https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93727
--- Comment #46 from dave.anglin at bell dot net --- On 2026-06-15 4:03 p.m., jvdelisle at gcc dot gnu.org wrote: > https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93727 > > --- Comment #45 from Jerry DeLisle <jvdelisle at gcc dot gnu.org> --- > (In reply to John David Anglin from comment #44) >> At r17-1532-ga514707ffd7, I see the following new fails: >> >> FAIL: gfortran.dg/EXformat_4.F90 -O0 execution test >> FAIL: gfortran.dg/EXformat_4.F90 -O1 execution test >> FAIL: gfortran.dg/EXformat_4.F90 -O2 execution test >> FAIL: gfortran.dg/EXformat_4.F90 -O3 -fomit-frame-pointer -funroll-loops >> -fpeel-loops -ftracer -finline-functions execution test >> FAIL: gfortran.dg/EXformat_4.F90 -O3 -g execution test >> FAIL: gfortran.dg/EXformat_4.F90 -Os execution test >> >> Fails are at stop 1 in test04. >> >> (gdb) p s >> $4 = '-0XF.FFFFFP+124', ' ' <repeats 25 times> >> (gdb) step >> 22 if (r4 /= -huge(1.0_4)) stop 1 >> (gdb) p r4 >> $6 = -0 > > I cant reproduce this here. Can you le me know what system you are on and can > you replace the stop with a print *, r4 so we can see what the -huge value is? I replaced stop 1 with print *, r4, -huge(1.0_4). bash-5.1$ ./EXformat_4.exe -0.00000000 -3.40282347E+38 STOP 2 It appears to me that the problem is with the read(s,'(EX40.0)') r4 statement. I stepped through the read operation but didn't see anything OS specific there. The write uses frexpf and scalbnf. Possibly s isn't correct. System is hppa64-hp-hpux11.11. There are also problems on powerpc64le-unknown-linux-gnu: https://gcc.gnu.org/pipermail/gcc-testresults/2026-June/879065.html I have hacked libgfortran to fix the caf.exp tests. Patches are posted to lists.
