[LFC][IFC] InlineFormattingState::addDetachingRule should accumulate rules.

[zalan]

instead of overwriting the old ones.

[zalan]

Created attachment 355180 [details]
Patch

[EWS Watchlist]

Comment on attachment 355180 [details]
Patch

Attachment 355180 [details] did not pass win-ews (win):
Output: https://webkit-queues.webkit.org/results/10031591

New failing tests:
fast/inline/simple-inline-with-out-of-flow-descendant2.html

[EWS Watchlist]

Created attachment 355190 [details]
Archive of layout-test-results from ews203 for win-future

The attached test failures were seen while running run-webkit-tests on the win-ews.
Bot: ews203  Port: win-future  Platform: CYGWIN_NT-6.1-2.9.0-0.318-5-3-x86_64-64bit

[Antti Koivisto]

Comment on attachment 355180 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=355180&action=review

> Source/WebCore/layout/inlineformatting/InlineFormattingState.cpp:58
> +    auto result = m_detachingRules.add(&layoutBox, detachingRule);
> +    if (result.isNewEntry)
> +        return;
> +    result.iterator->value.add(detachingRule);

m_detachingRules.add(&layoutBox, { }).iterator->value.add(detachingRule);

is probably equivalent and bit more compact?

> Source/WebCore/layout/inlineformatting/InlineFormattingState.h:77
> +    void setDetachingRules(const Box& layoutBox, DetachingRules detachingRules) { m_detachingRules.set(&layoutBox, detachingRules); }

Not from this patch but most OptionSet sites don't have type alias but simply write out the whole thing. 'OptionSet<DetachingRule>' is a less mysterious type than 'DetachingRules'.

[zalan]

Committed r238349: <https://trac.webkit.org/changeset/238349>

[Radar WebKit Bug Importer]

<rdar://problem/46150548>