4 # Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
5 # Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
6 # Claire Fousse <claire.fousse@ensimag.imag.fr>
7 # David Amouyal <david.amouyal@ensimag.imag.fr>
8 # Matthieu Moy <matthieu.moy@grenoble-inp.fr>
9 # License: GPL v2 or later
11 # Gateway between Git and MediaWiki.
12 # https://github.com/Bibzball/Git-Mediawiki/wiki
16 # - Poor performance in the best case: it takes forever to check
17 # whether we're up-to-date (on fetch or push) or to fetch a few
18 # revisions from a large wiki, because we use exclusively a
19 # page-based synchronization. We could switch to a wiki-wide
20 # synchronization when the synchronization involves few revisions
21 # but the wiki is large.
23 # - Git renames could be turned into MediaWiki renames (see TODO
26 # - login/password support requires the user to write the password
27 # cleartext in a file (see TODO below).
29 # - No way to import "one page, and all pages included in it"
31 # - Multiple remote MediaWikis have not been very well tested.
35 use DateTime
::Format
::ISO8601
;
37 # By default, use UTF-8 to communicate with Git and the user
38 binmode STDERR
, ":utf8";
39 binmode STDOUT
, ":utf8";
46 # Mediawiki filenames can contain forward slashes. This variable decides by which pattern they should be replaced
47 use constant SLASH_REPLACEMENT
=> "%2F";
49 # It's not always possible to delete pages (may require some
50 # priviledges). Deleted pages are replaced with this content.
51 use constant DELETED_CONTENT
=> "[[Category:Deleted]]\n";
53 # It's not possible to create empty pages. New empty files in Git are
54 # sent with this content instead.
55 use constant EMPTY_CONTENT
=> "<!-- empty page -->\n";
57 # used to reflect file creation or deletion in diff.
58 use constant NULL_SHA1
=> "0000000000000000000000000000000000000000";
60 my $remotename = $ARGV[0];
63 # Accept both space-separated and multiple keys in config file.
64 # Spaces should be written as _ anyway because we'll use chomp.
65 my @tracked_pages = split(/[ \n]/, run_git
("config --get-all remote.". $remotename .".pages"));
66 chomp(@tracked_pages);
68 # Just like @tracked_pages, but for MediaWiki categories.
69 my @tracked_categories = split(/[ \n]/, run_git
("config --get-all remote.". $remotename .".categories"));
70 chomp(@tracked_categories);
72 # Import media files too.
73 my $import_media = run_git
("config --get --bool remote.". $remotename .".mediaimport");
75 $import_media = ($import_media eq "true");
77 my $wiki_login = run_git
("config --get remote.". $remotename .".mwLogin");
78 # TODO: ideally, this should be able to read from keyboard, but we're
79 # inside a remote helper, so our stdin is connect to git, not to a
81 my $wiki_passwd = run_git
("config --get remote.". $remotename .".mwPassword");
82 my $wiki_domain = run_git
("config --get remote.". $remotename .".mwDomain");
87 # Import only last revisions (both for clone and fetch)
88 my $shallow_import = run_git
("config --get --bool remote.". $remotename .".shallow");
89 chomp($shallow_import);
90 $shallow_import = ($shallow_import eq "true");
92 # Dumb push: don't update notes and mediawiki ref to reflect the last push.
94 # Configurable with mediawiki.dumbPush, or per-remote with
95 # remote.<remotename>.dumbPush.
97 # This means the user will have to re-import the just-pushed
98 # revisions. On the other hand, this means that the Git revisions
99 # corresponding to MediaWiki revisions are all imported from the wiki,
100 # regardless of whether they were initially created in Git or from the
101 # web interface, hence all users will get the same history (i.e. if
102 # the push from Git to MediaWiki loses some information, everybody
103 # will get the history with information lost). If the import is
104 # deterministic, this means everybody gets the same sha1 for each
105 # MediaWiki revision.
106 my $dumb_push = run_git
("config --get --bool remote.$remotename.dumbPush");
107 unless ($dumb_push) {
108 $dumb_push = run_git
("config --get --bool mediawiki.dumbPush");
111 $dumb_push = ($dumb_push eq "true");
113 my $wiki_name = $url;
114 $wiki_name =~ s/[^\/]*:\/\
///;
115 # If URL is like http://user:password@example.com/, we clearly don't
116 # want the password in $wiki_name. While we're there, also remove user
117 # and '@' sign, to avoid author like MWUser@HTTPUser@host.com
118 $wiki_name =~ s/^.*@//;
126 if (defined($cmd[0])) {
128 if ($cmd[0] eq "capabilities") {
129 die("Too many arguments for capabilities") unless (!defined($cmd[1]));
131 } elsif ($cmd[0] eq "list") {
132 die("Too many arguments for list") unless (!defined($cmd[2]));
134 } elsif ($cmd[0] eq "import") {
135 die("Invalid arguments for import") unless ($cmd[1] ne "" && !defined($cmd[2]));
137 } elsif ($cmd[0] eq "option") {
138 die("Too many arguments for option") unless ($cmd[1] ne "" && $cmd[2] ne "" && !defined($cmd[3]));
139 mw_option
($cmd[1],$cmd[2]);
140 } elsif ($cmd[0] eq "push") {
143 print STDERR
"Unknown command. Aborting...\n";
147 # blank line: we should terminate
151 BEGIN { $| = 1 } # flush STDOUT, to make sure the previous
152 # command is fully processed.
