git-remote-mediawiki: import "File:" attachments
[git/git.git] / contrib / mw-to-git / git-remote-mediawiki
CommitLineData
428c995c
JN
1#! /usr/bin/perl
2
3# Copyright (C) 2011
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
10
11# Gateway between Git and MediaWiki.
12# https://github.com/Bibzball/Git-Mediawiki/wiki
13#
14# Known limitations:
15#
428c995c
JN
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.
22#
23# - Git renames could be turned into MediaWiki renames (see TODO
24# below)
25#
26# - login/password support requires the user to write the password
27# cleartext in a file (see TODO below).
28#
29# - No way to import "one page, and all pages included in it"
30#
31# - Multiple remote MediaWikis have not been very well tested.
32
33use strict;
34use MediaWiki::API;
35use DateTime::Format::ISO8601;
6a9e55b0 36use FileHandle;
428c995c 37
721a533f 38# By default, use UTF-8 to communicate with Git and the user
428c995c 39binmode STDERR, ":utf8";
721a533f 40binmode STDOUT, ":utf8";
428c995c
JN
41
42use URI::Escape;
a7271ad1
MM
43use IPC::Open2;
44
428c995c
JN
45use warnings;
46
47# Mediawiki filenames can contain forward slashes. This variable decides by which pattern they should be replaced
48use constant SLASH_REPLACEMENT => "%2F";
49
50# It's not always possible to delete pages (may require some
51# priviledges). Deleted pages are replaced with this content.
52use constant DELETED_CONTENT => "[[Category:Deleted]]\n";
53
54# It's not possible to create empty pages. New empty files in Git are
55# sent with this content instead.
56use constant EMPTY_CONTENT => "<!-- empty page -->\n";
57
58# used to reflect file creation or deletion in diff.
59use constant NULL_SHA1 => "0000000000000000000000000000000000000000";
60
61my $remotename = $ARGV[0];
62my $url = $ARGV[1];
63
64# Accept both space-separated and multiple keys in config file.
65# Spaces should be written as _ anyway because we'll use chomp.
66my @tracked_pages = split(/[ \n]/, run_git("config --get-all remote.". $remotename .".pages"));
67chomp(@tracked_pages);
68
69# Just like @tracked_pages, but for MediaWiki categories.
70my @tracked_categories = split(/[ \n]/, run_git("config --get-all remote.". $remotename .".categories"));
71chomp(@tracked_categories);
72
6a9e55b0
PV
73# Import media files too.
74my $import_media = run_git("config --get --bool remote.". $remotename .".mediaimport");
75chomp($import_media);
76$import_media = ($import_media eq "true");
77
428c995c
JN
78my $wiki_login = run_git("config --get remote.". $remotename .".mwLogin");
79# TODO: ideally, this should be able to read from keyboard, but we're
80# inside a remote helper, so our stdin is connect to git, not to a
81# terminal.
82my $wiki_passwd = run_git("config --get remote.". $remotename .".mwPassword");
1d6abac9 83my $wiki_domain = run_git("config --get remote.". $remotename .".mwDomain");
428c995c
JN
84chomp($wiki_login);
85chomp($wiki_passwd);
1d6abac9 86chomp($wiki_domain);
428c995c
JN
87
88# Import only last revisions (both for clone and fetch)
89my $shallow_import = run_git("config --get --bool remote.". $remotename .".shallow");
90chomp($shallow_import);
91$shallow_import = ($shallow_import eq "true");
92
6a9e55b0
PV
93# Cache for MediaWiki namespace ids.
94my %namespace_id;
95
93f0d338
MM
96# Dumb push: don't update notes and mediawiki ref to reflect the last push.
97#
98# Configurable with mediawiki.dumbPush, or per-remote with
99# remote.<remotename>.dumbPush.
100#
101# This means the user will have to re-import the just-pushed
102# revisions. On the other hand, this means that the Git revisions
103# corresponding to MediaWiki revisions are all imported from the wiki,
104# regardless of whether they were initially created in Git or from the
105# web interface, hence all users will get the same history (i.e. if
106# the push from Git to MediaWiki loses some information, everybody
107# will get the history with information lost). If the import is
108# deterministic, this means everybody gets the same sha1 for each
109# MediaWiki revision.
110my $dumb_push = run_git("config --get --bool remote.$remotename.dumbPush");
111unless ($dumb_push) {
112 $dumb_push = run_git("config --get --bool mediawiki.dumbPush");
113}
114chomp($dumb_push);
115$dumb_push = ($dumb_push eq "true");
116
428c995c
JN
117my $wiki_name = $url;
118$wiki_name =~ s/[^\/]*:\/\///;
9fb79503
MM
119# If URL is like http://user:password@example.com/, we clearly don't
120# want the password in $wiki_name. While we're there, also remove user
121# and '@' sign, to avoid author like MWUser@HTTPUser@host.com
122$wiki_name =~ s/^.*@//;
428c995c
JN
123
124# Commands parser
125my $entry;
126my @cmd;
127while (<STDIN>) {
128 chomp;
129 @cmd = split(/ /);
130 if (defined($cmd[0])) {
131 # Line not blank
132 if ($cmd[0] eq "capabilities") {
133 die("Too many arguments for capabilities") unless (!defined($cmd[1]));
134 mw_capabilities();
135 } elsif ($cmd[0] eq "list") {
136 die("Too many arguments for list") unless (!defined($cmd[2]));
137 mw_list($cmd[1]);
138 } elsif ($cmd[0] eq "import") {
139 die("Invalid arguments for import") unless ($cmd[1] ne "" && !defined($cmd[2]));
140 mw_import($cmd[1]);
141 } elsif ($cmd[0] eq "option") {
142 die("Too many arguments for option") unless ($cmd[1] ne "" && $cmd[2] ne "" && !defined($cmd[3]));
143 mw_option($cmd[1],$cmd[2]);
144 } elsif ($cmd[0] eq "push") {
145 mw_push($cmd[1]);
146 } else {
147 print STDERR "Unknown command. Aborting...\n";
148 last;
149 }
150 } else {
151 # blank line: we should terminate
152 last;
153 }
154
155 BEGIN { $| = 1 } # flush STDOUT, to make sure the previous
156 # command is fully processed.
