LGTM
I agree with your analysis here. See my comment on the bug. I think we
should
also guard against the case when it's not us who's sending the signal. We
can do
that by checking sampler->IsProfiling() next to sampler->IsActive().
In the future, when the locker/unlocker support for multiple isolates is
improved, this locked-by-current check should not be necessary, because the
thread that unlocked an isolate should have no current isolate.
http://codereview.chromium.org/6837028/
--
v8-dev mailing list
[email protected]
http://groups.google.com/group/v8-dev
