[dmd-internals] dmd commit, revision 657

[Sean Kelly]

On Sep 5, 2010, at 12:46 PM, Brad Roberts wrote:

> On 9/5/2010 7:15 AM, Sean Kelly wrote:
>> On Sep 4, 2010, at 9:13 PM, Sean Kelly wrote:
>> 
>>> On Sep 4, 2010, at 4:45 PM, Sean Kelly wrote:
>>> 
>>>> On Sep 4, 2010, at 3:36 PM, Brad Roberts wrote:
>>>> 
>>>>> On 9/4/2010 3:24 PM, Sean Kelly wrote:
>>>>>> Other, I thought Walter had reverted all of the auto-chaining code.
>>>>>> If not, that would explain the second bug. Easy fix.
>>>>>> 
>>>>> 
>>>>> I just ran the druntime unit test and the demangle tests fail there
>>>>> too.
>>>>> 
>>>>> http://d.puremagic.com/issues/show_bug.cgi?id=4795
>>>> 
>>>> That's what I get for not testing it with zero-length input.  I'll give
>>>> it a look tonight.
>>> 
>>> Darnit, I can't reproduce any of these on OSX.
>> 
>> Okay, I think the issue is that static data isn't in ROM on OSX, so throwing
>> classinfo.init for onOutOfMemoryError doesn't break the chaining code as-is.
>> I'll sort out chaining and with luck that will fix everything.
> 
> Looks like druntime r384 and 385 have fixed the test failures.  Clearly visible
> here:
> 
>    http://d.puremagic.com/test-results/index.ghtml
> 
> Do you consider the fix temporary or should bugs 4795 and 4797 be closed?

If they fix the issue then the bugs can be closed.  I just hadn't gotten around to testing the fix yet.