155 ########################## Functions ##############################
157 ## credential API management (generic functions)
159 sub credential_from_url
{
161 my $parsed = URI
->new($url);
164 if ($parsed->scheme) {
165 $credential{protocol
} = $parsed->scheme;
168 $credential{host
} = $parsed->host;
171 $credential{path
} = $parsed->path;
173 if ($parsed->userinfo) {
174 if ($parsed->userinfo =~ /([^:]*):(.*)/) {
175 $credential{username
} = $1;
176 $credential{password
} = $2;
178 $credential{username
} = $parsed->userinfo;
185 sub credential_read
{
190 my ($key, $value) = /([^=]*)=(.*)/;
191 if (not defined $key) {
192 die "ERROR receiving response from git credential $op:\n$_\n";
194 $credential{$key} = $value;
199 sub credential_write
{
200 my $credential = shift;
202 while (my ($key, $value) = each(%$credential) ) {
204 print $writer "$key=$value\n";
211 my $credential = shift;
212 my $pid = open2
(my $reader, my $writer, "git credential $op");
213 credential_write
($credential, $writer);
218 %$credential = credential_read
($reader, $op);
221 die "ERROR while running git credential $op:\n$_";
226 my $child_exit_status = $?
>> 8;
227 if ($child_exit_status != 0) {
228 die "'git credential $op' failed with code $child_exit_status.";
232 # MediaWiki API instance, created lazily.
235 sub mw_connect_maybe
{
239 $mediawiki = MediaWiki
::API
->new;
240 $mediawiki->{config
}->{api_url
} = "$url/api.php";
242 my %credential = credential_from_url
($url);
243 $credential{username
} = $wiki_login;
244 $credential{password
} = $wiki_passwd;
245 credential_run
("fill", \
%credential);
246 my $request = {lgname
=> $credential{username
},
247 lgpassword
=> $credential{password
},
248 lgdomain
=> $wiki_domain};
249 if ($mediawiki->login($request)) {
250 credential_run
("approve", \
%credential);
251 print STDERR
"Logged in mediawiki user \"$credential{username}\".\n";
253 print STDERR
"Failed to log in mediawiki user \"$credential{username}\" on $url\n";
254 print STDERR
" (error " .
255 $mediawiki->{error
}->{code
} . ': ' .
256 $mediawiki->{error
}->{details
} . ")\n";
257 credential_run
("reject", \
%credential);
263 ## Functions for listing pages on the remote wiki
264 sub get_mw_tracked_pages
{
266 get_mw_page_list
(\
@tracked_pages, $pages);
269 sub get_mw_page_list
{
270 my $page_list = shift;
272 my @some_pages = @
$page_list;
273 while (@some_pages) {
275 if ($#some_pages < $last) {
276 $last = $#some_pages;
278 my @slice = @some_pages[0..$last];
279 get_mw_first_pages
(\
@slice, $pages);
280 @some_pages = @some_pages[51..$#some_pages];
284 sub get_mw_tracked_categories
{
286 foreach my $category (@tracked_categories) {
287 if (index($category, ':') < 0) {
288 # Mediawiki requires the Category
289 # prefix, but let's not force the user
291 $category = "Category:" . $category;
293 my $mw_pages = $mediawiki->list( {
295 list
=> 'categorymembers',
296 cmtitle
=> $category,
298 || die $mediawiki->{error
}->{code
} . ': '
299 . $mediawiki->{error
}->{details
};
300 foreach my $page (@
{$mw_pages}) {
301 $pages->{$page->{title
}} = $page;
306 sub get_mw_all_pages
{
308 # No user-provided list, get the list of pages from the API.
309 my $mw_pages = $mediawiki->list({
314 if (!defined($mw_pages)) {
315 print STDERR
"fatal: could not get the list of wiki pages.\n";
316 print STDERR
"fatal: '$url' does not appear to be a mediawiki\n";
317 print STDERR
"fatal: make sure '$url/api.php' is a valid page.\n";
320 foreach my $page (@
{$mw_pages}) {
321 $pages->{$page->{title
}} = $page;
325 # queries the wiki for a set of pages. Meant to be used within a loop
326 # querying the wiki for slices of page list.
