For many years we have followed the same process where we file a bug for 
each failure where the summary is the exact error message (a few 
exceptions).

This results in a many:one relationship between bugs and testcases as 
testcases can fail for different reasons.

This current process makes it harder to:
1) understand how often the test case is really failing (which indicates 
noise or need for attention)
2) automatically annotate known failures (sheriffs/try pushes/etc.)

The change we will be implementing is to have a single meta bug for each 
test case (file in mozilla-central).  An example is:
https://bugzilla.mozilla.org/show_bug.cgi?id=1757074

If there is an investigation to fix the testcase or disable it on a given 
platform, then we can create a blocking bug on our meta bug for tracking 
purposes.  

We will start off small, here are the restrictions we will use:
1) only test failures, not crashes || timeouts || leaks || infra
2) only for xpcshell

The plan is to roll this out (in the next 2 weeks) and monitor it for a few 
weeks before expanding to other test suites.

Feel free to ask questions in Matrix: #test  (firefox-testing).

-- 
You received this message because you are subscribed to the Google Groups 
"[email protected]" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To view this discussion on the web visit 
https://groups.google.com/a/mozilla.org/d/msgid/dev-platform/a3d5ea68-b2c6-4898-8716-189d6acc063dn%40mozilla.org.

Reply via email to