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 # - Several strategies are provided to fetch modifications from the
17 # wiki, but no automatic heuristics is provided, the user has
18 # to understand and chose which strategy is appropriate for him.
20 # - Git renames could be turned into MediaWiki renames (see TODO
23 # - login/password support requires the user to write the password
24 # cleartext in a file (see TODO below).
26 # - No way to import "one page, and all pages included in it"
28 # - Multiple remote MediaWikis have not been very well tested.
32 use DateTime
::Format
::ISO8601
;
34 # By default, use UTF-8 to communicate with Git and the user
35 binmode STDERR
, ":utf8";
36 binmode STDOUT
, ":utf8";
43 # Mediawiki filenames can contain forward slashes. This variable decides by which pattern they should be replaced
44 use constant SLASH_REPLACEMENT
=> "%2F";
46 # It's not always possible to delete pages (may require some
47 # priviledges). Deleted pages are replaced with this content.
48 use constant DELETED_CONTENT
=> "[[Category:Deleted]]\n";
50 # It's not possible to create empty pages. New empty files in Git are
51 # sent with this content instead.
52 use constant EMPTY_CONTENT
=> "<!-- empty page -->\n";
54 # used to reflect file creation or deletion in diff.
55 use constant NULL_SHA1
=> "0000000000000000000000000000000000000000";
57 my $remotename = $ARGV[0];
60 # Accept both space-separated and multiple keys in config file.
61 # Spaces should be written as _ anyway because we'll use chomp.
62 my @tracked_pages = split(/[ \n]/, run_git
("config --get-all remote.". $remotename .".pages"));
63 chomp(@tracked_pages);
65 # Just like @tracked_pages, but for MediaWiki categories.
66 my @tracked_categories = split(/[ \n]/, run_git
("config --get-all remote.". $remotename .".categories"));
67 chomp(@tracked_categories);
69 # Import media files too.
70 my $import_media = run_git
("config --get --bool remote.". $remotename .".mediaimport");
72 $import_media = ($import_media eq "true");
74 my $wiki_login = run_git
("config --get remote.". $remotename .".mwLogin");
75 # TODO: ideally, this should be able to read from keyboard, but we're
76 # inside a remote helper, so our stdin is connect to git, not to a
78 my $wiki_passwd = run_git
("config --get remote.". $remotename .".mwPassword");
79 my $wiki_domain = run_git
("config --get remote.". $remotename .".mwDomain");
84 # Import only last revisions (both for clone and fetch)
85 my $shallow_import = run_git
("config --get --bool remote.". $remotename .".shallow");
86 chomp($shallow_import);
87 $shallow_import = ($shallow_import eq "true");
89 # Fetch (clone and pull) by revisions instead of by pages. This behavior
90 # is more efficient when we have a wiki with lots of pages and we fetch
91 # the revisions quite often so that they concern only few pages.
93 # - by_rev: perform one query per new revision on the remote wiki
94 # - by_page: query each tracked page for new revision
95 my $fetch_strategy = run_git
("config --get remote.$remotename.fetchStrategy");
96 unless ($fetch_strategy) {
97 $fetch_strategy = run_git
("config --get mediawiki.fetchStrategy");
99 chomp($fetch_strategy);
100 unless ($fetch_strategy) {
101 $fetch_strategy = "by_page";
104 # Dumb push: don't update notes and mediawiki ref to reflect the last push.
106 # Configurable with mediawiki.dumbPush, or per-remote with
107 # remote.<remotename>.dumbPush.
109 # This means the user will have to re-import the just-pushed
110 # revisions. On the other hand, this means that the Git revisions
111 # corresponding to MediaWiki revisions are all imported from the wiki,
112 # regardless of whether they were initially created in Git or from the
113 # web interface, hence all users will get the same history (i.e. if
114 # the push from Git to MediaWiki loses some information, everybody
115 # will get the history with information lost). If the import is
116 # deterministic, this means everybody gets the same sha1 for each
117 # MediaWiki revision.
118 my $dumb_push = run_git
("config --get --bool remote.$remotename.dumbPush");
119 unless ($dumb_push) {
120 $dumb_push = run_git
("config --get --bool mediawiki.dumbPush");
123 $dumb_push = ($dumb_push eq "true");
125 my $wiki_name = $url;
126 $wiki_name =~ s/[^\/]*:\/\
///;
127 # If URL is like http://user:password@example.com/, we clearly don't
128 # want the password in $wiki_name. While we're there, also remove user
129 # and '@' sign, to avoid author like MWUser@HTTPUser@host.com
130 $wiki_name =~ s/^.*@//;
138 if (defined($cmd[0])) {
140 if ($cmd[0] eq "capabilities") {
141 die("Too many arguments for capabilities") unless (!defined($cmd[1]));
143 } elsif ($cmd[0] eq "list") {
144 die("Too many arguments for list") unless (!defined($cmd[2]));
146 } elsif ($cmd[0] eq "import") {
147 die("Invalid arguments for import") unless ($cmd[1] ne "" && !defined($cmd[2]));
149 } elsif ($cmd[0] eq "option") {
150 die("Too many arguments for option") unless ($cmd[1] ne "" && $cmd[2] ne "" && !defined($cmd[3]));
151 mw_option
($cmd[1],$cmd[2]);
152 } elsif ($cmd[0] eq "push") {
155 print STDERR
"Unknown command. Aborting...\n";
159 # blank line: we should terminate
163 BEGIN { $| = 1 } # flush STDOUT, to make sure the previous
164 # command is fully processed.
