1694957 - abouthomecache/browser_process_crash.js fails in e10s if the process launch delay isn't 0ms

Status: NEW

(Core :: DOM: Content Processes, defect, P3)

Description

[Randell Jesup [:jesup] (needinfo me)]

Bug 1694036 fixes a bug where the dom.ipc.processLaunch.delayMs was only used at startup, instead of on all cases to refill the preallocation cache. This happens to break the restart simulation code used by abouthomecache/browser_process_crash.js code ("TypeError: can't access property "available", pageStream is undefined"). This is likely because it's observing process-creation events.

Comment 1

[BugBot [:suhaib / :marco/ :calixte]]

Set release status flags based on info from the regressing bug 1694036

Comment 2

[Chris Peterson [:cpeterson]]

kmag says this bug doesn't need to block Fission and that we can patch the test to set the pref to work around the problem for now.

P3

Comment 3

[Pascal Chevrel:pascalc (PTO until Sept 2)]

(In reply to Chris Peterson [:cpeterson] from comment #2)

kmag says this bug doesn't need to block Fission and that we can patch the test to set the pref to work around the problem for now.

P3

Is that work being done in this bug or somewhere else?

Comment 4

[Chris Peterson [:cpeterson]]

(In reply to Pascal Chevrel:pascalc from comment #3)

(In reply to Chris Peterson [:cpeterson] from comment #2)

kmag says this bug doesn't need to block Fission and that we can patch the test to set the pref to work around the problem for now.

Is that work being done in this bug or somewhere else?

I believe the intention was to add the test workaround in this bug.

This happens to break the restart simulation code used by abouthomecache/browser_process_crash.js code ("TypeError: can't access property "available", pageStream is undefined").

@ Jesup: is this test currently failing on Treeherder? Or does the test failure require some non-default value of dom.ipc.processLaunch.delayMs? How soon do we need to fix (or work around) this test issue?

Comment 5

[Randell Jesup [:jesup] (needinfo me)]

When I landed Bug 1694036 I included a patch to force the delay to 0 for the abouthome tests; that works around the problem, but it's not the correct solution (which would be to fix the tests themselves).