contrib/mw-to-git/git-remote-mediawiki.perl

   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 # Documentation & bugtracker: https://github.com/moy/Git-Mediawiki/
  13
  14 use strict;
  15 use MediaWiki::API;
  16 use Git;
  17 use DateTime::Format::ISO8601;
  18 use warnings;
  19
  20 # By default, use UTF-8 to communicate with Git and the user
  21 binmode STDERR, ":encoding(UTF-8)";
  22 binmode STDOUT, ":encoding(UTF-8)";
  23
  24 use URI::Escape;
  25
  26 # Mediawiki filenames can contain forward slashes. This variable decides by which pattern they should be replaced
  27 use constant SLASH_REPLACEMENT => "%2F";
  28
  29 # It's not always possible to delete pages (may require some
  30 # privileges). Deleted pages are replaced with this content.
  31 use constant DELETED_CONTENT => "[[Category:Deleted]]\n";
  32
  33 # It's not possible to create empty pages. New empty files in Git are
  34 # sent with this content instead.
  35 use constant EMPTY_CONTENT => "<!-- empty page -->\n";
  36
  37 # used to reflect file creation or deletion in diff.
  38 use constant NULL_SHA1 => "0000000000000000000000000000000000000000";
  39
  40 # Used on Git's side to reflect empty edit messages on the wiki
  41 use constant EMPTY_MESSAGE => '*Empty MediaWiki Message*';
  42
  43 my $remotename = $ARGV[0];
  44 my $url = $ARGV[1];
  45
  46 # Accept both space-separated and multiple keys in config file.
  47 # Spaces should be written as _ anyway because we'll use chomp.
  48 my @tracked_pages = split(/[ \n]/, run_git("config --get-all remote.". $remotename .".pages"));
  49 chomp(@tracked_pages);
  50
  51 # Just like @tracked_pages, but for MediaWiki categories.
  52 my @tracked_categories = split(/[ \n]/, run_git("config --get-all remote.". $remotename .".categories"));
  53 chomp(@tracked_categories);
  54
  55 # Import media files on pull
  56 my $import_media = run_git("config --get --bool remote.". $remotename .".mediaimport");
  57 chomp($import_media);
  58 $import_media = ($import_media eq "true");
  59
  60 # Export media files on push
  61 my $export_media = run_git("config --get --bool remote.". $remotename .".mediaexport");
  62 chomp($export_media);
  63 $export_media = !($export_media eq "false");
  64
  65 my $wiki_login = run_git("config --get remote.". $remotename .".mwLogin");
  66 # Note: mwPassword is discourraged. Use the credential system instead.
  67 my $wiki_passwd = run_git("config --get remote.". $remotename .".mwPassword");
  68 my $wiki_domain = run_git("config --get remote.". $remotename .".mwDomain");
  69 chomp($wiki_login);
  70 chomp($wiki_passwd);
  71 chomp($wiki_domain);
  72
  73 # Import only last revisions (both for clone and fetch)
  74 my $shallow_import = run_git("config --get --bool remote.". $remotename .".shallow");
  75 chomp($shallow_import);
  76 $shallow_import = ($shallow_import eq "true");
  77
  78 # Fetch (clone and pull) by revisions instead of by pages. This behavior
  79 # is more efficient when we have a wiki with lots of pages and we fetch
  80 # the revisions quite often so that they concern only few pages.
  81 # Possible values:
  82 # - by_rev: perform one query per new revision on the remote wiki
  83 # - by_page: query each tracked page for new revision
  84 my $fetch_strategy = run_git("config --get remote.$remotename.fetchStrategy");
  85 unless ($fetch_strategy) {
  86         $fetch_strategy = run_git("config --get mediawiki.fetchStrategy");
  87 }
  88 chomp($fetch_strategy);
  89 unless ($fetch_strategy) {
  90         $fetch_strategy = "by_page";
  91 }
  92
  93 # Remember the timestamp corresponding to a revision id.
  94 my %basetimestamps;
  95
  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.
 110 my $dumb_push = run_git("config --get --bool remote.$remotename.dumbPush");
 111 unless ($dumb_push) {
 112         $dumb_push = run_git("config --get --bool mediawiki.dumbPush");
 113 }
 114 chomp($dumb_push);
 115 $dumb_push = ($dumb_push eq "true");
 116
 117 my $wiki_name = $url;
 118 $wiki_name =~ s{[^/]*://}{};
 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 $wiki_name =~ s/^.*@//;
 123
 124 # Commands parser
 125 while (<STDIN>) {
 126         chomp;
 127
 128         if (!parse_command($_)) {
 129                 last;
 130         }
 131
 132         BEGIN { $| = 1 } # flush STDOUT, to make sure the previous
 133                          # command is fully processed.
 134 }
 135
 136 ########################## Functions ##############################
 137
 138 sub parse_command {
 139         my ($line) = @_;
 140         my @cmd = split(/ /, $line);
 141         if (!defined $cmd[0]) {
 142                 return 0;
 143         }
 144         if ($cmd[0] eq "capabilities") {
 145                 die("Too many arguments for capabilities\n")
 146                     if (defined($cmd[1]));
 147                 mw_capabilities();
 148         } elsif ($cmd[0] eq "list") {
 149                 die("Too many arguments for list\n") if (defined($cmd[2]));
 150                 mw_list($cmd[1]);
 151         } elsif ($cmd[0] eq "import") {
 152                 die("Invalid arguments for import\n")
 153                     if ($cmd[1] eq "" || defined($cmd[2]));
 154                 mw_import($cmd[1]);
 155         } elsif ($cmd[0] eq "option") {
 156                 die("Too many arguments for option\n")
 157                     if ($cmd[1] eq "" || $cmd[2] eq "" || defined($cmd[3]));
 158                 mw_option($cmd[1],$cmd[2]);
 159         } elsif ($cmd[0] eq "push") {
 160                 mw_push($cmd[1]);
 161         } else {
 162                 print STDERR "Unknown command. Aborting...\n";
 163                 return 0;
 164         }
 165         return 1;
 166 }
 167
 168 # MediaWiki API instance, created lazily.
 169 my $mediawiki;
 170
 171 sub mw_connect_maybe {
 172         if ($mediawiki) {
 173                 return;
 174         }
 175         $mediawiki = MediaWiki::API->new;
 176         $mediawiki->{config}->{api_url} = "$url/api.php";
 177         if ($wiki_login) {
 178                 my %credential = (
 179                         'url' => $url,
 180                         'username' => $wiki_login,
 181                         'password' => $wiki_passwd
 182                 );
 183                 Git::credential(\%credential);
 184                 my $request = {lgname => $credential{username},
 185                                lgpassword => $credential{password},
 186                                lgdomain => $wiki_domain};
 187                 if ($mediawiki->login($request)) {
 188                         Git::credential(\%credential, 'approve');
 189                         print STDERR "Logged in mediawiki user \"$credential{username}\".\n";
 190                 } else {
 191                         print STDERR "Failed to log in mediawiki user \"$credential{username}\" on $url\n";
 192                         print STDERR "  (error " .