167 ########################## Functions ##############################
169 ## credential API management (generic functions)
171 sub credential_from_url
{
173 my $parsed = URI
->new($url);
176 if ($parsed->scheme) {
177 $credential{protocol
} = $parsed->scheme;
180 $credential{host
} = $parsed->host;
183 $credential{path
} = $parsed->path;
185 if ($parsed->userinfo) {
186 if ($parsed->userinfo =~ /([^:]*):(.*)/) {
187 $credential{username
} = $1;
188 $credential{password
} = $2;
190 $credential{username
} = $parsed->userinfo;
197 sub credential_read
{
202 my ($key, $value) = /([^=]*)=(.*)/;
203 if (not defined $key) {
204 die "ERROR receiving response from git credential $op:\n$_\n";
206 $credential{$key} = $value;
211 sub credential_write
{
212 my $credential = shift;
214 while (my ($key, $value) = each(%$credential) ) {
216 print $writer "$key=$value\n";
223 my $credential = shift;
224 my $pid = open2
(my $reader, my $writer, "git credential $op");
225 credential_write
($credential, $writer);
230 %$credential = credential_read
($reader, $op);
233 die "ERROR while running git credential $op:\n$_";
238 my $child_exit_status = $?
>> 8;
239 if ($child_exit_status != 0) {
240 die "'git credential $op' failed with code $child_exit_status.";
244 # MediaWiki API instance, created lazily.
247 sub mw_connect_maybe
{
251 $mediawiki = MediaWiki
::API
->new;
252 $mediawiki->{config
}->{api_url
} = "$url/api.php";
254 my %credential = credential_from_url
($url);
255 $credential{username
} = $wiki_login;
256 $credential{password
} = $wiki_passwd;
257 credential_run
("fill", \
%credential);
258 my $request = {lgname
=> $credential{username
},
259 lgpassword
=> $credential{password
},
260 lgdomain
=> $wiki_domain};
261 if ($mediawiki->login($request)) {
262 credential_run
("approve", \
%credential);
263 print STDERR
"Logged in mediawiki user \"$credential{username}\".\n";
265 print STDERR
"Failed to log in mediawiki user \"$credential{username}\" on $url\n";
266 print STDERR
" (error " .
267 $mediawiki->{error
}->{code
} . ': ' .
268 $mediawiki->{error
}->{details
} . ")\n";
269 credential_run
("reject", \
%credential);
275 ## Functions for listing pages on the remote wiki
276 sub get_mw_tracked_pages
{
278 get_mw_page_list
(\
@tracked_pages, $pages);
281 sub get_mw_page_list
{
282 my $page_list = shift;
284 my @some_pages = @
$page_list;
285 while (@some_pages) {
287 if ($#some_pages < $last) {
288 $last = $#some_pages;
290 my @slice = @some_pages[0..$last];
291 get_mw_first_pages
(\
@slice, $pages);
292 @some_pages = @some_pages[51..$#some_pages];
296 sub get_mw_tracked_categories
{
298 foreach my $category (@tracked_categories) {
299 if (index($category, ':') < 0) {
300 # Mediawiki requires the Category
301 # prefix, but let's not force the user
303 $category = "Category:" . $category;
305 my $mw_pages = $mediawiki->list( {
307 list
=> 'categorymembers',
308 cmtitle
=> $category,
310 || die $mediawiki->{error
}->{code
} . ': '
311 . $mediawiki->{error
}->{details
};
312 foreach my $page (@
{$mw_pages}) {
313 $pages->{$page->{title
}} = $page;
318 sub get_mw_all_pages
{
320 # No user-provided list, get the list of pages from the API.
321 my $mw_pages = $mediawiki->list({
326 if (!defined($mw_pages)) {
327 print STDERR
"fatal: could not get the list of wiki pages.\n";
328 print STDERR
"fatal: '$url' does not appear to be a mediawiki\n";
329 print STDERR
"fatal: make sure '$url/api.php' is a valid page.\n";
332 foreach my $page (@
{$mw_pages}) {
333 $pages->{$page->{title
}} = $page;
337 # queries the wiki for a set of pages. Meant to be used within a loop
338 # querying the wiki for slices of page list.
339 sub get_mw_first_pages
{
340 my $some_pages = shift;
341 my @some_pages = @
{$some_pages};
345 # pattern 'page1|page2|...' required by the API
346 my $titles = join('|', @some_pages);
348 my $mw_pages = $mediawiki->api({
352 if (!defined($mw_pages)) {
353 print STDERR
"fatal: could not query the list of wiki pages.\n";
354 print STDERR
"fatal: '$url' does not appear to be a mediawiki\n";
355 print STDERR
"fatal: make sure '$url/api.php' is a valid page.\n";
358 while (my ($id, $page) = each(%{$mw_pages->{query
}->{pages
}})) {
360 print STDERR
"Warning: page $page->{title} not found on wiki\n";
362 $pages->{$page->{title
}} = $page;
367 # Get the list of pages to be fetched according to configuration.