327 sub get_mw_first_pages
{
328 my $some_pages = shift;
329 my @some_pages = @
{$some_pages};
333 # pattern 'page1|page2|...' required by the API
334 my $titles = join('|', @some_pages);
336 my $mw_pages = $mediawiki->api({
340 if (!defined($mw_pages)) {
341 print STDERR
"fatal: could not query the list of wiki pages.\n";
342 print STDERR
"fatal: '$url' does not appear to be a mediawiki\n";
343 print STDERR
"fatal: make sure '$url/api.php' is a valid page.\n";
346 while (my ($id, $page) = each(%{$mw_pages->{query
}->{pages
}})) {
348 print STDERR
"Warning: page $page->{title} not found on wiki\n";
350 $pages->{$page->{title
}} = $page;
355 # Get the list of pages to be fetched according to configuration.
359 my %pages; # hash on page titles to avoid duplicates
361 if (@tracked_pages) {
363 # The user provided a list of pages titles, but we
364 # still need to query the API to get the page IDs.
365 get_mw_tracked_pages
(\
%pages);
367 if (@tracked_categories) {
369 get_mw_tracked_categories
(\
%pages);
371 if (!$user_defined) {
372 get_mw_all_pages
(\
%pages);
375 print STDERR
"Getting media files for selected pages...\n";
377 get_linked_mediafiles
(\
%pages);
379 get_all_mediafiles
(\
%pages);
385 # usage: $out = run_git("command args");
386 # $out = run_git("command args", "raw"); # don't interpret output as UTF-8.
389 my $encoding = (shift || "encoding(UTF-8)");
390 open(my $git, "-|:$encoding", "git " . $args);
391 my $res = do { local $/; <$git> };
398 sub get_all_mediafiles
{
400 # Attach list of all pages for media files from the API,
401 # they are in a different namespace, only one namespace
402 # can be queried at the same moment
403 my $mw_pages = $mediawiki->list({
406 apnamespace
=> get_mw_namespace_id
("File"),
409 if (!defined($mw_pages)) {
410 print STDERR
"fatal: could not get the list of pages for media files.\n";
411 print STDERR
"fatal: '$url' does not appear to be a mediawiki\n";
412 print STDERR
"fatal: make sure '$url/api.php' is a valid page.\n";
415 foreach my $page (@
{$mw_pages}) {
416 $pages->{$page->{title
}} = $page;
420 sub get_linked_mediafiles
{
422 my @titles = map $_->{title
}, values(%{$pages});
424 # The query is split in small batches because of the MW API limit of
425 # the number of links to be returned (500 links max).
428 if ($#titles < $batch) {
431 my @slice = @titles[0..$batch];
433 # pattern 'page1|page2|...' required by the API
434 my $mw_titles = join('|', @slice);
436 # Media files could be included or linked from
437 # a page, get all related
440 prop
=> 'links|images',
441 titles
=> $mw_titles,
442 plnamespace
=> get_mw_namespace_id
("File"),
445 my $result = $mediawiki->api($query);
447 while (my ($id, $page) = each(%{$result->{query
}->{pages
}})) {
449 if (defined($page->{links
})) {
450 my @link_titles = map $_->{title
}, @
{$page->{links
}};
451 push(@media_titles, @link_titles);
453 if (defined($page->{images
})) {
454 my @image_titles = map $_->{title
}, @
{$page->{images
}};
455 push(@media_titles, @image_titles);
458 get_mw_page_list
(\
@media_titles, $pages);
462 @titles = @titles[($batch+1)..$#titles];
466 sub get_mw_mediafile_for_page_revision
{
467 # Name of the file on Wiki, with the prefix.
468 my $filename = shift;
469 my $timestamp = shift;
472 # Search if on a media file with given timestamp exists on
473 # MediaWiki. In that case download the file.
477 titles
=> "File:" . $filename,
478 iistart
=> $timestamp,
480 iiprop
=> 'timestamp|archivename|url',
483 my $result = $mediawiki->api($query);
485 my ($fileid, $file) = each( %{$result->{query
}->{pages
}} );
486 # If not defined it means there is no revision of the file for
488 if (defined($file->{imageinfo
})) {
489 $mediafile{title
} = $filename;
491 my $fileinfo = pop(@
{$file->{imageinfo
}});
492 $mediafile{timestamp
} = $fileinfo->{timestamp
};
493 # Mediawiki::API's download function doesn't support https URLs
494 # and can't download old versions of files.