 193                                 $mediawiki->{error}->{code} . ': ' .
 194                                 $mediawiki->{error}->{details} . ")\n";
 195                         Git::credential(\%credential, 'reject');
 196                         exit 1;
 197                 }
 198         }
 199         return;
 200 }
 201
 202 sub fatal_mw_error {
 203         my $action = shift;
 204         print STDERR "fatal: could not $action.\n";
 205         print STDERR "fatal: '$url' does not appear to be a mediawiki\n";
 206         if ($url =~ /^https/) {
 207                 print STDERR "fatal: make sure '$url/api.php' is a valid page\n";
 208                 print STDERR "fatal: and the SSL certificate is correct.\n";
 209         } else {
 210                 print STDERR "fatal: make sure '$url/api.php' is a valid page.\n";
 211         }
 212         print STDERR "fatal: (error " .
 213             $mediawiki->{error}->{code} . ': ' .
 214             $mediawiki->{error}->{details} . ")\n";
 215         exit 1;
 216 }
 217
 218 ## Functions for listing pages on the remote wiki
 219 sub get_mw_tracked_pages {
 220         my $pages = shift;
 221         get_mw_page_list(\@tracked_pages, $pages);
 222         return;
 223 }
 224
 225 sub get_mw_page_list {
 226         my $page_list = shift;
 227         my $pages = shift;
 228         my @some_pages = @$page_list;
 229         while (@some_pages) {
 230                 my $last_page = 50;
 231                 if ($#some_pages < $last_page) {
 232                         $last_page = $#some_pages;
 233                 }
 234                 my @slice = @some_pages[0..$last_page];
 235                 get_mw_first_pages(\@slice, $pages);
 236                 @some_pages = @some_pages[51..$#some_pages];
 237         }
 238         return;
 239 }
 240
 241 sub get_mw_tracked_categories {
 242         my $pages = shift;
 243         foreach my $category (@tracked_categories) {
 244                 if (index($category, ':') < 0) {
 245                         # Mediawiki requires the Category
 246                         # prefix, but let's not force the user
 247                         # to specify it.
 248                         $category = "Category:" . $category;
 249                 }
 250                 my $mw_pages = $mediawiki->list( {
 251                         action => 'query',
 252                         list => 'categorymembers',
 253                         cmtitle => $category,
 254                         cmlimit => 'max' } )
 255                         || die $mediawiki->{error}->{code} . ': '
 256                                 . $mediawiki->{error}->{details} . "\n";
 257                 foreach my $page (@{$mw_pages}) {
 258                         $pages->{$page->{title}} = $page;
 259                 }
 260         }
 261         return;
 262 }
 263
 264 sub get_mw_all_pages {
 265         my $pages = shift;
 266         # No user-provided list, get the list of pages from the API.
 267         my $mw_pages = $mediawiki->list({
 268                 action => 'query',
 269                 list => 'allpages',
 270                 aplimit => 'max'
 271         });
 272         if (!defined($mw_pages)) {
 273                 fatal_mw_error("get the list of wiki pages");
 274         }
 275         foreach my $page (@{$mw_pages}) {
 276                 $pages->{$page->{title}} = $page;
 277         }
 278         return;
 279 }
 280
 281 # queries the wiki for a set of pages. Meant to be used within a loop
 282 # querying the wiki for slices of page list.
 283 sub get_mw_first_pages {
 284         my $some_pages = shift;
 285         my @some_pages = @{$some_pages};
 286
 287         my $pages = shift;
 288
 289         # pattern 'page1|page2|...' required by the API
 290         my $titles = join('|', @some_pages);
 291
 292         my $mw_pages = $mediawiki->api({
 293                 action => 'query',
 294                 titles => $titles,
 295         });
 296         if (!defined($mw_pages)) {
 297                 fatal_mw_error("query the list of wiki pages");
 298         }
 299         while (my ($id, $page) = each(%{$mw_pages->{query}->{pages}})) {
 300                 if ($id < 0) {
 301                         print STDERR "Warning: page $page->{title} not found on wiki\n";
 302                 } else {
 303                         $pages->{$page->{title}} = $page;
 304                 }
 305         }
 306         return;
 307 }
 308
 309 # Get the list of pages to be fetched according to configuration.
 310 sub get_mw_pages {
 311         mw_connect_maybe();
 312
 313         print STDERR "Listing pages on remote wiki...\n";
 314
 315         my %pages; # hash on page titles to avoid duplicates
 316         my $user_defined;
 317         if (@tracked_pages) {
 318                 $user_defined = 1;
 319                 # The user provided a list of pages titles, but we
 320                 # still need to query the API to get the page IDs.
 321                 get_mw_tracked_pages(\%pages);
 322         }
 323         if (@tracked_categories) {
 324                 $user_defined = 1;
 325                 get_mw_tracked_categories(\%pages);
 326         }
 327         if (!$user_defined) {
 328                 get_mw_all_pages(\%pages);
 329         }
 330         if ($import_media) {
 331                 print STDERR "Getting media files for selected pages...\n";
 332                 if ($user_defined) {
 333                         get_linked_mediafiles(\%pages);
 334                 } else {
 335                         get_all_mediafiles(\%pages);
 336                 }
 337         }
 338         print STDERR (scalar keys %pages) . " pages found.\n";
 339         return %pages;
 340 }
 341
 342 # usage: $out = run_git("command args");
 343 #        $out = run_git("command args", "raw"); # don't interpret output as UTF-8.
 344 sub run_git {
 345         my $args = shift;
 346         my $encoding = (shift || "encoding(UTF-8)");
 347         open(my $git, "-|:$encoding", "git " . $args)
 348             or die "Unable to open: $!\n";
 349         my $res = do {
 350                 local $/ = undef;
 351                 <$git>
 352         };
 353         close($git);
 354
 355         return $res;
 356 }
 357
 358
 359 sub get_all_mediafiles {
 360         my $pages = shift;
 361         # Attach list of all pages for media files from the API,
 362         # they are in a different namespace, only one namespace
 363         # can be queried at the same moment
 364         my $mw_pages = $mediawiki->list({
 365                 action => 'query',
 366                 list => 'allpages',
 367                 apnamespace => get_mw_namespace_id("File"),
 368                 aplimit => 'max'
 369         });
 370         if (!defined($mw_pages)) {
 371                 print STDERR "fatal: could not get the list of pages for media files.\n";
 372                 print STDERR "fatal: '$url' does not appear to be a mediawiki\n";
 373                 print STDERR "fatal: make sure '$url/api.php' is a valid page.\n";
 374                 exit 1;
 375         }
 376         foreach my $page (@{$mw_pages}) {
 377                 $pages->{$page->{title}} = $page;
 378         }
 379         return;
 380 }
 381
 382 sub get_linked_mediafiles {
 383         my $pages = shift;
 384         my @titles = map { $_->{title} } values(%{$pages});
 385
 386         # The query is split in small batches because of the MW API limit of
 387         # the number of links to be returned (500 links max).