371 my %pages; # hash on page titles to avoid duplicates
373 if (@tracked_pages) {
375 # The user provided a list of pages titles, but we
376 # still need to query the API to get the page IDs.
377 get_mw_tracked_pages
(\
%pages);
379 if (@tracked_categories) {
381 get_mw_tracked_categories
(\
%pages);
383 if (!$user_defined) {
384 get_mw_all_pages
(\
%pages);
387 print STDERR
"Getting media files for selected pages...\n";
389 get_linked_mediafiles
(\
%pages);
391 get_all_mediafiles
(\
%pages);
397 # usage: $out = run_git("command args");
398 # $out = run_git("command args", "raw"); # don't interpret output as UTF-8.
401 my $encoding = (shift || "encoding(UTF-8)");
402 open(my $git, "-|:$encoding", "git " . $args);
403 my $res = do { local $/; <$git> };
410 sub get_all_mediafiles
{
412 # Attach list of all pages for media files from the API,
413 # they are in a different namespace, only one namespace
414 # can be queried at the same moment
415 my $mw_pages = $mediawiki->list({
418 apnamespace
=> get_mw_namespace_id
("File"),
421 if (!defined($mw_pages)) {
422 print STDERR
"fatal: could not get the list of pages for media files.\n";
423 print STDERR
"fatal: '$url' does not appear to be a mediawiki\n";
424 print STDERR
"fatal: make sure '$url/api.php' is a valid page.\n";
427 foreach my $page (@
{$mw_pages}) {
428 $pages->{$page->{title
}} = $page;
432 sub get_linked_mediafiles
{
434 my @titles = map $_->{title
}, values(%{$pages});
436 # The query is split in small batches because of the MW API limit of
437 # the number of links to be returned (500 links max).
440 if ($#titles < $batch) {
443 my @slice = @titles[0..$batch];
445 # pattern 'page1|page2|...' required by the API
446 my $mw_titles = join('|', @slice);
448 # Media files could be included or linked from
449 # a page, get all related
452 prop
=> 'links|images',
453 titles
=> $mw_titles,
454 plnamespace
=> get_mw_namespace_id
("File"),
457 my $result = $mediawiki->api($query);
459 while (my ($id, $page) = each(%{$result->{query
}->{pages
}})) {
461 if (defined($page->{links
})) {
462 my @link_titles = map $_->{title
}, @
{$page->{links
}};
463 push(@media_titles, @link_titles);
465 if (defined($page->{images
})) {
466 my @image_titles = map $_->{title
}, @
{$page->{images
}};
467 push(@media_titles, @image_titles);
470 get_mw_page_list
(\
@media_titles, $pages);
474 @titles = @titles[($batch+1)..$#titles];
478 sub get_mw_mediafile_for_page_revision
{
479 # Name of the file on Wiki, with the prefix.
480 my $filename = shift;
481 my $timestamp = shift;
484 # Search if on a media file with given timestamp exists on
485 # MediaWiki. In that case download the file.
489 titles
=> "File:" . $filename,
490 iistart
=> $timestamp,
492 iiprop
=> 'timestamp|archivename|url',
495 my $result = $mediawiki->api($query);
497 my ($fileid, $file) = each( %{$result->{query
}->{pages
}} );
498 # If not defined it means there is no revision of the file for
500 if (defined($file->{imageinfo
})) {
501 $mediafile{title
} = $filename;
503 my $fileinfo = pop(@
{$file->{imageinfo
}});
504 $mediafile{timestamp
} = $fileinfo->{timestamp
};
505 # Mediawiki::API's download function doesn't support https URLs
506 # and can't download old versions of files.
507 print STDERR
"\tDownloading file $mediafile{title}, version $mediafile{timestamp}\n";
508 $mediafile{content
} = download_mw_mediafile
($fileinfo->{url
});
513 sub download_mw_mediafile
{
516 my $response = $mediawiki->{ua
}->get($url);
517 if ($response->code == 200) {
518 return $response->decoded_content;
520 print STDERR
"Error downloading mediafile from :\n";
521 print STDERR
"URL: $url\n";
522 print STDERR
"Server response: " . $response->code . " " . $response->message . "\n";
527 sub get_last_local_revision
{
528 # Get note regarding last mediawiki revision
529 my $note = run_git
("notes --ref=$remotename/mediawiki show refs/mediawiki/$remotename/master 2>/dev/null");
530 my @note_info = split(/ /, $note);
532 my $lastrevision_number;
533 if (!(defined($note_info[0]) && $note_info[0] eq "mediawiki_revision:")) {
534 print STDERR
"No previous mediawiki revision found";
535 $lastrevision_number = 0;
537 # Notes are formatted : mediawiki_revision: #number
538 $lastrevision_number = $note_info[1];
539 chomp($lastrevision_number);
540 print STDERR
"Last local mediawiki revision found is $lastrevision_number";
542 return $lastrevision_number;
545 # Remember the timestamp corresponding to a revision id.
