On Wed, 2018-02-28 at 14:55 +0000, Daniel P. Berrangé wrote:
When building with CLang the structs that are emitted by pdwtags
appear
in a completely different order than with GCC, which causes the
comparison against expected data to fail.
Ideally the test would not be sensitive to the ordering, because even
future GCC could cause changes, but that's not easy to fix. So for now
just skip the test when using clang.
Signed-off-by: Daniel P. Berrangé <berrange(a)redhat.com>
---
src/Makefile.am | 10 +++++++++-
1 file changed, 9 insertions(+), 1 deletion(-)
Where have I seen this already? Oh, right :)
https://www.redhat.com/archives/libvir-list/2016-March/msg00514.html
diff --git a/src/Makefile.am b/src/Makefile.am
index 8d72f2f1e5..4207421986 100644
--- a/src/Makefile.am
+++ b/src/Makefile.am
@@ -664,8 +664,16 @@ struct_prefix = ($(libs_prefix)|$(other_prefix))
# remote/{,.libs/}libvirt_driver_remote_la-remote_protocol.o. We want
# the newest of the two, in case configure options changed and a stale
# file is left around from an earlier build.
+# The pdwtags output is completely different when building with clang
+# which causes the comparison against expected output to fail, so skip
+# if using clang as CC.
PDWTAGS = \
- $(AM_V_GEN)if (pdwtags --help) > /dev/null 2>&1; then \
+ $(AM_V_GEN)$CC -v 2>&1 | grep clang >/dev/null; \
+ if test $$? == 1; then \
+ echo 'WARNING: skipping pdwtags test with Clang' >&2; \
+ exit 0; \
+ fi; \
You could do
if $(CC) -v 2>&1 | grep -q clang; then
...
instead, it's shorter.
+ if (pdwtags --help) > /dev/null 2>&1; then \
o=`ls -t $(<:.lo=.$(OBJEXT)) \
$(subst /,/.libs/,$(<:.lo=.$(OBJEXT))) \
2>/dev/null | sed -n 1p`; \
Reviewed-by: Andrea Bolognani <abologna(a)redhat.com>
--
Andrea Bolognani / Red Hat / Virtualization