 388         my $batch = 10;
 389         while (@titles) {
 390                 if ($#titles < $batch) {
 391                         $batch = $#titles;
 392                 }
 393                 my @slice = @titles[0..$batch];
 394
 395                 # pattern 'page1|page2|...' required by the API
 396                 my $mw_titles = join('|', @slice);
 397
 398                 # Media files could be included or linked from
 399                 # a page, get all related
 400                 my $query = {
 401                         action => 'query',
 402                         prop => 'links|images',
 403                         titles => $mw_titles,
 404                         plnamespace => get_mw_namespace_id("File"),
 405                         pllimit => 'max'
 406                 };
 407                 my $result = $mediawiki->api($query);
 408
 409                 while (my ($id, $page) = each(%{$result->{query}->{pages}})) {
 410                         my @media_titles;
 411                         if (defined($page->{links})) {
 412                                 my @link_titles
 413                                     = map { $_->{title} } @{$page->{links}};
 414                                 push(@media_titles, @link_titles);
 415                         }
 416                         if (defined($page->{images})) {
 417                                 my @image_titles
 418                                     = map { $_->{title} } @{$page->{images}};
 419                                 push(@media_titles, @image_titles);
 420                         }
 421                         if (@media_titles) {
 422                                 get_mw_page_list(\@media_titles, $pages);
 423                         }
 424                 }
 425
 426                 @titles = @titles[($batch+1)..$#titles];
 427         }
 428         return;
 429 }
 430
 431 sub get_mw_mediafile_for_page_revision {
 432         # Name of the file on Wiki, with the prefix.
 433         my $filename = shift;
 434         my $timestamp = shift;
 435         my %mediafile;
 436
 437         # Search if on a media file with given timestamp exists on
 438         # MediaWiki. In that case download the file.
 439         my $query = {
 440                 action => 'query',
 441                 prop => 'imageinfo',
 442                 titles => "File:" . $filename,
 443                 iistart => $timestamp,
 444                 iiend => $timestamp,
 445                 iiprop => 'timestamp|archivename|url',
 446                 iilimit => 1
 447         };
 448         my $result = $mediawiki->api($query);
 449
 450         my ($fileid, $file) = each( %{$result->{query}->{pages}} );
 451         # If not defined it means there is no revision of the file for
 452         # given timestamp.
 453         if (defined($file->{imageinfo})) {
 454                 $mediafile{title} = $filename;
 455
 456                 my $fileinfo = pop(@{$file->{imageinfo}});
 457                 $mediafile{timestamp} = $fileinfo->{timestamp};
 458                 # Mediawiki::API's download function doesn't support https URLs
 459                 # and can't download old versions of files.
 460                 print STDERR "\tDownloading file $mediafile{title}, version $mediafile{timestamp}\n";
 461                 $mediafile{content} = download_mw_mediafile($fileinfo->{url});
 462         }
 463         return %mediafile;
 464 }
 465
 466 sub download_mw_mediafile {
 467         my $download_url = shift;
 468
 469         my $response = $mediawiki->{ua}->get($download_url);
 470         if ($response->code == 200) {
 471                 return $response->decoded_content;
 472         } else {
 473                 print STDERR "Error downloading mediafile from :\n";
 474                 print STDERR "URL: $download_url\n";
 475                 print STDERR "Server response: " . $response->code . " " . $response->message . "\n";
 476                 exit 1;
 477         }
 478 }
 479
 480 sub get_last_local_revision {
 481         # Get note regarding last mediawiki revision
 482         my $note = run_git("notes --ref=$remotename/mediawiki show refs/mediawiki/$remotename/master 2>/dev/null");
 483         my @note_info = split(/ /, $note);
 484
 485         my $lastrevision_number;
 486         if (!(defined($note_info[0]) && $note_info[0] eq "mediawiki_revision:")) {
 487                 print STDERR "No previous mediawiki revision found";
 488                 $lastrevision_number = 0;
 489         } else {
 490                 # Notes are formatted : mediawiki_revision: #number
 491                 $lastrevision_number = $note_info[1];
 492                 chomp($lastrevision_number);
 493                 print STDERR "Last local mediawiki revision found is $lastrevision_number";
 494         }
 495         return $lastrevision_number;
 496 }
 497
 498 # Get the last remote revision without taking in account which pages are
 499 # tracked or not. This function makes a single request to the wiki thus
 500 # avoid a loop onto all tracked pages. This is useful for the fetch-by-rev
 501 # option.
 502 sub get_last_global_remote_rev {
 503         mw_connect_maybe();
 504
 505         my $query = {
 506                 action => 'query',
 507                 list => 'recentchanges',
 508                 prop => 'revisions',
 509                 rclimit => '1',
 510                 rcdir => 'older',
 511         };
 512         my $result = $mediawiki->api($query);
 513         return $result->{query}->{recentchanges}[0]->{revid};
 514 }
 515
 516 # Get the last remote revision concerning the tracked pages and the tracked
 517 # categories.
 518 sub get_last_remote_revision {
 519         mw_connect_maybe();
 520
 521         my %pages_hash = get_mw_pages();
 522         my @pages = values(%pages_hash);
 523
 524         my $max_rev_num = 0;
 525
 526         print STDERR "Getting last revision id on tracked pages...\n";
 527
 528         foreach my $page (@pages) {
 529                 my $id = $page->{pageid};
 530
 531                 my $query = {
 532                         action => 'query',
 533                         prop => 'revisions',
 534                         rvprop => 'ids|timestamp',
 535                         pageids => $id,
 536                 };
 537
 538                 my $result = $mediawiki->api($query);
 539
 540                 my $lastrev = pop(@{$result->{query}->{pages}->{$id}->{revisions}});
 541
 542                 $basetimestamps{$lastrev->{revid}} = $lastrev->{timestamp};
 543
 544                 $max_rev_num = ($lastrev->{revid} > $max_rev_num ? $lastrev->{revid} : $max_rev_num);
 545         }
 546
 547         print STDERR "Last remote revision found is $max_rev_num.\n";
 548         return $max_rev_num;
 549 }
 550
 551 # Clean content before sending it to MediaWiki
 552 sub mediawiki_clean {
 553         my $string = shift;
 554         my $page_created = shift;
 555         # Mediawiki does not allow blank space at the end of a page and ends with a single \n.
