Closed
Bug 811793
Opened 12 years ago
Closed 12 years ago
Get at least a minimal subset of Android XPCShell tests running on tbpl
Categories
(Testing :: General, defect)
Tracking
(Not tracked)
RESOLVED
WORKSFORME
People
(Reporter: wlach, Assigned: gbrown)
References
Details
Currently these tests break quite often because we're not running them in continuous integration (the most recent example being bug 799863). In addition to being useful to make sure that new commits don't break stuff, my understanding is that the xpcshell harness is also useful for developing new features (since xpcshell tests have a faster turnaround time than the full firefox for android).
Although possibly not as high a priority as some other things right now, we should probably at least try to track this. The first step, of course, is to unbreak the existing tests. :)
Comment 1•12 years ago
|
||
Since it sounds like time to work on this is limited, could it make sense to prioritize getting the continuous integration running the xpcshell tests before fixing the currently broken ones? In other words, disable all the currently broken ones, and then turn on running under CI ASAP. This would have the advantages of staunching the bleeding of ongoing breakage of existing tests, and would likely make people more motivated to write new ones sooner. It seems like a good way to avoid allowing the perfect to be the enemy of the good.
Comment 2•12 years ago
|
||
maybe even the unittests running on tbpl until we get the remaining failures fixed or manifested off.
Reporter | ||
Comment 3•12 years ago
|
||
We need the xpcshell tests to run on tegras before this'll work. This is covered by bug 768491.
Depends on: 768491
Assignee | ||
Comment 4•12 years ago
|
||
I have few prospects for making progress on bug 768491. On the tegras, xpcshell crashes on exit for all tests. In fact, even if I remove the entire body of the xpcshell main, it still crashes on exit. My attempts to debug with gdb have provided no additional info - the SIGSEGV is detected, but I cannot get a meaningful backtrace. I thought I might be able to get better diagnostics on a different platform, but there is no crash-on-exit on any other device that I have tested.
:blassey suggested that I abandon work on bug 768491 and we just run xpcshell tests on pandas.
Assignee | ||
Updated•12 years ago
|
Assignee: nobody → gbrown
Assignee | ||
Comment 5•12 years ago
|
||
Surprisingly, bug 768491 has been resolved, allowing us to run on tegras; bug 827554 is a minor barrier to running on pandaboards.
Assignee | ||
Comment 6•12 years ago
|
||
I'm not actively working on this currently; just trying to wait patiently for bug 828069. :)
Assignee: gbrown → nobody
Assignee | ||
Comment 7•12 years ago
|
||
We are running 247+ xpcshell tests on tegras now. Panda runs are planned in bug 874965. Additional tests are being enabled in bug 865006. This bug seems superfluous.
Assignee: nobody → gbrown
Status: NEW → RESOLVED
Closed: 12 years ago
Resolution: --- → WORKSFORME
You need to log in
before you can comment on or make changes to this bug.
Description
•