157}
158
159########################## Functions ##############################
160
a7271ad1
MM
161## credential API management (generic functions)
162
163sub credential_from_url {
164 my $url = shift;
165 my $parsed = URI->new($url);
166 my %credential;
167
168 if ($parsed->scheme) {
169 $credential{protocol} = $parsed->scheme;
170 }
171 if ($parsed->host) {
172 $credential{host} = $parsed->host;
173 }
174 if ($parsed->path) {
175 $credential{path} = $parsed->path;
176 }
177 if ($parsed->userinfo) {
178 if ($parsed->userinfo =~ /([^:]*):(.*)/) {
179 $credential{username} = $1;
180 $credential{password} = $2;
181 } else {
182 $credential{username} = $parsed->userinfo;
183 }
184 }
185
186 return %credential;
187}
188
189sub credential_read {
190 my %credential;
191 my $reader = shift;
192 my $op = shift;
193 while (<$reader>) {
194 my ($key, $value) = /([^=]*)=(.*)/;
195 if (not defined $key) {
196 die "ERROR receiving response from git credential $op:\n$_\n";
197 }
198 $credential{$key} = $value;
199 }
200 return %credential;
201}
202
203sub credential_write {
204 my $credential = shift;
205 my $writer = shift;
206 while (my ($key, $value) = each(%$credential) ) {
207 if ($value) {
208 print $writer "$key=$value\n";
209 }
210 }
211}
212
213sub credential_run {
214 my $op = shift;
215 my $credential = shift;
216 my $pid = open2(my $reader, my $writer, "git credential $op");
217 credential_write($credential, $writer);
218 print $writer "\n";
219 close($writer);
220
221 if ($op eq "fill") {
222 %$credential = credential_read($reader, $op);
223 } else {
224 if (<$reader>) {
225 die "ERROR while running git credential $op:\n$_";
226 }
227 }
228 close($reader);
229 waitpid($pid, 0);
230 my $child_exit_status = $? >> 8;
231 if ($child_exit_status != 0) {
232 die "'git credential $op' failed with code $child_exit_status.";
233 }
234}
235
428c995c
JN
236# MediaWiki API instance, created lazily.
237my $mediawiki;
238
239sub mw_connect_maybe {
240 if ($mediawiki) {
a7271ad1 241 return;
428c995c
JN
242 }
243 $mediawiki = MediaWiki::API->new;
244 $mediawiki->{config}->{api_url} = "$url/api.php";
245 if ($wiki_login) {
a7271ad1
MM
246 my %credential = credential_from_url($url);
247 $credential{username} = $wiki_login;
248 $credential{password} = $wiki_passwd;
249 credential_run("fill", \%credential);
250 my $request = {lgname => $credential{username},
251 lgpassword => $credential{password},
252 lgdomain => $wiki_domain};
253 if ($mediawiki->login($request)) {
254 credential_run("approve", \%credential);
255 print STDERR "Logged in mediawiki user \"$credential{username}\".\n";
428c995c 256 } else {
a7271ad1
MM
257 print STDERR "Failed to log in mediawiki user \"$credential{username}\" on $url\n";
258 print STDERR " (error " .
259 $mediawiki->{error}->{code} . ': ' .
260 $mediawiki->{error}->{details} . ")\n";
261 credential_run("reject", \%credential);
262 exit 1;
428c995c
JN
263 }
264 }
265}
266
9cb74f36
MM
267## Functions for listing pages on the remote wiki
268sub get_mw_tracked_pages {
269 my $pages = shift;
270 my @some_pages = @tracked_pages;
271 while (@some_pages) {
272 my $last = 50;
273 if ($#some_pages < $last) {
274 $last = $#some_pages;
275 }
276 my @slice = @some_pages[0..$last];
277 get_mw_first_pages(\@slice, $pages);
278 @some_pages = @some_pages[51..$#some_pages];
279 }
280}
281
282sub get_mw_tracked_categories {
283 my $pages = shift;
284 foreach my $category (@tracked_categories) {
285 if (index($category, ':') < 0) {
286 # Mediawiki requires the Category
287 # prefix, but let's not force the user
288 # to specify it.
289 $category = "Category:" . $category;
290 }
291 my $mw_pages = $mediawiki->list( {
292 action => 'query',
293 list => 'categorymembers',
294 cmtitle => $category,
295 cmlimit => 'max' } )
296 || die $mediawiki->{error}->{code} . ': '
297 . $mediawiki->{error}->{details};
298 foreach my $page (@{$mw_pages}) {
299 $pages->{$page->{title}} = $page;
300 }
301 }
302}
303
304sub get_mw_all_pages {
305 my $pages = shift;
306 # No user-provided list, get the list of pages from the API.
307 my $mw_pages = $mediawiki->list({
308 action => 'query',
309 list => 'allpages',
310 aplimit => 'max'
311 });
312 if (!defined($mw_pages)) {
313 print STDERR "fatal: could not get the list of wiki pages.\n";
314 print STDERR "fatal: '$url' does not appear to be a mediawiki\n";
315 print STDERR "fatal: make sure '$url/api.php' is a valid page.\n";
316 exit 1;
317 }
318 foreach my $page (@{$mw_pages}) {
319 $pages->{$page->{title}} = $page;
320 }
321}
322
323# queries the wiki for a set of pages. Meant to be used within a loop
324# querying the wiki for slices of page list.
