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

   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 DateTime::Format::ISO8601;
  17
  18 # By default, use UTF-8 to communicate with Git and the user
  19 binmode STDERR, ":utf8";
  20 binmode STDOUT, ":utf8";
  21
  22 use URI::Escape;
  23 use IPC::Open2;
  24
  25 use warnings;
  26
  27 # Mediawiki filenames can contain forward slashes. This variable decides by which pattern they should be replaced
  28 use constant SLASH_REPLACEMENT => "%2F";
  29
  30 # It's not always possible to delete pages (may require some
  31 # priviledges). Deleted pages are replaced with this content.
  32 use constant DELETED_CONTENT => "[[Category:Deleted]]\n";
  33
  34 # It's not possible to create empty pages. New empty files in Git are
  35 # sent with this content instead.
  36 use constant EMPTY_CONTENT => "<!-- empty page -->\n";
  37
  38 # used to reflect file creation or deletion in diff.
  39 use constant NULL_SHA1 => "0000000000000000000000000000000000000000";
  40
  41 # Used on Git's side to reflect empty edit messages on the wiki
  42 use constant EMPTY_MESSAGE => '*Empty MediaWiki Message*';
  43
  44 my $remotename = $ARGV[0];
  45 my $url = $ARGV[1];
  46
  47 # Accept both space-separated and multiple keys in config file.
  48 # Spaces should be written as _ anyway because we'll use chomp.
  49 my @tracked_pages = split(/[ \n]/, run_git("config --get-all remote.". $remotename .".pages"));
  50 chomp(@tracked_pages);
  51
  52 # Just like @tracked_pages, but for MediaWiki categories.
  53 my @tracked_categories = split(/[ \n]/, run_git("config --get-all remote.". $remotename .".categories"));
  54 chomp(@tracked_categories);
  55
  56 # Import media files on pull
  57 my $import_media = run_git("config --get --bool remote.". $remotename .".mediaimport");
  58 chomp($import_media);
  59 $import_media = ($import_media eq "true");
  60
  61 # Export media files on push
  62 my $export_media = run_git("config --get --bool remote.". $remotename .".mediaexport");
  63 chomp($export_media);
  64 $export_media = !($export_media eq "false");
  65
  66 my $wiki_login = run_git("config --get remote.". $remotename .".mwLogin");
  67 # Note: mwPassword is discourraged. Use the credential system instead.
  68 my $wiki_passwd = run_git("config --get remote.". $remotename .".mwPassword");
  69 my $wiki_domain = run_git("config --get remote.". $remotename .".mwDomain");
  70 chomp($wiki_login);
  71 chomp($wiki_passwd);
  72 chomp($wiki_domain);
  73
  74 # Import only last revisions (both for clone and fetch)
  75 my $shallow_import = run_git("config --get --bool remote.". $remotename .".shallow");
  76 chomp($shallow_import);
  77 $shallow_import = ($shallow_import eq "true");
  78
  79 # Fetch (clone and pull) by revisions instead of by pages. This behavior
  80 # is more efficient when we have a wiki with lots of pages and we fetch
  81 # the revisions quite often so that they concern only few pages.
  82 # Possible values:
  83 # - by_rev: perform one query per new revision on the remote wiki
  84 # - by_page: query each tracked page for new revision
  85 my $fetch_strategy = run_git("config --get remote.$remotename.fetchStrategy");
  86 unless ($fetch_strategy) {
  87         $fetch_strategy = run_git("config --get mediawiki.fetchStrategy");
  88 }
  89 chomp($fetch_strategy);
  90 unless ($fetch_strategy) {
  91         $fetch_strategy = "by_page";
  92 }
  93
  94 # Dumb push: don't update notes and mediawiki ref to reflect the last push.
  95 #
  96 # Configurable with mediawiki.dumbPush, or per-remote with
  97 # remote.<remotename>.dumbPush.
  98 #
  99 # This means the user will have to re-import the just-pushed
 100 # revisions. On the other hand, this means that the Git revisions
 101 # corresponding to MediaWiki revisions are all imported from the wiki,
 102 # regardless of whether they were initially created in Git or from the
 103 # web interface, hence all users will get the same history (i.e. if
 104 # the push from Git to MediaWiki loses some information, everybody
 105 # will get the history with information lost). If the import is
 106 # deterministic, this means everybody gets the same sha1 for each
 107 # MediaWiki revision.
