This is caused by the API not escaping the & in the html encoded characters.
Mods can go around it but mistakes are bound to happen if it remains, if nothing else user might misunderstand and think they don't need to utf-8 to html convert the kanji signs.
I guess by API I mean CGI, sorry. The API works correct for these things since the only garbling it does is through the character set and we know about that already.
hm, it seems that this is browser dependend.
i.e. using opera the kanji characters are in deed shown in their decoded form, however opera will still submit the correctly encoded form to anidb.
but you're right of course, this is bound to confuse people.
exp wrote:hm, it seems that this is browser dependend.
i.e. using opera the kanji characters are in deed shown in their decoded form, however opera will still submit the correctly encoded form to anidb.
If Opera does that, it is quite badly broken with respect to the standards.
PetriW wrote:Hmm, my Opera didn't do that last time i checked. (Aka last time I like an idiot forgot to html encode my kanji!)
That's a different situation, and it might handle them differently if, for example, it only decodes character entities for display, leaving them in their original form in internal structures (which is clearly wrong, of course).