428c995c
JN
325sub get_mw_first_pages {
326 my $some_pages = shift;
327 my @some_pages = @{$some_pages};
328
329 my $pages = shift;
330
331 # pattern 'page1|page2|...' required by the API
332 my $titles = join('|', @some_pages);
333
334 my $mw_pages = $mediawiki->api({
335 action => 'query',
336 titles => $titles,
337 });
338 if (!defined($mw_pages)) {
339 print STDERR "fatal: could not query the list of wiki pages.\n";
340 print STDERR "fatal: '$url' does not appear to be a mediawiki\n";
341 print STDERR "fatal: make sure '$url/api.php' is a valid page.\n";
342 exit 1;
343 }
344 while (my ($id, $page) = each(%{$mw_pages->{query}->{pages}})) {
345 if ($id < 0) {
346 print STDERR "Warning: page $page->{title} not found on wiki\n";
347 } else {
348 $pages->{$page->{title}} = $page;
349 }
350 }
351}
352
9cb74f36 353# Get the list of pages to be fetched according to configuration.
428c995c
JN
354sub get_mw_pages {
355 mw_connect_maybe();
356
357 my %pages; # hash on page titles to avoid duplicates
358 my $user_defined;
359 if (@tracked_pages) {
360 $user_defined = 1;
361 # The user provided a list of pages titles, but we
362 # still need to query the API to get the page IDs.
9cb74f36 363 get_mw_tracked_pages(\%pages);
428c995c
JN
364 }
365 if (@tracked_categories) {
366 $user_defined = 1;
9cb74f36 367 get_mw_tracked_categories(\%pages);
428c995c
JN
368 }
369 if (!$user_defined) {
9cb74f36 370 get_mw_all_pages(\%pages);
428c995c 371 }
6a9e55b0
PV
372 if ($import_media) {
373 print STDERR "Getting media files for selected pages...\n";
374 if ($user_defined) {
375 get_linked_mediafiles(\%pages);
376 } else {
377 get_all_mediafiles(\%pages);
378 }
379 }
428c995c
JN
380 return values(%pages);
381}
382
b3d98595
NKT
383# usage: $out = run_git("command args");
384# $out = run_git("command args", "raw"); # don't interpret output as UTF-8.
428c995c 385sub run_git {
b3d98595
NKT
386 my $args = shift;
387 my $encoding = (shift || "encoding(UTF-8)");
388 open(my $git, "-|:$encoding", "git " . $args);
428c995c
JN
389 my $res = do { local $/; <$git> };
390 close($git);
391
392 return $res;
393}
394
395
6a9e55b0
PV
396sub get_all_mediafiles {
397 my $pages = shift;
398 # Attach list of all pages for media files from the API,
399 # they are in a different namespace, only one namespace
400 # can be queried at the same moment
401 my $mw_pages = $mediawiki->list({
402 action => 'query',
403 list => 'allpages',
404 apnamespace => get_mw_namespace_id("File"),
405 aplimit => 'max'
406 });
407 if (!defined($mw_pages)) {
408 print STDERR "fatal: could not get the list of pages for media files.\n";
409 print STDERR "fatal: '$url' does not appear to be a mediawiki\n";
410 print STDERR "fatal: make sure '$url/api.php' is a valid page.\n";
411 exit 1;
412 }
413 foreach my $page (@{$mw_pages}) {
414 $pages->{$page->{title}} = $page;
415 }
416}
417
418sub get_linked_mediafiles {
419 my $pages = shift;
420 my @titles = map $_->{title}, values(%{$pages});
421
422 # The query is split in small batches because of the MW API limit of
423 # the number of links to be returned (500 links max).
424 my $batch = 10;
425 while (@titles) {
426 if ($#titles < $batch) {
427 $batch = $#titles;
428 }
429 my @slice = @titles[0..$batch];
430
431 # pattern 'page1|page2|...' required by the API
432 my $mw_titles = join('|', @slice);
433
434 # Media files could be included or linked from
435 # a page, get all related
436 my $query = {
437 action => 'query',
438 prop => 'links|images',
439 titles => $mw_titles,
440 plnamespace => get_mw_namespace_id("File"),
441 pllimit => 'max'
442 };
443 my $result = $mediawiki->api($query);
444
445 while (my ($id, $page) = each(%{$result->{query}->{pages}})) {
446 my @titles;
447 if (defined($page->{links})) {
448 my @link_titles = map $_->{title}, @{$page->{links}};
449 push(@titles, @link_titles);
450 }
451 if (defined($page->{images})) {
452 my @image_titles = map $_->{title}, @{$page->{images}};
453 push(@titles, @image_titles);
454 }
455 if (@titles) {
456 get_mw_first_pages(\@titles, \%{$pages});
457 }
458 }
459
460 @titles = @titles[($batch+1)..$#titles];
461 }
462}
463
464sub get_mw_mediafile_for_page_revision {
465 # Name of the file on Wiki, with the prefix.
466 my $mw_filename = shift;
467 my $timestamp = shift;
468 my %mediafile;
469
470 # Search if on MediaWiki exists a media file with given
471 # timestamp. In that case download the file.
472 my $query = {
473 action => 'query',
474 prop => 'imageinfo',
475 titles => $mw_filename,
476 iistart => $timestamp,
477 iiend => $timestamp,
478 iiprop => 'timestamp|archivename|url',
479 iilimit => 1
480 };
481 my $result = $mediawiki->api($query);
482
483 my ($fileid, $file) = each ( %{$result->{query}->{pages}} );
484 # If not defined it means there is no revision of the file for
485 # given timestamp.
486 if (defined($file->{imageinfo})) {
487 # Get real name of media file.
488 my $filename;
489 if (index($mw_filename, 'File:') == 0) {
490 $filename = substr $mw_filename, 5;
491 } else {
492 $filename = substr $mw_filename, 6;
493 }
494 $mediafile{title} = $filename;
495
496 my $fileinfo = pop(@{$file->{imageinfo}});
497 $mediafile{timestamp} = $fileinfo->{timestamp};
498 # If this is an old version of the file, the file has to be
499 # obtained from the archive. Otherwise it can be downloaded
500 # by MediaWiki API download() function.