 108 my $dumb_push = run_git("config --get --bool remote.$remotename.dumbPush");
 109 unless ($dumb_push) {
 110         $dumb_push = run_git("config --get --bool mediawiki.dumbPush");
 111 }
 112 chomp($dumb_push);
 113 $dumb_push = ($dumb_push eq "true");
 114
 115 my $wiki_name = $url;
 116 $wiki_name =~ s/[^\/]*:\/\///;
 117 # If URL is like http://user:password@example.com/, we clearly don't
 118 # want the password in $wiki_name. While we're there, also remove user
 119 # and '@' sign, to avoid author like MWUser@HTTPUser@host.com
 120 $wiki_name =~ s/^.*@//;
 121
 122 # Commands parser
 123 my $entry;
 124 my @cmd;
 125 while (<STDIN>) {
 126         chomp;
 127         @cmd = split(/ /);
 128         if (defined($cmd[0])) {
 129                 # Line not blank
 130                 if ($cmd[0] eq "capabilities") {
 131                         die("Too many arguments for capabilities") unless (!defined($cmd[1]));
 132                         mw_capabilities();
 133                 } elsif ($cmd[0] eq "list") {
 134                         die("Too many arguments for list") unless (!defined($cmd[2]));
 135                         mw_list($cmd[1]);
 136                 } elsif ($cmd[0] eq "import") {
 137                         die("Invalid arguments for import") unless ($cmd[1] ne "" && !defined($cmd[2]));
 138                         mw_import($cmd[1]);
 139                 } elsif ($cmd[0] eq "option") {
 140                         die("Too many arguments for option") unless ($cmd[1] ne "" && $cmd[2] ne "" && !defined($cmd[3]));
 141                         mw_option($cmd[1],$cmd[2]);
 142                 } elsif ($cmd[0] eq "push") {
 143                         mw_push($cmd[1]);
 144                 } else {
 145                         print STDERR "Unknown command. Aborting...\n";
 146                         last;
 147                 }
 148         } else {
 149                 # blank line: we should terminate
 150                 last;
 151         }
 152
 153         BEGIN { $| = 1 } # flush STDOUT, to make sure the previous
 154                          # command is fully processed.
 155 }
 156
 157 ########################## Functions ##############################
 158
 159 ## credential API management (generic functions)
 160
 161 sub credential_read {
 162         my %credential;
 163         my $reader = shift;
 164         my $op = shift;
 165         while (<$reader>) {
 166                 my ($key, $value) = /([^=]*)=(.*)/;
 167                 if (not defined $key) {
 168                         die "ERROR receiving response from git credential $op:\n$_\n";
 169                 }
 170                 $credential{$key} = $value;
 171         }
 172         return %credential;
 173 }
 174
 175 sub credential_write {
 176         my $credential = shift;
 177         my $writer = shift;
 178         # url overwrites other fields, so it must come first
 179         print $writer "url=$credential->{url}\n" if exists $credential->{url};
 180         while (my ($key, $value) = each(%$credential) ) {
 181                 if (length $value && $key ne 'url') {
 182                         print $writer "$key=$value\n";
 183                 }
 184         }
 185 }
 186
 187 sub credential_run {
 188         my $op = shift;
 189         my $credential = shift;
 190         my $pid = open2(my $reader, my $writer, "git credential $op");
 191         credential_write($credential, $writer);
 192         print $writer "\n";
 193         close($writer);
 194
 195         if ($op eq "fill") {
 196                 %$credential = credential_read($reader, $op);
 197         } else {
 198                 if (<$reader>) {
 199                         die "ERROR while running git credential $op:\n$_";
 200                 }
 201         }
 202         close($reader);
 203         waitpid($pid, 0);
 204         my $child_exit_status = $? >> 8;
 205         if ($child_exit_status != 0) {
 206                 die "'git credential $op' failed with code $child_exit_status.";
 207         }
 208 }
 209
 210 # MediaWiki API instance, created lazily.
 211 my $mediawiki;
 212
 213 sub mw_connect_maybe {
 214         if ($mediawiki) {
 215                 return;
 216         }
 217         $mediawiki = MediaWiki::API->new;
 218         $mediawiki->{config}->{api_url} = "$url/api.php";
 219         if ($wiki_login) {
 220                 my %credential = (url => $url);
 221                 $credential{username} = $wiki_login;
 222                 $credential{password} = $wiki_passwd;
 223                 credential_run("fill", \%credential);
 224                 my $request = {lgname => $credential{username},
 225                                lgpassword => $credential{password},
 226                                lgdomain => $wiki_domain};
 227                 if ($mediawiki->login($request)) {
 228                         credential_run("approve", \%credential);
 229                         print STDERR "Logged in mediawiki user \"$credential{username}\".\n";
 230                 } else {
 231                         print STDERR "Failed to log in mediawiki user \"$credential{username}\" on $url\n";
 232                         print STDERR "  (error " .
 233                                 $mediawiki->{error}->{code} . ': ' .
 234                                 $mediawiki->{error}->{details} . ")\n";
 235                         credential_run("reject", \%credential);
 236                         exit 1;
 237                 }
 238         }
 239 }
 240
 241 ## Functions for listing pages on the remote wiki
 242 sub get_mw_tracked_pages {
 243         my $pages = shift;
 244         get_mw_page_list(\@tracked_pages, $pages);
 245 }
 246
 247 sub get_mw_page_list {
 248         my $page_list = shift;
 249         my $pages = shift;
 250         my @some_pages = @$page_list;
 251         while (@some_pages) {
 252                 my $last = 50;
 253                 if ($#some_pages < $last) {
 254                         $last = $#some_pages;
 255                 }
 256                 my @slice = @some_pages[0..$last];
 257                 get_mw_first_pages(\@slice, $pages);
 258                 @some_pages = @some_pages[51..$#some_pages];
 259         }
 260 }
 261
 262 sub get_mw_tracked_categories {
 263         my $pages = shift;
 264         foreach my $category (@tracked_categories) {
 265                 if (index($category, ':') < 0) {
 266                         # Mediawiki requires the Category
 267                         # prefix, but let's not force the user
 268                         # to specify it.
