Regression(252852@main) Games on kongregate.com are no longer loading

[cdumez]

b57dc1f

Regression(252852@main) Games on kongregate.com are no longer loading
https://bugs.webkit.org/show_bug.cgi?id=252636
rdar://104392542

Reviewed by Ryosuke Niwa.

Do not lazy load iframes that have no valid URL or "about:blank" as URL. Lazily
loading such iframes has no performance benefits and can actually cause breakage
as JS (like on kongregate.com) may expect them to load synchronously.

Note that we have a more general problem where we fail to create the Frame &
initial empty document when we decide to lazy load an iframe. This is not correct,
and we should only delay the navigation to the frame URL *after* the creation of
the initial empty document. However, fixing this is a larger change and this is
not a regression from 252852@main. As a result, I am merely adding a FIXME comment
in this patch and will address separately.

* LayoutTests/fast/frames/about-blank-frame-no-lazy-loading-expected.txt: Added.
* LayoutTests/fast/frames/about-blank-frame-no-lazy-loading.html: Added.
Add layout test coverage.

* Source/WebCore/html/HTMLFrameElementBase.cpp:
(WebCore::HTMLFrameElementBase::openURL):
* Source/WebCore/html/HTMLIFrameElement.cpp:
(WebCore::isFrameLazyLoadable):
(WebCore::HTMLIFrameElement::shouldLoadFrameLazily):

Canonical link: https://commits.webkit.org/260612@main

9b506f4

Misc	iOS, tvOS & watchOS	macOS	Linux	Windows
✅ 🧪 style	✅ 🛠 ios	✅ 🛠 mac	❌ 🛠 wpe	✅ 🛠 wincairo
✅ 🧪 bindings	✅ 🛠 ios-sim	✅ 🛠 mac-AS-debug	✅ 🛠 gtk
✅ 🧪 webkitperl	✅ 🧪 ios-wk2	✅ 🧪 api-mac	✅ 🧪 gtk-wk2
	✅ 🧪 api-ios	✅ 🧪 mac-wk1	✅ 🧪 api-gtk
	✅ 🛠 tv	✅ 🧪 mac-wk2
	✅ 🛠 tv-sim	✅ 🧪 mac-AS-debug-wk2
	✅ 🛠 watch	✅ 🧪 mac-wk2-stress
✅ 🛠 🧪 merge	✅ 🛠 watch-sim

[webkit-early-warning-system]

EWS run on current version of this PR (hash 9b506f4)

Details

Misc	iOS, tvOS & watchOS	macOS	Linux	Windows
✅ 🧪 style	✅ 🛠 ios	✅ 🛠 mac	❌ 🛠 wpe	✅ 🛠 wincairo
✅ 🧪 bindings	✅ 🛠 ios-sim	✅ 🛠 mac-AS-debug	✅ 🛠 gtk
✅ 🧪 webkitperl	✅ 🧪 ios-wk2	✅ 🧪 api-mac	✅ 🧪 gtk-wk2
	✅ 🧪 api-ios	✅ 🧪 mac-wk1	✅ 🧪 api-gtk
	✅ 🛠 tv	✅ 🧪 mac-wk2
	✅ 🛠 tv-sim	✅ 🧪 mac-AS-debug-wk2
	✅ 🛠 watch	✅ 🧪 mac-wk2-stress
✅ 🛠 🧪 merge	✅ 🛠 watch-sim

[rniwa]

Should we also file a spec bug?

[cdumez]

Should we also file a spec bug?

I don’t think a spec bug is needed. I think that if we fixed our real bug (the fact that we delay the construction of the initial empty document) then we wouldn’t have broken the site.

i will look into fixing this in a follow up. However, I went with the simpler partial revert here since this was a recent regression from my change and this will go to a branch.

[webkit-commit-queue]

Committed 260612@main (b57dc1f): https://commits.webkit.org/260612@main

Reviewed commits have been landed. Closing PR #10414 and removing active labels.