 }
4548
4549 my $localtime_format = '(%02d:%02d %s)';
4550 if ($date->{'hour_local'} < 6) {
4551 $localtime_format = '(<span class="atnight">%02d:%02d</span> %s)';
4552 }
4553 $strtime .= ' ' .
4554 sprintf($localtime_format,
4555 $date->{'hour_local'}, $date->{'minute_local'}, $date->{'tz_local'});
4556
4557 return $strtime;
4558 }
4559
4560 # Outputs the author name and date in long form
4561 sub git_print_authorship {
4562 my $co = shift;
4563 my %opts = @_;
4564 my $tag = $opts{-tag} || 'div';
4565 my $author = $co->{'author_name'};
4566
4567 my %ad = parse_date($co->{'author_epoch'}, $co->{'author_tz'});
4568 print "<$tag class=\"author_date\">" .
4569 format_search_author($author, "author", esc_html($author)) .
4570 " [".format_timestamp_html(\%ad)."]".
4571 git_get_avatar($co->{'author_email'}, -pad_before => 1) .
4572 "</$tag>\n";
4573 }
4574
4575 # Outputs table rows containing the full author or committer information,
4576 # in the format expected for 'commit' view (& similar).
4577 # Parameters are a commit hash reference, followed by the list of people
4578 # to output information for. If the list is empty it defaults to both
4579 # author and committer.
4580 sub git_print_authorship_rows {
4581 my $co = shift;
4582 # too bad we can't use @people = @_ || ('author', 'committer')
4583 my @people = @_;
4584 @people = ('author', 'committer') unless @people;
4585 foreach my $who (@people) {
4586 my %wd = parse_date($co->{"${who}_epoch"}, $co->{"${who}_tz"});
4587 print "<tr><td>$who</td><td>" .
4588 format_search_author($co->{"${who}_name"}, $who,
4589 esc_html($co->{"${who}_name"})) . " " .
4590 format_search_author($co->{"${who}_email"}, $who,
4591 esc_html("<" . $co->{"${who}_email"} . ">")) .
4592 "</td><td rowspan=\"2\">" .
4593 git_get_avatar($co->{"${who}_email"}, -size => 'double') .
4594 "</td></tr>\n" .
4595 "<tr>" .
4596 "<td></td><td>" .
4597 format_timestamp_html(\%wd) .
4598 "</td>" .
4599 "</tr>\n";
4600 }
4601 }
4602
4603 sub git_print_page_path {
4604 my $name = shift;
4605 my $type = shift;
4606 my $hb = shift;
4607
4608
4609 print "<div class=\"page_path\">";
4610 print $cgi->a({-href => href(action=>"tree", hash_base=>$hb),
4611 -title => 'tree root'}, to_utf8("[$project]"));
4612 print " / ";
4613 if (defined $name) {
4614 my @dirname = split '/', $name;
4615 my $basename = pop @dirname;
4616 my $fullname = '';
4617
4618 foreach my $dir (@dirname) {
4619 $fullname .= ($fullname ? '/' : '') . $dir;
4620 print $cgi->a({-href => href(action=>"tree", file_name=>$fullname,
4621 hash_base=>$hb),
4622 -title => $fullname}, esc_path($dir));
4623 print " / ";
4624 }
4625 if (defined $type && $type eq 'blob') {
4626 print $cgi->a({-href => href(action=>"blob_plain", file_name=>$file_name,
4627 hash_base=>$hb),
4628 -title => $name}, esc_path($basename));
4629 } elsif (defined $type && $type eq 'tree') {
4630 print $cgi->a({-href => href(action=>"tree", file_name=>$file_name,
4631 hash_base=>$hb),
4632 -title => $name}, esc_path($basename));
4633 print " / ";
4634 } else {
4635 print esc_path($basename);
4636 }
4637 }
4638 print "<br/></div>\n";
4639 }
4640
4641 sub git_print_log {
4642 my $log = shift;
4643 my %opts = @_;
4644
4645 if ($opts{'-remove_title'}) {
4646 # remove title, i.e. first line of log
4647 shift @$log;
4648 }
4649 # remove leading empty lines
4650 while (defined $log->[0] && $log->[0] eq "") {
4651 shift @$log;
4652 }
4653
4654 # print log
4655 my $skip_blank_line = 0;
4656 foreach my $line (@$log) {
4657 if ($line =~ m/^\s*([A-Z][-A-Za-z]*-([Bb]y|[Tt]o)|C[Cc]|(Clos|Fix)es): /) {
4658 if (! $opts{'-remove_signoff'}) {
4659 print "<span class=\"signoff\">" . esc_html($line) . "</span><br/>\n";
4660 $skip_blank_line = 1;
4661 }
4662 next;
4663 }
4664
4665 if ($line =~ m,\s*([a-z]*link): (https?://\S+),i) {
4666 if (! $opts{'-remove_signoff'}) {
4667 print "<span class=\"signoff\">" . esc_html($1) . ": " .
