Fix memory leak in unittests with new Mojo EDK.

Fix memory leak in unittests with new Mojo EDK.

BUG=561803
Committed: https://crrev.com/2f282b01c7b51c3e87668a9bc00b53900af07038
Cr-Commit-Position: refs/heads/master@{#366026}

[jam, 2015-12-18 01:48:32]

jam@chromium.org changed reviewers:
+ msw@chromium.org

[msw, 2015-12-18 01:54:27]

lgtm with nit and q

https://codereview.chromium.org/1534013003/diff/20001/content/app/mojo/mojo_i...
File content/app/mojo/mojo_init.cc (left):

https://codereview.chromium.org/1534013003/diff/20001/content/app/mojo/mojo_i...
content/app/mojo/mojo_init.cc:29: if (0 && process_type.empty() &&
!command_line.HasSwitch("use-old-edk")) {
q: did you intend to enable the new edk by default in this CL?

https://codereview.chromium.org/1534013003/diff/20001/mojo/edk/system/message...
File mojo/edk/system/message_pipe_dispatcher.cc (right):

https://codereview.chromium.org/1534013003/diff/20001/mojo/edk/system/message...
mojo/edk/system/message_pipe_dispatcher.cc:940: call_release = true;
nit: could you actually just call Release() here and remove |call_release|?

[jam, 2015-12-18 02:03:47]

https://codereview.chromium.org/1534013003/diff/20001/content/app/mojo/mojo_i...
File content/app/mojo/mojo_init.cc (left):

https://codereview.chromium.org/1534013003/diff/20001/content/app/mojo/mojo_i...
content/app/mojo/mojo_init.cc:29: if (0 && process_type.empty() &&
!command_line.HasSwitch("use-old-edk")) {
On 2015/12/18 01:54:26, msw wrote:
> q: did you intend to enable the new edk by default in this CL?

this was for the try runs, I'll land it without this line :) sorry wasn't clear

https://codereview.chromium.org/1534013003/diff/20001/mojo/edk/system/message...
File mojo/edk/system/message_pipe_dispatcher.cc (right):

https://codereview.chromium.org/1534013003/diff/20001/mojo/edk/system/message...
mojo/edk/system/message_pipe_dispatcher.cc:940: call_release = true;
On 2015/12/18 01:54:26, msw wrote:
> nit: could you actually just call Release() here and remove |call_release|?

that's what I did in patchset 1 but that failed since if this was the last
reference, the autolock's destructor above would cause a UAF

[jam, 2015-12-18 03:55:50]

The earlier fix didn't fix all the leaks. Here's a simpler change that fixes all
(verified locally). Thanks.

[jam, 2015-12-18 04:29:58]

Patchset #5 (id:80001) has been deleted

[Fady Samuel, 2015-12-18 04:40:37]

fsamuel@chromium.org changed reviewers:
+ fsamuel@chromium.org

[Fady Samuel, 2015-12-18 04:40:38]

lgtm

[jam, 2015-12-18 04:53:26]

The CQ bit was checked by jam@chromium.org

[jam, 2015-12-18 04:53:26]

The patchset sent to the CQ was uploaded after l-g-t-m from msw@chromium.org
Link to the patchset: https://codereview.chromium.org/1534013003/#ps100001
(title: "revert change to enable new edk")

[commit-bot: I haz the power, 2015-12-18 04:54:00]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1534013003/100001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1534013003/100001

[commit-bot: I haz the power, 2015-12-18 05:59:57]

Committed patchset #5 (id:100001)

[commit-bot: I haz the power, 2015-12-18 06:00:30]

Description was changed from

==========
Fix memory leak in unittests with new Mojo EDK.

BUG=561803
==========

to

==========
Fix memory leak in unittests with new Mojo EDK.

BUG=561803

Committed: https://crrev.com/2f282b01c7b51c3e87668a9bc00b53900af07038
Cr-Commit-Position: refs/heads/master@{#366026}
==========

[commit-bot: I haz the power, 2015-12-18 06:00:31]

Patchset 5 (id:??) landed as
https://crrev.com/2f282b01c7b51c3e87668a9bc00b53900af07038
Cr-Commit-Position: refs/heads/master@{#366026}