[Facets-devel] PETSC ERROR

[Barry Smith]

    If they want it optional then they should pass  
PETSC_NULL_CHARACTER in PetscInitialize() and then call  
PetscOptionsInsertFile() afterwards with the appropriate flag.
So, no change to current approach.

    Barry

In the future we may just wish to ditch the file argument from  
PETScInitialize()


On Jun 4, 2009, at 3:37 PM, Satish Balay wrote:

> Barry/Matt,
>
> Should we make optfile in PetscIntialize(optfile) optional?
>
> Current code is:
>
> PetscIntialize("file") -> PetscOptionsInsert(optfile) ->
> PetscOptionsInsertFile(optfile,PETSC_TRUE)
>
> i.e it gives error if the "optfile" is not found.
>
> Should the code be changed to use:
> PetscOptionsInsertFile(optfile,PETSC_FALSE)?
>
>
> Alternatives would be to expose the flag - a bit higher [i.e use
> PetscOptionsInsert(file,flg)] or expose it further up to
> PetscIntialize(optfile,flg). But changing PetscIntialize() might not
> be desireable.
>
>
> I'm inclined to use:
> PetscIntialize("file") -> PetscOptionsInsert(optfile,PETSC_FALSE) ->
> PetscOptionsInsertFile(optfile,flg)
>
> Satish
>
> On Thu, 4 Jun 2009, John R. Cary wrote:
>
>> In the UEDGE tests we get
>>
>> oxygen.cary$ grep PETSC */*.out | grep petscopt
>> setup/alloc-cxxpy.out:[0]PETSC ERROR: Unable to open Options File  
>> petscopt!
>> tokgeo_HffCnog_ss/tokHffCnog-cxxpy.out:[0]PETSC ERROR: Unable to open
>> Options File petscopt!
>> tokgeo_HmsCnog_ss/tokHmsCnog-cxxpy.out:[0]PETSC ERROR: Unable to open
>> Options File petscopt!
>> tokgeo_H/tokHsm-cxxpy.out:[0]PETSC ERROR: Unable to open Options File
>> petscopt!
>>
>>
>> Is this really an error?  The code seems to run fine, but those error
>> messages would be disconcerting to a user.
>>
>> John
>> _______________________________________________
>> Facets-devel mailing list
>> Facets-devel at ice.txcorp.com
>> https://ice.txcorp.com/mailman/listinfo/facets-devel
>>
>