[petsc-dev] [petsc-maint #53815] problem configuring petsc-dev w/ML on CIG buildbots

[Barry Smith]

On Oct 1, 2010, at 11:30 AM, Satish Balay wrote:

> Hm - looks like the new ML is using cmake for build - so it might take
> a bit more figureouing before --download-ml can be made to work with
> it..

   Yikes! Guess we have learn the basics of cmake :-(

   Barry

> 
> satish
> 
> 
> On Fri, 1 Oct 2010, Hong Zhang wrote:
> 
>> Satish,
>> 
>> I downloaded the latest trilinos-10.6.0-Source.tar.
>> It is located under /homes/hzhang/tmp/trilinos-10.6.0-Source.tar
>> 
>> Checking trilinos-10.6.0-Source/packages/ml/ChangeLog-ML:
>> 
>> ML -- History of visible changes
>> ================================
>> 
>> This file reports (some of) the visible changes to ML.
>> Please submit ML bugs at software.sandia.gov/bugzilla, or
>> report bugs to {jhu|rstumin}@sandia.gov.
>> 
>> August 2008
>> ...
>> 
>> Shall we check the changes made on trilinos-10.6.0?
>> 
>> Hong
>> 
>> On Fri, Oct 1, 2010 at 11:01 AM, Satish Balay <petsc-maint at mcs.anl.gov> wrote:
>>> Perhaps we should check for latest ML? - hopefully they fixed the
>>> issue - and we don't need these workarrounds?
>>> 
>>> [if someone is already registered - they can do the download
>>> and check]
>>> http://trilinos.sandia.gov/download/trilinos-10.6.html
>>> 
>>> satish
>>> 
>>> On Fri, 1 Oct 2010, Barry Smith wrote:
>>> 
>>>> 
>>>> (1) the C++ bindings of MPI are being depreciated
>>>> 
>>>> (2) no one uses the C++ bindings of MPI, in particular ML doesn't
>>>> 
>>>> (3) therefor ALL compiles (and use of mpicxx -E)  in BuildSystem of C++ should be passed the flags  -DOMPI_SKIP_MPICXX  -DMPICH_SKIP_MPICXX  automatically ALWAYS
>>>> 
>>>>   Matt,
>>>> 
>>>>     Can you fix this ASAP? Either by always putting these #defines in the config.h files you generate that you include or by passing them as -D Hell you can always put them in, even for C code tests if that makes it easier.
>>>> 
>>>>   Satish,
>>>> 
>>>>     You knew about this problem for years you should have told us so Matt could have fixed it years ago.
>>>> 
>>>>    Barry
>>>> 
>>>> 
>>>> 
>>>> 
>>>> 
>>>> On Sep 30, 2010, at 10:50 PM, Brad Aagaard wrote:
>>>> 
>>>>> We are having trouble building petsc-dev w/ML on the CIG buildbots. ML
>>>>> is configured and built but the test for using the ML library fails due to
>>>>> 
>>>>> #error "SEEK_SET is #defined but must not be for the C++ binding of MPI
>>>>> 
>>>>> even though we have used -DMPICH_IGNORE_CXX_SEEK to avoid this error.
>>>>> The logs are available at http://www.geodynamics.org:8009/?project=PETSc
>>>>> (second column x86_64 linux).
>>>>> 
>>>>> Any insight you can provide on what might be the cause of this error
>>>>> would be appreciated.
>>>>> 
>>>>> Thanks,
>>>>> Brad
>>>>> 
>>>> 
>>>> 
>>>> 
>>> 
>>> 
>>> 
>> 
>> 
>