Now that HashMap works with OwnPtr mapped values, we want to go through all the sites using deleteAllValues on a HashMap and change them to use OwnPtr instead. Once we’ve done that, we may decide to remove the deleteAllValues function.
Anders, did you finish the job on this? Should I mark it as resolved?
Looks like there’s a handful of uses left.
Here’s all that’s left: Source/WebCore/platform/blackberry/CookieMap.cpp: deleteAllValues(m_subdomains); Source/WebCore/platform/network/blackberry/rss/RSSParserBase.cpp: deleteAllValues(m_items); Source/WebCore/platform/win/WCDataObject.cpp: WTF::deleteAllValues(m_formats); Source/WebKit/blackberry/WebKitSupport/InPageSearchManager.cpp: deleteAllValues(m_deferredScopingWork); Source/WebKit2/Shared/Plugins/NPRemoteObjectMap.cpp: deleteAllValues(messageReceivers); Tools/DumpRenderTree/win/DRTDataObject.cpp: WTF::deleteAllValues(m_formats); Tools/DumpRenderTree/win/UIDelegate.cpp: ~DRTUndoStack() { deleteAllValues(m_undoVector); } Tools/DumpRenderTree/win/UIDelegate.cpp: void clear() { deleteAllValues(m_undoVector); m_undoVector.clear(); } And then we are done.
We are now down to 5 uses of this. One in NPRemoteObjectMap.cpp, and the other 4 all in Windows-specific code.
(In reply to comment #4) > We are now down to 5 uses of this. One in NPRemoteObjectMap.cpp, and the other 4 all in Windows-specific code. I'll see about those Windows files ASAP.
Created attachment 229519 [details] Patch
Created attachment 229542 [details] Patch
The build failure on Windows EWS looks like the bot has a zombie build process holding onto a file. I'll ask Lucas to reboot the bot.
Comment on attachment 229542 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=229542&action=review r=me > Source/WTF/wtf/Vector.h:-1216 > -} Yay!
Committed r167432: <http://trac.webkit.org/changeset/167432>