On Fri, Nov 7, 2008 at 12:54 PM, Scott Blum <[EMAIL PROTECTED]> wrote: > Two qq: > 1) Is this for trunk or 1.6?
It's against trunk, which was my initial impression of where it would probably end up going. It can easily go against 1.6, though, if desired. Just note that the patch will have to be semi-manually replayed, because it affects ControlFlowAnalyzer. > 2) Does this do anything to fix issue #3064? Seems like that bug is highly > related, and might be a compelling enough reason to put this into 1.6. I've been thinking a lot about issue 3064. They are related in that they both deal with method override. I'm not yet sure the best way to address 3064, but the main ideas I've had should be independent of this patch. In general, GenerateJavaScriptAST could insert vtable entries analogously to how Java creates bridge methods. The same method (JS function) could end up being referenced by multiple vtable entries. That leaves open the question of how GenerateJavaScriptAST knows when to insert these entries. One way would be to literally add a bridge JMethod in the appropriate classes. All it would do is call its overridden method. Another way would be to add an extra field to JClassType (or maybe one of its superclasses) that lists the extra vtable entries needed. I haven't gotten further than that, but they look independent so far. Lex --~--~---------~--~----~------------~-------~--~----~ http://groups.google.com/group/Google-Web-Toolkit-Contributors -~----------~----~----~----~------~----~------~--~---
