On Fri, Dec 12, 2008 at 2:19 AM, Paul Lindner <plind...@hi5.com> wrote:
> > On Dec 12, 2008, at 2:14 AM, Kevin Brown wrote: > > On Thu, Dec 11, 2008 at 9:39 PM, Paul Lindner <plind...@hi5.com> wrote: >> >> Thanks for parsing the jslint output. >>> >>> FWIW this is the very same JS code that is running in most opensocial >>> production instances today. To say it's not "production ready" just not >>> true. The truth is that this code fails to pass an arbitrary set of >>> guidelines. Try running PMD or findbugs on the Java code and you'll get >>> a >>> simliar list of scary errors that need to be fixed right now! >>> >>> Also, jslint is just stupid in some sections. For example: >>> >>> >>> /home/henning/ning/shindig/git/shindig/features/src/main/javascript/features/core/legacy.js:190:63:'escape' >>>> is undefined. >>>> return window.encodeURIComponent ? encodeURIComponent(str) : >>>> escape(str); >>>> >>>> >>> >>> This is code that attempts to use encodeURIComponent on browsers that >>> support it and falls back to escape() for ancient browsers. This is >>> probably overkill since I doubt that OpenSocial performs very well on >>> Netscape Navigator or IE4 :) >>> >>> In any case, I'll see about cleaning up some of these if it pushes things >>> along. >>> >>> For the future it might be nice to enable the jslint maven plugin for the >>> release target. That combined with standard criteria might serve us well >>> for future revs. >>> >> >> >> We do have it enabled, but the default configuration is worthless, with >> warnings like "only one var statement should be used per scope" polluting >> any useful output. >> > > > That's actually the yui-compressor output... I'm guessing the jslint could > be configured in a more fine-grained manner. > > d'oh, yeah, you're right.