Re: [Patch, Fortran] PING: PR fortran/37746: -fbounds-check for string lengths of dummy arguments

[Daniel Kraft <d at domob dot eu>]

Steve Kargl wrote:
> On Sat, Mar 28, 2009 at 10:57:50PM +0100, Daniel Kraft wrote:
> > Hi,
> >
> > PING for the patch below.
> >
> > No regressions with a fresh test.  Ok for trunk?
>
> A couple of comments.
>
> First, your testcases contain MODULEs.  Do you need to cleanup
> any *.mod files.  That is, add
>
> !{ dg-final { cleanup-modules "module names here" } }
>
> where needed.

Ups, then I forgot about these; will check.

> > +static void
> > +add_argument_checking (stmtblock_t* block, gfc_symbol* sym)
>
> Elsewhere in gfortran one finds for whitespace about * either
>
>    add_argument_checking (stmtblock_t * block, gfc_symbol * sym)
>
> or
>
>    add_argument_checking (stmtblock_t *block, gfc_symbol *sym)
>
> > +{
> > +  gfc_formal_arglist* formal;
>
> The preferred style is 'gfc_formal_arglist *formal'.

Personally, I prefer the "gfc_formal_arglist* formal" style as in my 
opinion "gfc_formal_arglist*" is the type and formal the variable of 
this type; I'm well aware that most C programmers prefer it the other 
way round.  So far no one has commented on a general requirement for 
those, but if that's the guideline I will of course adapt it.

Daniel

--
Done:  Arc-Bar-Cav-Ran-Rog-Sam-Tou-Val-Wiz
To go: Hea-Kni-Mon-Pri