 556         # This function right trims a string and adds a \n at the end to follow this rule
 557         $string =~ s/\s+$//;
 558         if ($string eq "" && $page_created) {
 559                 # Creating empty pages is forbidden.
 560                 $string = EMPTY_CONTENT;
 561         }
 562         return $string."\n";
 563 }
 564
 565 # Filter applied on MediaWiki data before adding them to Git
 566 sub mediawiki_smudge {
 567         my $string = shift;
 568         if ($string eq EMPTY_CONTENT) {
 569                 $string = "";
 570         }
 571         # This \n is important. This is due to mediawiki's way to handle end of files.
 572         return $string."\n";
 573 }
 574
 575 sub mediawiki_clean_filename {
 576         my $filename = shift;
 577         $filename =~ s{@{[SLASH_REPLACEMENT]}}{/}g;
 578         # [, ], |, {, and } are forbidden by MediaWiki, even URL-encoded.
 579         # Do a variant of URL-encoding, i.e. looks like URL-encoding,
 580         # but with _ added to prevent MediaWiki from thinking this is
 581         # an actual special character.
 582         $filename =~ s/[\[\]\{\}\|]/sprintf("_%%_%x", ord($&))/ge;
 583         # If we use the uri escape before
 584         # we should unescape here, before anything
 585
 586         return $filename;
 587 }
 588
 589 sub mediawiki_smudge_filename {
 590         my $filename = shift;
 591         $filename =~ s{/}{@{[SLASH_REPLACEMENT]}}g;
 592         $filename =~ s/ /_/g;
 593         # Decode forbidden characters encoded in mediawiki_clean_filename
 594         $filename =~ s/_%_([0-9a-fA-F][0-9a-fA-F])/sprintf("%c", hex($1))/ge;
 595         return $filename;
 596 }
 597
 598 sub literal_data {
 599         my ($content) = @_;
 600         print STDOUT "data ", bytes::length($content), "\n", $content;
 601         return;
 602 }
 603
 604 sub literal_data_raw {
 605         # Output possibly binary content.
 606         my ($content) = @_;
 607         # Avoid confusion between size in bytes and in characters
 608         utf8::downgrade($content);
 609         binmode STDOUT, ":raw";
 610         print STDOUT "data ", bytes::length($content), "\n", $content;
 611         binmode STDOUT, ":encoding(UTF-8)";
 612         return;
 613 }
 614
 615 sub mw_capabilities {
 616         # Revisions are imported to the private namespace
 617         # refs/mediawiki/$remotename/ by the helper and fetched into
 618         # refs/remotes/$remotename later by fetch.
 619         print STDOUT "refspec refs/heads/*:refs/mediawiki/$remotename/*\n";
 620         print STDOUT "import\n";
 621         print STDOUT "list\n";
 622         print STDOUT "push\n";
 623         print STDOUT "\n";
 624         return;
 625 }
 626
 627 sub mw_list {
 628         # MediaWiki do not have branches, we consider one branch arbitrarily
 629         # called master, and HEAD pointing to it.
 630         print STDOUT "? refs/heads/master\n";
 631         print STDOUT "\@refs/heads/master HEAD\n";
 632         print STDOUT "\n";
 633         return;
 634 }
 635
 636 sub mw_option {
 637         print STDERR "remote-helper command 'option $_[0]' not yet implemented\n";
 638         print STDOUT "unsupported\n";
 639         return;
 640 }
 641
 642 sub fetch_mw_revisions_for_page {
 643         my $page = shift;
 644         my $id = shift;
 645         my $fetch_from = shift;
 646         my @page_revs = ();
 647         my $query = {
 648                 action => 'query',
 649                 prop => 'revisions',
 650                 rvprop => 'ids',
 651                 rvdir => 'newer',
 652                 rvstartid => $fetch_from,
 653                 rvlimit => 500,
 654                 pageids => $id,
 655         };
 656
 657         my $revnum = 0;
 658         # Get 500 revisions at a time due to the mediawiki api limit
 659         while (1) {
 660                 my $result = $mediawiki->api($query);
 661
 662                 # Parse each of those 500 revisions
 663                 foreach my $revision (@{$result->{query}->{pages}->{$id}->{revisions}}) {
 664                         my $page_rev_ids;
 665                         $page_rev_ids->{pageid} = $page->{pageid};
 666                         $page_rev_ids->{revid} = $revision->{revid};
 667                         push(@page_revs, $page_rev_ids);
 668                         $revnum++;
 669                 }
 670                 last unless $result->{'query-continue'};
 671                 $query->{rvstartid} = $result->{'query-continue'}->{revisions}->{rvstartid};
 672         }
 673         if ($shallow_import && @page_revs) {
 674                 print STDERR "  Found 1 revision (shallow import).\n";
 675                 @page_revs = sort {$b->{revid} <=> $a->{revid}} (@page_revs);
 676                 return $page_revs[0];
 677         }
 678         print STDERR "  Found ", $revnum, " revision(s).\n";
 679         return @page_revs;
 680 }
 681
 682 sub fetch_mw_revisions {
 683         my $pages = shift; my @pages = @{$pages};
 684         my $fetch_from = shift;
 685
 686         my @revisions = ();
 687         my $n = 1;
 688         foreach my $page (@pages) {
 689                 my $id = $page->{pageid};
 690
 691                 print STDERR "page $n/", scalar(@pages), ": ". $page->{title} ."\n";
 692                 $n++;
 693                 my @page_revs = fetch_mw_revisions_for_page($page, $id, $fetch_from);
 694                 @revisions = (@page_revs, @revisions);
 695         }
 696
 697         return ($n, @revisions);
 698 }
 699
 700 sub fe_escape_path {
 701     my $path = shift;
 702     $path =~ s/\\/\\\\/g;
 703     $path =~ s/"/\\"/g;
 704     $path =~ s/\n/\\n/g;
 705     return '"' . $path . '"';
 706 }
 707
 708 sub import_file_revision {
 709         my $commit = shift;
 710         my %commit = %{$commit};
 711         my $full_import = shift;
 712         my $n = shift;
 713         my $mediafile = shift;
 714         my %mediafile;
 715         if ($mediafile) {
 716                 %mediafile = %{$mediafile};
 717         }
 718
 719         my $title = $commit{title};
 720         my $comment = $commit{comment};
 721         my $content = $commit{content};
 722         my $author = $commit{author};
 723         my $date = $commit{date};
 724
 725         print STDOUT "commit refs/mediawiki/$remotename/master\n";
 726         print STDOUT "mark :$n\n";
 727         print STDOUT "committer $author <$author\@$wiki_name> ", $date->epoch, " +0000\n";
 728         literal_data($comment);
 729
 730         # If it's not a clone, we need to know where to start from
 731         if (!$full_import && $n == 1) {
 732                 print STDOUT "from refs/mediawiki/$remotename/master^0\n";
 733         }
 734         if ($content ne DELETED_CONTENT) {
 735                 print STDOUT "M 644 inline " .
