Re: [eigen] Re: Scheduling beta3 on sunday february 6 ?

[ Thread Index | Date Index ]


For some of these issues, this problem here could be relevant:
http://support.microsoft.com/kb/2280741

The next update to Visual Studio should contain the fix (SP1 afaik)

Cheers
Benjamin

On 02/09/2011 04:19 PM, Benoit Jacob wrote:
> Another update. Beta3 is delayed again by MSVC compilation issues. The
> remaining blockers are:
> 
>    165: [MSVC] Inverse unit test crashes in release mode
>    179: [MSVC] mixingtypes 3 (complex) segfaults since changeset 3285
> 
> Many msvc fixes have gone in lately, it just seems that a few more are needed...
> 
> Non-msvc users may safely treat the current revision as 'beta3' as it
> seems that nothing else than msvc fixes is needed.
> 
> Benoit
> 
> 2011/2/5 Benoit Jacob <jacob.benoit.1@xxxxxxxxx>:
>> I'm requesting a couple more days to release beta3. Probably around
>> tuesday/wednesday.
>>
>> I've not had as much time as I hoped lately, and many compilation
>> issues have been discovered in recent testing (thanks everybody)
>> making a release tomorrow seem too optimistic. Also, Gael has had the
>> great idea of triaging all our old bugs and has unearthed a few
>> important ones.
>>
>> Current blocker list:
>>
>>       45: ei_random<Integer type> should be uniform distribution on
>> the whole integer range
>>
>> --> easy fix!
>>
>> The rest is compiler issues, mostly MSVC, bug 145 is GCC-on-OSX:
>>
>>       83: Compilation error in Visual Studio 2008
>>       89: householder_1 and _2 are continuously failing with segfaults
>>       145: internal compiler error for complex float matrix-vector
>> multiplication
>>       156: Several CDash Tests fail under MSVC 2008 + 2010
>>       164: Assignment: multiple overloads have similar conversions
>>       165: Inverse unit test crashes (release mode/msvc)
>>
>> Benoit
>>
>> 2011/2/4 Rhys Ulerich <rhys.ulerich@xxxxxxxxx>:
>>>> It's also time to start testing the heck out of it.
>>>> ...
>>>> cmake -DEIGEN_DEFAULT_TO_ROW_MAJOR=ON
>>>> cmake  -DEIGEN_TEST_OPENMP=ON
>>>
>>> Found several of interesting things for Intel 11.1 in bug #176.
>>>
>>>> cmake -DEIGEN_TEST_EIGEN2=ON
>>>
>>> Additionally beyond #176 I saw a couple of eigen2 test failures:
>>>
>>> Above failure plus some failures in
>>>        600 - eigen2_eigensolver_2 (Failed)
>>>        602 - eigen2_eigensolver_4 (Failed)
>>>        603 - eigen2_eigensolver_5 (Failed)
>>>
>>> Filed in bug #177.
>>>
>>> - Rhys
>>>
>>>
>>>
>>
> 
> 




Mail converted by MHonArc 2.6.18 http://listengine.tuxfamily.org/