Code coverage false negatives
Categories
(Thunderbird :: Mail Window Front End, defect)
Tracking
(Not tracked)
People
(Reporter: darktrojan, Assigned: darktrojan)
References
Details
(Keywords: regression)
Attachments
(2 files)
About a week ago the code coverage score dropped quite a bit, and lots of code we know to have test coverage was listed as not covered. I assume this has been triggered by some change to the core code, but as with previous false-negatives, it's because some window objects are not being released properly and never record their coverage data. I've managed to trace this back to some faulty code in the folder quota display, which adds a TabSelect
listener and never removes it.
Assignee | ||
Comment 1•6 months ago
|
||
Assignee | ||
Updated•6 months ago
|
Pushed by geoff@darktrojan.net:
https://hg.mozilla.org/comm-central/rev/1c39d7ab86a0
Clear event listener when about:3pane window is unloaded. r=mkmelin
Assignee | ||
Comment 3•6 months ago
|
||
There must be more to this, because while that patch fixed some things, we're still missing coverage data that should exist. I'm not overly surprised, just disappointed. I'm going to reuse this bug to continue hunting.
Assignee | ||
Comment 4•6 months ago
|
||
We're missing a bunch of coverage data if the test finishes before the message loads. The data is not
from the loading window, but one of its ancestors, so I'm not entirely sure what's going on here.
Assignee | ||
Updated•6 months ago
|
Pushed by geoff@darktrojan.net:
https://hg.mozilla.org/comm-central/rev/46e94949ddf9
Stop browser_preferDisplayName.js breaking code coverage by waiting for messages to load. r=aleca
Description
•