Kelly, since you have experience with this, I'd like you to be the decider (i.e. Freeland is now waiting on your LGTM).
On Fri, Mar 27, 2009 at 2:16 PM, Freeland Abbott <gwt.team.fabb...@gmail.com > wrote: > I think the argument is more for "unnecessary" rather than "bad"... >> although without JsArrayBase (we can make it package-protected, and call it >> JsArrayImpl if anyone cares), we duplicate the JSNI implementation for a >> couple trivial methods. I thought refactoring them into one place was nice, >> but trivial enough that I'm not fighting over it. Revised patch attached; I >> can go either way on this. >> >> >> >> >> >> On Fri, Mar 27, 2009 at 2:06 PM, Scott Blum <sco...@google.com> wrote: >> >>> I'm going to punt this review to Bruce & Kelly, 'cause I have no idea why >>> having JsArrayBase would be bad. :) >>> >>> On Fri, Mar 27, 2009 at 1:28 PM, Freeland Abbott < >>> gwt.team.fabb...@gmail.com> wrote: >>> >>>> Scott, we already talked about this, but here's the patch for public >>>> review. >>>> >>>> The basic goal is to surface the native length, sort, push, and shift >>>> operators for JsArrays... I know you mentioned that IE6's push may be >>>> slower >>>> than indexed extension, and thus a candidate for deferred binding, but I >>>> wanted to get a basic implementation in first. >>>> >>>> There should be only checkstyle changes from what we discussed (though >>>> that obviously doesn't help the rest GWTC). I also added some checkstyle >>>> fixes to JavaScriptObject, introduced by my c5082. >>>> >>> >>> >> > --~--~---------~--~----~------------~-------~--~----~ http://groups.google.com/group/Google-Web-Toolkit-Contributors -~----------~----~----~----~------~----~------~--~---