548 # Get the last remote revision without taking in account which pages are
549 # tracked or not. This function makes a single request to the wiki thus
550 # avoid a loop onto all tracked pages. This is useful for the fetch-by-rev
552 sub get_last_global_remote_rev
{
557 list
=> 'recentchanges',
562 my $result = $mediawiki->api($query);
563 return $result->{query
}->{recentchanges
}[0]->{revid
};
566 # Get the last remote revision concerning the tracked pages and the tracked
568 sub get_last_remote_revision
{
571 my %pages_hash = get_mw_pages
();
572 my @pages = values(%pages_hash);
576 foreach my $page (@pages) {
577 my $id = $page->{pageid
};
582 rvprop
=> 'ids|timestamp',
586 my $result = $mediawiki->api($query);
588 my $lastrev = pop(@
{$result->{query
}->{pages
}->{$id}->{revisions
}});
590 $basetimestamps{$lastrev->{revid
}} = $lastrev->{timestamp
};
592 $max_rev_num = ($lastrev->{revid
} > $max_rev_num ?
$lastrev->{revid
} : $max_rev_num);
595 print STDERR
"Last remote revision found is $max_rev_num.\n";
599 # Clean content before sending it to MediaWiki
600 sub mediawiki_clean
{
602 my $page_created = shift;
603 # Mediawiki does not allow blank space at the end of a page and ends with a single \n.
604 # This function right trims a string and adds a \n at the end to follow this rule
606 if ($string eq "" && $page_created) {
607 # Creating empty pages is forbidden.
608 $string = EMPTY_CONTENT
;
613 # Filter applied on MediaWiki data before adding them to Git
614 sub mediawiki_smudge
{
616 if ($string eq EMPTY_CONTENT
) {
619 # This \n is important. This is due to mediawiki's way to handle end of files.
623 sub mediawiki_clean_filename
{
624 my $filename = shift;
625 $filename =~ s/@{[SLASH_REPLACEMENT]}/\//g
;
626 # [, ], |, {, and } are forbidden by MediaWiki, even URL-encoded.
627 # Do a variant of URL-encoding, i.e. looks like URL-encoding,
628 # but with _ added to prevent MediaWiki from thinking this is
629 # an actual special character.
630 $filename =~ s/[\[\]\{\}\|]/sprintf("_%%_%x", ord($&))/ge;
631 # If we use the uri escape before
632 # we should unescape here, before anything
637 sub mediawiki_smudge_filename
{
638 my $filename = shift;
639 $filename =~ s/\//@
{[SLASH_REPLACEMENT
]}/g
;
640 $filename =~ s/ /_/g;
641 # Decode forbidden characters encoded in mediawiki_clean_filename
642 $filename =~ s/_%_([0-9a-fA-F][0-9a-fA-F])/sprintf("%c", hex($1))/ge;
648 print STDOUT
"data ", bytes
::length($content), "\n", $content;
651 sub literal_data_raw
{
652 # Output possibly binary content.
654 # Avoid confusion between size in bytes and in characters
655 utf8
::downgrade
($content);
656 binmode STDOUT
, ":raw";
657 print STDOUT
"data ", bytes
::length($content), "\n", $content;
658 binmode STDOUT
, ":utf8";
661 sub mw_capabilities
{
662 # Revisions are imported to the private namespace
663 # refs/mediawiki/$remotename/ by the helper and fetched into
664 # refs/remotes/$remotename later by fetch.
665 print STDOUT
"refspec refs/heads/*:refs/mediawiki/$remotename/*\n";
666 print STDOUT
"import\n";
667 print STDOUT
"list\n";
668 print STDOUT
"push\n";
673 # MediaWiki do not have branches, we consider one branch arbitrarily
674 # called master, and HEAD pointing to it.