4668 "<a href=\"" . esc_html($2) . "\">" . esc_html($2) . "</a>" .
4669 "</span><br/>\n";
4670 $skip_blank_line = 1;
4671 }
4672 next;
4673 }
4674
4675 # print only one empty line
4676 # do not print empty line after signoff
4677 if ($line eq "") {
4678 next if ($skip_blank_line);
4679 $skip_blank_line = 1;
4680 } else {
4681 $skip_blank_line = 0;
4682 }
4683
4684 print format_log_line_html($line) . "<br/>\n";
4685 }
4686
4687 if ($opts{'-final_empty_line'}) {
4688 # end with single empty line
4689 print "<br/>\n" unless $skip_blank_line;
4690 }
4691 }
4692
4693 # return link target (what link points to)
4694 sub git_get_link_target {
4695 my $hash = shift;
4696 my $link_target;
4697
4698 # read link
4699 open my $fd, "-|", git_cmd(), "cat-file", "blob", $hash
4700 or return;
4701 {
4702 local $/ = undef;
4703 $link_target = <$fd>;
4704 }
4705 close $fd
4706 or return;
4707
4708 return $link_target;
4709 }
4710
4711 # given link target, and the directory (basedir) the link is in,
4712 # return target of link relative to top directory (top tree);
4713 # return undef if it is not possible (including absolute links).
4714 sub normalize_link_target {
4715 my ($link_target, $basedir) = @_;
4716
4717 # absolute symlinks (beginning with '/') cannot be normalized
4718 return if (substr($link_target, 0, 1) eq '/');
4719
4720 # normalize link target to path from top (root) tree (dir)
4721 my $path;
4722 if ($basedir) {
4723 $path = $basedir . '/' . $link_target;
4724 } else {
4725 # we are in top (root) tree (dir)
4726 $path = $link_target;
4727 }
4728
4729 # remove //, /./, and /../
4730 my @path_parts;
4731 foreach my $part (split('/', $path)) {
4732 # discard '.' and ''
4733 next if (!$part || $part eq '.');
4734 # handle '..'
4735 if ($part eq '..') {
4736 if (@path_parts) {
4737 pop @path_parts;
4738 } else {
4739 # link leads outside repository (outside top dir)
4740 return;
4741 }
4742 } else {
4743 push @path_parts, $part;
4744 }
4745 }
4746 $path = join('/', @path_parts);
4747
4748 return $path;
4749 }
4750
4751 # print tree entry (row of git_tree), but without encompassing <tr> element
4752 sub git_print_tree_entry {
4753 my ($t, $basedir, $hash_base, $have_blame) = @_;
4754
4755 my %base_key = ();
4756 $base_key{'hash_base'} = $hash_base if defined $hash_base;
4757
4758 # The format of a table row is: mode list link. Where mode is
4759 # the mode of the entry, list is the name of the entry, an href,
4760 # and link is the action links of the entry.
4761
4762 print "<td class=\"mode\">" . mode_str($t->{'mode'}) . "</td>\n";
4763 if (exists $t->{'size'}) {
4764 print "<td class=\"size\">$t->{'size'}</td>\n";
4765 }
4766 if ($t->{'type'} eq "blob") {
4767 print "<td class=\"list\">" .
4768 $cgi->a({-href => href(action=>"blob", hash=>$t->{'hash'},
4769 file_name=>"$basedir$t->{'name'}", %base_key),
4770 -class => "list"}, esc_path($t->{'name'}));
4771 if (S_ISLNK(oct $t->{'mode'})) {
4772 my $link_target = git_get_link_target($t->{'hash'});
4773 if ($link_target) {
4774 my $norm_target = normalize_link_target($link_target, $basedir);
4775 if (defined $norm_target) {
4776 print " -> " .
