Page MenuHomePhabricator

A link to a title with two quotes is not parsed correctly
Closed, ResolvedPublic

Description

If a page has two quote symbols in its name, an interwiki link to it will not be
shown properly in the other languages section.

Example: [[nap:Torre d''o Grieco]] is shown as text instead of an interwiki link
to the nap: article.
Workaround: [[nap:Torre d<nowiki>''</nowiki>o Grieco]] solves it, but later all
the bots remove that as improper link.

See the URL link for the specific example.


Version: unspecified
Severity: normal
URL: http://en.wikipedia.org/w/index.php?title=Torre_del_Greco&diff=45981821&oldid=45970406

Details

Reference
bz5454

Event Timeline

bzimport raised the priority of this task from to Medium.Nov 21 2014, 9:10 PM
bzimport added a project: MediaWiki-Parser.
bzimport set Reference to bz5454.
bzimport added a subscriber: Unknown Object (MLST).

Just to point out that the link which fails in the wiki works here, whilst the hacked-up link which works in
the wiki fails here.

Maybe there's something in whatever renders links here which might be of help :-)

Wiki.Melancholie wrote:

Instead of using <nowiki> you could replace one or both apostrophes by: &#x0027;
I do not know whether bots will replace this, too; but I don't think so.

Wiki.Melancholie wrote:

@Yuri: Could you test it with your bot at [[Torre del Greco]] again?

The bot was designed to resolve all &...; numbers to make them readable to the
users. Before utf-8 all links had to be entered that way, and it was a mess.
Now, bot converts any such numbers to readable symbols, which means that
"&#x0027;'" would get converted to "''", unless we introduce some ugly hack.
Regardless, this is clearly a bug in the mediawiki, so hopefully we can get it
resolved.

Most likely we should just ban '' and ''' in title names, as we've banned %[0-9a-f]{2}
to avoid ambiguous and breaky markup-in-titles.

I think banning such titles might be a good solution, but, unlike the %XX, its
not breaking the syntax or causes ambiguity - it is just a bug in the parser
that can be fixed. Maybe this bug will help us structure the parser better? :)

gangleri wrote:

Hallo!

I assume this is a duplicate of
Bug 4598: Order of title recognition and bold / italic rendering should be reversed

Search for such titles will fail; see
Bug 4601: Exact search for titles using php meta / escape characters

However I see this as valid titles. When bug 4598 no escaping will be / should
be required.

best regards reinhardt [[user:gangleri]]

This is exactly a duplicate. A patch is pending review in #4598

  • This bug has been marked as a duplicate of 4598 ***