Issue 2507: Oversized allocation does not fail correctly

>
> http://code.google.com/p/pharo/issues/detail?id=2507
>
> The following test in KernelTests-Exception fails in PharoCore1.1
>
> AllocationTest>>testOneGigAllocation
>
> This test tries to allocate one gig of memory and verify that it either succeeds with the right size of array, or
> allocation should fail. Instead we get an array of size 0.
>
> Curiously:
>
> (Array new: 1024*1024*1023) size. -> "Warning! Pharo is almost out of memory!"
> (Array new: 1024*1024*1024) size. -> 0
> (Array new: 1024*1024*1025) size. -> 1048576
>
> Array>>new: runs primitive 71, so this appears to be a VM bug.
>
> Oscar Nierstrasz
>
>
> _______________________________________________
> Pharo-project mailing list
> [hidden email]
> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

>
> Fixed in 4.2.4 VM.
>
> - on
>
> On 5 Jun 2010, at 11:39, Oscar Nierstrasz wrote:
>
> >
> > http://code.google.com/p/pharo/issues/detail?id=2507
> >
> > The following test in KernelTests-Exception fails in PharoCore1.1
> >
> > AllocationTest>>testOneGigAllocation
> >
> > This test tries to allocate one gig of memory and verify that it either succeeds with the right size of array, or
> > allocation should fail. Instead we get an array of size 0.
> >
> > Curiously:
> >
> > (Array new: 1024*1024*1023) size. -> "Warning! Pharo is almost out of memory!"
> > (Array new: 1024*1024*1024) size. -> 0
> > (Array new: 1024*1024*1025) size. -> 1048576
> >
> > Array>>new: runs primitive 71, so this appears to be a VM bug.
> >
> > Oscar Nierstrasz

> I suggested a couple of times to update the download link of the Mac vm of
> the website to 4.2.4.
>
> Adrian, can we do this?
>
> Cheers
>
> Mariano
>
> On Sat, Jun 5, 2010 at 12:39 PM, Oscar Nierstrasz <[hidden email]>wrote:
>
>>
>> Fixed in 4.2.4 VM.
>>
>> - on
>>
>> On 5 Jun 2010, at 11:39, Oscar Nierstrasz wrote:
>>
>>>
>>> http://code.google.com/p/pharo/issues/detail?id=2507
>>>
>>> The following test in KernelTests-Exception fails in PharoCore1.1
>>>
>>> AllocationTest>>testOneGigAllocation
>>>
>>> This test tries to allocate one gig of memory and verify that it either
>> succeeds with the right size of array, or
>>> allocation should fail. Instead we get an array of size 0.
>>>
>>> Curiously:
>>>
>>> (Array new: 1024*1024*1023) size. -> "Warning! Pharo is almost out of
>> memory!"
>>> (Array new: 1024*1024*1024) size. -> 0
>>> (Array new: 1024*1024*1025) size. -> 1048576
>>>
>>> Array>>new: runs primitive 71, so this appears to be a VM bug.
>>>
>>> Oscar Nierstrasz
>>>
>>>
>>> _______________________________________________
>>> Pharo-project mailing list
>>> [hidden email]
>>> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
>>
>>
>> _______________________________________________
>> Pharo-project mailing list
>> [hidden email]
>> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
>>
> _______________________________________________
> Pharo-project mailing list
> [hidden email]
> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project