4777 $cgi->a({-href => href(action=>"object", hash_base=>$hash_base,
4778 file_name=>$norm_target),
4779 -title => $norm_target}, esc_path($link_target));
4780 } else {
4781 print " -> " . esc_path($link_target);
4782 }
4783 }
4784 }
4785 print "</td>\n";
4786 print "<td class=\"link\">";
4787 print $cgi->a({-href => href(action=>"blob", hash=>$t->{'hash'},
4788 file_name=>"$basedir$t->{'name'}", %base_key)},
4789 "blob");
4790 if ($have_blame) {
4791 print " | " .
4792 $cgi->a({-href => href(action=>"blame", hash=>$t->{'hash'},
4793 file_name=>"$basedir$t->{'name'}", %base_key)},
4794 "blame");
4795 }
4796 if (defined $hash_base) {
4797 print " | " .
4798 $cgi->a({-href => href(action=>"history", hash_base=>$hash_base,
4799 hash=>$t->{'hash'}, file_name=>"$basedir$t->{'name'}")},
4800 "history");
4801 }
4802 print " | " .
4803 $cgi->a({-href => href(action=>"blob_plain", hash_base=>$hash_base,
4804 file_name=>"$basedir$t->{'name'}")},
4805 "raw");
4806 print "</td>\n";
4807
4808 } elsif ($t->{'type'} eq "tree") {
4809 print "<td class=\"list\">";
4810 print $cgi->a({-href => href(action=>"tree", hash=>$t->{'hash'},
4811 file_name=>"$basedir$t->{'name'}",
4812 %base_key)},
4813 esc_path($t->{'name'}));
4814 print "</td>\n";
4815 print "<td class=\"link\">";
4816 print $cgi->a({-href => href(action=>"tree", hash=>$t->{'hash'},
4817 file_name=>"$basedir$t->{'name'}",
4818 %base_key)},
4819 "tree");
4820 if (defined $hash_base) {
4821 print " | " .
4822 $cgi->a({-href => href(action=>"history", hash_base=>$hash_base,
4823 file_name=>"$basedir$t->{'name'}")},
4824 "history");
4825 }
4826 print "</td>\n";
4827 } else {
4828 # unknown object: we can only present history for it
4829 # (this includes 'commit' object, i.e. submodule support)
4830 print "<td class=\"list\">" .
4831 esc_path($t->{'name'}) .
4832 "</td>\n";
4833 print "<td class=\"link\">";
4834 if (defined $hash_base) {
4835 print $cgi->a({-href => href(action=>"history",
4836 hash_base=>$hash_base,
4837 file_name=>"$basedir$t->{'name'}")},
4838 "history");
4839 }
4840 print "</td>\n";
4841 }
4842 }
4843
4844 ## ......................................................................
4845 ## functions printing large fragments of HTML
4846
4847 # get pre-image filenames for merge (combined) diff
4848 sub fill_from_file_info {
4849 my ($diff, @parents) = @_;
4850
4851 $diff->{'from_file'} = [ ];
4852 $diff->{'from_file'}[$diff->{'nparents'} - 1] = undef;
4853 for (my $i = 0; $i < $diff->{'nparents'}; $i++) {
4854 if ($diff->{'status'}[$i] eq 'R' ||
4855 $diff->{'status'}[$i] eq 'C') {
4856 $diff->{'from_file'}[$i] =
4857 git_get_path_by_hash($parents[$i], $diff->{'from_id'}[$i]);
4858 }
4859 }
4860
4861 return $diff;
4862 }
4863
4864 # is current raw difftree line of file deletion
4865 sub is_deleted {
4866 my $diffinfo = shift;
4867
4868 return $diffinfo->{'to_id'} eq ('0' x 40) || $diffinfo->{'to_id'} eq ('0' x 64);
4869 }
4870
4871 # does patch correspond to [previous] difftree raw line
4872 # $diffinfo - hashref of parsed raw diff format
4873 # $patchinfo - hashref of parsed patch diff format
4874 # (the same keys as in $diffinfo)
4875 sub is_patch_split {
4876 my ($diffinfo, $patchinfo) = @_;
4877
4878 return defined $diffinfo && defined $patchinfo
4879 && $diffinfo->{'to_file'} eq $patchinfo->{'to_file'};
4880 }
4881
4882
4883 sub git_difftree_body {
4884 my ($difftree, $hash, @parents) = @_;
4885 my ($parent) = $parents[0];
4886 my $have_blame = gitweb_check_feature('blame');
4887 print "<div class=\"list_head\">\n";
4888 if ($#{$difftree} > 10) {
4889 print(($#{$difftree} + 1) . " files changed:\n");
4890 }
4891 print "</div>\n";
4892
4893 print "<table class=\"" .