 269                         $category = "Category:" . $category;
 270                 }
 271                 my $mw_pages = $mediawiki->list( {
 272                         action => 'query',
 273                         list => 'categorymembers',
 274                         cmtitle => $category,
 275                         cmlimit => 'max' } )
 276                         || die $mediawiki->{error}->{code} . ': '
 277                                 . $mediawiki->{error}->{details};
 278                 foreach my $page (@{$mw_pages}) {
 279                         $pages->{$page->{title}} = $page;
 280                 }
 281         }
 282 }
 283
 284 sub get_mw_all_pages {
 285         my $pages = shift;
 286         # No user-provided list, get the list of pages from the API.
 287         my $mw_pages = $mediawiki->list({
 288                 action => 'query',
 289                 list => 'allpages',
 290                 aplimit => 'max'
 291         });
 292         if (!defined($mw_pages)) {
 293                 print STDERR "fatal: could not get the list of wiki pages.\n";
 294                 print STDERR "fatal: '$url' does not appear to be a mediawiki\n";
 295                 print STDERR "fatal: make sure '$url/api.php' is a valid page.\n";
 296                 exit 1;
 297         }
 298         foreach my $page (@{$mw_pages}) {
 299                 $pages->{$page->{title}} = $page;
 300         }
 301 }
 302
 303 # queries the wiki for a set of pages. Meant to be used within a loop
 304 # querying the wiki for slices of page list.
 305 sub get_mw_first_pages {
 306         my $some_pages = shift;
 307         my @some_pages = @{$some_pages};
 308
 309         my $pages = shift;
 310
 311         # pattern 'page1|page2|...' required by the API
 312         my $titles = join('|', @some_pages);
 313
 314         my $mw_pages = $mediawiki->api({
 315                 action => 'query',
 316                 titles => $titles,
 317         });
 318         if (!defined($mw_pages)) {
 319                 print STDERR "fatal: could not query the list of wiki pages.\n";
 320                 print STDERR "fatal: '$url' does not appear to be a mediawiki\n";
 321                 print STDERR "fatal: make sure '$url/api.php' is a valid page.\n";
 322                 exit 1;
 323         }
 324         while (my ($id, $page) = each(%{$mw_pages->{query}->{pages}})) {
 325                 if ($id < 0) {
 326                         print STDERR "Warning: page $page->{title} not found on wiki\n";
 327                 } else {
 328                         $pages->{$page->{title}} = $page;
 329                 }
 330         }
 331 }
 332
 333 # Get the list of pages to be fetched according to configuration.
 334 sub get_mw_pages {
 335         mw_connect_maybe();
 336
 337         print STDERR "Listing pages on remote wiki...\n";
 338
 339         my %pages; # hash on page titles to avoid duplicates
 340         my $user_defined;
 341         if (@tracked_pages) {
 342                 $user_defined = 1;
 343                 # The user provided a list of pages titles, but we
 344                 # still need to query the API to get the page IDs.
 345                 get_mw_tracked_pages(\%pages);
 346         }
 347         if (@tracked_categories) {
 348                 $user_defined = 1;
 349                 get_mw_tracked_categories(\%pages);
 350         }
 351         if (!$user_defined) {
 352                 get_mw_all_pages(\%pages);
 353         }
 354         if ($import_media) {
 355                 print STDERR "Getting media files for selected pages...\n";
 356                 if ($user_defined) {
 357                         get_linked_mediafiles(\%pages);
 358                 } else {
 359                         get_all_mediafiles(\%pages);
 360                 }
 361         }
 362         print STDERR (scalar keys %pages) . " pages found.\n";
 363         return %pages;
 364 }
 365
 366 # usage: $out = run_git("command args");
 367 #        $out = run_git("command args", "raw"); # don't interpret output as UTF-8.
 368 sub run_git {
 369         my $args = shift;
 370         my $encoding = (shift || "encoding(UTF-8)");
 371         open(my $git, "-|:$encoding", "git " . $args);
 372         my $res = do { local $/; <$git> };
 373         close($git);
 374
 375         return $res;
 376 }
 377
 378
 379 sub get_all_mediafiles {
 380         my $pages = shift;
 381         # Attach list of all pages for media files from the API,
 382         # they are in a different namespace, only one namespace
 383         # can be queried at the same moment
 384         my $mw_pages = $mediawiki->list({
 385                 action => 'query',
 386                 list => 'allpages',
 387                 apnamespace => get_mw_namespace_id("File"),
 388                 aplimit => 'max'
 389         });
 390         if (!defined($mw_pages)) {
 391                 print STDERR "fatal: could not get the list of pages for media files.\n";
 392                 print STDERR "fatal: '$url' does not appear to be a mediawiki\n";
 393                 print STDERR "fatal: make sure '$url/api.php' is a valid page.\n";
 394                 exit 1;
 395         }
 396         foreach my $page (@{$mw_pages}) {
 397                 $pages->{$page->{title}} = $page;
 398         }
 399 }
 400
 401 sub get_linked_mediafiles {
 402         my $pages = shift;
 403         my @titles = map $_->{title}, values(%{$pages});
 404
 405         # The query is split in small batches because of the MW API limit of
 406         # the number of links to be returned (500 links max).