 736                     fe_escape_path($title . ".mw") . "\n";
 737                 literal_data($content);
 738                 if (%mediafile) {
 739                         print STDOUT "M 644 inline "
 740                             . fe_escape_path($mediafile{title}) . "\n";
 741                         literal_data_raw($mediafile{content});
 742                 }
 743                 print STDOUT "\n\n";
 744         } else {
 745                 print STDOUT "D " . fe_escape_path($title . ".mw") . "\n";
 746         }
 747
 748         # mediawiki revision number in the git note
 749         if ($full_import && $n == 1) {
 750                 print STDOUT "reset refs/notes/$remotename/mediawiki\n";
 751         }
 752         print STDOUT "commit refs/notes/$remotename/mediawiki\n";
 753         print STDOUT "committer $author <$author\@$wiki_name> ", $date->epoch, " +0000\n";
 754         literal_data("Note added by git-mediawiki during import");
 755         if (!$full_import && $n == 1) {
 756                 print STDOUT "from refs/notes/$remotename/mediawiki^0\n";
 757         }
 758         print STDOUT "N inline :$n\n";
 759         literal_data("mediawiki_revision: " . $commit{mw_revision});
 760         print STDOUT "\n\n";
 761         return;
 762 }
 763
 764 # parse a sequence of
 765 # <cmd> <arg1>
 766 # <cmd> <arg2>
 767 # \n
 768 # (like batch sequence of import and sequence of push statements)
 769 sub get_more_refs {
 770         my $cmd = shift;
 771         my @refs;
 772         while (1) {
 773                 my $line = <STDIN>;
 774                 if ($line =~ /^$cmd (.*)$/) {
 775                         push(@refs, $1);
 776                 } elsif ($line eq "\n") {
 777                         return @refs;
 778                 } else {
 779                         die("Invalid command in a '$cmd' batch: $_\n");
 780                 }
 781         }
 782         return;
 783 }
 784
 785 sub mw_import {
 786         # multiple import commands can follow each other.
 787         my @refs = (shift, get_more_refs("import"));
 788         foreach my $ref (@refs) {
 789                 mw_import_ref($ref);
 790         }
 791         print STDOUT "done\n";
 792         return;
 793 }
 794
 795 sub mw_import_ref {
 796         my $ref = shift;
 797         # The remote helper will call "import HEAD" and
 798         # "import refs/heads/master".
 799         # Since HEAD is a symbolic ref to master (by convention,
 800         # followed by the output of the command "list" that we gave),
 801         # we don't need to do anything in this case.
 802         if ($ref eq "HEAD") {
 803                 return;
 804         }
 805
 806         mw_connect_maybe();
 807
 808         print STDERR "Searching revisions...\n";
 809         my $last_local = get_last_local_revision();
 810         my $fetch_from = $last_local + 1;
 811         if ($fetch_from == 1) {
 812                 print STDERR ", fetching from beginning.\n";
 813         } else {
 814                 print STDERR ", fetching from here.\n";
 815         }
 816
 817         my $n = 0;
 818         if ($fetch_strategy eq "by_rev") {
 819                 print STDERR "Fetching & writing export data by revs...\n";
 820                 $n = mw_import_ref_by_revs($fetch_from);
 821         } elsif ($fetch_strategy eq "by_page") {
 822                 print STDERR "Fetching & writing export data by pages...\n";
 823                 $n = mw_import_ref_by_pages($fetch_from);
 824         } else {
 825                 print STDERR "fatal: invalid fetch strategy \"$fetch_strategy\".\n";
 826                 print STDERR "Check your configuration variables remote.$remotename.fetchStrategy and mediawiki.fetchStrategy\n";
 827                 exit 1;
 828         }
 829
 830         if ($fetch_from == 1 && $n == 0) {
 831                 print STDERR "You appear to have cloned an empty MediaWiki.\n";
 832                 # Something has to be done remote-helper side. If nothing is done, an error is
 833                 # thrown saying that HEAD is referring to unknown object 0000000000000000000
 834                 # and the clone fails.
 835         }
 836         return;
 837 }
 838
 839 sub mw_import_ref_by_pages {
 840
 841         my $fetch_from = shift;
 842         my %pages_hash = get_mw_pages();
 843         my @pages = values(%pages_hash);
 844
 845         my ($n, @revisions) = fetch_mw_revisions(\@pages, $fetch_from);
 846
 847         @revisions = sort {$a->{revid} <=> $b->{revid}} @revisions;
 848         my @revision_ids = map { $_->{revid} } @revisions;
 849
 850         return mw_import_revids($fetch_from, \@revision_ids, \%pages_hash);
 851 }
 852
 853 sub mw_import_ref_by_revs {
 854
 855         my $fetch_from = shift;
 856         my %pages_hash = get_mw_pages();
 857
 858         my $last_remote = get_last_global_remote_rev();
 859         my @revision_ids = $fetch_from..$last_remote;
 860         return mw_import_revids($fetch_from, \@revision_ids, \%pages_hash);
 861 }
 862
 863 # Import revisions given in second argument (array of integers).
 864 # Only pages appearing in the third argument (hash indexed by page titles)
 865 # will be imported.
 866 sub mw_import_revids {
 867         my $fetch_from = shift;
 868         my $revision_ids = shift;
 869         my $pages = shift;
 870
 871         my $n = 0;
 872         my $n_actual = 0;
 873         my $last_timestamp = 0; # Placeholer in case $rev->timestamp is undefined
 874
 875         foreach my $pagerevid (@$revision_ids) {
 876                 # Count page even if we skip it, since we display
 877                 # $n/$total and $total includes skipped pages.
 878                 $n++;
 879
 880                 # fetch the content of the pages
 881                 my $query = {
 882                         action => 'query',
 883                         prop => 'revisions',
 884                         rvprop => 'content|timestamp|comment|user|ids',
 885                         revids => $pagerevid,
 886                 };
 887
 888                 my $result = $mediawiki->api($query);
 889
 890                 if (!$result) {
 891                         die "Failed to retrieve modified page for revision $pagerevid\n";
 892                 }
 893
 894                 if (defined($result->{query}->{badrevids}->{$pagerevid})) {
 895                         # The revision id does not exist on the remote wiki.