675 print STDOUT
"? refs/heads/master\n";
676 print STDOUT
"\@refs/heads/master HEAD\n";
681 print STDERR
"remote-helper command 'option $_[0]' not yet implemented\n";
682 print STDOUT
"unsupported\n";
685 sub fetch_mw_revisions_for_page
{
688 my $fetch_from = shift;
695 rvstartid
=> $fetch_from,
701 # Get 500 revisions at a time due to the mediawiki api limit
703 my $result = $mediawiki->api($query);
705 # Parse each of those 500 revisions
706 foreach my $revision (@
{$result->{query
}->{pages
}->{$id}->{revisions
}}) {
708 $page_rev_ids->{pageid
} = $page->{pageid
};
709 $page_rev_ids->{revid
} = $revision->{revid
};
710 push(@page_revs, $page_rev_ids);
713 last unless $result->{'query-continue'};
714 $query->{rvstartid
} = $result->{'query-continue'}->{revisions
}->{rvstartid
};
716 if ($shallow_import && @page_revs) {
717 print STDERR
" Found 1 revision (shallow import).\n";
718 @page_revs = sort {$b->{revid
} <=> $a->{revid
}} (@page_revs);
719 return $page_revs[0];
721 print STDERR
" Found ", $revnum, " revision(s).\n";
725 sub fetch_mw_revisions
{
726 my $pages = shift; my @pages = @
{$pages};
727 my $fetch_from = shift;
731 foreach my $page (@pages) {
732 my $id = $page->{pageid
};
734 print STDERR
"page $n/", scalar(@pages), ": ". $page->{title
} ."\n";
736 my @page_revs = fetch_mw_revisions_for_page
($page, $id, $fetch_from);
737 @revisions = (@page_revs, @revisions);
740 return ($n, @revisions);
743 sub import_file_revision
{
745 my %commit = %{$commit};
746 my $full_import = shift;
748 my $mediafile = shift;
751 %mediafile = %{$mediafile};
754 my $title = $commit{title
};
755 my $comment = $commit{comment
};
756 my $content = $commit{content
};
757 my $author = $commit{author
};
758 my $date = $commit{date
};
760 print STDOUT
"commit refs/mediawiki/$remotename/master\n";
761 print STDOUT
"mark :$n\n";
762 print STDOUT
"committer $author <$author\@$wiki_name> ", $date->epoch, " +0000\n";
763 literal_data
($comment);
765 # If it's not a clone, we need to know where to start from
766 if (!$full_import && $n == 1) {
767 print STDOUT
"from refs/mediawiki/$remotename/master^0\n";
769 if ($content ne DELETED_CONTENT
) {
770 print STDOUT
"M 644 inline $title.mw\n";
771 literal_data
($content);
773 print STDOUT
"M 644 inline $mediafile{title}\n";
774 literal_data_raw
($mediafile{content
});
778 print STDOUT
"D $title.mw\n";
781 # mediawiki revision number in the git note
782 if ($full_import && $n == 1) {
783 print STDOUT
"reset refs/notes/$remotename/mediawiki\n";
785 print STDOUT
"commit refs/notes/$remotename/mediawiki\n";
786 print STDOUT
"committer $author <$author\@$wiki_name> ", $date->epoch, " +0000\n";
787 literal_data
("Note added by git-mediawiki during import");
788 if (!$full_import && $n == 1) {
789 print STDOUT
"from refs/notes/$remotename/mediawiki^0\n";
791 print STDOUT
"N inline :$n\n";
792 literal_data
("mediawiki_revision: " . $commit{mw_revision
});
796 # parse a sequence of
800 # (like batch sequence of import and sequence of push statements)
806 if ($line =~ m/^$cmd (.*)$/) {
808 } elsif ($line eq "\n") {
811 die("Invalid command in a '$cmd' batch: ". $_);
817 # multiple import commands can follow each other.
818 my @refs = (shift, get_more_refs
("import"));
819 foreach my $ref (@refs) {
822 print STDOUT
"done\n";
827 # The remote helper will call "import HEAD" and
828 # "import refs/heads/master".
829 # Since HEAD is a symbolic ref to master (by convention,
830 # followed by the output of the command "list" that we gave),
831 # we don't need to do anything in this case.
832 if ($ref eq "HEAD") {
838 print STDERR
"Searching revisions...\n";
839 my $last_local = get_last_local_revision
();
840 my $fetch_from = $last_local + 1;
841 if ($fetch_from == 1) {
842 print STDERR
", fetching from beginning.\n";
844 print STDERR
", fetching from here.\n";
848 if ($fetch_strategy eq "by_rev") {
849 print STDERR
"Fetching & writing export data by revs...\n";
850 $n = mw_import_ref_by_revs
($fetch_from);
851 } elsif ($fetch_strategy eq "by_page") {
852 print STDERR
"Fetching & writing export data by pages...\n";
853 $n = mw_import_ref_by_pages
($fetch_from);
855 print STDERR
"fatal: invalid fetch strategy \"$fetch_strategy\".\n";
856 print STDERR
"Check your configuration variables remote.$remotename.fetchStrategy and mediawiki.fetchStrategy\n";
860 if ($fetch_from == 1 && $n == 0) {
861 print STDERR
"You appear to have cloned an empty MediaWiki.\n";
862 # Something has to be done remote-helper side. If nothing is done, an error is
863 # thrown saying that HEAD is refering to unknown object 0000000000000000000
864 # and the clone fails.
868 sub mw_import_ref_by_pages
{
870 my $fetch_from = shift;
871 my %pages_hash = get_mw_pages
();
872 my @pages = values(%pages_hash);
874 my ($n, @revisions) = fetch_mw_revisions
(\
@pages, $fetch_from);
876 @revisions = sort {$a->{revid
} <=> $b->{revid
}} @revisions;
877 my @revision_ids = map $_->{revid
}, @revisions;
879 return mw_import_revids
($fetch_from, \
@revision_ids, \
%pages_hash);
882 sub mw_import_ref_by_revs
{
884 my $fetch_from = shift;
885 my %pages_hash = get_mw_pages
();
887 my $last_remote = get_last_global_remote_rev
();
888 my @revision_ids = $fetch_from..$last_remote;
889 return mw_import_revids
($fetch_from, \
@revision_ids, \
%pages_hash);
892 # Import revisions given in second argument (array of integers).
