Page MenuHomePhabricator

Using English namespace names shouldn't cause HTTP redirect only
Closed, DeclinedPublic

Description

Author: marco

Description:
When I call the URL "http://de.wikipedia.org/wiki/Image:Beispiel.jpg", I recieve
an HTTP 301 Moved Permanently with "Location:
http://de.wikipedia.org/wiki/Bild:Beispiel.jpg". This is bad for bots as they
have to know the local namespace names and so need a special engine parsing the
HTTP redirs. A workaround would be to simply echo the page under the HTTP
redirect so that a normal browser forwards, but a bot can see the image (or
other namespace).


Version: unspecified
Severity: normal
OS: Windows XP
Platform: PC

Details

Reference
bz6014

Event Timeline

bzimport raised the priority of this task from to Lowest.Nov 21 2014, 9:16 PM
bzimport set Reference to bz6014.
bzimport added a subscriber: Unknown Object (MLST).

Uh, whatever HTTP client lib you are using should be able to cope with
redirects, either implicitely or on demand. Whatever "special engine" are you
talking about? Outputting the full page in adition to the redirect would cause
server load that is useles for 99.9% of all requests. Depending on the client
used, it may also load that even if following the redirect, wasting bandwidth on
both ends.

In short: when you use HTTP, make sure you support the HTTP standard.

marco wrote:

Maybe self-written libraries?

robchur wrote:

Then fix them. As comment 1 points out, it's madness for us to do this.

Add to your bot the ability of, on getting a Location: header, resend the latest
code to the new url.

You can get all the namespace names on Special:Export.