 896                         next;
 897                 }
 898
 899                 if (!defined($result->{query}->{pages})) {
 900                         die "Invalid revision $pagerevid.\n";
 901                 }
 902
 903                 my @result_pages = values(%{$result->{query}->{pages}});
 904                 my $result_page = $result_pages[0];
 905                 my $rev = $result_pages[0]->{revisions}->[0];
 906
 907                 my $page_title = $result_page->{title};
 908
 909                 if (!exists($pages->{$page_title})) {
 910                         print STDERR "$n/", scalar(@$revision_ids),
 911                                 ": Skipping revision #$rev->{revid} of $page_title\n";
 912                         next;
 913                 }
 914
 915                 $n_actual++;
 916
 917                 my %commit;
 918                 $commit{author} = $rev->{user} || 'Anonymous';
 919                 $commit{comment} = $rev->{comment} || EMPTY_MESSAGE;
 920                 $commit{title} = mediawiki_smudge_filename($page_title);
 921                 $commit{mw_revision} = $rev->{revid};
 922                 $commit{content} = mediawiki_smudge($rev->{'*'});
 923
 924                 if (!defined($rev->{timestamp})) {
 925                         $last_timestamp++;
 926                 } else {
 927                         $last_timestamp = $rev->{timestamp};
 928                 }
 929                 $commit{date} = DateTime::Format::ISO8601->parse_datetime($last_timestamp);
 930
 931                 # Differentiates classic pages and media files.
 932                 my ($namespace, $filename) = $page_title =~ /^([^:]*):(.*)$/;
 933                 my %mediafile;
 934                 if ($namespace) {
 935                         my $id = get_mw_namespace_id($namespace);
 936                         if ($id && $id == get_mw_namespace_id("File")) {
 937                                 %mediafile = get_mw_mediafile_for_page_revision($filename, $rev->{timestamp});
 938                         }
 939                 }
 940                 # If this is a revision of the media page for new version
 941                 # of a file do one common commit for both file and media page.
 942                 # Else do commit only for that page.
 943                 print STDERR "$n/", scalar(@$revision_ids), ": Revision #$rev->{revid} of $commit{title}\n";
 944                 import_file_revision(\%commit, ($fetch_from == 1), $n_actual, \%mediafile);
 945         }
 946
 947         return $n_actual;
 948 }
 949
 950 sub error_non_fast_forward {
 951         my $advice = run_git("config --bool advice.pushNonFastForward");
 952         chomp($advice);
 953         if ($advice ne "false") {
 954                 # Native git-push would show this after the summary.
 955                 # We can't ask it to display it cleanly, so print it
 956                 # ourselves before.
 957                 print STDERR "To prevent you from losing history, non-fast-forward updates were rejected\n";
 958                 print STDERR "Merge the remote changes (e.g. 'git pull') before pushing again. See the\n";
 959                 print STDERR "'Note about fast-forwards' section of 'git push --help' for details.\n";
 960         }
 961         print STDOUT "error $_[0] \"non-fast-forward\"\n";
 962         return 0;
 963 }
 964
 965 sub mw_upload_file {
 966         my $complete_file_name = shift;
 967         my $new_sha1 = shift;
 968         my $extension = shift;
 969         my $file_deleted = shift;
 970         my $summary = shift;
 971         my $newrevid;
 972         my $path = "File:" . $complete_file_name;
 973         my %hashFiles = get_allowed_file_extensions();
 974         if (!exists($hashFiles{$extension})) {
 975                 print STDERR "$complete_file_name is not a permitted file on this wiki.\n";
 976                 print STDERR "Check the configuration of file uploads in your mediawiki.\n";
 977                 return $newrevid;
 978         }
 979         # Deleting and uploading a file requires a priviledged user
 980         if ($file_deleted) {
 981                 mw_connect_maybe();
 982                 my $query = {
 983                         action => 'delete',
 984                         title => $path,
 985                         reason => $summary
 986                 };
 987                 if (!$mediawiki->edit($query)) {
 988                         print STDERR "Failed to delete file on remote wiki\n";
 989                         print STDERR "Check your permissions on the remote site. Error code:\n";
 990                         print STDERR $mediawiki->{error}->{code} . ':' . $mediawiki->{error}->{details};
 991                         exit 1;
 992                 }
 993         } else {
 994                 # Don't let perl try to interpret file content as UTF-8 => use "raw"
 995                 my $content = run_git("cat-file blob $new_sha1", "raw");
 996                 if ($content ne "") {
 997                         mw_connect_maybe();
 998                         $mediawiki->{config}->{upload_url} =
 999                                 "$url/index.php/Special:Upload";
1000                         $mediawiki->edit({
1001                                 action => 'upload',
1002                                 filename => $complete_file_name,
1003                                 comment => $summary,
1004                                 file => [undef,
1005                                          $complete_file_name,
1006                                          Content => $content],
1007                                 ignorewarnings => 1,
1008                         }, {
1009                                 skip_encoding => 1
1010                         } ) || die $mediawiki->{error}->{code} . ':'
1011                                  . $mediawiki->{error}->{details} . "\n";
1012                         my $last_file_page = $mediawiki->get_page({title => $path});
1013                         $newrevid = $last_file_page->{revid};
1014                         print STDERR "Pushed file: $new_sha1 - $complete_file_name.\n";
1015                 } else {
1016                         print STDERR "Empty file $complete_file_name not pushed.\n";
1017                 }
1018         }
1019         return $newrevid;
1020 }
1021
1022 sub mw_push_file {
1023         my $diff_info = shift;
1024         # $diff_info contains a string in this format:
1025         # 100644 100644 <sha1_of_blob_before_commit> <sha1_of_blob_now> <status>
1026         my @diff_info_split = split(/[ \t]/, $diff_info);
1027
1028         # Filename, including .mw extension
1029         my $complete_file_name = shift;
1030         # Commit message
1031         my $summary = shift;
1032         # MediaWiki revision number. Keep the previous one by default,
1033         # in case there's no edit to perform.