495 print STDERR
"\tDownloading file $mediafile{title}, version $mediafile{timestamp}\n";
496 $mediafile{content
} = download_mw_mediafile
($fileinfo->{url
});
501 sub download_mw_mediafile
{
504 my $response = $mediawiki->{ua
}->get($url);
505 if ($response->code == 200) {
506 return $response->decoded_content;
508 print STDERR
"Error downloading mediafile from :\n";
509 print STDERR
"URL: $url\n";
510 print STDERR
"Server response: " . $response->code . " " . $response->message . "\n";
515 sub get_last_local_revision
{
516 # Get note regarding last mediawiki revision
517 my $note = run_git
("notes --ref=$remotename/mediawiki show refs/mediawiki/$remotename/master 2>/dev/null");
518 my @note_info = split(/ /, $note);
520 my $lastrevision_number;
521 if (!(defined($note_info[0]) && $note_info[0] eq "mediawiki_revision:")) {
522 print STDERR
"No previous mediawiki revision found";
523 $lastrevision_number = 0;
525 # Notes are formatted : mediawiki_revision: #number
526 $lastrevision_number = $note_info[1];
527 chomp($lastrevision_number);
528 print STDERR
"Last local mediawiki revision found is $lastrevision_number";
530 return $lastrevision_number;
533 # Remember the timestamp corresponding to a revision id.
536 sub get_last_remote_revision
{
539 my %pages_hash = get_mw_pages
();
540 my @pages = values(%pages_hash);
544 foreach my $page (@pages) {
545 my $id = $page->{pageid
};
550 rvprop
=> 'ids|timestamp',
554 my $result = $mediawiki->api($query);
556 my $lastrev = pop(@
{$result->{query
}->{pages
}->{$id}->{revisions
}});
558 $basetimestamps{$lastrev->{revid
}} = $lastrev->{timestamp
};
560 $max_rev_num = ($lastrev->{revid
} > $max_rev_num ?
$lastrev->{revid
} : $max_rev_num);
563 print STDERR
"Last remote revision found is $max_rev_num.\n";
567 # Clean content before sending it to MediaWiki
568 sub mediawiki_clean
{
570 my $page_created = shift;
571 # Mediawiki does not allow blank space at the end of a page and ends with a single \n.
572 # This function right trims a string and adds a \n at the end to follow this rule
574 if ($string eq "" && $page_created) {
575 # Creating empty pages is forbidden.
576 $string = EMPTY_CONTENT
;
581 # Filter applied on MediaWiki data before adding them to Git
582 sub mediawiki_smudge
{
584 if ($string eq EMPTY_CONTENT
) {
587 # This \n is important. This is due to mediawiki's way to handle end of files.
591 sub mediawiki_clean_filename
{
592 my $filename = shift;
593 $filename =~ s/@{[SLASH_REPLACEMENT]}/\//g
;
594 # [, ], |, {, and } are forbidden by MediaWiki, even URL-encoded.
595 # Do a variant of URL-encoding, i.e. looks like URL-encoding,
596 # but with _ added to prevent MediaWiki from thinking this is
597 # an actual special character.
598 $filename =~ s/[\[\]\{\}\|]/sprintf("_%%_%x", ord($&))/ge;
599 # If we use the uri escape before
600 # we should unescape here, before anything
605 sub mediawiki_smudge_filename
{
606 my $filename = shift;
607 $filename =~ s/\//@
{[SLASH_REPLACEMENT
]}/g
;
608 $filename =~ s/ /_/g;
609 # Decode forbidden characters encoded in mediawiki_clean_filename
610 $filename =~ s/_%_([0-9a-fA-F][0-9a-fA-F])/sprintf("%c", hex($1))/ge;
616 print STDOUT
"data ", bytes
::length($content), "\n", $content;
619 sub literal_data_raw
{
620 # Output possibly binary content.
622 # Avoid confusion between size in bytes and in characters
623 utf8
::downgrade
($content);
624 binmode STDOUT
, ":raw";
625 print STDOUT
"data ", bytes
::length($content), "\n", $content;
626 binmode STDOUT
, ":utf8";
629 sub mw_capabilities
{
630 # Revisions are imported to the private namespace
631 # refs/mediawiki/$remotename/ by the helper and fetched into
632 # refs/remotes/$remotename later by fetch.
633 print STDOUT
"refspec refs/heads/*:refs/mediawiki/$remotename/*\n";
634 print STDOUT
"import\n";
635 print STDOUT
"list\n";
636 print STDOUT
"push\n";
641 # MediaWiki do not have branches, we consider one branch arbitrarily
642 # called master, and HEAD pointing to it.
