REGRESSION(r230681) Do not use stored credentials if WKBundlePageResourceLoadClient.shouldUseCredentialStorage returns false
Created attachment 367788 [details] Patch
Created attachment 367805 [details] Patch
Comment on attachment 367805 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=367805&action=review > Source/WebKit/UIProcess/WebProcessPool.cpp:1724 > +void WebProcessPool::removeCredential(WebCore::Credential&& credential, WebCore::ProtectionSpace&& protectionSpace, CompletionHandler<void()>&& completionHandler) Do we really need Credential&& and ProtectionSpace&&?
(In reply to youenn fablet from comment #3) > Comment on attachment 367805 [details] > Patch > > View in context: > https://bugs.webkit.org/attachment.cgi?id=367805&action=review > > > Source/WebKit/UIProcess/WebProcessPool.cpp:1724 > > +void WebProcessPool::removeCredential(WebCore::Credential&& credential, WebCore::ProtectionSpace&& protectionSpace, CompletionHandler<void()>&& completionHandler) > > Do we really need Credential&& and ProtectionSpace&&? It is coming from IPC so I think this is common practice.
> > > Source/WebKit/UIProcess/WebProcessPool.cpp:1724 > > > +void WebProcessPool::removeCredential(WebCore::Credential&& credential, WebCore::ProtectionSpace&& protectionSpace, CompletionHandler<void()>&& completionHandler) > > > > Do we really need Credential&& and ProtectionSpace&&? > > It is coming from IPC so I think this is common practice. NetworkProcess::removeCredential is coming from IPC and maybe in the future we might need r-values. WebProcessPool::removeCredential is called directly from SPI and does not seem to need r-values.
(In reply to youenn fablet from comment #5) > > > > Source/WebKit/UIProcess/WebProcessPool.cpp:1724 > > > > +void WebProcessPool::removeCredential(WebCore::Credential&& credential, WebCore::ProtectionSpace&& protectionSpace, CompletionHandler<void()>&& completionHandler) > > > > > > Do we really need Credential&& and ProtectionSpace&&? > > > > It is coming from IPC so I think this is common practice. > > NetworkProcess::removeCredential is coming from IPC and maybe in the future > we might need r-values. > WebProcessPool::removeCredential is called directly from SPI and does not > seem to need r-values. Oh, right. I misread which one you were commenting on. Sorry.
https://trac.webkit.org/changeset/244521/webkit
<rdar://problem/50111468>
Reopening to attach new patch.
Created attachment 373556 [details] Patch
The patch is not for this bug...