Prevent a worker that loops forever from preventing shutdown of the worker pr...

Prevent a worker that loops forever from preventing shutdown of the worker process.

BUG=11672
Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=16662

[Matt Perry, 2009-05-21 19:59:31]

http://codereview.chromium.org/112049/diff/1/2
File chrome/worker/webworkerclient_proxy.cc (right):

http://codereview.chromium.org/112049/diff/1/2#newcode34
Line 34: proxy_->workerContextDestroyed();
What if the proxy is destroyed before this runs?  Will this crash the process?

http://codereview.chromium.org/112049/diff/1/2#newcode121
Line 121: if (message.type() == WorkerMsg_TerminateWorkerContext::ID) {
This pattern is kind of ugly in my opinion.  Can you do another
IPC_BEGIN_MESSAGE_MAP that handles the TerminateWorkerContext message here?

http://codereview.chromium.org/112049/diff/1/2#newcode126
Line 126: NOTREACHED();
why can't this be reached?

[jam, 2009-05-21 20:19:26]

http://codereview.chromium.org/112049/diff/1/2
File chrome/worker/webworkerclient_proxy.cc (right):

http://codereview.chromium.org/112049/diff/1/2#newcode34
Line 34: proxy_->workerContextDestroyed();
On 2009/05/21 19:59:32, Matt Perry wrote:
> What if the proxy is destroyed before this runs?  Will this crash the process?

The process would have already quit, since this code runs when there's one
worker per process only.

http://codereview.chromium.org/112049/diff/1/2#newcode121
Line 121: if (message.type() == WorkerMsg_TerminateWorkerContext::ID) {
On 2009/05/21 19:59:32, Matt Perry wrote:
> This pattern is kind of ugly in my opinion.  Can you do another
> IPC_BEGIN_MESSAGE_MAP that handles the TerminateWorkerContext message here?

done

http://codereview.chromium.org/112049/diff/1/2#newcode126
Line 126: NOTREACHED();
On 2009/05/21 19:59:32, Matt Perry wrote:
> why can't this be reached?

I guess I meant something more like NOTIMPLEMENTED(), done.

[Matt Perry, 2009-05-21 20:32:52]

LGTM