Closed
Bug 1127942
Opened 9 years ago
Closed 6 years ago
devicemanagerADB fails to retry on timeout and fails silently
Categories
(Testing :: Mozbase, defect)
Tracking
(Not tracked)
RESOLVED
WORKSFORME
People
(Reporter: gbrown, Assigned: gbrown)
References
Details
See http://hg.mozilla.org/mozilla-central/annotate/4fab18f1ff55/testing/mozbase/mozdevice/mozdevice/devicemanagerADB.py#l601. devicemanagerADB._checkCmd uses mozprocess.ProcessHandler to execute adb and wait for its completion. If the wait returns None, then _checkCmd retries the command; if none of the retries succeeds, we raise a DMError, "Timeout exceeded for _checkCmd call after %d retries." In bug 1127082, on the Android 4.4 emulator on aws, pushDir() frequently exceeds its timeout, mozprocess kills adb (I think), and then proc.wait() returns -9; since that is not None, no retries are triggered, no DMError is thrown, _checkCmd returns -9, but that triggers no error handling in pushDir, and the failure is silently ignored. I think the intention in the existing code is to retry on timeout, and that's not working. I am not sure that retrying on timeout is a great idea, but maybe that is useful sometimes? Certainly the silent failure is a bad thing. Should _checkCmd test for a 0 return code?
|
Assignee | |
Comment 3•6 years ago
|
||
Much better now.
Assignee: nobody → gbrown
Status: NEW → RESOLVED
Closed: 6 years ago
Resolution: --- → WORKSFORME
You need to log in
before you can comment on or make changes to this bug.
Description
•