[PATCH] Fortran: parameter inquiries of constant complex arrays [PR102599,PR114022]

2025-05-30 Thread Harald Anlauf
Dear all, here's a patch fixing the handling of parameter inquiries of constant complex arrays. It profits from previous fixes for inquiries of substrings and essentially adds only the simplification of %re/%im applies to complex arrays - and fixes a minor frontend memleak encountered on the way

Re: [PATCH] Fortran: parameter inquiries of constant complex arrays [PR102599,PR114022]

2025-05-30 Thread Harald Anlauf
On 5/30/25 19:44, Steve Kargl wrote: On Fri, May 30, 2025 at 07:37:49PM +0200, Harald Anlauf wrote: here's a patch fixing the handling of parameter inquiries of constant complex arrays. It profits from previous fixes for inquiries of substrings and essentially adds only the simplification of %

[patch, Fortran, committed] Fix PR 120355, regression with global functions

2025-05-30 Thread Thomas Koenig
Hello world, I have just committed the attached patch as obvious and simple after regression-testing. Will backport to gcc 15 soonish. This fixes a 15/16 regression where the code was looking at the typespec of the symbol which we don't set if there is a RESULT clause. Thanks to Harald for poin

Re: [PATCH] Fortran: parameter inquiries of constant complex arrays [PR102599,PR114022]

2025-05-30 Thread Steve Kargl
On Fri, May 30, 2025 at 07:37:49PM +0200, Harald Anlauf wrote: > > here's a patch fixing the handling of parameter inquiries of > constant complex arrays. It profits from previous fixes for > inquiries of substrings and essentially adds only the simplification > of %re/%im applies to complex arra

Execution time for gfortran regression testing

2025-05-30 Thread Harald Anlauf
Dear all, the time needed for running the gfortran testsuite has increased so much that I am looking for options to get this down to something that is more bearable when working on a Fortran-only patch. I am already building with --disable-bootstrap, as I cannot afford a full bootstrap. It is a

Re: Execution time for gfortran regression testing

2025-05-30 Thread Steve Kargl
On Fri, May 30, 2025 at 09:03:27PM +0200, Harald Anlauf wrote: > > the time needed for running the gfortran testsuite has increased so > much that I am looking for options to get this down to something > that is more bearable when working on a Fortran-only patch. > > I am already building with --

Re: Execution time for gfortran regression testing

2025-05-30 Thread Harald Anlauf
When I'm working on a particular area of gfortran, I tend to use RUNTESTFLAGS to limit what is tested.  For example, I just fixed SPREAD() for scalar source and ncopies < 1. I do % cd obj % gmake % cd gcc % gmake check-fortran RUNTESTFLAGS="dg.exp=spread\*.f90" This only runs 63 tests.  Of cours

Re: Execution time for gfortran regression testing

2025-05-30 Thread Jerry Delisle
You could possibly modify the dg.exp file. These are basically scripts. It's a bit of a pain. On Fri, May 30, 2025, 1:29 PM Harald Anlauf wrote: > > When I'm working on a particular area of gfortran, I tend > > to use RUNTESTFLAGS to limit what is tested. For example, > > I just fixed SPREAD()