[Vm-dev] latest VMMaker breaks tests in Pharo

[Nicolas Cellier]

Just to confirm what Esteban said: it did not work because an image side
change was missing.
I reopened issue 18034, tested on bleeding edge VM provided by Esteban and
provided a new slice.
The gain is about 30% on [2038074743 isPrime] bench.

2016-04-22 0:25 GMT+02:00 Nicolas Cellier <
nicolas.cellier.aka.nice at gmail.com>:

>
>
> 2016-04-21 20:42 GMT+02:00 Clément Bera <bera.clement at gmail.com>:
>
>>
>> I have just run those test on the latest VM + latest Pharo 5 and had no
>> issue.
>>
>> Seemingly Nicolas fixed it yesterday or today.
>>
>> Could it be possible that you are out of sync with the svn and / or the
>> VMMaker package version ? I remember Nicolas precising that one needed to
>> update the svn repo to make a commit work.
>>
>>
> And now I have reverted the svn changes (rev. 3675), but this will require
> syncing with VMMaker.oscog-nice.1830.
> Apologies for the jerk.
>
>
>>
>> On Mon, Apr 18, 2016 at 4:44 PM, Esteban Lorenzano <estebanlm at gmail.com>
>> wrote:
>>
>>>
>>> Hi,
>>>
>>> In concrete this ones:
>>>
>>> IntegerTest>>#testIsPrime
>>> IntegerTest>>#testIsPrime2
>>> IntegerTest>>#testIsProbablyPrime
>>> AlienSunit>>#testLong
>>> AlienSunit>>#testLongAccessors
>>>
>>> I suspect this is a common failure and not a Pharo specific issue (since
>>> nothing we change could remotely affect longs, I think)
>>>
>>> any idea?
>>>
>>> cheers,
>>> Esteban
>>
>>
>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.squeakfoundation.org/pipermail/vm-dev/attachments/20160422/fd4151ca/attachment-0001.htm