Closed Bug 947259 Opened 7 years ago Closed 7 years ago
Script Finished() has no effect in performance _helper _atom .
marionetteScriptFinished() has no effect in performance_helper_atom. This is a followup from bug 915156 In tests/performance/app.js: waitForPerfEvents(), we call executeAsyncScript(). The atom is in ./tests/performance/performance_helper_atom.js, when the function finish() is called (we checked it is), marionetteScriptFinished() is called, but does nothing. Upon closer inspection, it seems that in marionette-listener.js (Gecko) the commandId as passed to asyncComplete is different from the global asyncTestCommandId. Note sure exactly what's happening yet, but this cause perf test to succeed on a 50s timeout.
It seems that it is a problem with order of things, aka a misunderstanding on how it works. See pull request.
Comment on attachment 8346717 [details] [review] Pull request cancelling review due to some things I forgot.
Comment on attachment 8346717 [details] [review] Pull request This time we should be good. Look ! No timeout !
Remove blocker. Isn't really blocked.
No longer depends on: 949209
Comment on attachment 8346717 [details] [review] Pull request Added some cleanup comments on the PR. I admit I still don't understand how marionetteScriptFinished works but I'll trust you on this ;)
I have updated the PR addressing the cleanup comments. Travis is only perma-orange.
Comment on attachment 8346717 [details] [review] Pull request r=me with the 2 additional comments
Attachment #8346717 - Flags: review?(felash) → review+
Status: ASSIGNED → RESOLVED
Closed: 7 years ago
Resolution: --- → FIXED
shall we backport to 1.4? should be easy and would allow the test to run better on 1.4
blocking-b2g: --- → 1.4?
I meant 1.3?
blocking-b2g: 1.4? → 1.3?
Let's just do it, it's a=npotb (not part of the build). John, could you assist please?
blocking-b2g: 1.3? → ---
[v1.3 81af9e2] Merge pull request #14621 from hfiguiere/bug947259
You need to log in before you can comment on or make changes to this bug.