BTW There is a four year old patch from Steve K in PR105594. I will push it as obvious immediately after 88632 unless you have any objections.
Since the weather forecast is so atrocious, my aim is to try to empty out the submodule meta-bug in the next week or so. Then back to PDTs. Cheers Paul On Mon, 1 Jun 2026 at 21:53, Paul Richard Thomas <[email protected]> wrote: > > Hi Harald, > > Do you know what? I stared at that spelling correction for a while and > decided that I must be wrong :about it -) Thanks for reassuring me > about my spelling instincts and making me suspicious of Claude's. > > Thanks for the review. > > > Paul > > On Mon, 1 Jun 2026 at 19:58, Harald Anlauf <[email protected]> wrote: > > > > Hi Paul! > > > > Am 01.06.26 um 2:03 PM schrieb Paul Richard Thomas: > > > Hello All, > > > > > > The attached patch is relatively trivial. It regtests OK on Fedora > > > 44/x86_64. > > > > > > OK for mainline and for backporting to 16-branch? > > > > Ehm, the first hunk > > > > --- a/gcc/fortran/module.cc > > +++ b/gcc/fortran/module.cc > > @@ -4372,7 +4372,7 @@ mio_full_f2k_derived (gfc_symbol *sym) > > /* PDT templates make use of the mechanisms for formal args > > and so the parameter symbols are stored in the formal > > namespace. Transfer the sym_root to f2k_derived and then > > - free the formal namespace since it is unneeded. */ > > + free the formal namespace since it is uneeded. */ > > if (sym->attr.pdt_template && sym->formal && sym->formal->sym) > > { > > ns = sym->formal->sym->ns; > > > > seems incorrect to me. And my mail reader agrees with me... > > > > The rest of the patch LGTM. > > > > Thanks for the patch! > > > > Harald > > > > > Regards > > > > > > Paul > >
