[petsc-dev] sor smoothers

[Barry Smith]

On Sep 9, 2013, at 8:57 PM, Jed Brown <jedbrown at mcs.anl.gov> wrote:

> Barry Smith <bsmith at mcs.anl.gov> writes:
>>   Take a look at MatCreate_SeqAIJ. These fill up all the function
>>   points in one swoop with the big function table, this is why there
>>   is not "base class setup".
> 
> Those MatOps_Values are a perpetual maintenance hassle.  I would replace
> them all with a boring list of assignment statements for any implemented
> functions, as in MatCreate_Nest.

   Someone broke all your fingers :-)

>  Then we wouldn't need any "dummy"
> entries, the order wouldn't matter, etc.
> 
> If Microsoft would implement C99, we could use designated initializers:
> 
> static struct _MatOps MatOps_Values = {
>  .setvalues = MatSetValues_SeqAIJ,
>  .mult      = MatMult_SeqAIJ,
>  .....
> };