899883 - Media Recording - Recording media with no timeslice in a mochitest leaks memory

Status: RESOLVED DUPLICATE of bug 896353

(Core :: Audio/Video: Recording, defect)

Description

[Jason Smith [:jsmith]]

The mochitest on bug 899878 currently will pass all specified tests, but will leak memory. See https://bugzilla.mozilla.org/attachment.cgi?id=783530 for the output of the test that reveals that we've leaked memory.

Note - this blocks being able to land bug 899878.

Comment 1

[Jason Smith [:jsmith]]

Randy indicates he is fixing this as part of bug 896353, so I'm duping against that bug.

Comment 2

[Jason Smith [:jsmith]]

Looks like bug 896353 didn't not fix all of the leaks involved, as the recent try run with that patch steals reveals leaks. Reopening as such.

Comment 3

[Jason Smith [:jsmith]]

Latest try run for analysis here - https://tbpl.mozilla.org/?tree=Try&rev=9fb7b34fc18b

Comment 4

[Randy Lin [:rlin] - Be Mozillian on 2015/01/26]

Do you run the hg pull to the latest one? 
https://tbpl.mozilla.org/?tree=Try&rev=f5fe2c48718d
BTW, I have commented out the state check after stop method.
Right now the state would change to inactive while fire the onstop event.
It's a little tricy to check the state right after the stop method or onstop event.
This bug would solve this problem. 
Bug 903758 - Media Recording - setting the recorder state to inActive on the stop method

Comment 5

[Jason Smith [:jsmith]]

(In reply to Randy Lin [:rlin] from comment #4)
> Do you run the hg pull to the latest one? 
> https://tbpl.mozilla.org/?tree=Try&rev=f5fe2c48718d
> BTW, I have commented out the state check after stop method.
> Right now the state would change to inactive while fire the onstop event.
> It's a little tricy to check the state right after the stop method or onstop
> event.
> This bug would solve this problem. 
> Bug 903758 - Media Recording - setting the recorder state to inActive on the
> stop method

Ah, looks like I didn't kick off the try run correctly. Okay, reduping then.