 407         my $batch = 10;
 408         while (@titles) {
 409                 if ($#titles < $batch) {
 410                         $batch = $#titles;
 411                 }
 412                 my @slice = @titles[0..$batch];
 413
 414                 # pattern 'page1|page2|...' required by the API
 415                 my $mw_titles = join('|', @slice);
 416
 417                 # Media files could be included or linked from
 418                 # a page, get all related
 419                 my $query = {
 420                         action => 'query',
 421                         prop => 'links|images',
 422                         titles => $mw_titles,
 423                         plnamespace => get_mw_namespace_id("File"),
 424                         pllimit => 'max'
 425                 };
 426                 my $result = $mediawiki->api($query);
 427
 428                 while (my ($id, $page) = each(%{$result->{query}->{pages}})) {
 429                         my @media_titles;
 430                         if (defined($page->{links})) {
 431                                 my @link_titles = map $_->{title}, @{$page->{links}};
 432                                 push(@media_titles, @link_titles);
 433                         }
 434                         if (defined($page->{images})) {
 435                                 my @image_titles = map $_->{title}, @{$page->{images}};
 436                                 push(@media_titles, @image_titles);
 437                         }
 438                         if (@media_titles) {
 439                                 get_mw_page_list(\@media_titles, $pages);
 440                         }
 441                 }
 442
 443                 @titles = @titles[($batch+1)..$#titles];
 444         }
 445 }
 446
 447 sub get_mw_mediafile_for_page_revision {
 448         # Name of the file on Wiki, with the prefix.
 449         my $filename = shift;
 450         my $timestamp = shift;
 451         my %mediafile;
 452
 453         # Search if on a media file with given timestamp exists on
 454         # MediaWiki. In that case download the file.
 455         my $query = {
 456                 action => 'query',
 457                 prop => 'imageinfo',
 458                 titles => "File:" . $filename,
 459                 iistart => $timestamp,
 460                 iiend => $timestamp,
 461                 iiprop => 'timestamp|archivename|url',
 462                 iilimit => 1
 463         };
 464         my $result = $mediawiki->api($query);
 465
 466         my ($fileid, $file) = each( %{$result->{query}->{pages}} );
 467         # If not defined it means there is no revision of the file for
 468         # given timestamp.
 469         if (defined($file->{imageinfo})) {
 470                 $mediafile{title} = $filename;
 471
 472                 my $fileinfo = pop(@{$file->{imageinfo}});
 473                 $mediafile{timestamp} = $fileinfo->{timestamp};
 474                 # Mediawiki::API's download function doesn't support https URLs
 475                 # and can't download old versions of files.
 476                 print STDERR "\tDownloading file $mediafile{title}, version $mediafile{timestamp}\n";
 477                 $mediafile{content} = download_mw_mediafile($fileinfo->{url});
 478         }
 479         return %mediafile;
 480 }
 481
 482 sub download_mw_mediafile {
 483         my $url = shift;
 484
 485         my $response = $mediawiki->{ua}->get($url);
 486         if ($response->code == 200) {
 487                 return $response->decoded_content;
 488         } else {
 489                 print STDERR "Error downloading mediafile from :\n";
 490                 print STDERR "URL: $url\n";
 491                 print STDERR "Server response: " . $response->code . " " . $response->message . "\n";
 492                 exit 1;
 493         }
 494 }
 495
 496 sub get_last_local_revision {
 497         # Get note regarding last mediawiki revision
 498         my $note = run_git("notes --ref=$remotename/mediawiki show refs/mediawiki/$remotename/master 2>/dev/null");
 499         my @note_info = split(/ /, $note);
 500
 501         my $lastrevision_number;
 502         if (!(defined($note_info[0]) && $note_info[0] eq "mediawiki_revision:")) {
 503                 print STDERR "No previous mediawiki revision found";
 504                 $lastrevision_number = 0;
 505         } else {
 506                 # Notes are formatted : mediawiki_revision: #number
 507                 $lastrevision_number = $note_info[1];
 508                 chomp($lastrevision_number);
 509                 print STDERR "Last local mediawiki revision found is $lastrevision_number";
 510         }
 511         return $lastrevision_number;
 512 }
 513
 514 # Remember the timestamp corresponding to a revision id.
 515 my %basetimestamps;
 516
 517 # Get the last remote revision without taking in account which pages are
 518 # tracked or not. This function makes a single request to the wiki thus
 519 # avoid a loop onto all tracked pages. This is useful for the fetch-by-rev
 520 # option.
 521 sub get_last_global_remote_rev {
 522         mw_connect_maybe();
 523
 524         my $query = {
 525                 action => 'query',
 526                 list => 'recentchanges',
 527                 prop => 'revisions',
 528                 rclimit => '1',
 529                 rcdir => 'older',
 530         };
 531         my $result = $mediawiki->api($query);
 532         return $result->{query}->{recentchanges}[0]->{revid};
 533 }
 534
 535 # Get the last remote revision concerning the tracked pages and the tracked
 536 # categories.
