Use WTF::Variant for WebPreferencesStore::Value
Created attachment 346358 [details] Patch
Created attachment 346393 [details] Patch
Created attachment 346394 [details] Patch
Created attachment 346395 [details] Patch
Comment on attachment 346395 [details] Patch Looks good. I think this would be a great opportunity to add tests to TestWebKitAPI for ArgumentCoder. As you will see, the name TestWebKitAPI is a bit misleading, as we have many tests for WTF and WebCore types that aren't actually exposed as API, and are really just good old unit tests.
Comment on attachment 346395 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=346395&action=review > Source/WebKit/ChangeLog:8 > + It's basically a Variant now. This just uses templates. This is not a very clear commit message. Please actually state what you did. Also, per-function information would be appreciated.
I committed the WTF part in http://trac.webkit.org/r234514
(In reply to Sam Weinig from comment #5) > Comment on attachment 346395 [details] > Patch > > Looks good. I think this would be a great opportunity to add tests to > TestWebKitAPI for ArgumentCoder. As you will see, the name TestWebKitAPI is > a bit misleading, as we have many tests for WTF and WebCore types that > aren't actually exposed as API, and are really just good old unit tests. This requires significant refactoring because we have checks that HAVE and PLATFORM macros aren't exported in WebKit.framework's headers, and Encoder/Decoder and dependencies use those macros. I filed https://bugs.webkit.org/show_bug.cgi?id=188278
Created attachment 346418 [details] Patch
http://trac.webkit.org/r234516
<rdar://problem/42874231>
Comment on attachment 346418 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=346418&action=review > Source/WebKit/Platform/IPC/ArgumentCoders.h:564 > + ASSERT(static_cast<unsigned>(variant.index()) == variant.index()); I removed this assertion in http://trac.webkit.org/r234519 because it broke some builds.
Comment on attachment 346418 [details] Patch Not new to the patch: this code seems to misspell "overridden" repeatedly.