On Tue, May 20, 2025 at 11:44:43PM +0200, Alejandro Colomar wrote:
> Hi Jakub,
>
> On Tue, May 20, 2025 at 11:20:27PM +0200, Jakub Jelinek wrote:
> > On Tue, May 20, 2025 at 11:12:38PM +0200, Alejandro Colomar wrote:
> > > Okay; how about this?
> > >
> > > gcc/c-family/ChangeLog:
> > >
> > > * c-common.h: Add _Countof operator.
> > > * c-common.def: Likewise.
> > > * c-common.cc (c_countof_type): Likewise.
> >
> > No, that doesn't describe what you've changed and how.
>
> Well, it does. All the changes I've applied to those files are all
> to implement the new _Countof operator, and only for that. That is,
> they're sufficient and necessary. So, saying I've added the _Countof
> operator is correct. I could go and talk about the specific changes to
> each file, but then I don't see the value in that change log over the
> actual diff.
>
> >
> > So probably something like:
> >
> > * c-common.h (enum rid): Add RID_COUNTOF.
> > * c-common.def (COUNTOF_EXPR): New tree.
> > * c-common.cc (c_common_reswords): Add RID_COUNTOF entry.
> > (c_countof_type): New function.
>
> I'm honestly unsure about the usefulness of going too low level in the
> changelog as to listing newly added functions as added functions,
No, that is exactly the level all others fill in and people grep that etc.
> instead of talking high-level about what they're for. But if that's
> what you want, then okay.
>
> I think
>
> (c_countof_type): New function.
>
> is an example of what I think is useless bureaucracy. Could you please
> confirm that's what you want?
Yes, we want exactly that.
Jakub
