While reviewing SubPageList to see if it was good enough quality to install on a wiki of mine, I came round to Validator double checking it (since SubPageList uses Validator).
From the looks of the code, Validator, and various Validator based extensions appear to be using parser->parse() inside of hooks where they are supposed to be using ->recursiveTagParse with a proper frame. The Validator extension's api appears to provoke this bad practice because I don't see the frame in the arguments render methods are using. ~Daniel Friesen (Dantman, Nadir-Seen-Fire) [http://daniel.friesen.name] On 11-01-11 04:23 PM, Platonides wrote: > Well, I just had an issue with Validator, so I am not too sympatatic > with your extension right now ;) > After grepping for setHook, it turns out that an extension like Maps, > that has zero matches, sets parser hooks indirectly via Validator > extension. And not only that, but it also sets a hook for a different > name. It seems to set a hook for<display_map> but actually sets it for > <display map> (Why??) so that even looking for the full tag name doesn't > give you any result. > > It makes things more complex. On the maps case, the parameters need a > deal more validation, but for most cases $parser->setHook() is clearer > than registering a hook to class::staticInit() for a class which extends > ParserHook and has some functions returning the hook configuration. _______________________________________________ Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
