1.) Source/WebCore/dom/Node.cpp should include ExceptionCode.h HIERARCHY_REQUEST_ERR, NOT_FOUND_ERR, NOT_SUPPORTED_ERR, ... are defined in ExceptionCode.h In normal WebKit build ExceptionCode.h is included indirectly, but --minimal build fails. 2.) Source/WebCore/page/Page.cpp shouldn't call the undefined PlatformMediaSessionManager::updateNowPlayingInfoIfNecessary() It is defined on Mac always and on !Mac inside ENABLE(VIDEO) || ENABLE(WEB_AUDIO) guards. 3.) emptyString() can't be converted to ExceptionOr<String> in Source/WebCore/testing/Internals.cpp ../../Source/WebCore/testing/Internals.cpp: In member function 'WebCore::ExceptionOr<WTF::String> WebCore::Internals::captionsStyleSheetOverride()': ../../Source/WebCore/testing/Internals.cpp:2511:23: error: could not convert 'WTF::emptyString()' from 'const WTF::String' to 'WebCore::ExceptionOr<WTF::String>'
Created attachment 294058 [details] Patch
Comment on attachment 294058 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=294058&action=review I am OK with these build fixes, but the ones in Page.cpp could be done a little better. > Source/WebCore/page/Page.cpp:1450 > +#if PLATFORM(MAC) || ENABLE(VIDEO) || ENABLE(WEB_AUDIO) If we keep this conditional, there is no reason for it to mention #if PLATFORM(MAC). Just ENABLE(VIDEO) || ENABLE(WEB_AUDIO) should do. > Source/WebCore/page/Page.cpp:1473 > +#if PLATFORM(MAC) || ENABLE(VIDEO) || ENABLE(WEB_AUDIO) > if (wasVisibleAndActive != isVisibleAndActive()) > PlatformMediaSessionManager::updateNowPlayingInfoIfNecessary(); > +#endif Ditto. But maybe this conditional is not needed. Instead, perhaps PlatformMediaSessionManager::updateNowPlayingInfoIfNecessary should be defined even when ENABLE(VIDEO) || ENABLE(WEB_AUDIO) is false, in PlatformMediaSessionManager.cpp. Since the header is not conditional, it seems to me that the function body shouldn’t be either. It‘s not like we need to optimize this case to avoid the function call. > Source/WebCore/testing/Internals.cpp:2511 > - return emptyString(); > + return String(); If we want to return an empty string and not null, then it would be better to return: return String { emptyString() }; But I guess it’s OK your way.
Created attachment 294699 [details] patch for landing
Created attachment 294701 [details] patch for landing
Comment on attachment 294701 [details] patch for landing Clearing flags on attachment: 294701 Committed r208682: <http://trac.webkit.org/changeset/208682>