643 print STDOUT
"? refs/heads/master\n";
644 print STDOUT
"\@refs/heads/master HEAD\n";
649 print STDERR
"remote-helper command 'option $_[0]' not yet implemented\n";
650 print STDOUT
"unsupported\n";
653 sub fetch_mw_revisions_for_page
{
656 my $fetch_from = shift;
663 rvstartid
=> $fetch_from,
669 # Get 500 revisions at a time due to the mediawiki api limit
671 my $result = $mediawiki->api($query);
673 # Parse each of those 500 revisions
674 foreach my $revision (@
{$result->{query
}->{pages
}->{$id}->{revisions
}}) {
676 $page_rev_ids->{pageid
} = $page->{pageid
};
677 $page_rev_ids->{revid
} = $revision->{revid
};
678 push(@page_revs, $page_rev_ids);
681 last unless $result->{'query-continue'};
682 $query->{rvstartid
} = $result->{'query-continue'}->{revisions
}->{rvstartid
};
684 if ($shallow_import && @page_revs) {
685 print STDERR
" Found 1 revision (shallow import).\n";
686 @page_revs = sort {$b->{revid
} <=> $a->{revid
}} (@page_revs);
687 return $page_revs[0];
689 print STDERR
" Found ", $revnum, " revision(s).\n";
693 sub fetch_mw_revisions
{
694 my $pages = shift; my @pages = @
{$pages};
695 my $fetch_from = shift;
699 foreach my $page (@pages) {
700 my $id = $page->{pageid
};
702 print STDERR
"page $n/", scalar(@pages), ": ". $page->{title
} ."\n";
704 my @page_revs = fetch_mw_revisions_for_page
($page, $id, $fetch_from);
705 @revisions = (@page_revs, @revisions);
708 return ($n, @revisions);
711 sub import_file_revision
{
713 my %commit = %{$commit};
714 my $full_import = shift;
716 my $mediafile = shift;
719 %mediafile = %{$mediafile};
722 my $title = $commit{title
};
723 my $comment = $commit{comment
};
724 my $content = $commit{content
};
725 my $author = $commit{author
};
726 my $date = $commit{date
};
728 print STDOUT
"commit refs/mediawiki/$remotename/master\n";
729 print STDOUT
"mark :$n\n";
730 print STDOUT
"committer $author <$author\@$wiki_name> ", $date->epoch, " +0000\n";
731 literal_data
($comment);
733 # If it's not a clone, we need to know where to start from
734 if (!$full_import && $n == 1) {
735 print STDOUT
"from refs/mediawiki/$remotename/master^0\n";
737 if ($content ne DELETED_CONTENT
) {
738 print STDOUT
"M 644 inline $title.mw\n";
739 literal_data
($content);
741 print STDOUT
"M 644 inline $mediafile{title}\n";
742 literal_data_raw
($mediafile{content
});
746 print STDOUT
"D $title.mw\n";
749 # mediawiki revision number in the git note
750 if ($full_import && $n == 1) {
751 print STDOUT
"reset refs/notes/$remotename/mediawiki\n";
753 print STDOUT
"commit refs/notes/$remotename/mediawiki\n";
754 print STDOUT
"committer $author <$author\@$wiki_name> ", $date->epoch, " +0000\n";
755 literal_data
("Note added by git-mediawiki during import");
756 if (!$full_import && $n == 1) {
757 print STDOUT
"from refs/notes/$remotename/mediawiki^0\n";
759 print STDOUT
"N inline :$n\n";
760 literal_data
("mediawiki_revision: " . $commit{mw_revision
});
764 # parse a sequence of
768 # (like batch sequence of import and sequence of push statements)
774 if ($line =~ m/^$cmd (.*)$/) {
776 } elsif ($line eq "\n") {
779 die("Invalid command in a '$cmd' batch: ". $_);
785 # multiple import commands can follow each other.
786 my @refs = (shift, get_more_refs
("import"));
787 foreach my $ref (@refs) {
790 print STDOUT
"done\n";
795 # The remote helper will call "import HEAD" and
796 # "import refs/heads/master".
797 # Since HEAD is a symbolic ref to master (by convention,
798 # followed by the output of the command "list" that we gave),
799 # we don't need to do anything in this case.
800 if ($ref eq "HEAD") {
806 my %pages_hash = get_mw_pages
();
807 my @pages = values(%pages_hash);
809 print STDERR
"Searching revisions...\n";
810 my $last_local = get_last_local_revision
();
811 my $fetch_from = $last_local + 1;
812 if ($fetch_from == 1) {
813 print STDERR
", fetching from beginning.\n";
815 print STDERR
", fetching from here.\n";
817 my ($n, @revisions) = fetch_mw_revisions
(\
@pages, $fetch_from);
819 # Creation of the fast-import stream
820 print STDERR
"Fetching & writing export data...\n";
822 @revisions = sort {$a->{revid
} <=> $b->{revid
}} @revisions;
823 my @revision_ids = map $_->{revid
}, @revisions;
825 return mw_import_revids
($fetch_from, \
@revision_ids, \
%pages_hash);
828 sub mw_import_revids
{
829 my $fetch_from = shift;
830 my $revision_ids = shift;
834 my $last_timestamp = 0; # Placeholer in case $rev->timestamp is undefined
836 foreach my $pagerevid (@
$revision_ids) {
837 # fetch the content of the pages
841 rvprop
=> 'content|timestamp|comment|user|ids',
842 revids
=> $pagerevid,
845 my $result = $mediawiki->api($query);
847 my @result_pages = values(%{$result->{query
}->{pages
}});
848 my $result_page = $result_pages[0];
849 my $rev = $result_pages[0]->{revisions
}->[0];
853 my $page_title = $result_page->{title
};
855 $commit{author
} = $rev->{user
} || 'Anonymous';
856 $commit{comment
} = $rev->{comment
} || '*Empty MediaWiki Message*';
857 $commit{title
} = mediawiki_smudge_filename
($page_title);
858 $commit{mw_revision
} = $rev->{revid
};
859 $commit{content
} = mediawiki_smudge
($rev->{'*'});
861 if (!defined($rev->{timestamp
})) {
864 $last_timestamp = $rev->{timestamp
};
866 $commit{date
} = DateTime
::Format
::ISO8601
->parse_datetime($last_timestamp);
868 # Differentiates classic pages and media files.
