GNOME Bugzilla – Bug 582794
rganalysis unit test fails with git core
Last modified: 2009-05-16 10:24:54 UTC
The unit test checks for message->src == element, which now fails since gst_element_found_tags_for_pad() now sets the src object to the pad and no longer to the element. Options: [ ] do nothing [ ] remove assertion from unit test [ ] use gst_element_found_tags() instead of gst_element_found_tags_for_pads() [ ] revert core change Pick one. Depends if we want the -good 0.10.15 release to successfully 'make check' against the next core release, I guess.
The whole thing about changing our messages to have pads as senders scares me. I'm OK with going along with it if people think it's cool, but I'm not particularly comfortable yet that we won't break the world for someone.
> The whole thing about changing our messages to have pads as senders scares me. > > I'm OK with going along with it if people think it's cool, but I'm not > particularly comfortable yet that we won't break the world for someone. Maybe, I don't know. It has always been a GstObject and always been wrong, and we've never really cared much about obviously wrong things in the past. I think the thing in git is more of a trial balloon, to see if something happens. We can still revert it later if we want. In any case, I think we should just remove that particular assertion from the unit test and be done with it. It doesn't really serve any particular purpose in the context of the test and leaves all options open for later.
commit fa15557bf58b6e22f828ab8d9af50898ac7ce980 Author: Jan Schmidt <thaytan@noraisin.net> Date: Sat May 16 01:21:34 2009 +0100 check: Remove assertion that breaks check again git master Remove the assertion that the sender of the tags message is the element until we decide whether that's going to be true or not.
Can't this bug be closed then or shall it be used as a tracker for the core change?