Hi Jerry, As usual, thanks for the review and testing :-)
Pushed as r16-4383. Regards Paul On Sat, 11 Oct 2025 at 17:46, Jerry D <[email protected]> wrote: > > On 10/11/25 5:17 AM, Paul Richard Thomas wrote: > > Hi All, > > > > These fixes are minor on the scale of things and fix no fewer than > > three problems (2 PRs): > > > > (i) PR95543 exposed two separate bugs. The first was that default > > initializer expressions that are not constants, were not being > > expanded so that the name of the PDT instance could not be generated > > and an ICE was caused downstream. This is fixed by the first two > > chunks in the patch. Once this was fixed, comparison with other brands > > showed the testcase failing at the declaration " type(t()) :: z" > > because the type specification list is empty, contrary to F2023 R755: > > type-param-spec is [ keyword = ] type-param-value. This is fixed > > by the chunk in primary.cc. > > > > (ii) PR103748 was due to type spec lists being applied to derived > > types that are not parameterized. This is fixed by the third chunk. > > > > Regtests with FC42/x86_64 - OK for mainline? > > > > Paul > > Looks good to go Paul. Tests OK here as well. > > Thanks, > > Jerry