1034         my $oldrevid = shift;
1035         my $newrevid;
1036
1037         if ($summary eq EMPTY_MESSAGE) {
1038                 $summary = '';
1039         }
1040
1041         my $new_sha1 = $diff_info_split[3];
1042         my $old_sha1 = $diff_info_split[2];
1043         my $page_created = ($old_sha1 eq NULL_SHA1);
1044         my $page_deleted = ($new_sha1 eq NULL_SHA1);
1045         $complete_file_name = mediawiki_clean_filename($complete_file_name);
1046
1047         my ($title, $extension) = $complete_file_name =~ /^(.*)\.([^\.]*)$/;
1048         if (!defined($extension)) {
1049                 $extension = "";
1050         }
1051         if ($extension eq "mw") {
1052                 my $ns = get_mw_namespace_id_for_page($complete_file_name);
1053                 if ($ns && $ns == get_mw_namespace_id("File") && (!$export_media)) {
1054                         print STDERR "Ignoring media file related page: $complete_file_name\n";
1055                         return ($oldrevid, "ok");
1056                 }
1057                 my $file_content;
1058                 if ($page_deleted) {
1059                         # Deleting a page usually requires
1060                         # special privileges. A common
1061                         # convention is to replace the page
1062                         # with this content instead:
1063                         $file_content = DELETED_CONTENT;
1064                 } else {
1065                         $file_content = run_git("cat-file blob $new_sha1");
1066                 }
1067
1068                 mw_connect_maybe();
1069
1070                 my $result = $mediawiki->edit( {
1071                         action => 'edit',
1072                         summary => $summary,
1073                         title => $title,
1074                         basetimestamp => $basetimestamps{$oldrevid},
1075                         text => mediawiki_clean($file_content, $page_created),
1076                                   }, {
1077                                           skip_encoding => 1 # Helps with names with accentuated characters
1078                                   });
1079                 if (!$result) {
1080                         if ($mediawiki->{error}->{code} == 3) {
1081                                 # edit conflicts, considered as non-fast-forward
1082                                 print STDERR 'Warning: Error ' .
1083                                     $mediawiki->{error}->{code} .
1084                                     ' from mediwiki: ' . $mediawiki->{error}->{details} .
1085                                     ".\n";
1086                                 return ($oldrevid, "non-fast-forward");
1087                         } else {
1088                                 # Other errors. Shouldn't happen => just die()
1089                                 die 'Fatal: Error ' .
1090                                     $mediawiki->{error}->{code} .
1091                                     ' from mediwiki: ' . $mediawiki->{error}->{details} . "\n";
1092                         }
1093                 }
1094                 $newrevid = $result->{edit}->{newrevid};
1095                 print STDERR "Pushed file: $new_sha1 - $title\n";
1096         } elsif ($export_media) {
1097                 $newrevid = mw_upload_file($complete_file_name, $new_sha1,
1098                                            $extension, $page_deleted,
1099                                            $summary);
1100         } else {
1101                 print STDERR "Ignoring media file $title\n";
1102         }
1103         $newrevid = ($newrevid or $oldrevid);
1104         return ($newrevid, "ok");
1105 }
1106
1107 sub mw_push {
1108         # multiple push statements can follow each other
1109         my @refsspecs = (shift, get_more_refs("push"));
1110         my $pushed;
1111         for my $refspec (@refsspecs) {
1112                 my ($force, $local, $remote) = $refspec =~ /^(\+)?([^:]*):([^:]*)$/
1113                     or die("Invalid refspec for push. Expected <src>:<dst> or +<src>:<dst>\n");
1114                 if ($force) {
1115                         print STDERR "Warning: forced push not allowed on a MediaWiki.\n";
1116                 }
1117                 if ($local eq "") {
1118                         print STDERR "Cannot delete remote branch on a MediaWiki\n";
1119                         print STDOUT "error $remote cannot delete\n";
1120                         next;
1121                 }
1122                 if ($remote ne "refs/heads/master") {
1123                         print STDERR "Only push to the branch 'master' is supported on a MediaWiki\n";
1124                         print STDOUT "error $remote only master allowed\n";
1125                         next;
1126                 }
1127                 if (mw_push_revision($local, $remote)) {
1128                         $pushed = 1;
1129                 }
1130         }
1131
1132         # Notify Git that the push is done
1133         print STDOUT "\n";
1134
1135         if ($pushed && $dumb_push) {
1136                 print STDERR "Just pushed some revisions to MediaWiki.\n";
1137                 print STDERR "The pushed revisions now have to be re-imported, and your current branch\n";
1138                 print STDERR "needs to be updated with these re-imported commits. You can do this with\n";
1139                 print STDERR "\n";
1140                 print STDERR "  git pull --rebase\n";
1141                 print STDERR "\n";
1142         }
1143         return;
1144 }
1145
1146 sub mw_push_revision {
1147         my $local = shift;
1148         my $remote = shift; # actually, this has to be "refs/heads/master" at this point.
1149         my $last_local_revid = get_last_local_revision();
1150         print STDERR ".\n"; # Finish sentence started by get_last_local_revision()
1151         my $last_remote_revid = get_last_remote_revision();
1152         my $mw_revision = $last_remote_revid;
1153
1154         # Get sha1 of commit pointed by local HEAD
1155         my $HEAD_sha1 = run_git("rev-parse $local 2>/dev/null"); chomp($HEAD_sha1);
1156         # Get sha1 of commit pointed by remotes/$remotename/master
1157         my $remoteorigin_sha1 = run_git("rev-parse refs/remotes/$remotename/master 2>/dev/null");
1158         chomp($remoteorigin_sha1);
1159
1160         if ($last_local_revid > 0 &&
1161             $last_local_revid < $last_remote_revid) {
1162                 return error_non_fast_forward($remote);
1163         }
1164
1165         if ($HEAD_sha1 eq $remoteorigin_sha1) {
1166                 # nothing to push
1167                 return 0;
1168         }
1169
1170         # Get every commit in between HEAD and refs/remotes/origin/master,
1171         # including HEAD and refs/remotes/origin/master
1172         my @commit_pairs = ();
1173         if ($last_local_revid > 0) {
1174                 my $parsed_sha1 = $remoteorigin_sha1;
1175                 # Find a path from last MediaWiki commit to pushed commit
1176                 print STDERR "Computing path from local to remote ...\n";
1177                 my @local_ancestry = split(/\n/, run_git("rev-list --boundary --parents $local ^$parsed_sha1"));
1178                 my %local_ancestry;
1179                 foreach my $line (@local_ancestry) {
1180                         if (my ($child, $parents) = $line =~ /^-?([a-f0-9]+) ([a-f0-9 ]+)/) {
1181                                 foreach my $parent (split(/ /, $parents)) {
1182                                         $local_ancestry{$parent} = $child;
1183                                 }
1184                         } elsif (!$line =~ /^([a-f0-9]+)/) {
1185                                 die "Unexpected output from git rev-list: $line\n";
1186                         }
1187                 }
1188                 while ($parsed_sha1 ne $HEAD_sha1) {
1189                         my $child = $local_ancestry{$parsed_sha1};
1190                         if (!$child) {
1191                                 printf STDERR "Cannot find a path in history from remote commit to last commit\n";
1192                                 return error_non_fast_forward($remote);
1193                         }
1194                         push(@commit_pairs, [$parsed_sha1, $child]);
1195                         $parsed_sha1 = $child;
1196                 }
1197         } else {
1198                 # No remote mediawiki revision. Export the whole
1199                 # history (linearized with --first-parent)
1200                 print STDERR "Warning: no common ancestor, pushing complete history\n";
1201                 my $history = run_git("rev-list --first-parent --children $local");
1202                 my @history = split(/\n/, $history);
1203                 @history = @history[1..$#history];
1204                 foreach my $line (reverse @history) {
1205                         my @commit_info_split = split(/[ \n]/, $line);
1206                         push(@commit_pairs, \@commit_info_split);
1207                 }
1208         }
1209
1210         foreach my $commit_info_split (@commit_pairs) {
1211                 my $sha1_child = @{$commit_info_split}[0];
1212                 my $sha1_commit = @{$commit_info_split}[1];
1213                 my $diff_infos = run_git("diff-tree -r --raw -z $sha1_child $sha1_commit");
1214                 # TODO: we could detect rename, and encode them with a #redirect on the wiki.
