Infoslicer does not tolerate wiki redirects

[greenfeld]

1. Start Infoslicer
2. Attempt to download the Gato article from Spanish Wikipedia (using an Internet connection).
3. You will get a blank entry. ​http://es.wikipedia.org/wiki/Gato is only a redirect to ​http://es.wikipedia.org/wiki/Felis_silvestris_catus , but does not get followed.

Seen in XO-1.75 os43/Infoslicer-9.

[humitos]

I can reproduce this in Fedora 16 with the git version of the source code.

After typing "Gato" I get a blank screen where it should be the article.

[walter]

aargh. I am loath to dig into this code... but perhaps the time has come. I still haven't removed the hippo dependency either. :P

[godiard]

No problem, I think me or humitos can look at this issue.

[humitos]

Take a look at the content of this page:

• ​http://es.wikipedia.org/w/api.php?action=query&titles=gato&redirects&format=xml

What should we do?

I saw that there is a library called infoslicer/processing/MediaWiki_Helper.py but it seems to be quite obsolete (Copyright (C) IBM Corporation 2008). I didn't find a webpage pointing to the repository of this library or something like that.

Maybe we should update this API for a newer one:

• ​http://www.mediawiki.org/wiki/Alternative_parsers

[humitos]

There problem was that MediaWiki_Helper was following the redirects but when it was calling the redirect it always used the default wiki (English Wikipedia) so it was falling.

I change this behavior to use the same wiki in every redirect. See the attached patch.

[humitos]

Pushed:

• ​http://git.sugarlabs.org/infoslicer/mainline/commit/1bbcccf7e5f24af88f843cb84b845a79e2298407

This will be available in version 15.