[Vm-dev] Re: a problem merging latest changes

[Eliot Miranda]

Oops again.  Failed to deal with nilling of method on initialization.  Fix
uploading. Thanks!

On Mon, Feb 20, 2012 at 11:01 AM, Esteban Lorenzano <estebanlm at gmail.com>wrote:

> cool, thanks :)
>
> btw... I don't know if this helps you, but I compiled a version
> with MULTIPLEBYTECODESETS := true and it compiles at first instance, but
> throws a bad access in gcc3x-cointerp.c:
>
> #  if MULTIPLEBYTECODESETS
> bytecodeSetSelector = ((((headerOf(GIV(method))) >> 1)) < 0
> ? 256
> : 0);
>
> #  endif /* MULTIPLEBYTECODESETS */
>
> (more accurate: at headerOf call)
>
> best,
> Esteban
>
>
> El 20/02/2012, a las 2:55p.m., Eliot Miranda escribió:
>
> Oops.  Fix is uploading now.
>
> On Mon, Feb 20, 2012 at 7:34 AM, Esteban Lorenzano <estebanlm at gmail.com>wrote:
>
>> Hi Eliot, list
>>
>> I'm merging latest changes with my work and I have a problem with some of
>> your changes:
>>
>> StackInterpreter>>#declareCVarsIn:
>>
>> now ads this code:
>>
>> aCCodeGenerator addConstantForBinding: (self bindingOf:
>> #MULTIPLEBYTECODESETS).
>> MULTIPLEBYTECODESETS ifFalse:
>>  [aCCodeGenerator
>> var: #bytecodeSetSelector
>> declareC: '#define bytecodeSetSelector 0'].
>>
>>
>> and MULTIPLEBYTECODESETS is initialized to false
>>
>> now, problem is that
>>
>> aCCodeGenerator
>> var: #bytecodeSetSelector
>>  declareC: '#define bytecodeSetSelector 0'
>>
>> will generate
>>
>> #define bytecodeSetSelector 0;
>>
>> and semicolon with generate a compiler error... so... there is a problem
>> there.
>> How can I change that so it works?
>>
>> btw, if I initialize MULTIPLEBYTECODESETS as true problem gone, but I
>> don't know if that change is appropiate (and also, we need a fix for bad
>> generated code)
>>
>> cheers,
>> Esteban
>>
>>
>
>
> --
> best,
> Eliot
>
>
>


-- 
best,
Eliot
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.squeakfoundation.org/pipermail/vm-dev/attachments/20120220/063d87f0/attachment.htm