501 if (defined($fileinfo->{archivename})) {
502 $mediafile{content} = download_mw_mediafile_from_archive($fileinfo->{url});
503 } else {
504 $mediafile{content} = download_mw_mediafile($mw_filename);
505 }
506 }
507 return %mediafile;
508}
509
510sub download_mw_mediafile_from_archive {
511 my $url = shift;
512 my $file;
513
514 my $ua = LWP::UserAgent->new;
515 my $response = $ua->get($url);
516 if ($response->code) {
517 $file = $response->decoded_content;
518 } else {
519 print STDERR "Error downloading a file from archive.\n";
520 }
521
522 return $file;
523}
524
525sub download_mw_mediafile {
526 my $filename = shift;
527
528 $mediawiki->{config}->{files_url} = $url;
529
530 my $file_content = $mediawiki->download( { title => $filename } );
531 if (!defined($file_content)) {
532 print STDERR "\tFile \'$filename\' could not be downloaded.\n";
533 exit 1;
534 } elsif ($file_content eq "") {
535 print STDERR "\tFile \'$filename\' does not exist on the wiki.\n";
536 exit 1;
537 } else {
538 return $file_content;
539 }
540}
541
428c995c
JN
542sub get_last_local_revision {
543 # Get note regarding last mediawiki revision
544 my $note = run_git("notes --ref=$remotename/mediawiki show refs/mediawiki/$remotename/master 2>/dev/null");
545 my @note_info = split(/ /, $note);
546
547 my $lastrevision_number;
548 if (!(defined($note_info[0]) && $note_info[0] eq "mediawiki_revision:")) {
549 print STDERR "No previous mediawiki revision found";
550 $lastrevision_number = 0;
551 } else {
552 # Notes are formatted : mediawiki_revision: #number
553 $lastrevision_number = $note_info[1];
554 chomp($lastrevision_number);
555 print STDERR "Last local mediawiki revision found is $lastrevision_number";
556 }
557 return $lastrevision_number;
558}
559
3c1ed90e
MM
560# Remember the timestamp corresponding to a revision id.
561my %basetimestamps;
562
428c995c
JN
563sub get_last_remote_revision {
564 mw_connect_maybe();
565
566 my @pages = get_mw_pages();
567
568 my $max_rev_num = 0;
569
570 foreach my $page (@pages) {
571 my $id = $page->{pageid};
572
573 my $query = {
574 action => 'query',
575 prop => 'revisions',
3c1ed90e 576 rvprop => 'ids|timestamp',
428c995c
JN
577 pageids => $id,
578 };
579
580 my $result = $mediawiki->api($query);
581
582 my $lastrev = pop(@{$result->{query}->{pages}->{$id}->{revisions}});
583
3c1ed90e
MM
584 $basetimestamps{$lastrev->{revid}} = $lastrev->{timestamp};
585
428c995c
JN
586 $max_rev_num = ($lastrev->{revid} > $max_rev_num ? $lastrev->{revid} : $max_rev_num);
587 }
588
589 print STDERR "Last remote revision found is $max_rev_num.\n";
590 return $max_rev_num;
591}
592
593# Clean content before sending it to MediaWiki
594sub mediawiki_clean {
595 my $string = shift;
596 my $page_created = shift;
597 # Mediawiki does not allow blank space at the end of a page and ends with a single \n.
598 # This function right trims a string and adds a \n at the end to follow this rule
599 $string =~ s/\s+$//;
600 if ($string eq "" && $page_created) {
601 # Creating empty pages is forbidden.
602 $string = EMPTY_CONTENT;
603 }
604 return $string."\n";
605}
606
607# Filter applied on MediaWiki data before adding them to Git
608sub mediawiki_smudge {
609 my $string = shift;
610 if ($string eq EMPTY_CONTENT) {
611 $string = "";
612 }
613 # This \n is important. This is due to mediawiki's way to handle end of files.
614 return $string."\n";
615}
616
617sub mediawiki_clean_filename {
618 my $filename = shift;
619 $filename =~ s/@{[SLASH_REPLACEMENT]}/\//g;
620 # [, ], |, {, and } are forbidden by MediaWiki, even URL-encoded.
621 # Do a variant of URL-encoding, i.e. looks like URL-encoding,
622 # but with _ added to prevent MediaWiki from thinking this is
623 # an actual special character.
624 $filename =~ s/[\[\]\{\}\|]/sprintf("_%%_%x", ord($&))/ge;
625 # If we use the uri escape before
626 # we should unescape here, before anything
627
628 return $filename;
629}
630
631sub mediawiki_smudge_filename {
632 my $filename = shift;
633 $filename =~ s/\//@{[SLASH_REPLACEMENT]}/g;
634 $filename =~ s/ /_/g;
635 # Decode forbidden characters encoded in mediawiki_clean_filename
636 $filename =~ s/_%_([0-9a-fA-F][0-9a-fA-F])/sprintf("%c", hex($1))/ge;
637 return $filename;
638}
639
640sub literal_data {
641 my ($content) = @_;
642 print STDOUT "data ", bytes::length($content), "\n", $content;
643}
644
6a9e55b0
PV
645sub literal_data_raw {
646 # Output possibly binary content.
647 my ($content) = @_;
648 # Avoid confusion between size in bytes and in characters
649 utf8::downgrade($content);
650 binmode STDOUT, ":raw";
651 print STDOUT "data ", bytes::length($content), "\n", $content;
652 binmode STDOUT, ":utf8";
653}
654
428c995c
JN
655sub mw_capabilities {
656 # Revisions are imported to the private namespace
657 # refs/mediawiki/$remotename/ by the helper and fetched into
658 # refs/remotes/$remotename later by fetch.
659 print STDOUT "refspec refs/heads/*:refs/mediawiki/$remotename/*\n";
660 print STDOUT "import\n";
661 print STDOUT "list\n";
662 print STDOUT "push\n";
663 print STDOUT "\n";
664}
665
666sub mw_list {
667 # MediaWiki do not have branches, we consider one branch arbitrarily
668 # called master, and HEAD pointing to it.
