Re: [PATCH] PR125430 Fixmodule-contained PRIVATE procedures must have global, ELF linkage
Jerry D <[email protected]>
| Newsgroups | gmane.comp.gcc.patches,gmane.comp.gcc.fortran |
|---|---|
| Message-ID | <[email protected]> |
Committed. The master branch has been updated by Jerry DeLisle <[email protected]>: https://gcc.gnu.org/g:1cf8c5913bfe8016b8b2436cc17a7577be1b295f commit r17-913-g1cf8c5913bfe8016b8b2436cc17a7577be1b295f Author: Jerry DeLisle <[email protected]> Date: Fri May 22 21:56:34 2026 -0700 fortran: module-contained PRIVATE procedures must have global ELF linkage [PR125430] On 5/28/26 10:32 AM, Jerry D wrote: > On 5/28/26 12:34 AM, Paul Richard Thomas wrote: >> Hi Jerry, >> >> Contrary to my remarks on Mattermost, this patch fixes PRs 88632 and >> 104630 with the assistance of my tentative patch for pr125429 was >> totally spurious: It does it on its own! It regression tests without >> errors. >> >> I would consider losing the first comment in trans-decl.cc. It is >> essentially repeated in the second one and its second sentence says >> what any fortran developer must know already. >> >> Now that I see the substance of the patch, it looks rather 'obvious'. >> >> OK for mainline and, I would suggest, immediate backport to 16-branch. >> >> Regards >> >> Paul >> >> On Wed, 27 May 2026 at 16:21, Jerry D <[email protected]> wrote: >>> >>> See attached patch. This bug found when trying to build the Formal package. >>> >>> Developed with with Assistance from Claude Sonnet 4.6. Two new testsuite files >>> provided. The change wiggled on a few exiting test cases. >>> >>> Regression tested on x86_64 Linux. >>> >>> OK for mainline? What about backport? >>> >>> Regards, >>> >>> Jerry >>> > > Thanks Paul, I am cleaning it up per your comments and will push. > > Jerry