 537 sub get_last_remote_revision {
 538         mw_connect_maybe();
 539
 540         my %pages_hash = get_mw_pages();
 541         my @pages = values(%pages_hash);
 542
 543         my $max_rev_num = 0;
 544
 545         print STDERR "Getting last revision id on tracked pages...\n";
 546
 547         foreach my $page (@pages) {
 548                 my $id = $page->{pageid};
 549
 550                 my $query = {
 551                         action => 'query',
 552                         prop => 'revisions',
 553                         rvprop => 'ids|timestamp',
 554                         pageids => $id,
 555                 };
 556
 557                 my $result = $mediawiki->api($query);
 558
 559                 my $lastrev = pop(@{$result->{query}->{pages}->{$id}->{revisions}});
 560
 561                 $basetimestamps{$lastrev->{revid}} = $lastrev->{timestamp};
 562
 563                 $max_rev_num = ($lastrev->{revid} > $max_rev_num ? $lastrev->{revid} : $max_rev_num);
 564         }
 565
 566         print STDERR "Last remote revision found is $max_rev_num.\n";
 567         return $max_rev_num;
 568 }
 569
 570 # Clean content before sending it to MediaWiki
 571 sub mediawiki_clean {
 572         my $string = shift;
 573         my $page_created = shift;
 574         # Mediawiki does not allow blank space at the end of a page and ends with a single \n.
 575         # This function right trims a string and adds a \n at the end to follow this rule
 576         $string =~ s/\s+$//;
 577         if ($string eq "" && $page_created) {
 578                 # Creating empty pages is forbidden.
 579                 $string = EMPTY_CONTENT;
 580         }
 581         return $string."\n";
 582 }
 583
 584 # Filter applied on MediaWiki data before adding them to Git
 585 sub mediawiki_smudge {
 586         my $string = shift;
 587         if ($string eq EMPTY_CONTENT) {
 588                 $string = "";
 589         }
 590         # This \n is important. This is due to mediawiki's way to handle end of files.
 591         return $string."\n";
 592 }
 593
 594 sub mediawiki_clean_filename {
 595         my $filename = shift;
 596         $filename =~ s/@{[SLASH_REPLACEMENT]}/\//g;
 597         # [, ], |, {, and } are forbidden by MediaWiki, even URL-encoded.
 598         # Do a variant of URL-encoding, i.e. looks like URL-encoding,
 599         # but with _ added to prevent MediaWiki from thinking this is
 600         # an actual special character.
 601         $filename =~ s/[\[\]\{\}\|]/sprintf("_%%_%x", ord($&))/ge;
 602         # If we use the uri escape before
 603         # we should unescape here, before anything
 604
 605         return $filename;
 606 }
 607
 608 sub mediawiki_smudge_filename {
 609         my $filename = shift;
 610         $filename =~ s/\//@{[SLASH_REPLACEMENT]}/g;
 611         $filename =~ s/ /_/g;
 612         # Decode forbidden characters encoded in mediawiki_clean_filename
 613         $filename =~ s/_%_([0-9a-fA-F][0-9a-fA-F])/sprintf("%c", hex($1))/ge;
 614         return $filename;
 615 }
 616
 617 sub literal_data {
 618         my ($content) = @_;
 619         print STDOUT "data ", bytes::length($content), "\n", $content;
 620 }
 621
 622 sub literal_data_raw {
 623         # Output possibly binary content.
 624         my ($content) = @_;
 625         # Avoid confusion between size in bytes and in characters
 626         utf8::downgrade($content);
 627         binmode STDOUT, ":raw";
 628         print STDOUT "data ", bytes::length($content), "\n", $content;
 629         binmode STDOUT, ":utf8";
 630 }
 631
 632 sub mw_capabilities {
 633         # Revisions are imported to the private namespace
 634         # refs/mediawiki/$remotename/ by the helper and fetched into
 635         # refs/remotes/$remotename later by fetch.
 636         print STDOUT "refspec refs/heads/*:refs/mediawiki/$remotename/*\n";
 637         print STDOUT "import\n";
 638         print STDOUT "list\n";
 639         print STDOUT "push\n";
 640         print STDOUT "\n";
 641 }
 642
 643 sub mw_list {
 644         # MediaWiki do not have branches, we consider one branch arbitrarily
 645         # called master, and HEAD pointing to it.
