dolfin team mailing list archive

["Garth N. Wells" <gnw20@xxxxxxxxx>]

Anders Logg wrote:
> On Fri, Jan 29, 2010 at 03:06:01PM +0100, Johannes Ring wrote:
>> On Fri, Jan 29, 2010 at 1:16 PM, Johannes Ring <johannr@xxxxxxxxx> wrote:
>>> On Fri, Jan 29, 2010 at 12:36 PM, Anders Logg <logg@xxxxxxxxx> wrote:
>>>> On Fri, Jan 29, 2010 at 12:08:21PM +0100, Johannes Ring wrote:
>>>>> [resending to include list]
>>>>>
>>>>> On Fri, Jan 29, 2010 at 11:29 AM, Anders Logg <logg@xxxxxxxxx> wrote:
>>>>>> Are there still any HAS_CGAL issues to fix or has this been sorted
>>>>>> out?
>>>>> I'm not sure but the C++ function demo fails on the Mac buildbot with
>>>>> the following message:
>>>>>
>>>>>  $ ./test
>>>>>  Solving linear variational problem
>>>>>    Matrix of size 4913 x 4913 has 126785 nonzero entries.
>>>>>    Solving linear system of size 4913 x 4913 (PETSc LU solver, umfpack).
>>>>>  ##Failure Location unknown## : Error
>>>>>  Test name: Eval::testArbitraryEval
>>>>>  uncaught exception of unknown type
>>>>>
>>>>>  Failures !!!
>>>>>  Run: 1   Failure total: 1   Failures: 0   Errors: 1
>>>>>
>>>>> Could it be because CGAL is missing?
>>>> I don't know. I saw some strange errors yesterday in testArbitraryEval
>>>> which went away after I did instant-clean. Is that done on the
>>>> buildbots before they run?
>>> It is only done nightly when a rebuild of Instant is triggered.
>>> However, one can force instant-clean to run by forcing a rebuild of
>>> Instant. But I don't think that is the problem here since I just
>>> logged into the Mac buildbot and ran instant-clean and then the
>>> function test again and I got the same error.
>> The problem is now gone after I got CGAL working so I guess an #ifdef
>> HAS_CGAL is needed somewhere in the C++ function test. I'm not sure
>> where to add it so I leave that to someone else.
>>
>> BTW, the problem with CGAL was resolved by removing GMP 4.2.4 and
>> build GMP 4.3.1 instead. I also had to apply Harish's dorsal patch and
>> turn off PDB in CGAL.
>>
>> Johannes
> 
> ok, I have tried adding one. We currently have double #ifdefs for CGAL
> and GTS. GTS should be removed soon (replaced by CGAL) so then the
> extra #ifdef should disappear.
> 

Perhaps when the change is made we should make CGAL a dependency. It's
key to some of the recent and the upcoming DOLFIN functionality.

Garth



> I'll restart the Windows buildbot and see what happens.
> 
> --
> Anders
> 
> 
> ------------------------------------------------------------------------
> 
> _______________________________________________
> Mailing list: https://launchpad.net/~dolfin
> Post to     : dolfin@xxxxxxxxxxxxxxxxxxx
> Unsubscribe : https://launchpad.net/~dolfin
> More help   : https://help.launchpad.net/ListHelp