On 2010/10/05 01:02:01, chirag wrote: http://codereview.appspot.com/2119043/diff/63001/java/common/conf/shindig.properties
File java/common/conf/shindig.properties (right):
http://codereview.appspot.com/2119043/diff/63001/java/common/conf/shindig.properties#newcode155
java/common/conf/shindig.properties:155: shindig.html.parser=neko Doesn't guice offer the same solution? Currently, we can switch between parsers by updating
ParseModule#configure with
bind(GadgetHtmlParser.class).to(NekoSimplifiedHtmlParser.class);
When you bind GadgetHtmlParser to say NekoSimplifiedHtmlParser, you lose the ability to use neko for say proxy servlet (because existing gadgets are using it) and caja for say accel servlet. Also, adding the router opens more possibilities like: 1) Parsing dom for encoding detection: You have around 1024 bytes of the initial content of a page and you want to determine the encoding. You can try to parse the html and look for <meta http-equiv="Content-Type"> To make this parsing more robust, you could cycle through all the available parsers and choose the one which does not throw an exception. http://codereview.appspot.com/2119043/
