renaming in PC

Barry Smith bsmith at mcs.anl.gov
Tue Oct 9 15:36:22 CDT 2007


  Well yes. There are probably others as well :-(

  Barry


On Tue, 9 Oct 2007, Lisandro Dalcin wrote:

> Good point...
> 
> So we should rename
> 
> PetscErrorCode PCHasApplyTranspose(PC,PetscTruth*)
> 
> to
> 
> PetscErrorCode PCApplyTransposeExists(PC,PetscTruth*)
> 
> ???
> 
> 
> On 10/9/07, Barry Smith <bsmith at mcs.anl.gov> wrote:
> >
> >   HasApply is more grammatical. But I like having the naming
> > follow the pattern of
> >
> > PCXXYY
> > PCXXYYzz
> > PCXXYYww
> >
> > rather than PCzzXXYY, or PCwwXXYY etc.
> >
> > That is as much of the strings should start the same way for
> > related routines. Hence PCApplyRichardsonExists() is
> > near PCApplyRichardson() while PCHasApplyRichardson()
> > is kilometers away.
> >
> >   Barry
> >
> > Hence we have
> > KSPMonitor()
> > KSPMonitorSet()
> > KSPMonitorCancel()
> > KSPMonitorSingularValue()
> > KSPMonitorDefault()
> > I realize the KSPMonitorSet() instead of KSPSetMonitor() is non-standard but
> > that doesn't mean it is wrong :-).
> >
> > Essentially when you sort the function names, I'd like the related ones
> > together.
> >
> >
> > On Tue, 9 Oct 2007, Lisandro Dalcin wrote:
> >
> > > Which form should be the preferred one? I'm inclined for the fist...
> > >
> > > PetscErrorCode PCHasApplyTranspose(PC,PetscTruth*)
> > > PetscErrorCode PCApplyRichardsonExists(PC,PetscTruth*)
> > >
> > >
> > >
> >
> >
> 
> 
> 




More information about the petsc-dev mailing list