869 my ($namespace, $filename) = $page_title =~ /^([^:]*):(.*)$/;
871 if ($namespace && get_mw_namespace_id
($namespace) == get_mw_namespace_id
("File")) {
872 %mediafile = get_mw_mediafile_for_page_revision
($filename, $rev->{timestamp
});
874 # If this is a revision of the media page for new version
875 # of a file do one common commit for both file and media page.
876 # Else do commit only for that page.
877 print STDERR
"$n/", scalar(@
$revision_ids), ": Revision #$rev->{revid} of $commit{title}\n";
878 import_file_revision
(\
%commit, ($fetch_from == 1), $n, \
%mediafile);
881 if ($fetch_from == 1 && $n == 0) {
882 print STDERR
"You appear to have cloned an empty MediaWiki.\n";
883 # Something has to be done remote-helper side. If nothing is done, an error is
884 # thrown saying that HEAD is refering to unknown object 0000000000000000000
885 # and the clone fails.
891 sub error_non_fast_forward
{
892 my $advice = run_git
("config --bool advice.pushNonFastForward");
894 if ($advice ne "false") {
895 # Native git-push would show this after the summary.
896 # We can't ask it to display it cleanly, so print it
898 print STDERR
"To prevent you from losing history, non-fast-forward updates were rejected\n";
899 print STDERR
"Merge the remote changes (e.g. 'git pull') before pushing again. See the\n";
900 print STDERR
"'Note about fast-forwards' section of 'git push --help' for details.\n";
902 print STDOUT
"error $_[0] \"non-fast-forward\"\n";
907 my $complete_file_name = shift;
908 my $new_sha1 = shift;
909 my $extension = shift;
910 my $file_deleted = shift;
913 my $path = "File:" . $complete_file_name;
914 my %hashFiles = get_allowed_file_extensions
();
915 if (!exists($hashFiles{$extension})) {
916 print STDERR
"$complete_file_name is not a permitted file on this wiki.\n";
917 print STDERR
"Check the configuration of file uploads in your mediawiki.\n";
920 # Deleting and uploading a file requires a priviledged user
928 if (!$mediawiki->edit($query)) {
929 print STDERR
"Failed to delete file on remote wiki\n";
930 print STDERR
"Check your permissions on the remote site. Error code:\n";
931 print STDERR
$mediawiki->{error
}->{code
} . ':' . $mediawiki->{error
}->{details
};
935 # Don't let perl try to interpret file content as UTF-8 => use "raw"
936 my $content = run_git
("cat-file blob $new_sha1", "raw");
937 if ($content ne "") {
939 $mediawiki->{config
}->{upload_url
} =
940 "$url/index.php/Special:Upload";
943 filename
=> $complete_file_name,
947 Content
=> $content],
951 } ) || die $mediawiki->{error
}->{code
} . ':'
952 . $mediawiki->{error
}->{details
};
953 my $last_file_page = $mediawiki->get_page({title
=> $path});
954 $newrevid = $last_file_page->{revid
};
955 print STDERR
"Pushed file: $new_sha1 - $complete_file_name.\n";
957 print STDERR
"Empty file $complete_file_name not pushed.\n";
964 my $diff_info = shift;
965 # $diff_info contains a string in this format:
966 # 100644 100644 <sha1_of_blob_before_commit> <sha1_of_blob_now> <status>
967 my @diff_info_split = split(/[ \t]/, $diff_info);
969 # Filename, including .mw extension
970 my $complete_file_name = shift;
973 # MediaWiki revision number. Keep the previous one by default,
974 # in case there's no edit to perform.
