Re: [Patch, fortran] PRs123545, 123673 & 122949 - [PDT] Fix faults in fiats runtime
Jerry D <[email protected]>
| Newsgroups | gmane.comp.gcc.patches,gmane.comp.gcc.fortran |
|---|---|
| Message-ID | <[email protected]> |
On 2/5/26 9:57 AM, Paul Richard Thomas wrote: > Hello All, > > PRs 123545 and 123673 are somewhat intertwined. The patch fixes them > both. PR122949 has been fixed by another patch somewhere along the way > and the testcase ensures that it remains fixed. > > The first chunk in the patch is completely obvious and it is rather > surprising that it wasn't picked up earlier (mea maxima culpa!). > > The second chunk fixes the problems caused by moving or deleting > chunks of code in the modules and, where appropriate , repositioning > in the main program. The faults result from the unusual way in which > pdt_types are generated from their template and are triggered when the > pdt_type is first encountered in a component in a contained procedure. > > The third chunk delays the freeing of the destination component until > after it has been made use of by being deep copied downstream. > > Regtested on FC43/x86_64. OK for mainline? > > Paul Yes, this is ok Paul. Thanks, Jerry