669 print STDOUT "? refs/heads/master\n";
670 print STDOUT "\@refs/heads/master HEAD\n";
671 print STDOUT "\n";
672}
673
674sub mw_option {
675 print STDERR "remote-helper command 'option $_[0]' not yet implemented\n";
676 print STDOUT "unsupported\n";
677}
678
679sub fetch_mw_revisions_for_page {
680 my $page = shift;
681 my $id = shift;
682 my $fetch_from = shift;
683 my @page_revs = ();
684 my $query = {
685 action => 'query',
686 prop => 'revisions',
687 rvprop => 'ids',
688 rvdir => 'newer',
689 rvstartid => $fetch_from,
690 rvlimit => 500,
691 pageids => $id,
692 };
693
694 my $revnum = 0;
695 # Get 500 revisions at a time due to the mediawiki api limit
696 while (1) {
697 my $result = $mediawiki->api($query);
698
699 # Parse each of those 500 revisions
700 foreach my $revision (@{$result->{query}->{pages}->{$id}->{revisions}}) {
701 my $page_rev_ids;
702 $page_rev_ids->{pageid} = $page->{pageid};
703 $page_rev_ids->{revid} = $revision->{revid};
704 push(@page_revs, $page_rev_ids);
705 $revnum++;
706 }
707 last unless $result->{'query-continue'};
708 $query->{rvstartid} = $result->{'query-continue'}->{revisions}->{rvstartid};
709 }
710 if ($shallow_import && @page_revs) {
711 print STDERR " Found 1 revision (shallow import).\n";
712 @page_revs = sort {$b->{revid} <=> $a->{revid}} (@page_revs);
713 return $page_revs[0];
714 }
715 print STDERR " Found ", $revnum, " revision(s).\n";
716 return @page_revs;
717}
718
719sub fetch_mw_revisions {
720 my $pages = shift; my @pages = @{$pages};
721 my $fetch_from = shift;
722
723 my @revisions = ();
724 my $n = 1;
725 foreach my $page (@pages) {
726 my $id = $page->{pageid};
727
728 print STDERR "page $n/", scalar(@pages), ": ". $page->{title} ."\n";
729 $n++;
730 my @page_revs = fetch_mw_revisions_for_page($page, $id, $fetch_from);
731 @revisions = (@page_revs, @revisions);
732 }
733
734 return ($n, @revisions);
735}
736
737sub import_file_revision {
738 my $commit = shift;
739 my %commit = %{$commit};
740 my $full_import = shift;
741 my $n = shift;
6a9e55b0
PV
742 my $mediafile = shift;
743 my %mediafile;
744 if ($mediafile) {
745 %mediafile = %{$mediafile};
746 }
428c995c
JN
747
748 my $title = $commit{title};
749 my $comment = $commit{comment};
750 my $content = $commit{content};
751 my $author = $commit{author};
752 my $date = $commit{date};
753
754 print STDOUT "commit refs/mediawiki/$remotename/master\n";
755 print STDOUT "mark :$n\n";
756 print STDOUT "committer $author <$author\@$wiki_name> ", $date->epoch, " +0000\n";
757 literal_data($comment);
758
759 # If it's not a clone, we need to know where to start from
760 if (!$full_import && $n == 1) {
761 print STDOUT "from refs/mediawiki/$remotename/master^0\n";
762 }
763 if ($content ne DELETED_CONTENT) {
764 print STDOUT "M 644 inline $title.mw\n";
765 literal_data($content);
6a9e55b0
PV
766 if (%mediafile) {
767 print STDOUT "M 644 inline $mediafile{title}\n";
768 literal_data_raw($mediafile{content});
769 }
428c995c
JN
770 print STDOUT "\n\n";
771 } else {
772 print STDOUT "D $title.mw\n";
773 }
774
775 # mediawiki revision number in the git note
776 if ($full_import && $n == 1) {
777 print STDOUT "reset refs/notes/$remotename/mediawiki\n";
778 }
779 print STDOUT "commit refs/notes/$remotename/mediawiki\n";
780 print STDOUT "committer $author <$author\@$wiki_name> ", $date->epoch, " +0000\n";
781 literal_data("Note added by git-mediawiki during import");
782 if (!$full_import && $n == 1) {
783 print STDOUT "from refs/notes/$remotename/mediawiki^0\n";
784 }
785 print STDOUT "N inline :$n\n";
786 literal_data("mediawiki_revision: " . $commit{mw_revision});
787 print STDOUT "\n\n";
788}
789
790# parse a sequence of
791# <cmd> <arg1>
792# <cmd> <arg2>
793# \n
794# (like batch sequence of import and sequence of push statements)
795sub get_more_refs {
796 my $cmd = shift;
797 my @refs;
798 while (1) {
799 my $line = <STDIN>;
800 if ($line =~ m/^$cmd (.*)$/) {
801 push(@refs, $1);
802 } elsif ($line eq "\n") {
803 return @refs;
804 } else {
805 die("Invalid command in a '$cmd' batch: ". $_);
806 }
807 }
808}
809
810sub mw_import {
811 # multiple import commands can follow each other.
812 my @refs = (shift, get_more_refs("import"));
813 foreach my $ref (@refs) {
814 mw_import_ref($ref);
815 }
816 print STDOUT "done\n";
817}
818
819sub mw_import_ref {
820 my $ref = shift;
821 # The remote helper will call "import HEAD" and
822 # "import refs/heads/master".
823 # Since HEAD is a symbolic ref to master (by convention,
824 # followed by the output of the command "list" that we gave),
825 # we don't need to do anything in this case.
