Steven Snyder <> changed:

           What    |Removed                     |Added
                 CC|                            |

--- Comment #6 from Steven Snyder <> 2010-10-07 00:06:57 
UTC ---
I'm having a related problem. I'm trying to use interwiki links on a private
wiki to make a request from a database table. The table name has a double
underscore in it like: "INVENTORY__MATERIALS_METAL". The interwiki link looks
like: [[mydatabase:INVENTORY__MATERIALS_METAL]] but the interwiki input parser
removes the double underscore, so it breaks the link.

Passing a double underscore to a wiki site doesn't have any ill effects, so why
not just let the target site process the link itself?

As it is, the documentation (
seems to be wrong, since it says "iw_url is the target of the link; the page
name is substituted for $1"... but in fact, $1 isn't directly substituted, as
it's modified by some parsing code.

If the input string were parsed directly, it wouldn't break the existing "meant
for wikis" purpose, but would at least be more predictable and perhaps more
useful for linking to other sites.

On a side note, "interwikis are meant for wikis" isn't documented; the manual
just says they're shorthand for long links. Further adding to the confusion is
the fact that a MediaWiki install comes by default with interwiki links to
sites like google.

Is there at least a way to disable the interwiki input string processor so I
can get the raw string into my link?

Configure bugmail:
------- You are receiving this mail because: -------
You are on the CC list for the bug.

Wikibugs-l mailing list

Reply via email to