 646         print STDOUT "? refs/heads/master\n";
 647         print STDOUT "\@refs/heads/master HEAD\n";
 648         print STDOUT "\n";
 649 }
 650
 651 sub mw_option {
 652         print STDERR "remote-helper command 'option $_[0]' not yet implemented\n";
 653         print STDOUT "unsupported\n";
 654 }
 655
 656 sub fetch_mw_revisions_for_page {
 657         my $page = shift;
 658         my $id = shift;
 659         my $fetch_from = shift;
 660         my @page_revs = ();
 661         my $query = {
 662                 action => 'query',
 663                 prop => 'revisions',
 664                 rvprop => 'ids',
 665                 rvdir => 'newer',
 666                 rvstartid => $fetch_from,
 667                 rvlimit => 500,
 668                 pageids => $id,
 669         };
 670
 671         my $revnum = 0;
 672         # Get 500 revisions at a time due to the mediawiki api limit
 673         while (1) {
 674                 my $result = $mediawiki->api($query);
 675
 676                 # Parse each of those 500 revisions
 677                 foreach my $revision (@{$result->{query}->{pages}->{$id}->{revisions}}) {
 678                         my $page_rev_ids;
 679                         $page_rev_ids->{pageid} = $page->{pageid};
 680                         $page_rev_ids->{revid} = $revision->{revid};
 681                         push(@page_revs, $page_rev_ids);
 682                         $revnum++;
 683                 }
 684                 last unless $result->{'query-continue'};
 685                 $query->{rvstartid} = $result->{'query-continue'}->{revisions}->{rvstartid};
 686         }
 687         if ($shallow_import && @page_revs) {
 688                 print STDERR "  Found 1 revision (shallow import).\n";
 689                 @page_revs = sort {$b->{revid} <=> $a->{revid}} (@page_revs);
 690                 return $page_revs[0];
 691         }
 692         print STDERR "  Found ", $revnum, " revision(s).\n";
 693         return @page_revs;
 694 }
 695
 696 sub fetch_mw_revisions {
 697         my $pages = shift; my @pages = @{$pages};
 698         my $fetch_from = shift;
 699
 700         my @revisions = ();
 701         my $n = 1;
 702         foreach my $page (@pages) {
 703                 my $id = $page->{pageid};
 704
 705                 print STDERR "page $n/", scalar(@pages), ": ". $page->{title} ."\n";
 706                 $n++;
 707                 my @page_revs = fetch_mw_revisions_for_page($page, $id, $fetch_from);
 708                 @revisions = (@page_revs, @revisions);
 709         }
 710
 711         return ($n, @revisions);
 712 }
 713
 714 sub import_file_revision {
 715         my $commit = shift;
 716         my %commit = %{$commit};
 717         my $full_import = shift;
 718         my $n = shift;
 719         my $mediafile = shift;
 720         my %mediafile;
 721         if ($mediafile) {
 722                 %mediafile = %{$mediafile};
 723         }
 724
 725         my $title = $commit{title};
 726         my $comment = $commit{comment};
 727         my $content = $commit{content};
 728         my $author = $commit{author};
 729         my $date = $commit{date};
 730
 731         print STDOUT "commit refs/mediawiki/$remotename/master\n";
 732         print STDOUT "mark :$n\n";
 733         print STDOUT "committer $author <$author\@$wiki_name> ", $date->epoch, " +0000\n";
 734         literal_data($comment);
 735
 736         # If it's not a clone, we need to know where to start from
 737         if (!$full_import && $n == 1) {
 738                 print STDOUT "from refs/mediawiki/$remotename/master^0\n";
 739         }
 740         if ($content ne DELETED_CONTENT) {
 741                 print STDOUT "M 644 inline $title.mw\n";
 742                 literal_data($content);
 743                 if (%mediafile) {
 744                         print STDOUT "M 644 inline $mediafile{title}\n";
 745                         literal_data_raw($mediafile{content});
 746                 }
 747                 print STDOUT "\n\n";
 748         } else {
 749                 print STDOUT "D $title.mw\n";
 750         }
 751
 752         # mediawiki revision number in the git note
 753         if ($full_import && $n == 1) {
 754                 print STDOUT "reset refs/notes/$remotename/mediawiki\n";
 755         }
 756         print STDOUT "commit refs/notes/$remotename/mediawiki\n";
 757         print STDOUT "committer $author <$author\@$wiki_name> ", $date->epoch, " +0000\n";
 758         literal_data("Note added by git-mediawiki during import");
 759         if (!$full_import && $n == 1) {
 760                 print STDOUT "from refs/notes/$remotename/mediawiki^0\n";
 761         }
 762         print STDOUT "N inline :$n\n";
 763         literal_data("mediawiki_revision: " . $commit{mw_revision});
 764         print STDOUT "\n\n";
 765 }
 766
 767 # parse a sequence of
 768 # <cmd> <arg1>
 769 # <cmd> <arg2>
 770 # \n
 771 # (like batch sequence of import and sequence of push statements)
 772 sub get_more_refs {
 773         my $cmd = shift;
 774         my @refs;
 775         while (1) {
 776                 my $line = <STDIN>;
 777                 if ($line =~ m/^$cmd (.*)$/) {
 778                         push(@refs, $1);
 779                 } elsif ($line eq "\n") {
 780                         return @refs;
 781                 } else {
 782                         die("Invalid command in a '$cmd' batch: ". $_);
 783                 }
 784         }
 785 }
 786
 787 sub mw_import {
 788         # multiple import commands can follow each other.
 789         my @refs = (shift, get_more_refs("import"));
 790         foreach my $ref (@refs) {
 791                 mw_import_ref($ref);
 792         }
 793         print STDOUT "done\n";
 794 }
 795
 796 sub mw_import_ref {
 797         my $ref = shift;
 798         # The remote helper will call "import HEAD" and
 799         # "import refs/heads/master".
 800         # Since HEAD is a symbolic ref to master (by convention,
 801         # followed by the output of the command "list" that we gave),
 802         # we don't need to do anything in this case.
