ASSERT(m_state = Open); is bad news bears
Created attachment 69586 [details]
Comment on attachment 69586 [details]
With gcc 4.5 would write a plugin to inspect basic blocks for ASSERTs... Good catch!!!!
Committed r68986: <http://trac.webkit.org/changeset/68986>
Good to fix, but nearly harmless; the effect is local to the assertion. In a debug build we simply lose the assertion; but if the assertion was true then the statement has no effect. In a release build the expression is not even compiled.
Sure, but the bad case is if the assertion is false. Then the statement masks the problem in Debug builds, but Release builds behave differently.