532221 - Get BloatView working on child under e10s xpcshell tests

Status: RESOLVED FIXED

(Testing :: XPCShell Harness, defect)

Description

[Jason Duell]

So BloatView prints out a nice leak listing for xpcshell tests, but it only prints info for the chrome process, not the content.  It would be very helpful to have both.

Comment 1

[Jason Duell]

Attachment: Finds BloatView logs for tab process(es) and prints them too

OK, I think this fixes it, provided that it's OK for runxpcshell to print multiple BloatView logs for a single test (are there any tools that rely on just a single log being printed?).

Comment 2

[(not currently active) Ted Mielczarek]

The only thing that deals with the leak logs is runxpcshelltests.py itself. The other test harnesses (Mochitest, Reftest) actually analyze the log to ensure that we're not leaking anything, but we haven't gotten to that point with xpcshell yet (bug 469523).

That makes me think, did we handle the leak logging of child processes in bug 523208?

Comment 3

[Jonathan Griffin (:jgriffin)]

> That makes me think, did we handle the leak logging of child processes in bug
> 523208?

Yes, child process leaks are logged per bug 519570.

Comment 4

[Jason Duell]

Yet another patch that's ready to land in /projects/electrolysis.

In my patch queue I've got the following order of patches: bug 521922, bug 528145, and then this one.  The latter two may apply in either order, but why take chances?

Maybe it's time I got commit access...

Comment 5

[Jason Duell]

Attachment: Fix bitrot on approved patch: ready for commit

This just fixes a line of context that broke the patch.   Still ready to commit to the e10s tree.

Comment 6

[Jason Duell]

Attachment: yet more bitrot fixed

Comment 7

[Jason Duell]

http://hg.mozilla.org/projects/electrolysis/rev/500750dc6acc