Opened 10 years ago
Closed 4 years ago
#4455 closed defect (fixed)
Better dependency tracking for ffprobe FATE tests needed
Reported by: | Timothy Gu | Owned by: | |
---|---|---|---|
Priority: | normal | Component: | undetermined |
Version: | git-master | Keywords: | fate |
Cc: | Blocked By: | ||
Blocking: | Reproduced by developer: | yes | |
Analyzed by developer: | no |
Description
Right now, ffprobe tests uses tests/data/ffprobe-test.nut
, which depends on ffmpeg$(EXESUF)
. However, ffprobe tests are not disabled with ffmpeg is disabled, like when --disable-avdevice
is passed. This leads to errors when testing ffprobe.
There are two variants:
http://fate.ffmpeg.org/log.cgi?time=latest&log=test&slot=x86_64-archlinux-gcc-disable-everything:
GEN tests/data/ffprobe-test.nut /home/ux/fate/ffmpeg/tests/Makefile:47: recipe for target 'tests/data/ffprobe-test.nut' failed make: *** [tests/data/ffprobe-test.nut] Error 1
http://fate.ffmpeg.org/log.cgi?time=latest&log=test&slot=x86_64-archlinux-gcc-disableavfilter:
make: *** No rule to make target 'ffmpeg', needed by 'tests/data/ffprobe-test.nut'.
Attachments (1)
Change History (4)
comment:1 by , 10 years ago
Reproduced by developer: | set |
---|---|
Status: | new → open |
Version: | unspecified → git-master |
by , 10 years ago
Attachment: | patchfate.diff added |
---|
comment:2 by , 10 years ago
Fixed an error in my first comment and sent a patch to change the test file so the audio stream gets disposition default.
comment:3 by , 4 years ago
Resolution: | → fixed |
---|---|
Status: | open → closed |
Fixed in 4845f6687d4f1ddcb1ed57b09b09f284e8500fc1 (it added a dependency on avdevice; one could do this in a more fine-grained manner, but it fixes the issue as reported).
FFprobe tests are disabled if ffmpeg is disabled,ffmpeg does not depend on avdevice.The FFprobe tests depend on a file that can currently only be created if avdevice was enabled, I will attach a patch that works around this issue by not using avdevice when creating the file. Unfortunately, it slightly changes the output file because the audio stream gets a default stream (this may or may not be related to ticket #3622).
A solution would be to first change the test file so the audio stream will be set to default and then apply my patch.