https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64712
vries at gcc dot gnu.org changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|WAITING |UNCONFIRMED
Ever confirmed|1 |0
--- Comment #4 from vries at gcc dot gnu.org ---
I can reproduce the FAIL on the command line, between 1/2 and 1/3 of the runs.
[ When I filed the bug, I didn't realize it didn't always reproduce. I saw the
failure for both the bootstrap and non-bootstrap build, and assumed it was
stable. ]
Furthermore, I copied gnat/dg.exp to gnat/dg-gnat.exp, and modified it like
this:
...
for {set i 0} {$i < 100} {incr i} {
dg-runtest [lsort [glob -nocomplain
$srcdir/$subdir/unchecked_convert1.adb]] \
"" $DEFAULT_CFLAGS
}
...
And managed to reproduce using just unix/-m32:
...
$ make -k -j9 check-gcc 'RUNTESTFLAGS=dg-gnat.exp --target_board=unix/-m32'
$ grep -c 'FAIL:.*execution test' nobootstrap/build/gcc/testsuite/gnat/gnat.log
47
vries@linux1:~/gcc_versions/data/ref-master-15-01-21$ grep -c 'PASS:.*execution
test' nobootstrap/build/gcc/testsuite/gnat/gnat.log
53
...
