HashMapImpl needs to m_buffer.clear() in its constructor
authorfpizlo@apple.com <fpizlo@apple.com@268f45cc-cd09-0410-ab3c-d52691b4dbfc>
Fri, 9 Sep 2016 16:50:10 +0000 (16:50 +0000)
committerfpizlo@apple.com <fpizlo@apple.com@268f45cc-cd09-0410-ab3c-d52691b4dbfc>
Fri, 9 Sep 2016 16:50:10 +0000 (16:50 +0000)
commitadd16fee5fccc89da6154759922292ba9977bee5
treeb253f893e11c74f5c0963a4b0b09f6c01cc10a2d
parent6fa62ab0ee701b2e8005c45992bdfeda6e3d862c
HashMapImpl needs to m_buffer.clear() in its constructor
https://bugs.webkit.org/show_bug.cgi?id=161796

Reviewed by Keith Miller.

This is the second time that I'm fixing a bug because AuxiliaryBarrier does not initialize
itself. That seemed like a good idea because maybe sometimes the user knows better how to
initialize it. But, it's not worth it if it's a constant source of bugs.

So, I'm fixing it for good by making AuxiliaryBarrier::AuxiliaryBarrier() initialize its
m_value.

* runtime/AuxiliaryBarrier.h:
(JSC::AuxiliaryBarrier::AuxiliaryBarrier):
* runtime/DirectArguments.cpp:
(JSC::DirectArguments::DirectArguments):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@205746 268f45cc-cd09-0410-ab3c-d52691b4dbfc
Source/JavaScriptCore/ChangeLog
Source/JavaScriptCore/runtime/AuxiliaryBarrier.h
Source/JavaScriptCore/runtime/DirectArguments.cpp