Remove static initializer in dummy_event

Remove static initializer in dummy_event


BUG=94925
TEST=sizes


Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=120405

[Tyler Breisacher (Chromium), 2012-01-27 05:32:28]

Not sure if this is the right way to do this...

[jam, 2012-01-27 18:19:04]

http://codereview.chromium.org/9297009/diff/1/ipc/ipc_sync_message.cc
File ipc/ipc_sync_message.cc (right):

http://codereview.chromium.org/9297009/diff/1/ipc/ipc_sync_message.cc#newcode51
ipc/ipc_sync_message.cc:51: static base::WaitableEvent* dummy_event = new
base::WaitableEvent(true, true);
I don't believe this is thread safe, i.e. if it's called by two threads at the
same time two objects can be created.

[Tyler Breisacher (Chromium), 2012-01-30 22:58:03]

http://codereview.chromium.org/9297009/diff/1/ipc/ipc_sync_message.cc
File ipc/ipc_sync_message.cc (right):

http://codereview.chromium.org/9297009/diff/1/ipc/ipc_sync_message.cc#newcode51
ipc/ipc_sync_message.cc:51: static base::WaitableEvent* dummy_event = new
base::WaitableEvent(true, true);
Right, sorry. What do you think is the right way to handle this case?

[jam, 2012-02-01 17:01:20]

http://codereview.chromium.org/9297009/diff/1/ipc/ipc_sync_message.cc
File ipc/ipc_sync_message.cc (right):

http://codereview.chromium.org/9297009/diff/1/ipc/ipc_sync_message.cc#newcode51
ipc/ipc_sync_message.cc:51: static base::WaitableEvent* dummy_event = new
base::WaitableEvent(true, true);
On 2012/01/30 22:58:03, tbreisacher wrote:
> Right, sorry. What do you think is the right way to handle this case?

use LazyInstance. it's thread-safe

[Tyler Breisacher (Chromium), 2012-02-01 20:00:40]

I haven't worked with LazyInstance before, so I could be missing something, but
it looks like LazyInstance can only call the no-arg constructor for the type
it's a lazy instance of. Unfortunately, WaitableEvent doesn't have a no-arg
constructor. Should we add one?

[jam, 2012-02-03 00:32:55]

On 2012/02/01 20:00:40, Tyler Breisacher wrote:
> I haven't worked with LazyInstance before, so I could be missing something,
but
> it looks like LazyInstance can only call the no-arg constructor for the type
> it's a lazy instance of. Unfortunately, WaitableEvent doesn't have a no-arg
> constructor. Should we add one?

see
http://code.google.com/codesearch#OAMlx_jo-ck/src/content/browser/renderer_ho...
for an examle

[Tyler Breisacher (Chromium), 2012-02-03 05:03:52]

Thank you! I think this should work now.

[jam, 2012-02-03 07:54:04]

lgtm

[commit-bot: I haz the power, 2012-02-03 07:59:00]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/tbreisacher@chromium.org/9297009/8002

[commit-bot: I haz the power, 2012-02-03 09:15:05]

Try job failure for 9297009-8002 (retry) on linux_rel for step "ui_tests".
It's a second try, previously, step "ui_tests" failed.
http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=linux_rel&...

[commit-bot: I haz the power, 2012-02-03 19:06:56]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/tbreisacher@chromium.org/9297009/8004

[commit-bot: I haz the power, 2012-02-03 21:48:45]

Change committed as 120405

[Nico, 2012-02-03 23:02:16]

http://codereview.chromium.org/9297009/diff/8004/ipc/ipc_sync_message.cc
File ipc/ipc_sync_message.cc (right):

http://codereview.chromium.org/9297009/diff/8004/ipc/ipc_sync_message.cc#newc...
ipc/ipc_sync_message.cc:28: static base::LazyInstance<base::WaitableEvent,
WaitableEventLazyInstanceTraits>
Can you make this ::Lazy? As is, this adds an exit time destructor call to at
atexit() time, which is kind of like a static initializer in reverse.

[Tyler Breisacher (Chromium), 2012-02-03 23:38:17]

We discussed this over IM but just to put it here for future reference: We can't
really make it leaky because LeakyLazyInstanceTraits is in an internal
namespace. We could just use LazyInstance<WaitableEvent>::Leaky, except that
calls WaitableEvent's no-arg constructor. But WaitableEvent doesn't have a
no-arg constructor so we will leave this as is for now.