975 my $oldrevid = shift;
978 my $new_sha1 = $diff_info_split[3];
979 my $old_sha1 = $diff_info_split[2];
980 my $page_created = ($old_sha1 eq NULL_SHA1
);
981 my $page_deleted = ($new_sha1 eq NULL_SHA1
);
982 $complete_file_name = mediawiki_clean_filename
($complete_file_name);
984 my ($title, $extension) = $complete_file_name =~ /^(.*)\.([^\.]*)$/;
985 if (!defined($extension)) {
988 if ($extension eq "mw") {
991 # Deleting a page usually requires
992 # special priviledges. A common
993 # convention is to replace the page
994 # with this content instead:
995 $file_content = DELETED_CONTENT
;
997 $file_content = run_git
("cat-file blob $new_sha1");
1002 my $result = $mediawiki->edit( {
1004 summary
=> $summary,
1006 basetimestamp
=> $basetimestamps{$oldrevid},
1007 text
=> mediawiki_clean
($file_content, $page_created),
1009 skip_encoding
=> 1 # Helps with names with accentuated characters
1012 if ($mediawiki->{error
}->{code
} == 3) {
1013 # edit conflicts, considered as non-fast-forward
1014 print STDERR
'Warning: Error ' .
1015 $mediawiki->{error
}->{code
} .
1016 ' from mediwiki: ' . $mediawiki->{error
}->{details
} .
1018 return ($oldrevid, "non-fast-forward");
1020 # Other errors. Shouldn't happen => just die()
1021 die 'Fatal: Error ' .
1022 $mediawiki->{error
}->{code
} .
1023 ' from mediwiki: ' . $mediawiki->{error
}->{details
};
1026 $newrevid = $result->{edit
}->{newrevid
};
1027 print STDERR
"Pushed file: $new_sha1 - $title\n";
1029 $newrevid = mw_upload_file
($complete_file_name, $new_sha1,
1030 $extension, $page_deleted,
1033 $newrevid = ($newrevid or $oldrevid);
1034 return ($newrevid, "ok");
1038 # multiple push statements can follow each other
1039 my @refsspecs = (shift, get_more_refs
("push"));
1041 for my $refspec (@refsspecs) {
1042 my ($force, $local, $remote) = $refspec =~ /^(\+)?([^:]*):([^:]*)$/
1043 or die("Invalid refspec for push. Expected <src>:<dst> or +<src>:<dst>");
1045 print STDERR
"Warning: forced push not allowed on a MediaWiki.\n";
1048 print STDERR
"Cannot delete remote branch on a MediaWiki\n";
1049 print STDOUT
"error $remote cannot delete\n";
1052 if ($remote ne "refs/heads/master") {
1053 print STDERR
"Only push to the branch 'master' is supported on a MediaWiki\n";
1054 print STDOUT
"error $remote only master allowed\n";
1057 if (mw_push_revision
($local, $remote)) {
1062 # Notify Git that the push is done
1065 if ($pushed && $dumb_push) {
1066 print STDERR
"Just pushed some revisions to MediaWiki.\n";
1067 print STDERR
"The pushed revisions now have to be re-imported, and your current branch\n";
1068 print STDERR
"needs to be updated with these re-imported commits. You can do this with\n";
1070 print STDERR
" git pull --rebase\n";
1075 sub mw_push_revision
{
1077 my $remote = shift; # actually, this has to be "refs/heads/master" at this point.
1078 my $last_local_revid = get_last_local_revision
();
1079 print STDERR
".\n"; # Finish sentence started by get_last_local_revision()
1080 my $last_remote_revid = get_last_remote_revision
();
1081 my $mw_revision = $last_remote_revid;
1083 # Get sha1 of commit pointed by local HEAD
1084 my $HEAD_sha1 = run_git
("rev-parse $local 2>/dev/null"); chomp($HEAD_sha1);
1085 # Get sha1 of commit pointed by remotes/$remotename/master
1086 my $remoteorigin_sha1 = run_git
("rev-parse refs/remotes/$remotename/master 2>/dev/null");
1087 chomp($remoteorigin_sha1);
1089 if ($last_local_revid > 0 &&
1090 $last_local_revid < $last_remote_revid) {
1091 return error_non_fast_forward
($remote);
1094 if ($HEAD_sha1 eq $remoteorigin_sha1) {
1099 # Get every commit in between HEAD and refs/remotes/origin/master,
1100 # including HEAD and refs/remotes/origin/master
1101 my @commit_pairs = ();
1102 if ($last_local_revid > 0) {
1103 my $parsed_sha1 = $remoteorigin_sha1;
1104 # Find a path from last MediaWiki commit to pushed commit
1105 while ($parsed_sha1 ne $HEAD_sha1) {
1106 my @commit_info = grep(/^$parsed_sha1/, split(/\n/, run_git
("rev-list --children $local")));
1107 if (!@commit_info) {
1108 return error_non_fast_forward
($remote);
1110 my @commit_info_split = split(/ |\n/, $commit_info[0]);
1111 # $commit_info_split[1] is the sha1 of the commit to export
1112 # $commit_info_split[0] is the sha1 of its direct child
1113 push(@commit_pairs, \
@commit_info_split);
1114 $parsed_sha1 = $commit_info_split[1];
1117 # No remote mediawiki revision. Export the whole
1118 # history (linearized with --first-parent)
1119 print STDERR
"Warning: no common ancestor, pushing complete history\n";
1120 my $history = run_git
("rev-list --first-parent --children $local");
1121 my @history = split('\n', $history);
1122 @history = @history[1..$#history];
1123 foreach my $line (reverse @history) {
1124 my @commit_info_split = split(/ |\n/, $line);
1125 push(@commit_pairs, \
@commit_info_split);
1129 foreach my $commit_info_split (@commit_pairs) {
1130 my $sha1_child = @
{$commit_info_split}[0];
1131 my $sha1_commit = @
{$commit_info_split}[1];
1132 my $diff_infos = run_git
("diff-tree -r --raw -z $sha1_child $sha1_commit");
1133 # TODO: we could detect rename, and encode them with a #redirect on the wiki.
