GNOME Bugzilla – Bug 697693
Some tests fail when CK_FORK=no
Last modified: 2018-05-04 12:41:40 UTC
Lack of forking has a number of side-effects. For example, pipelines/vorbisenc.c does this: GstElement *sink = gst_bin_get_by_name (GST_BIN (bin), "fakesink0"); more than once, but the number on the element "fakesink" will be "1" for the second test of the batch, not "0". Please run the testsuite with CK_FORK=no on a platform where all tests pass, and fix the failures introduced by CK_FORK=no. Otherwise running the testsuite on platforms that have no fork (such as W32) is completely useless, since it's difficult to tell apart nofork-related problems and true w32-only bugs (also, tests may hang up).
CK_FORK=no make pipelines/vorbisenc.check passes for me. but other do indeed break. I'll close this as wontfix. Please re-open if you plan to fix this and then we can discuss about enabling this on the CI, for example by enabling tests on a future Windows CI.