The SEGV in test-driver of course has nothing to do with your patches. The behavior of logging now respects cdio_loglevel_default more pervasively so the behavior changed. I think for the better here. So that test was wrong.
Commit 311455eefaebcf483a75e4134511888ec1cfd55a fixes this along wiith your automake patch applied. I had intended to make these two separate commits, but that slipped by. Your other patch to remove the executable bit also has been applied. On Tue, Jul 28, 2015 at 4:21 AM, Adrian Reber <[email protected]> wrote: > I got following Fedora bug report: > > https://bugzilla.redhat.com/show_bug.cgi?id=1247351 > > which points out that there are files with the executable bit set, > which do not really need it. So I have a patch to remove some of the > unneeded executable bits: > > > https://lisas.de/git/?p=libcdio.git/git;a=commit;h=a3435e6f74e01e7c02e774101d68f6c246a2fb7a > > and another change which silences an automake warning: > > > https://lisas.de/git/?p=libcdio.git/git;a=commit;h=d83906e0319b99b9056fdfc9f549c045b3df3b7b > > > Checking if my changes break anything I found out that > the commit 8fcd4ce909f4f04e77bee78f573a241d0e9f2c34 breaks make check. > > commit 8fcd4ce909f4f04e77bee78f573a241d0e9f2c34 > Author: R. Bernstein <[email protected]> > Date: Tue May 26 08:40:40 2015 -0400 > > Recursion checking in cdio_logv() can't use cdio_assert_not_reached() > which causes more recursion. To reduce the possibility concurrent > log writes, we'll eliminate those that are ignored earlier. > sr #108751: > > After that commit it fails with: > > ../../test-driver: line 107: 24504 Segmentation fault (core dumped) > "$@" > $log_file 2>&1 > FAIL: logger > > > Adrian > >
