Running a single failing xpcshell test for debugging shouldn't run it twice
Categories
(Testing :: XPCShell Harness, defect, P2)
Tracking
(firefox67 fixed)
| Tracking | Status | |
|---|---|---|
| firefox67 | --- | fixed |
People
(Reporter: florian, Assigned: gbrown)
References
Details
Attachments
(1 file)
When attempting to debug an xpcshell test to debug it using ./mach test <path>.js, the test runs and fails, and then is ran a second time after the "Retrying tests that failed when run in parallel" message.
Apparently this was already fixed once in bug 969633.
|
Assignee | |
Comment 1•5 years ago
|
||
The earlier fix looks appropriate. I think it was accidentally lost in bug 1209312.
|
|
Assignee | |
Updated•5 years ago
|
|
|
Assignee | |
Comment 2•5 years ago
|
||
In the harness, sequential mode is set when self.singleFile is set. self.singleFile has
been set correctly when a single test is specified as a test_path (mach xpcshell-test <test>)
but it has not been set correctly if a single test is specified in a manifest (mach test <test);
this patch corrects for that by setting self.singleFile after all test specifications have
set self.alltests.
An unintended consequence of the code rearrangement is that the chunk_by_slice filter is
now used even when self.singleFile would be set -- but only if totalChunks > 1; this
situation seems very unlikely, and the consequences minimal.
Pushed by gbrown@mozilla.com: https://hg.mozilla.org/integration/autoland/rev/8130ba1701bb Use xpcshell sequential mode consistently when running a single test; r=bc
|
||
Comment 4•5 years ago
|
||
| bugherder | ||
Description
•