1134 # TODO: for now, it's just a delete+add
1135 my @diff_info_list = split(/\0/, $diff_infos);
1136 # Keep the subject line of the commit message as mediawiki comment for the revision
1137 my $commit_msg = run_git
("log --no-walk --format=\"%s\" $sha1_commit");
1140 while (@diff_info_list) {
1142 # git diff-tree -z gives an output like
1143 # <metadata>\0<filename1>\0
1144 # <metadata>\0<filename2>\0
1145 # and we've split on \0.
1146 my $info = shift(@diff_info_list);
1147 my $file = shift(@diff_info_list);
1148 ($mw_revision, $status) = mw_push_file
($info, $file, $commit_msg, $mw_revision);
1149 if ($status eq "non-fast-forward") {
1150 # we may already have sent part of the
1151 # commit to MediaWiki, but it's too
1152 # late to cancel it. Stop the push in
1153 # the middle, but still give an
1154 # accurate error message.
1155 return error_non_fast_forward
($remote);
1157 if ($status ne "ok") {
1158 die("Unknown error from mw_push_file()");
1161 unless ($dumb_push) {
1162 run_git
("notes --ref=$remotename/mediawiki add -m \"mediawiki_revision: $mw_revision\" $sha1_commit");
1163 run_git
("update-ref -m \"Git-MediaWiki push\" refs/mediawiki/$remotename/master $sha1_commit $sha1_child");
1167 print STDOUT
"ok $remote\n";
1171 sub get_allowed_file_extensions
{
1177 siprop
=> 'fileextensions'
1179 my $result = $mediawiki->api($query);
1180 my @file_extensions= map $_->{ext
},@
{$result->{query
}->{fileextensions
}};
1181 my %hashFile = map {$_ => 1}@file_extensions;
1186 # In memory cache for MediaWiki namespace ids.
1189 # Namespaces whose id is cached in the configuration file
1190 # (to avoid duplicates)
1191 my %cached_mw_namespace_id;
1193 # Return MediaWiki id for a canonical namespace name.
1194 # Ex.: "File", "Project".
1195 sub get_mw_namespace_id
{
1199 if (!exists $namespace_id{$name}) {
1200 # Look at configuration file, if the record for that namespace is
1201 # already cached. Namespaces are stored in form:
1202 # "Name_of_namespace:Id_namespace", ex.: "File:6".
1203 my @temp = split(/[ \n]/, run_git
("config --get-all remote."
1204 . $remotename .".namespaceCache"));
1206 foreach my $ns (@temp) {
1207 my ($n, $id) = split(/:/, $ns);
1208 $namespace_id{$n} = $id;
1209 $cached_mw_namespace_id{$n} = 1;
1213 if (!exists $namespace_id{$name}) {
1214 print STDERR
"Namespace $name not found in cache, querying the wiki ...\n";
1215 # NS not found => get namespace id from MW and store it in
1216 # configuration file.
1220 siprop
=> 'namespaces'
1222 my $result = $mediawiki->api($query);
1224 while (my ($id, $ns) = each(%{$result->{query
}->{namespaces
}})) {
1225 if (defined($ns->{id
}) && defined($ns->{canonical
})) {
1226 $namespace_id{$ns->{canonical
}} = $ns->{id
};
1228 # alias (e.g. french Fichier: as alias for canonical File:)
1229 $namespace_id{$ns->{'*'}} = $ns->{id
};
1235 my $id = $namespace_id{$name};
1238 # Store explicitely requested namespaces on disk
1239 if (!exists $cached_mw_namespace_id{$name}) {
1240 run_git
("config --add remote.". $remotename
1241 .".namespaceCache \"". $name .":". $id ."\"");
1242 $cached_mw_namespace_id{$name} = 1;
1246 die "No such namespace $name on MediaWiki.";