[petsc-dev] wrong makefile variables?

[Barry Smith]

   Ok, config/PETSc/Configure.py will automatically put PETSC_LIB_BASIC = -lpetsc into $PETC_ARCH/conf/petscvariables when run with single library.

    Barry

   BTW: PETSC_LIB_VARIABLE was just an internal variable in the makefile system for making other variables, since I didn't know you used it I didn't try to preserve it. I didn't mess it up just to bug you.

On Oct 16, 2010, at 2:32 PM, Lisandro Dalcin wrote:

> On 16 October 2010 22:21, Barry Smith <bsmith at mcs.anl.gov> wrote:
>> 
>>  So all you need is a single variable
>> 
>> PETSC_LIB_BASIC = -lpetsc
>> 
>> or
>> 
>> PETSC_LIB_BASIC = -lpetscts .... -lpetscsys
>> 
>> ?
>> I'll put that in today if that is what you need.
>> 
> 
> Yes, that's what I need, and that was available a week ago. So yes,
> please restore it.
> 
>>   Barry
>> 
>> 
>> On Oct 16, 2010, at 1:34 PM, Lisandro Dalcin wrote:
>> 
>>> On 16 October 2010 19:17, Barry Smith <bsmith at mcs.anl.gov> wrote:
>>>> 
>>>>  Maybe if you told/showed me exactly how you use it we can come up with the optimal set of variables you want?
>>>> 
>>> 
>>> Take a look here to what I'm currently using:
>>> 
>>> http://code.google.com/p/petsc4py/source/browse/conf/core/confcore.py#144
>>> 
>>> I'm not saying that this is optimal, I've always used what PETSc provides.
>>> 
>>>>   Do you want a flag that says single library or multiple libraries?
>>>> 
>>> 
>>> Perhaps that could work, but...
>>> 
>>>>   Do you want a variable with the list of the PETSc libraries?
>>>> 
>>> 
>>> That's what I was used to get with PETSC_LIB_BASIC. That variable
>>> expanded to "-lpetsc" for single-lib builds, and "-lpetscts
>>> -lpetscsnes ..." for the multi-lib case.
>>> 
>>>>   Something else?
>>>> 
>>> 
>>> If you could provide the correct list of PETSc libs in
>>> PETSC_LIB_BASIC, that would be nice, because my code would continue
>>> working with previous 3.0.0 and 3.1 releases.
>>> 
>>> 
>>>>   Barry
>>>> 
>>>> On Oct 16, 2010, at 9:44 AM, Lisandro Dalcin wrote:
>>>> 
>>>>> On 16 October 2010 16:23, Barry Smith <bsmith at mcs.anl.gov> wrote:
>>>>>> 
>>>>>> On Oct 16, 2010, at 3:43 AM, Lisandro Dalcin wrote:
>>>>>> 
>>>>>>> It seems that after changes to BuildSystem, the makefile variables
>>>>>>> PETSC_LIB_BASIC and PETSC_EXTERNAL_LIB_BASIC are not properly set for
>>>>>>> single-library builds.
>>>>>> 
>>>>>>   This is a change.  ./configure now generates PETSC_EXTERNAL_LIB_BASIC  and PETSC_LIB_BASIC is never used
>>>>>> 
>>>>>>  Barry
>>>>>> 
>>>>> 
>>>>> How should I figure if PETSc is using single vs. multiple library
>>>>> builds (remember petsc4py inspect makefile variables to get PETSc
>>>>> stuff)?
>>>>> 
>>>>> Any chance you could still generate the correct PETSC_LIB_BASIC, even
>>>>> if core PETSc does not use it?
>>>>> 
>>>>>>> 
>>>>>>> A few days ago we had:
>>>>>>> 
>>>>>>> $ grep _BASIC $PETSC_ARCH/conf/petscvariables
>>>>>>> PETSC_LIB_BASIC = -lpetsc
>>>>>>> PETSC_CONTRIB_BASIC = -lpetsc
>>>>>>> PETSC_SNES_LIB_BASIC = -lpetsc
>>>>>>> PETSC_KSP_LIB_BASIC = -lpetsc
>>>>>>> PETSC_MAT_LIB_BASIC = -lpetsc
>>>>>>> PETSC_VEC_LIB_BASIC = -lpetsc
>>>>>>> PETSC_DM_LIB_BASIC = -lpetsc
>>>>>>> PETSC_TS_LIB_BASIC = -lpetsc
>>>>>>> PETSC_SYS_LIB_BASIC = -lpetsc
>>>>>>> 
>>>>>>> but now:
>>>>>>> 
>>>>>>> $ grep _BASIC $PETSC_ARCH/conf/petscvariables
>>>>>>> SL_LINKER_LIBS = ${PETSC_EXTERNAL_LIB_BASIC}
>>>>>>> PETSC_EXTERNAL_LIB_BASIC = -lX11 -llapack -lblas -lm
>>>>>>> -Wl,-rpath,/usr/lib64/mpich2 -L/usr/lib64/mpich2
>>>>>>> -Wl,-rpath,/usr/lib/gcc/x86_64-redhat-linux/4.4.1
>>>>>>> -L/usr/lib/gcc/x86_64-redhat-linux/4.4.1 -ldl -lmpich -lopa -lpthread
>>>>>>> -lrt -lgcc_s -lmpichf90 -lgfortran -lm -lm -ldl -lmpich -lopa
>>>>>>> -lpthread -lrt -lgcc_s -ldl
>>>>>>> 
>>>>>>> 
>>>>>>> Could any of you check and fix this?
>>>>>>> 
>>>>>>> --
>>>>>>> Lisandro Dalcin
>>>>>>> ---------------
>>>>>>> CIMEC (INTEC/CONICET-UNL)
>>>>>>> Predio CONICET-Santa Fe
>>>>>>> Colectora RN 168 Km 472, Paraje El Pozo
>>>>>>> Tel: +54-342-4511594 (ext 1011)
>>>>>>> Tel/Fax: +54-342-4511169
>>>>>> 
>>>>>> 
>>>>> 
>>>>> 
>>>>> 
>>>>> --
>>>>> Lisandro Dalcin
>>>>> ---------------
>>>>> CIMEC (INTEC/CONICET-UNL)
>>>>> Predio CONICET-Santa Fe
>>>>> Colectora RN 168 Km 472, Paraje El Pozo
>>>>> Tel: +54-342-4511594 (ext 1011)
>>>>> Tel/Fax: +54-342-4511169
>>>> 
>>>> 
>>> 
>>> 
>>> 
>>> --
>>> Lisandro Dalcin
>>> ---------------
>>> CIMEC (INTEC/CONICET-UNL)
>>> Predio CONICET-Santa Fe
>>> Colectora RN 168 Km 472, Paraje El Pozo
>>> Tel: +54-342-4511594 (ext 1011)
>>> Tel/Fax: +54-342-4511169
>> 
>> 
> 
> 
> 
> -- 
> Lisandro Dalcin
> ---------------
> CIMEC (INTEC/CONICET-UNL)
> Predio CONICET-Santa Fe
> Colectora RN 168 Km 472, Paraje El Pozo
> Tel: +54-342-4511594 (ext 1011)
> Tel/Fax: +54-342-4511169