remote-mediawiki: skip virtual namespaces

Virtual namespaces do not correspond to pages in the database and are
automatically generated by MediaWiki. It makes little sense,
therefore, to fetch pages from those namespaces and the MW API doesn't
support listing those pages.

According to the documentation, those virtual namespaces are currently
"Special" (-1) and "Media" (-2) but we treat all negative namespaces
as "virtual" as a future-proofing mechanism.

Signed-off-by: Antoine Beaupré <anarcat@debian.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>

diff --git a/contrib/mw-to-git/git-remote-mediawiki.perl b/contrib/mw-to-git/git-remote-mediawiki.perl
index 5e88458..611a04c 100755 (executable)
--- a/contrib/mw-to-git/git-remote-mediawiki.perl
+++ b/contrib/mw-to-git/git-remote-mediawiki.perl
@@ -264,10 +264,13 @@ sub get_mw_tracked_categories {
 sub get_mw_tracked_namespaces {
     my $pages = shift;
     foreach my $local_namespace (@tracked_namespaces) {
+        my $namespace_id = get_mw_namespace_id($local_namespace);
+        # virtual namespaces don't support allpages
+        next if !defined($namespace_id) || $namespace_id < 0;
         my $mw_pages = $mediawiki->list( {
             action => 'query',
             list => 'allpages',
-            apnamespace => get_mw_namespace_id($local_namespace),
+            apnamespace => $namespace_id,
             aplimit => 'max' } )
             || die $mediawiki->{error}->{code} . ': '
                 . $mediawiki->{error}->{details} . "\n";