Looking at the patch file, looks good, though I didn't patch it into a copy
of your branch.
One thing: are you doing a linear search GenerateJavaAST multiplied by the
number of functions in the program?

On Tue, Oct 7, 2008 at 2:26 PM, Scott Blum <[EMAIL PROTECTED]> wrote:

> Will review.
>
>
> On Tue, Oct 7, 2008 at 12:22 PM, Lex Spoon <[EMAIL PROTECTED]> wrote:
>
>> Okay, here is a patch that gives the bootstrap entry method an
>> enclosing class.  It went together simply; I simply updated all
>> references to JProgram.entryMethods.  The one non-trivial part was
>> GenerateJavaScriptAST, which used to pick up entry methods by looking
>> at the first (or last?) thing that a JProgram visits.  Now the entry
>> methods get seen at difficult to predict times during the traversal of
>> the JProgram, so GenerateJavaScriptAST has an array to record them as
>> they come by.  It turns out this is the only method I could find that
>> does not have an enclosing class, so the patch includes an assert.
>>
>> I ended up not adding a single entry class, because it would cause
>> awkwardness for the runAsync entry methods.  Each runAsync call
>> generates a class with multiple methods, and one of those methods is
>> an entry method.  To have a single entry class, that method would have
>> to be moved over to the entry class, and so the standard generator
>> framework wouldn't work out of the box.  So, I left it be for now.
>>
>> Can you review it, Scott, or pass it to someone?
>>
>>
>> -Lex
>>
>
>

--~--~---------~--~----~------------~-------~--~----~
http://groups.google.com/group/Google-Web-Toolkit-Contributors
-~----------~----~----~----~------~----~------~--~---

Reply via email to