1215                 # TODO: for now, it's just a delete+add
1216                 my @diff_info_list = split(/\0/, $diff_infos);
1217                 # Keep the subject line of the commit message as mediawiki comment for the revision
1218                 my $commit_msg = run_git("log --no-walk --format=\"%s\" $sha1_commit");
1219                 chomp($commit_msg);
1220                 # Push every blob
1221                 while (@diff_info_list) {
1222                         my $status;
1223                         # git diff-tree -z gives an output like
1224                         # <metadata>\0<filename1>\0
1225                         # <metadata>\0<filename2>\0
1226                         # and we've split on \0.
1227                         my $info = shift(@diff_info_list);
1228                         my $file = shift(@diff_info_list);
1229                         ($mw_revision, $status) = mw_push_file($info, $file, $commit_msg, $mw_revision);
1230                         if ($status eq "non-fast-forward") {
1231                                 # we may already have sent part of the
1232                                 # commit to MediaWiki, but it's too
1233                                 # late to cancel it. Stop the push in
1234                                 # the middle, but still give an
1235                                 # accurate error message.
1236                                 return error_non_fast_forward($remote);
1237                         }
1238                         if ($status ne "ok") {
1239                                 die("Unknown error from mw_push_file()\n");
1240                         }
1241                 }
1242                 unless ($dumb_push) {
1243                         run_git("notes --ref=$remotename/mediawiki add -f -m \"mediawiki_revision: $mw_revision\" $sha1_commit");
1244                         run_git("update-ref -m \"Git-MediaWiki push\" refs/mediawiki/$remotename/master $sha1_commit $sha1_child");
1245                 }
1246         }
1247
1248         print STDOUT "ok $remote\n";
1249         return 1;
1250 }
1251
1252 sub get_allowed_file_extensions {
1253         mw_connect_maybe();
1254
1255         my $query = {
1256                 action => 'query',
1257                 meta => 'siteinfo',
1258                 siprop => 'fileextensions'
1259                 };
1260         my $result = $mediawiki->api($query);
1261         my @file_extensions = map { $_->{ext}} @{$result->{query}->{fileextensions}};
1262         my %hashFile = map { $_ => 1 } @file_extensions;
1263
1264         return %hashFile;
1265 }
1266
1267 # In memory cache for MediaWiki namespace ids.
1268 my %namespace_id;
1269
1270 # Namespaces whose id is cached in the configuration file
1271 # (to avoid duplicates)
1272 my %cached_mw_namespace_id;
1273
1274 # Return MediaWiki id for a canonical namespace name.
1275 # Ex.: "File", "Project".
1276 sub get_mw_namespace_id {
1277         mw_connect_maybe();
1278         my $name = shift;
1279
1280         if (!exists $namespace_id{$name}) {
1281                 # Look at configuration file, if the record for that namespace is
1282                 # already cached. Namespaces are stored in form:
1283                 # "Name_of_namespace:Id_namespace", ex.: "File:6".
1284                 my @temp = split(/\n/, run_git("config --get-all remote."
1285                                                 . $remotename .".namespaceCache"));
1286                 chomp(@temp);
1287                 foreach my $ns (@temp) {
1288                         my ($n, $id) = split(/:/, $ns);
1289                         if ($id eq 'notANameSpace') {
1290                                 $namespace_id{$n} = {is_namespace => 0};
1291                         } else {
1292                                 $namespace_id{$n} = {is_namespace => 1, id => $id};
1293                         }
1294                         $cached_mw_namespace_id{$n} = 1;
1295                 }
1296         }
1297
1298         if (!exists $namespace_id{$name}) {
1299                 print STDERR "Namespace $name not found in cache, querying the wiki ...\n";
1300                 # NS not found => get namespace id from MW and store it in
1301                 # configuration file.
1302                 my $query = {
1303                         action => 'query',
1304                         meta => 'siteinfo',
1305                         siprop => 'namespaces'
1306                 };
1307                 my $result = $mediawiki->api($query);
1308
1309                 while (my ($id, $ns) = each(%{$result->{query}->{namespaces}})) {
1310                         if (defined($ns->{id}) && defined($ns->{canonical})) {
1311                                 $namespace_id{$ns->{canonical}} = {is_namespace => 1, id => $ns->{id}};
1312                                 if ($ns->{'*'}) {
1313                                         # alias (e.g. french Fichier: as alias for canonical File:)
1314                                         $namespace_id{$ns->{'*'}} = {is_namespace => 1, id => $ns->{id}};
1315                                 }
1316                         }
1317                 }
1318         }
1319
1320         my $ns = $namespace_id{$name};
1321         my $id;
1322
1323         unless (defined $ns) {
1324                 print STDERR "No such namespace $name on MediaWiki.\n";
1325                 $ns = {is_namespace => 0};
1326                 $namespace_id{$name} = $ns;
1327         }
1328
1329         if ($ns->{is_namespace}) {
1330                 $id = $ns->{id};
1331         }
1332
1333         # Store "notANameSpace" as special value for inexisting namespaces
1334         my $store_id = ($id || 'notANameSpace');
1335
1336         # Store explicitely requested namespaces on disk
1337         if (!exists $cached_mw_namespace_id{$name}) {
1338                 run_git("config --add remote.". $remotename
1339                         .".namespaceCache \"". $name .":". $store_id ."\"");
1340                 $cached_mw_namespace_id{$name} = 1;
1341         }
1342         return $id;
1343 }
1344
1345 sub get_mw_namespace_id_for_page {
1346         my $namespace = shift;
1347         if ($namespace =~ /^([^:]*):/) {
1348                 return get_mw_namespace_id($namespace);
1349         } else {
1350                 return;
1351         }
1352 }