826 if ($ref eq "HEAD") {
827 return;
828 }
829
830 mw_connect_maybe();
831
832 my @pages = get_mw_pages();
833
834 print STDERR "Searching revisions...\n";
835 my $last_local = get_last_local_revision();
836 my $fetch_from = $last_local + 1;
837 if ($fetch_from == 1) {
838 print STDERR ", fetching from beginning.\n";
839 } else {
840 print STDERR ", fetching from here.\n";
841 }
842 my ($n, @revisions) = fetch_mw_revisions(\@pages, $fetch_from);
843
844 # Creation of the fast-import stream
845 print STDERR "Fetching & writing export data...\n";
846
847 $n = 0;
848 my $last_timestamp = 0; # Placeholer in case $rev->timestamp is undefined
849
850 foreach my $pagerevid (sort {$a->{revid} <=> $b->{revid}} @revisions) {
851 # fetch the content of the pages
852 my $query = {
853 action => 'query',
854 prop => 'revisions',
855 rvprop => 'content|timestamp|comment|user|ids',
856 revids => $pagerevid->{revid},
857 };
858
859 my $result = $mediawiki->api($query);
860
861 my $rev = pop(@{$result->{query}->{pages}->{$pagerevid->{pageid}}->{revisions}});
862
863 $n++;
864
6a9e55b0 865 my $page_title = $result->{query}->{pages}->{$pagerevid->{pageid}}->{title};
428c995c
JN
866 my %commit;
867 $commit{author} = $rev->{user} || 'Anonymous';
868 $commit{comment} = $rev->{comment} || '*Empty MediaWiki Message*';
6a9e55b0 869 $commit{title} = mediawiki_smudge_filename($page_title);
428c995c
JN
870 $commit{mw_revision} = $pagerevid->{revid};
871 $commit{content} = mediawiki_smudge($rev->{'*'});
872
873 if (!defined($rev->{timestamp})) {
874 $last_timestamp++;
875 } else {
876 $last_timestamp = $rev->{timestamp};
877 }
878 $commit{date} = DateTime::Format::ISO8601->parse_datetime($last_timestamp);
879
6a9e55b0
PV
880 # Differentiates classic pages and media files.
881 my @prefix = split(":", $page_title);
428c995c 882
6a9e55b0
PV
883 my %mediafile;
884 if ($prefix[0] eq "File" || $prefix[0] eq "Image") {
885 # The name of the file is the same as the media page.
886 my $filename = $page_title;
887 %mediafile = get_mw_mediafile_for_page_revision($filename, $rev->{timestamp});
888 }
889 # If this is a revision of the media page for new version
890 # of a file do one common commit for both file and media page.
891 # Else do commit only for that page.
892 print STDERR "$n/", scalar(@revisions), ": Revision #$pagerevid->{revid} of $commit{title}\n";
893 if (%mediafile) {
894 print STDERR "\tDownloading file $mediafile{title}, version $mediafile{timestamp}\n";
895 import_file_revision(\%commit, ($fetch_from == 1), $n, \%mediafile);
896 } else {
897 import_file_revision(\%commit, ($fetch_from == 1), $n);
898 }
428c995c
JN
899 }
900
901 if ($fetch_from == 1 && $n == 0) {
902 print STDERR "You appear to have cloned an empty MediaWiki.\n";
903 # Something has to be done remote-helper side. If nothing is done, an error is
904 # thrown saying that HEAD is refering to unknown object 0000000000000000000
905 # and the clone fails.
906 }
907}
908
909sub error_non_fast_forward {
fd47d7b9
MM
910 my $advice = run_git("config --bool advice.pushNonFastForward");
911 chomp($advice);
912 if ($advice ne "false") {
913 # Native git-push would show this after the summary.
914 # We can't ask it to display it cleanly, so print it
915 # ourselves before.
916 print STDERR "To prevent you from losing history, non-fast-forward updates were rejected\n";
917 print STDERR "Merge the remote changes (e.g. 'git pull') before pushing again. See the\n";
918 print STDERR "'Note about fast-forwards' section of 'git push --help' for details.\n";
919 }
428c995c
JN
920 print STDOUT "error $_[0] \"non-fast-forward\"\n";
921 return 0;
922}
923
b3d98595
NKT
924sub mw_upload_file {
925 my $complete_file_name = shift;
926 my $new_sha1 = shift;
927 my $extension = shift;
928 my $file_deleted = shift;
929 my $summary = shift;
930 my $newrevid;
931 my $path = "File:" . $complete_file_name;
932 my %hashFiles = get_allowed_file_extensions();
933 if (!exists($hashFiles{$extension})) {
934 print STDERR "$complete_file_name is not a permitted file on this wiki.\n";
935 print STDERR "Check the configuration of file uploads in your mediawiki.\n";
936 return $newrevid;
937 }
938 # Deleting and uploading a file requires a priviledged user
939 if ($file_deleted) {
940 mw_connect_maybe();
941 my $query = {
942 action => 'delete',
943 title => $path,
944 reason => $summary
945 };
946 if (!$mediawiki->edit($query)) {
947 print STDERR "Failed to delete file on remote wiki\n";
948 print STDERR "Check your permissions on the remote site. Error code:\n";
949 print STDERR $mediawiki->{error}->{code} . ':' . $mediawiki->{error}->{details};
950 exit 1;
951 }
952 } else {
953 # Don't let perl try to interpret file content as UTF-8 => use "raw"
954 my $content = run_git("cat-file blob $new_sha1", "raw");
955 if ($content ne "") {
956 mw_connect_maybe();
957 $mediawiki->{config}->{upload_url} =
958 "$url/index.php/Special:Upload";
959 $mediawiki->edit({
960 action => 'upload',
961 filename => $complete_file_name,
962 comment => $summary,
963 file => [undef,
964 $complete_file_name,
965 Content => $content],
966 ignorewarnings => 1,
967 }, {
968 skip_encoding => 1
969 } ) || die $mediawiki->{error}->{code} . ':'
970 . $mediawiki->{error}->{details};
971 my $last_file_page = $mediawiki->get_page({title => $path});
972 $newrevid = $last_file_page->{revid};
973 print STDERR "Pushed file: $new_sha1 - $complete_file_name.\n";
974 } else {
975 print STDERR "Empty file $complete_file_name not pushed.\n";
976 }
977 }
978 return $newrevid;
979}
980
428c995c
JN
981sub mw_push_file {
982 my $diff_info = shift;
983 # $diff_info contains a string in this format:
984 # 100644 100644 <sha1_of_blob_before_commit> <sha1_of_blob_now> <status>
985 my @diff_info_split = split(/[ \t]/, $diff_info);
986
987 # Filename, including .mw extension
988 my $complete_file_name = shift;
989 # Commit message
990 my $summary = shift;
93f0d338
MM
991 # MediaWiki revision number. Keep the previous one by default,
992 # in case there's no edit to perform.
