On Fri, Nov 28, 2014 at 06:48:01PM +0100, Thomas Graf wrote: > On 11/28/14 at 11:01am, Flavio Leitner wrote: > > On Fri, Nov 28, 2014 at 01:50:02PM +0100, Thomas Graf wrote: > > > This reduces the total travis build time significantly. > > > > > > Signed-off-by: Thomas Graf <tg...@noironetworks.com> > > > --- > > > .travis/build.sh | 2 +- > > > 1 file changed, 1 insertion(+), 1 deletion(-) > > > > > > diff --git a/.travis/build.sh b/.travis/build.sh > > > index 6e99c63..990d263 100755 > > > --- a/.travis/build.sh > > > +++ b/.travis/build.sh > > > @@ -80,7 +80,7 @@ else > > > make CFLAGS="$CFLAGS" C=1 > > > fi > > > > > > -if [ $TESTSUITE ]; then > > > +if [ $TESTSUITE ] && [ $CC != "clang" ]; then > > > if ! make distcheck; then > > > # testsuite.log is necessary for debugging. > > > cat */_build/tests/testsuite.log > > > > I like the idea but my concern is that clang could > > generate a slightly different code that helps to find > > a subtle problem. > > Completely agree with the point. I should have written this in the > commit message: We had several builds fail because they exceeded the > enforced time limit. It's usually due to increased load on travis > which slows down the tests. This should help to have less spurious > failures. I agree, it's a trade off and we can go either way.
I don't know travis but I suspect there is no way to pass extra options like compile this time with gcc and another time with clang. Acked-by: Flavio Leitner <f...@redhat.com> _______________________________________________ dev mailing list dev@openvswitch.org http://openvswitch.org/mailman/listinfo/dev
