This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Have regression tester continue after bootstrap comparison failure
- From: gkeating at apple dot com (Geoff Keating)
- To: gcc-patches at gcc dot gnu dot org
- Date: Wed, 2 Sep 2009 13:26:07 -0700 (PDT)
- Subject: Have regression tester continue after bootstrap comparison failure
The regression tester is supposed to treat a bootstrap comparison
failure as a regression, not a build failure, so it can go ahead and
run the testsuite. This hasn't been working, because when the build
was re-started it would run the comparison again and of course fail
again right away. The obvious solution is to touch the 'compare' file
so that make won't run the comparison again.
--
- Geoffrey Keating <geoffk@apple.com>
===File ~/patches/regress-comparefail.patch=================
2009-09-02 Geoff Keating <geoffk@apple.com>
* btest-gcc.sh: Continue after bootstrap comparison failure.
Index: btest-gcc.sh
===================================================================
--- btest-gcc.sh (revision 151348)
+++ btest-gcc.sh (working copy)
@@ -127,6 +127,7 @@
if ! make $dashj bootstrap ; then
[ -s .bad_compare ] || exit 1
cat .bad_compare >> $REGRESS || exit 1
+ touch compare || exit 1 # Prevent the comparison from running again
make $dashj all || exit 1
fi
else
============================================================