b3d98595
NKT
993 my $oldrevid = shift;
994 my $newrevid;
428c995c
JN
995
996 my $new_sha1 = $diff_info_split[3];
997 my $old_sha1 = $diff_info_split[2];
998 my $page_created = ($old_sha1 eq NULL_SHA1);
999 my $page_deleted = ($new_sha1 eq NULL_SHA1);
1000 $complete_file_name = mediawiki_clean_filename($complete_file_name);
1001
b3d98595
NKT
1002 my ($title, $extension) = $complete_file_name =~ /^(.*)\.([^\.]*)$/;
1003 if (!defined($extension)) {
1004 $extension = "";
1005 }
1006 if ($extension eq "mw") {
428c995c
JN
1007 my $file_content;
1008 if ($page_deleted) {
1009 # Deleting a page usually requires
1010 # special priviledges. A common
1011 # convention is to replace the page
1012 # with this content instead:
1013 $file_content = DELETED_CONTENT;
1014 } else {
1015 $file_content = run_git("cat-file blob $new_sha1");
1016 }
1017
1018 mw_connect_maybe();
1019
1020 my $result = $mediawiki->edit( {
1021 action => 'edit',
1022 summary => $summary,
1023 title => $title,
b3d98595 1024 basetimestamp => $basetimestamps{$oldrevid},
428c995c
JN
1025 text => mediawiki_clean($file_content, $page_created),
1026 }, {
1027 skip_encoding => 1 # Helps with names with accentuated characters
3c1ed90e
MM
1028 });
1029 if (!$result) {
1030 if ($mediawiki->{error}->{code} == 3) {
1031 # edit conflicts, considered as non-fast-forward
1032 print STDERR 'Warning: Error ' .
1033 $mediawiki->{error}->{code} .
1034 ' from mediwiki: ' . $mediawiki->{error}->{details} .
1035 ".\n";
b3d98595 1036 return ($oldrevid, "non-fast-forward");
3c1ed90e
MM
1037 } else {
1038 # Other errors. Shouldn't happen => just die()
1039 die 'Fatal: Error ' .
1040 $mediawiki->{error}->{code} .
1041 ' from mediwiki: ' . $mediawiki->{error}->{details};
1042 }
1043 }
93f0d338 1044 $newrevid = $result->{edit}->{newrevid};
ac86ec0f 1045 print STDERR "Pushed file: $new_sha1 - $title\n";
428c995c 1046 } else {
b3d98595
NKT
1047 $newrevid = mw_upload_file($complete_file_name, $new_sha1,
1048 $extension, $page_deleted,
1049 $summary);
428c995c 1050 }
b3d98595 1051 $newrevid = ($newrevid or $oldrevid);
3c1ed90e 1052 return ($newrevid, "ok");
428c995c
JN
1053}
1054
1055sub mw_push {
1056 # multiple push statements can follow each other
1057 my @refsspecs = (shift, get_more_refs("push"));
428c995c
JN
1058 my $pushed;
1059 for my $refspec (@refsspecs) {
1060 my ($force, $local, $remote) = $refspec =~ /^(\+)?([^:]*):([^:]*)$/
1061 or die("Invalid refspec for push. Expected <src>:<dst> or +<src>:<dst>");
1062 if ($force) {
1063 print STDERR "Warning: forced push not allowed on a MediaWiki.\n";
1064 }
1065 if ($local eq "") {
1066 print STDERR "Cannot delete remote branch on a MediaWiki\n";
1067 print STDOUT "error $remote cannot delete\n";
1068 next;
1069 }
1070 if ($remote ne "refs/heads/master") {
1071 print STDERR "Only push to the branch 'master' is supported on a MediaWiki\n";
1072 print STDOUT "error $remote only master allowed\n";
1073 next;
1074 }
1075 if (mw_push_revision($local, $remote)) {
1076 $pushed = 1;
1077 }
1078 }
1079
1080 # Notify Git that the push is done
1081 print STDOUT "\n";
1082
93f0d338 1083 if ($pushed && $dumb_push) {
428c995c
JN
1084 print STDERR "Just pushed some revisions to MediaWiki.\n";
1085 print STDERR "The pushed revisions now have to be re-imported, and your current branch\n";
1086 print STDERR "needs to be updated with these re-imported commits. You can do this with\n";
1087 print STDERR "\n";
1088 print STDERR " git pull --rebase\n";
1089 print STDERR "\n";
1090 }
1091}
1092
1093sub mw_push_revision {
1094 my $local = shift;
1095 my $remote = shift; # actually, this has to be "refs/heads/master" at this point.