4894 (@parents > 1 ? "combined " : "") .
4895 "diff_tree\">\n";
4896
4897 # header only for combined diff in 'commitdiff' view
4898 my $has_header = @$difftree && @parents > 1 && $action eq 'commitdiff';
4899 if ($has_header) {
4900 # table header
4901 print "<thead><tr>\n" .
4902 "<th></th><th></th>\n"; # filename, patchN link
4903 for (my $i = 0; $i < @parents; $i++) {
4904 my $par = $parents[$i];
4905 print "<th>" .
4906 $cgi->a({-href => href(action=>"commitdiff",
4907 hash=>$hash, hash_parent=>$par),
4908 -title => 'commitdiff to parent number ' .
4909 ($i+1) . ': ' . substr($par,0,7)},
4910 $i+1) .
4911 "&nbsp;</th>\n";
4912 }
4913 print "</tr></thead>\n<tbody>\n";
4914 }
4915
4916 my $alternate = 1;
4917 my $patchno = 0;
4918 foreach my $line (@{$difftree}) {
4919 my $diff = parsed_difftree_line($line);
4920
4921 if ($alternate) {
4922 print "<tr class=\"dark\">\n";
4923 } else {
4924 print "<tr class=\"light\">\n";
4925 }
4926 $alternate ^= 1;
4927
4928 if (exists $diff->{'nparents'}) { # combined diff
4929
4930 fill_from_file_info($diff, @parents)
4931 unless exists $diff->{'from_file'};
4932
4933 if (!is_deleted($diff)) {
4934 # file exists in the result (child) commit
4935 print "<td>" .
4936 $cgi->a({-href => href(action=>"blob", hash=>$diff->{'to_id'},
4937 file_name=>$diff->{'to_file'},
4938 hash_base=>$hash),
4939 -class => "list"}, esc_path($diff->{'to_file'})) .
4940 "</td>\n";
4941 } else {
4942 print "<td>" .
4943 esc_path($diff->{'to_file'}) .
4944 "</td>\n";
4945 }
4946
4947 if ($action eq 'commitdiff') {
4948 # link to patch
4949 $patchno++;
4950 print "<td class=\"link\">" .
4951 $cgi->a({-href => href(-anchor=>"patch$patchno")},
4952 "patch") .
4953 " | " .
4954 "</td>\n";
4955 }
4956
4957 my $has_history = 0;
4958 my $not_deleted = 0;
4959 for (my $i = 0; $i < $diff->{'nparents'}; $i++) {
4960 my $hash_parent = $parents[$i];
4961 my $from_hash = $diff->{'from_id'}[$i];
4962 my $from_path = $diff->{'from_file'}[$i];
4963 my $status = $diff->{'status'}[$i];
4964
4965 $has_history ||= ($status ne 'A');
4966 $not_deleted ||= ($status ne 'D');
4967
4968 if ($status eq 'A') {
4969 print "<td class=\"link\" align=\"right\"> | </td>\n";
4970 } elsif ($status eq 'D') {
4971 print "<td class=\"link\">" .
4972 $cgi->a({-href => href(action=>"blob",
4973 hash_base=>$hash,
4974 hash=>$from_hash,
4975 file_name=>$from_path)},
4976 "blob" . ($i+1)) .
4977 " | </td>\n";
4978 } else {
4979 if ($diff->{'to_id'} eq $from_hash) {
4980 print "<td class=\"link nochange\">";
4981 } else {
4982 print "<td class=\"link\">";
4983 }
4984 print $cgi->a({-href => href(action=>"blobdiff",
4985 hash=>$diff->{'to_id'},
4986 hash_parent=>$from_hash,
4987 hash_base=>$hash,
4988 hash_parent_base=>$hash_parent,
4989 file_name=>$diff->{'to_file'},
4990 file_parent=>$from_path)},
4991 "diff" . ($i+1)) .
4992 " | </td>\n";
4993 }
4994 }
4995
4996 print "<td class=\"link\">";
4997 if ($not_deleted) {
4998 print $cgi->a({-href => href(action=>"blob",
4999 hash=>$diff->{'to_id'},
5000 file_name=>$diff->{'to_file'},
Showing first 5,000 of 8,499 lines. View raw