Bugzilla – Bug 14549
cannot parse some opml file with bad characters
Last modified: 2010-02-17 09:43:21 UTC
I'm using this opml list to get french radio podcast http://www.netsecureweb.com/radiofr.opml podcast url are generated automatically and some channels sent bad characters. For example there are many "" (it seems unicode 0019 character: "^Y" in ascii traces) instead of single quote "'" <outline text="" title="On va sgêner" xmlUrl="http://www.europe1.fr/rss_export/feed/on-va-s-gener-laurent-ruquier" type="rss"/> I know the issue is mostly on this file but to prevent error it should be good to put a workaround in xmlToHash, which does already contain some encoding workaround I do not understand why utf8decode_guess does not correct encoding issue here are my traces before and after utf8decode_guess call 09-10-03 15:21:23.6144] Slim::Formats::XML::xmlToHash (630) before: <outline text="" title="On va s^YgM-CM-*ner" xmlUrl="http://www.europe1.fr/rss_export/feed/on-va-s-gener-laurent-ruquier" type="rss"/> [09-10-03 15:21:23.6147] Slim::Formats::XML::xmlToHash (632) after: <outline text="" title="On va s^YgM-CM-*ner" xmlUrl="http://www.europe1.fr/rss_export/feed/on-va-s-gener-laurent-ruquier" type="rss"/>
Andy/Michael: your thoughts on this?
I've tested on 7.3.X because squeezescrobbler plugin does not work well on 7.4.X, and the issue is the same on this version. It's clear that issue is in opml file but to prevent browsing failed in opml file, it may be converted to right charset or just ignore related item
OPML must be valid XML.