1096 my $last_local_revid = get_last_local_revision();
1097 print STDERR ".\n"; # Finish sentence started by get_last_local_revision()
1098 my $last_remote_revid = get_last_remote_revision();
93f0d338 1099 my $mw_revision = $last_remote_revid;
428c995c
JN
1100
1101 # Get sha1 of commit pointed by local HEAD
1102 my $HEAD_sha1 = run_git("rev-parse $local 2>/dev/null"); chomp($HEAD_sha1);
1103 # Get sha1 of commit pointed by remotes/$remotename/master
1104 my $remoteorigin_sha1 = run_git("rev-parse refs/remotes/$remotename/master 2>/dev/null");
1105 chomp($remoteorigin_sha1);
1106
1107 if ($last_local_revid > 0 &&
1108 $last_local_revid < $last_remote_revid) {
1109 return error_non_fast_forward($remote);
1110 }
1111
1112 if ($HEAD_sha1 eq $remoteorigin_sha1) {
1113 # nothing to push
1114 return 0;
1115 }
1116
1117 # Get every commit in between HEAD and refs/remotes/origin/master,
1118 # including HEAD and refs/remotes/origin/master
1119 my @commit_pairs = ();
1120 if ($last_local_revid > 0) {
1121 my $parsed_sha1 = $remoteorigin_sha1;
1122 # Find a path from last MediaWiki commit to pushed commit
1123 while ($parsed_sha1 ne $HEAD_sha1) {
1124 my @commit_info = grep(/^$parsed_sha1/, split(/\n/, run_git("rev-list --children $local")));
1125 if (!@commit_info) {
1126 return error_non_fast_forward($remote);
1127 }
1128 my @commit_info_split = split(/ |\n/, $commit_info[0]);
1129 # $commit_info_split[1] is the sha1 of the commit to export
1130 # $commit_info_split[0] is the sha1 of its direct child
1131 push(@commit_pairs, \@commit_info_split);
1132 $parsed_sha1 = $commit_info_split[1];
1133 }
1134 } else {
1135 # No remote mediawiki revision. Export the whole
1136 # history (linearized with --first-parent)
1137 print STDERR "Warning: no common ancestor, pushing complete history\n";
1138 my $history = run_git("rev-list --first-parent --children $local");
1139 my @history = split('\n', $history);
1140 @history = @history[1..$#history];
1141 foreach my $line (reverse @history) {
1142 my @commit_info_split = split(/ |\n/, $line);
1143 push(@commit_pairs, \@commit_info_split);
1144 }
1145 }
1146
1147 foreach my $commit_info_split (@commit_pairs) {
1148 my $sha1_child = @{$commit_info_split}[0];
1149 my $sha1_commit = @{$commit_info_split}[1];
1150 my $diff_infos = run_git("diff-tree -r --raw -z $sha1_child $sha1_commit");
1151 # TODO: we could detect rename, and encode them with a #redirect on the wiki.
1152 # TODO: for now, it's just a delete+add
1153 my @diff_info_list = split(/\0/, $diff_infos);
28c24bd7
MM
1154 # Keep the subject line of the commit message as mediawiki comment for the revision
1155 my $commit_msg = run_git("log --no-walk --format=\"%s\" $sha1_commit");
428c995c
JN
1156 chomp($commit_msg);
1157 # Push every blob
1158 while (@diff_info_list) {
3c1ed90e 1159 my $status;
428c995c
JN
1160 # git diff-tree -z gives an output like
1161 # <metadata>\0<filename1>\0
1162 # <metadata>\0<filename2>\0
1163 # and we've split on \0.
1164 my $info = shift(@diff_info_list);
1165 my $file = shift(@diff_info_list);
3c1ed90e
MM
1166 ($mw_revision, $status) = mw_push_file($info, $file, $commit_msg, $mw_revision);
1167 if ($status eq "non-fast-forward") {
1168 # we may already have sent part of the
1169 # commit to MediaWiki, but it's too
1170 # late to cancel it. Stop the push in
1171 # the middle, but still give an
1172 # accurate error message.
1173 return error_non_fast_forward($remote);
1174 }
1175 if ($status ne "ok") {
1176 die("Unknown error from mw_push_file()");
1177 }
93f0d338
MM
1178 }
1179 unless ($dumb_push) {
1180 run_git("notes --ref=$remotename/mediawiki add -m \"mediawiki_revision: $mw_revision\" $sha1_commit");
1181 run_git("update-ref -m \"Git-MediaWiki push\" refs/mediawiki/$remotename/master $sha1_commit $sha1_child");
428c995c
JN
1182 }
1183 }
1184
1185 print STDOUT "ok $remote\n";
1186 return 1;
1187}
b3d98595
NKT
1188
1189sub get_allowed_file_extensions {
1190 mw_connect_maybe();
1191
1192 my $query = {
1193 action => 'query',
1194 meta => 'siteinfo',
1195 siprop => 'fileextensions'
1196 };
1197 my $result = $mediawiki->api($query);
1198 my @file_extensions= map $_->{ext},@{$result->{query}->{fileextensions}};
1199 my %hashFile = map {$_ => 1}@file_extensions;
1200
1201 return %hashFile;
1202}
6a9e55b0
PV
1203
1204# Return MediaWiki id for a canonical namespace name.
1205# Ex.: "File", "Project".
1206# Looks for the namespace id in the local configuration
1207# variables, if it is not found asks MW API.
1208sub get_mw_namespace_id {
1209 mw_connect_maybe();
1210 my $name = shift;
1211
1212 if (!exists $namespace_id{$name}) {
1213 # Look at configuration file, if the record for that namespace is
1214 # already stored. Namespaces are stored in form:
1215 # "Name_of_namespace:Id_namespace", ex.: "File:6".
1216 my @temp = split(/[ \n]/, run_git("config --get-all remote."
1217 . $remotename .".namespaces"));
1218 chomp(@temp);
1219 foreach my $ns (@temp) {
1220 my ($n, $s) = split(/:/, $ns);
1221 $namespace_id{$n} = $s;
1222 }
1223 }
1224
1225 if (!exists $namespace_id{$name}) {
1226 # NS not found => get namespace id from MW and store it in
1227 # configuration file.
1228 my $query = {
1229 action => 'query',
1230 meta => 'siteinfo',
1231 siprop => 'namespaces'
1232 };
1233 my $result = $mediawiki->api($query);
1234
1235 while (my ($id, $ns) = each(%{$result->{query}->{namespaces}})) {
1236 if (defined($ns->{canonical}) && ($ns->{canonical} eq $name)) {
1237 run_git("config --add remote.". $remotename
1238 .".namespaces ". $name .":". $ns->{id});
1239 $namespace_id{$name} = $ns->{id};
1240 }
1241 }
1242 }
1243
1244 if (exists $namespace_id{$name}) {
1245 return $namespace_id{$name};
1246 } else {
1247 die "No such namespace $name on MediaWiki.";
1248 }
1249}