> Von: "John Peterson" <[email protected]>
>
>> On Fri, Apr 1, 2011 at 2:14 PM, <[email protected]> wrote:
>>>
>>> nm --defined-only 
>>> /home/bucher/soft/petsc-3.1-p7/linux-gnu-debug/lib/libpetsc* | grep 
>>> Petsc_MPI_Abort
>>>
>>> results in
>>> 000002f8 T __Z22Petsc_MPI_AbortOnErrorPiS_
>>> 000006b1 T _Petsc_MPI_Abort
>>
>> Hmm, then I really don't know why your linker says its not found. I
>> wonder if the preceding underscore has something to do with it...?

What I'm reading suggests that the Windows ABI requires a leading
underscore - Cygwin gcc doesn't even honor the
"-f-no-leading-underscore" flag.  So then the question is why would
the linker be asking for anything else...

It's also very weird that the errors are coming from rb_base.*.o - the
linker usually works in alphabetical order, and should have hit a
missing petsc symbol in the numerics/p* objects before it found one in
a systems/r* object.

Not sure how to make use of that difference for the diagnosis,
though... unlike with the gzstreams weirdness, we don't have any
--disable-rbOOmit option to turn off the offending objects this time.
And if you just remove rb_base manually it'll remove a dependency for
a ton of other classes.

Hey, David, would it be much work for you to add a --disable-rbOOmit
option?
---
Roy

------------------------------------------------------------------------------
Xperia(TM) PLAY
It's a major breakthrough. An authentic gaming
smartphone on the nation's most reliable network.
And it wants your games.
http://p.sf.net/sfu/verizon-sfdev
_______________________________________________
Libmesh-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/libmesh-users

Reply via email to