missing a tc_control from a test directory results hard to decipher failure
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
UTAH |
Fix Released
|
High
|
Andy Doan |
Bug Description
I accidentally forgot to populate a test with a tc_control file and I got the following error:
+ sleep 2
+ utah -d -o /var/lib/
2013-06-13 16:31:25,550 - DEBUG - Namespace(
2013-06-13 16:31:25,558 - INFO - ('Linux', 'ubuntu-phablet', '3.4.0-1-mako', '#8-Ubuntu SMP PREEMPT Mon Jun 10 14:28:29 UTC 2013', 'armv7l', 'armv7l')
DEVELOPMENT
utah_tests_
+ [ -f /var/lib/
+ python -c import xmlrpclib; s=xmlrpclib.
utah: completed /etc/utah/
..so I got a return code of zero (success??) and none of the other tests got run. Took me a while to figure out why it passed yet failed to run all my existing tests when I added a new test without a tc_control file.
Perhaps there should be some better sanity checking and error handling if users make dumb mistakes like I did.
Related branches
- Max Brustkern (community): Approve
-
Diff: 56 lines (+6/-3)4 files modifiedclient.py (+1/-0)
debian/changelog (+1/-0)
utah/client/common.py (+2/-2)
utah/client/testcase.py (+2/-1)
Changed in utah: | |
status: | Fix Committed → Fix Released |
there are two parts to this bug:
1) the utah-client's error message stinks. its practically hidden: tests_sample/ idle_max_ brightness/ tc_control"
"utah_
I'm fixing this ASAP as well as a couple of others I've uncovered.
2) I need to dig a bit in my new phablet script. One part of the problem is that adb-shell always returns 0. but we should be able to figure out what's really happening.