Fix valgrind memory leak in session_manager_unittest....

Fix valgrind memory leak in session_manager_unittest.

BUG=46678, 53277
TEST=remove 2 suppressions, run valgrind linux


Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=58496

[awong, 2010-08-26 20:39:06]

LGTM

Nice catch!

http://codereview.chromium.org/3170049/diff/1/2
File remoting/host/session_manager_unittest.cc (right):

http://codereview.chromium.org/3170049/diff/1/2#newcode71
remoting/host/session_manager_unittest.cc:71: delete arg2;
Stupid non-self-deleting callbacks.

[dhollowa, 2010-08-26 21:18:25]

LGTM.  Thanks.  Now to see how the bots like it...

[garykac, 2010-08-30 21:43:13]

On 2010/08/26 21:18:25, dhollowa wrote:
> LGTM.  Thanks.  Now to see how the bots like it...

The bots are happy except for mac_valgrind, which seems to be completely broken:
http://build.chromium.org/buildbot/try-server/builders/mac_valgrind/builds/285

I've run them through the mac valgrind bots 7 times since Friday and, AFAICT,
they don't seem to be returning useful info.

[garykac, 2010-08-31 00:57:12]

On 2010/08/30 21:43:13, garykac wrote:
> On 2010/08/26 21:18:25, dhollowa wrote:
> > LGTM.  Thanks.  Now to see how the bots like it...
> 
> The bots are happy except for mac_valgrind, which seems to be completely
broken:
> http://build.chromium.org/buildbot/try-server/builders/mac_valgrind/builds/285
> 
> I've run them through the mac valgrind bots 7 times since Friday and, AFAICT,
> they don't seem to be returning useful info.

Message from MA Ruel Re: mac try issues:

ld: in /b/slave/mac_valgrind/build/src/xcodebuild/Debug/libwebcore.a, can't map
file, errno=12
collect2: ld returned 1 exit status
It should be fixed at r57905. I worked the last 4 days to fix that f***ing bug.

So I'm running it through again...