893 # Only pages appearing in the third argument (hash indexed by page titles)
895 sub mw_import_revids
{
896 my $fetch_from = shift;
897 my $revision_ids = shift;
902 my $last_timestamp = 0; # Placeholer in case $rev->timestamp is undefined
904 foreach my $pagerevid (@
$revision_ids) {
905 # fetch the content of the pages
909 rvprop
=> 'content|timestamp|comment|user|ids',
910 revids
=> $pagerevid,
913 my $result = $mediawiki->api($query);
915 my @result_pages = values(%{$result->{query
}->{pages
}});
916 my $result_page = $result_pages[0];
917 my $rev = $result_pages[0]->{revisions
}->[0];
919 # Count page even if we skip it, since we display
920 # $n/$total and $total includes skipped pages.
923 my $page_title = $result_page->{title
};
925 if (!exists($pages->{$page_title})) {
926 print STDERR
"$n/", scalar(@
$revision_ids),
927 ": Skipping revision #$rev->{revid} of $page_title\n";
934 $commit{author
} = $rev->{user
} || 'Anonymous';
935 $commit{comment
} = $rev->{comment
} || '*Empty MediaWiki Message*';
936 $commit{title
} = mediawiki_smudge_filename
($page_title);
937 $commit{mw_revision
} = $rev->{revid
};
938 $commit{content
} = mediawiki_smudge
($rev->{'*'});
940 if (!defined($rev->{timestamp
})) {
943 $last_timestamp = $rev->{timestamp
};
945 $commit{date
} = DateTime
::Format
::ISO8601
->parse_datetime($last_timestamp);
947 # Differentiates classic pages and media files.
948 my ($namespace, $filename) = $page_title =~ /^([^:]*):(.*)$/;
950 if ($namespace && get_mw_namespace_id
($namespace) == get_mw_namespace_id
("File")) {
951 %mediafile = get_mw_mediafile_for_page_revision
($filename, $rev->{timestamp
});
953 # If this is a revision of the media page for new version
954 # of a file do one common commit for both file and media page.
955 # Else do commit only for that page.
956 print STDERR
"$n/", scalar(@
$revision_ids), ": Revision #$rev->{revid} of $commit{title}\n";
957 import_file_revision
(\
%commit, ($fetch_from == 1), $n_actual, \
%mediafile);
963 sub error_non_fast_forward
{
964 my $advice = run_git
("config --bool advice.pushNonFastForward");
966 if ($advice ne "false") {
967 # Native git-push would show this after the summary.
968 # We can't ask it to display it cleanly, so print it
970 print STDERR
"To prevent you from losing history, non-fast-forward updates were rejected\n";
971 print STDERR
"Merge the remote changes (e.g. 'git pull') before pushing again. See the\n";
972 print STDERR
"'Note about fast-forwards' section of 'git push --help' for details.\n";
974 print STDOUT
"error $_[0] \"non-fast-forward\"\n";
979 my $complete_file_name = shift;
980 my $new_sha1 = shift;
981 my $extension = shift;
982 my $file_deleted = shift;
985 my $path = "File:" . $complete_file_name;
986 my %hashFiles = get_allowed_file_extensions
();
987 if (!exists($hashFiles{$extension})) {
988 print STDERR
"$complete_file_name is not a permitted file on this wiki.\n";
989 print STDERR
"Check the configuration of file uploads in your mediawiki.\n";
992 # Deleting and uploading a file requires a priviledged user
1000 if (!$mediawiki->edit($query)) {
1001 print STDERR
"Failed to delete file on remote wiki\n";
1002 print STDERR
"Check your permissions on the remote site. Error code:\n";
1003 print STDERR
$mediawiki->{error
}->{code
} . ':' . $mediawiki->{error
}->{details
};
1007 # Don't let perl try to interpret file content as UTF-8 => use "raw"
1008 my $content = run_git
("cat-file blob $new_sha1", "raw");
1009 if ($content ne "") {
1011 $mediawiki->{config
}->{upload_url
} =
1012 "$url/index.php/Special:Upload";
1015 filename
=> $complete_file_name,
1016 comment
=> $summary,
1018 $complete_file_name,
1019 Content
=> $content],
1020 ignorewarnings
=> 1,
1023 } ) || die $mediawiki->{error
}->{code
} . ':'
1024 . $mediawiki->{error
}->{details
};
1025 my $last_file_page = $mediawiki->get_page({title
=> $path});
1026 $newrevid = $last_file_page->{revid
};
1027 print STDERR
"Pushed file: $new_sha1 - $complete_file_name.\n";
1029 print STDERR
"Empty file $complete_file_name not pushed.\n";
1036 my $diff_info = shift;
1037 # $diff_info contains a string in this format:
1038 # 100644 100644 <sha1_of_blob_before_commit> <sha1_of_blob_now> <status>
1039 my @diff_info_split = split(/[ \t]/, $diff_info);
1041 # Filename, including .mw extension
1042 my $complete_file_name = shift;
1044 my $summary = shift;
1045 # MediaWiki revision number. Keep the previous one by default,
1046 # in case there's no edit to perform.
