Ensure V8 is initialized before locking and unlocking threads.

Ensure V8 is initialized before locking and unlocking threads.

Committed: http://code.google.com/p/v8/source/detail?r=3008

[Erik Corry, 2009-10-02 11:12:07]

LGTM with comments

http://codereview.chromium.org/242123/diff/1/2
File src/v8threads.cc (right):

http://codereview.chromium.org/242123/diff/1/2#newcode72
Line 72: // with deserialization, so we must ensure initialization occurs here.
I'd like this to be moved into the if above so it doesn't waste time on
recursive locks and it would also be good to put it in an 'if
(!internal::V8::IsRunning()))' to further save time in the common case.

http://codereview.chromium.org/242123/diff/1/2#newcode72
Line 72: // with deserialization, so we must ensure initialization occurs here.
Can we mention the 'number of roots' issue explicitly in the comment.  I think
we're not the first to work this one out, but the last person didn't explain
fully what the issue was.

[William Hesse, 2009-10-02 11:29:29]

http://codereview.chromium.org/242123/diff/1/2
File src/v8threads.cc (right):

http://codereview.chromium.org/242123/diff/1/2#newcode72
Line 72: // with deserialization, so we must ensure initialization occurs here.
On 2009/10/02 11:12:07, Erik Corry wrote:
> I'd like this to be moved into the if above so it doesn't waste time on
> recursive locks and it would also be good to put it in an 'if
> (!internal::V8::IsRunning()))' to further save time in the common case.

Done.

http://codereview.chromium.org/242123/diff/1/2#newcode72
Line 72: // with deserialization, so we must ensure initialization occurs here.
On 2009/10/02 11:12:07, Erik Corry wrote:
> Can we mention the 'number of roots' issue explicitly in the comment.  I think
> we're not the first to work this one out, but the last person didn't explain
> fully what the issue was.

Done.