I've been looking at the test itself. It's checking the get_caps method on one of these GstTypeFindFactory objects.
> > > # Failed test 'The thing isa GStreamer::Caps' > > > # at t/GstTypeFindFactory.t line 16. > > > # The thing isn't defined Upstream have already removed a test from this script for ->get_extensions() with a comment that they cannot guarantee the output will not be NULL. On my machine, I dumped the content of ->get_caps() for every element from ->get_list() (not just the first), and there were 2 out of 164 which weren't returning a GStreamer::Caps object. I haven't tracked these down further. On some machines it might be possible that one of these gets put at the start of the list, and the test will fail. It looks like the registration will happen in the same order each time, but I don't know what controls that ordering. I don't know why the build would succeed for lucas on tmpfs, either. :) Reading the gstreamer source (gsttypefind.c), it looks like both extensions and capabilities are *optional* arguments to gst_type_find_register. So I reckon the right thing to do is to disable this test as well, and tell upstream (who might well have more knowledge of this). -- Tim Retout <t...@retout.co.uk> -- To UNSUBSCRIBE, email to debian-bugs-rc-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
