Hi!

On 2023-03-30T04:00:03-0300, Alexandre Oliva <oliva@adacore.com> wrote:
> On Mar 29, 2023, Thomas Schwinge <thomas@codesourcery.com> wrote:
>> But surely the curly braces in:
>
>>     -UNSUPPORTED: g++.dg/modules/explicit-bool-1_a.H -std=c++17
>
>>     +UNSUPPORTED: {g++.dg/modules/explicit-bool-1_a.H -std=c++17}
>
>> ... are not intentional?  (Alexandre?)
>
> Unintended indeed, will look, thanks for letting me know
>
>
>> But worse, the latter also "bleeds into" all other testing
>
> Eeek
>
> Yeah, that's a much bigger problem indeed.
>
>> ..., this isn't sufficient.  Instead, we should undo the 'rename' at the
>> end of 'g++.dg/modules/modules.exp'.  OK to push the attached
>> "'g++.dg/modules/modules.exp': don't leak local 'unsupported' proc [PR108899]"
>> after proper testing?
>
> Ooh, nice, I didn't know how to drop the renaming after we were done
> with it, and hoped the end of the .exp would have accomplished that by
> ending a scope.  Jakub had already pointed out this wasn't the case, but
> I didn't realize, when he did, that this would carry over onto other
> modules.
>
> If we're dropping the renaming, I suppose we could also revert Jakub's
> change.

Yes, my plan was to push a 'git revert' of Jakub's change as a follow-up
(clean-up) *after* my proposed
"'g++.dg/modules/modules.exp': don't leak local 'unsupported' proc [PR108899]",
see attached again.

My testing has completed without issues; OK to push that one?

> +# Restore the original unsupported proc, lest it will affect
> +# subsequent test runs, or even fail renaming if we run modules.exp
> +# for multiple targets/multilibs/options.
> +rename unsupported {}
> +rename modules-saved-unsupported unsupported

Should I incorporate that comment instead of my simpler one?


Grüße
 Thomas


-----------------
Siemens Electronic Design Automation GmbH; Anschrift: Arnulfstraße 201, 80634 München; Gesellschaft mit beschränkter Haftung; Geschäftsführer: Thomas Heurung, Frank Thürauf; Sitz der Gesellschaft: München; Registergericht München, HRB 106955