https://gcc.gnu.org/bugzilla/show_bug.cgi?id=117763
--- Comment #24 from Paul Thomas <pault at gcc dot gnu.org> --- (In reply to Richard Biener from comment #23) > The new testcase FAILs on the gcc-12 branch (it works on the 13 branch) on > x86_64-linux: > > FAIL: gfortran.dg/pr117763.f90 -O scan-tree-dump-times original > "__result->span = [12].." 1 > > the only __result->span assignment is in model_data_get_field_array_ptr > and that has > > __result->span = model->_data->field.span; > > there's some assignments to a span that might match: > > model->_data->field.span = 272; > > ptrtemp.124.span = 272; > > I suppose the testcase is a bit fragile? For a wrong-code testcase I'd > have expected dg-do run rather than relying on a dump scan. I'll take a look. I suspect that some other patch has not been backported to 12-branch and so the most likely outcome is that I revert the backport. As to the use of a dump scan, I am very conscious of just how long gfortran.dg takes to run since my development machine has a rather humble specification. I base the choice on the nature of the bug. Regards Paul