1047 my $oldrevid = shift;
1050 my $new_sha1 = $diff_info_split[3];
1051 my $old_sha1 = $diff_info_split[2];
1052 my $page_created = ($old_sha1 eq NULL_SHA1
);
1053 my $page_deleted = ($new_sha1 eq NULL_SHA1
);
1054 $complete_file_name = mediawiki_clean_filename
($complete_file_name);
1056 my ($title, $extension) = $complete_file_name =~ /^(.*)\.([^\.]*)$/;
1057 if (!defined($extension)) {
1060 if ($extension eq "mw") {
1062 if ($page_deleted) {
1063 # Deleting a page usually requires
1064 # special priviledges. A common
1065 # convention is to replace the page
1066 # with this content instead:
1067 $file_content = DELETED_CONTENT
;
1069 $file_content = run_git
("cat-file blob $new_sha1");
1074 my $result = $mediawiki->edit( {
1076 summary
=> $summary,
1078 basetimestamp
=> $basetimestamps{$oldrevid},
1079 text
=> mediawiki_clean
($file_content, $page_created),
1081 skip_encoding
=> 1 # Helps with names with accentuated characters
1084 if ($mediawiki->{error
}->{code
} == 3) {
1085 # edit conflicts, considered as non-fast-forward
1086 print STDERR
'Warning: Error ' .
1087 $mediawiki->{error
}->{code
} .
1088 ' from mediwiki: ' . $mediawiki->{error
}->{details
} .
1090 return ($oldrevid, "non-fast-forward");
1092 # Other errors. Shouldn't happen => just die()
1093 die 'Fatal: Error ' .
1094 $mediawiki->{error
}->{code
} .
1095 ' from mediwiki: ' . $mediawiki->{error
}->{details
};
1098 $newrevid = $result->{edit
}->{newrevid
};
1099 print STDERR
"Pushed file: $new_sha1 - $title\n";
1101 $newrevid = mw_upload_file
($complete_file_name, $new_sha1,
1102 $extension, $page_deleted,
1105 $newrevid = ($newrevid or $oldrevid);
1106 return ($newrevid, "ok");
1110 # multiple push statements can follow each other
1111 my @refsspecs = (shift, get_more_refs
("push"));
1113 for my $refspec (@refsspecs) {
1114 my ($force, $local, $remote) = $refspec =~ /^(\+)?([^:]*):([^:]*)$/
1115 or die("Invalid refspec for push. Expected <src>:<dst> or +<src>:<dst>");
1117 print STDERR
"Warning: forced push not allowed on a MediaWiki.\n";
1120 print STDERR
"Cannot delete remote branch on a MediaWiki\n";
1121 print STDOUT
"error $remote cannot delete\n";
1124 if ($remote ne "refs/heads/master") {
1125 print STDERR
"Only push to the branch 'master' is supported on a MediaWiki\n";
1126 print STDOUT
"error $remote only master allowed\n";
1129 if (mw_push_revision
($local, $remote)) {
1134 # Notify Git that the push is done
1137 if ($pushed && $dumb_push) {
1138 print STDERR
"Just pushed some revisions to MediaWiki.\n";
1139 print STDERR
"The pushed revisions now have to be re-imported, and your current branch\n";
1140 print STDERR
"needs to be updated with these re-imported commits. You can do this with\n";
1142 print STDERR
" git pull --rebase\n";
1147 sub mw_push_revision
{
1149 my $remote = shift; # actually, this has to be "refs/heads/master" at this point.
1150 my $last_local_revid = get_last_local_revision
();
1151 print STDERR
".\n"; # Finish sentence started by get_last_local_revision()
1152 my $last_remote_revid = get_last_remote_revision
();
1153 my $mw_revision = $last_remote_revid;
1155 # Get sha1 of commit pointed by local HEAD
1156 my $HEAD_sha1 = run_git
("rev-parse $local 2>/dev/null"); chomp($HEAD_sha1);
1157 # Get sha1 of commit pointed by remotes/$remotename/master
1158 my $remoteorigin_sha1 = run_git
("rev-parse refs/remotes/$remotename/master 2>/dev/null");
1159 chomp($remoteorigin_sha1);
1161 if ($last_local_revid > 0 &&
1162 $last_local_revid < $last_remote_revid) {
1163 return error_non_fast_forward
($remote);
1166 if ($HEAD_sha1 eq $remoteorigin_sha1) {
1171 # Get every commit in between HEAD and refs/remotes/origin/master,
1172 # including HEAD and refs/remotes/origin/master
1173 my @commit_pairs = ();
1174 if ($last_local_revid > 0) {
1175 my $parsed_sha1 = $remoteorigin_sha1;
1176 # Find a path from last MediaWiki commit to pushed commit
1177 while ($parsed_sha1 ne $HEAD_sha1) {
1178 my @commit_info = grep(/^$parsed_sha1/, split(/\n/, run_git
("rev-list --children $local")));
1179 if (!@commit_info) {
1180 return error_non_fast_forward
($remote);
1182 my @commit_info_split = split(/ |\n/, $commit_info[0]);
1183 # $commit_info_split[1] is the sha1 of the commit to export
1184 # $commit_info_split[0] is the sha1 of its direct child
1185 push(@commit_pairs, \
@commit_info_split);
1186 $parsed_sha1 = $commit_info_split[1];
1189 # No remote mediawiki revision. Export the whole
1190 # history (linearized with --first-parent)
1191 print STDERR
"Warning: no common ancestor, pushing complete history\n";
1192 my $history = run_git
("rev-list --first-parent --children $local");
1193 my @history = split('\n', $history);
1194 @history = @history[1..$#history];
1195 foreach my $line (reverse @history) {
1196 my @commit_info_split = split(/ |\n/, $line);
1197 push(@commit_pairs, \
@commit_info_split);
1201 foreach my $commit_info_split (@commit_pairs) {
1202 my $sha1_child = @
{$commit_info_split}[0];
1203 my $sha1_commit = @
{$commit_info_split}[1];
1204 my $diff_infos = run_git
("diff-tree -r --raw -z $sha1_child $sha1_commit");
1205 # TODO: we could detect rename, and encode them with a #redirect on the wiki.