 803         if ($ref eq "HEAD") {
 804                 return;
 805         }
 806
 807         mw_connect_maybe();
 808
 809         print STDERR "Searching revisions...\n";
 810         my $last_local = get_last_local_revision();
 811         my $fetch_from = $last_local + 1;
 812         if ($fetch_from == 1) {
 813                 print STDERR ", fetching from beginning.\n";
 814         } else {
 815                 print STDERR ", fetching from here.\n";
 816         }
 817
 818         my $n = 0;
 819         if ($fetch_strategy eq "by_rev") {
 820                 print STDERR "Fetching & writing export data by revs...\n";
 821                 $n = mw_import_ref_by_revs($fetch_from);
 822         } elsif ($fetch_strategy eq "by_page") {
 823                 print STDERR "Fetching & writing export data by pages...\n";
 824                 $n = mw_import_ref_by_pages($fetch_from);
 825         } else {
 826                 print STDERR "fatal: invalid fetch strategy \"$fetch_strategy\".\n";
 827                 print STDERR "Check your configuration variables remote.$remotename.fetchStrategy and mediawiki.fetchStrategy\n";
 828                 exit 1;
 829         }
 830
 831         if ($fetch_from == 1 && $n == 0) {
 832                 print STDERR "You appear to have cloned an empty MediaWiki.\n";
 833                 # Something has to be done remote-helper side. If nothing is done, an error is
 834                 # thrown saying that HEAD is refering to unknown object 0000000000000000000
 835                 # and the clone fails.
 836         }
 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 $pagerevid";
 892                 }
 893
 894                 if (defined($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 $pagerevid.";
 901                 }
 902
 903                 my @result_pages = values(%{$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};
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 priviledges. 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};
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>");
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 }
1144
1145 sub mw_push_revision {
1146         my $local = shift;
1147         my $remote = shift; # actually, this has to be "refs/heads/master" at this point.
1148         my $last_local_revid = get_last_local_revision();
1149         print STDERR ".\n"; # Finish sentence started by get_last_local_revision()
1150         my $last_remote_revid = get_last_remote_revision();
1151         my $mw_revision = $last_remote_revid;
1152
1153         # Get sha1 of commit pointed by local HEAD
1154         my $HEAD_sha1 = run_git("rev-parse $local 2>/dev/null"); chomp($HEAD_sha1);
1155         # Get sha1 of commit pointed by remotes/$remotename/master
1156         my $remoteorigin_sha1 = run_git("rev-parse refs/remotes/$remotename/master 2>/dev/null");
1157         chomp($remoteorigin_sha1);
1158
1159         if ($last_local_revid > 0 &&
1160             $last_local_revid < $last_remote_revid) {
1161                 return error_non_fast_forward($remote);
1162         }
1163
1164         if ($HEAD_sha1 eq $remoteorigin_sha1) {
1165                 # nothing to push
1166                 return 0;
1167         }
1168
1169         # Get every commit in between HEAD and refs/remotes/origin/master,
1170         # including HEAD and refs/remotes/origin/master
1171         my @commit_pairs = ();
1172         if ($last_local_revid > 0) {
1173                 my $parsed_sha1 = $remoteorigin_sha1;
1174                 # Find a path from last MediaWiki commit to pushed commit
1175                 print STDERR "Computing path from local to remote ...\n";
1176                 my @local_ancestry = split(/\n/, run_git("rev-list --boundary --parents $local ^$parsed_sha1"));
1177                 my %local_ancestry;
1178                 foreach my $line (@local_ancestry) {
1179                         if (my ($child, $parents) = $line =~ m/^-?([a-f0-9]+) ([a-f0-9 ]+)/) {
1180                                 foreach my $parent (split(' ', $parents)) {
1181                                         $local_ancestry{$parent} = $child;
1182                                 }
1183                         } elsif (!$line =~ m/^([a-f0-9]+)/) {
1184                                 die "Unexpected output from git rev-list: $line";
1185                         }
1186                 }
1187                 while ($parsed_sha1 ne $HEAD_sha1) {
1188                         my $child = $local_ancestry{$parsed_sha1};
1189                         if (!$child) {
1190                                 printf STDERR "Cannot find a path in history from remote commit to last commit\n";
1191                                 return error_non_fast_forward($remote);
1192                         }
1193                         push(@commit_pairs, [$parsed_sha1, $child]);
1194                         $parsed_sha1 = $child;
1195                 }
1196         } else {
1197                 # No remote mediawiki revision. Export the whole
1198                 # history (linearized with --first-parent)
1199                 print STDERR "Warning: no common ancestor, pushing complete history\n";
1200                 my $history = run_git("rev-list --first-parent --children $local");
1201                 my @history = split('\n', $history);
1202                 @history = @history[1..$#history];
1203                 foreach my $line (reverse @history) {
1204                         my @commit_info_split = split(/ |\n/, $line);
1205                         push(@commit_pairs, \@commit_info_split);
1206                 }
1207         }
1208
1209         foreach my $commit_info_split (@commit_pairs) {
1210                 my $sha1_child = @{$commit_info_split}[0];
1211                 my $sha1_commit = @{$commit_info_split}[1];
1212                 my $diff_infos = run_git("diff-tree -r --raw -z $sha1_child $sha1_commit");
1213                 # TODO: we could detect rename, and encode them with a #redirect on the wiki.
1214                 # TODO: for now, it's just a delete+add
1215                 my @diff_info_list = split(/\0/, $diff_infos);
1216                 # Keep the subject line of the commit message as mediawiki comment for the revision
1217                 my $commit_msg = run_git("log --no-walk --format=\"%s\" $sha1_commit");
1218                 chomp($commit_msg);
1219                 # Push every blob
1220                 while (@diff_info_list) {
1221                         my $status;
1222                         # git diff-tree -z gives an output like
1223                         # <metadata>\0<filename1>\0
1224                         # <metadata>\0<filename2>\0
1225                         # and we've split on \0.
1226                         my $info = shift(@diff_info_list);
1227                         my $file = shift(@diff_info_list);
1228                         ($mw_revision, $status) = mw_push_file($info, $file, $commit_msg, $mw_revision);
1229                         if ($status eq "non-fast-forward") {
1230                                 # we may already have sent part of the
1231                                 # commit to MediaWiki, but it's too
1232                                 # late to cancel it. Stop the push in
1233                                 # the middle, but still give an
1234                                 # accurate error message.
1235                                 return error_non_fast_forward($remote);
1236                         }
1237                         if ($status ne "ok") {
1238                                 die("Unknown error from mw_push_file()");
1239                         }
1240                 }
1241                 unless ($dumb_push) {
1242                         run_git("notes --ref=$remotename/mediawiki add -f -m \"mediawiki_revision: $mw_revision\" $sha1_commit");
1243                         run_git("update-ref -m \"Git-MediaWiki push\" refs/mediawiki/$remotename/master $sha1_commit $sha1_child");
1244                 }
1245         }
1246
1247         print STDOUT "ok $remote\n";
1248         return 1;
1249 }
1250
1251 sub get_allowed_file_extensions {
1252         mw_connect_maybe();
1253
1254         my $query = {
1255                 action => 'query',
1256                 meta => 'siteinfo',
1257                 siprop => 'fileextensions'
1258                 };
1259         my $result = $mediawiki->api($query);
1260         my @file_extensions= map $_->{ext},@{$result->{query}->{fileextensions}};
1261         my %hashFile = map {$_ => 1}@file_extensions;
1262
1263         return %hashFile;
1264 }
1265
1266 # In memory cache for MediaWiki namespace ids.
1267 my %namespace_id;
1268
1269 # Namespaces whose id is cached in the configuration file
1270 # (to avoid duplicates)
1271 my %cached_mw_namespace_id;
1272
1273 # Return MediaWiki id for a canonical namespace name.
1274 # Ex.: "File", "Project".
1275 sub get_mw_namespace_id {
1276         mw_connect_maybe();
1277         my $name = shift;
1278
1279         if (!exists $namespace_id{$name}) {
1280                 # Look at configuration file, if the record for that namespace is
1281                 # already cached. Namespaces are stored in form:
1282                 # "Name_of_namespace:Id_namespace", ex.: "File:6".
1283                 my @temp = split(/[\n]/, run_git("config --get-all remote."
1284                                                 . $remotename .".namespaceCache"));
1285                 chomp(@temp);
1286                 foreach my $ns (@temp) {
1287                         my ($n, $id) = split(/:/, $ns);
1288                         if ($id eq 'notANameSpace') {
1289                                 $namespace_id{$n} = {is_namespace => 0};
1290                         } else {
1291                                 $namespace_id{$n} = {is_namespace => 1, id => $id};
1292                         }
1293                         $cached_mw_namespace_id{$n} = 1;
1294                 }
1295         }
1296
1297         if (!exists $namespace_id{$name}) {
1298                 print STDERR "Namespace $name not found in cache, querying the wiki ...\n";
1299                 # NS not found => get namespace id from MW and store it in
1300                 # configuration file.
1301                 my $query = {
1302                         action => 'query',
1303                         meta => 'siteinfo',
1304                         siprop => 'namespaces'
1305                 };
1306                 my $result = $mediawiki->api($query);
1307
1308                 while (my ($id, $ns) = each(%{$result->{query}->{namespaces}})) {
1309                         if (defined($ns->{id}) && defined($ns->{canonical})) {
1310                                 $namespace_id{$ns->{canonical}} = {is_namespace => 1, id => $ns->{id}};
1311                                 if ($ns->{'*'}) {
1312                                         # alias (e.g. french Fichier: as alias for canonical File:)
1313                                         $namespace_id{$ns->{'*'}} = {is_namespace => 1, id => $ns->{id}};
1314                                 }
1315                         }
1316                 }
1317         }
1318
1319         my $ns = $namespace_id{$name};
1320         my $id;
1321
1322         unless (defined $ns) {
1323                 print STDERR "No such namespace $name on MediaWiki.\n";
1324                 $ns = {is_namespace => 0};
1325                 $namespace_id{$name} = $ns;
1326         }
1327
1328         if ($ns->{is_namespace}) {
1329                 $id = $ns->{id};
1330         }
1331
1332         # Store "notANameSpace" as special value for inexisting namespaces
1333         my $store_id = ($id || 'notANameSpace');
1334
1335         # Store explicitely requested namespaces on disk
1336         if (!exists $cached_mw_namespace_id{$name}) {
1337                 run_git("config --add remote.". $remotename
1338                         .".namespaceCache \"". $name .":". $store_id ."\"");
1339                 $cached_mw_namespace_id{$name} = 1;
1340         }
1341         return $id;
1342 }
1343
1344 sub get_mw_namespace_id_for_page {
1345         if (my ($namespace) = $_[0] =~ /^([^:]*):/) {
1346                 return get_mw_namespace_id($namespace);
1347         } else {
1348                 return;
1349         }
1350 }