1206 # TODO: for now, it's just a delete+add
1207 my @diff_info_list = split(/\0/, $diff_infos);
1208 # Keep the subject line of the commit message as mediawiki comment for the revision
1209 my $commit_msg = run_git
("log --no-walk --format=\"%s\" $sha1_commit");
1212 while (@diff_info_list) {
1214 # git diff-tree -z gives an output like
1215 # <metadata>\0<filename1>\0
1216 # <metadata>\0<filename2>\0
1217 # and we've split on \0.
1218 my $info = shift(@diff_info_list);
1219 my $file = shift(@diff_info_list);
1220 ($mw_revision, $status) = mw_push_file
($info, $file, $commit_msg, $mw_revision);
1221 if ($status eq "non-fast-forward") {
1222 # we may already have sent part of the
1223 # commit to MediaWiki, but it's too
1224 # late to cancel it. Stop the push in
1225 # the middle, but still give an
1226 # accurate error message.
1227 return error_non_fast_forward
($remote);
1229 if ($status ne "ok") {
1230 die("Unknown error from mw_push_file()");
1233 unless ($dumb_push) {
1234 run_git
("notes --ref=$remotename/mediawiki add -m \"mediawiki_revision: $mw_revision\" $sha1_commit");
1235 run_git
("update-ref -m \"Git-MediaWiki push\" refs/mediawiki/$remotename/master $sha1_commit $sha1_child");
1239 print STDOUT
"ok $remote\n";
1243 sub get_allowed_file_extensions
{
1249 siprop
=> 'fileextensions'
1251 my $result = $mediawiki->api($query);
1252 my @file_extensions= map $_->{ext
},@
{$result->{query
}->{fileextensions
}};
1253 my %hashFile = map {$_ => 1}@file_extensions;
1258 # In memory cache for MediaWiki namespace ids.
1261 # Namespaces whose id is cached in the configuration file
1262 # (to avoid duplicates)
1263 my %cached_mw_namespace_id;
1265 # Return MediaWiki id for a canonical namespace name.
1266 # Ex.: "File", "Project".
1267 sub get_mw_namespace_id
{
1271 if (!exists $namespace_id{$name}) {
1272 # Look at configuration file, if the record for that namespace is
1273 # already cached. Namespaces are stored in form:
1274 # "Name_of_namespace:Id_namespace", ex.: "File:6".
1275 my @temp = split(/[ \n]/, run_git
("config --get-all remote."
1276 . $remotename .".namespaceCache"));
1278 foreach my $ns (@temp) {
1279 my ($n, $id) = split(/:/, $ns);
1280 $namespace_id{$n} = $id;
1281 $cached_mw_namespace_id{$n} = 1;
1285 if (!exists $namespace_id{$name}) {
1286 print STDERR
"Namespace $name not found in cache, querying the wiki ...\n";
1287 # NS not found => get namespace id from MW and store it in
1288 # configuration file.
1292 siprop
=> 'namespaces'
1294 my $result = $mediawiki->api($query);
1296 while (my ($id, $ns) = each(%{$result->{query
}->{namespaces
}})) {
1297 if (defined($ns->{id
}) && defined($ns->{canonical
})) {
1298 $namespace_id{$ns->{canonical
}} = $ns->{id
};
1300 # alias (e.g. french Fichier: as alias for canonical File:)
1301 $namespace_id{$ns->{'*'}} = $ns->{id
};
1307 my $id = $namespace_id{$name};
1310 # Store explicitely requested namespaces on disk
1311 if (!exists $cached_mw_namespace_id{$name}) {
1312 run_git
("config --add remote.". $remotename
1313 .".namespaceCache \"". $name .":". $id ."\"");
1314 $cached_mw_namespace_id{$name} = 1;
1318 die "No such namespace $name on MediaWiki.";