See https://cmex-30.github.io/Bopomofo_on_Web/testpage/index.html In the last box, these two characters are not placed correctly: - ˫ U+02EB MODIFIER LETTER YANG DEPARTING TONE MARK - ˪ U+02EA MODIFIER LETTER YIN DEPARTING TONE MARK Compare the positioning to Chrome.
<rdar://problem/87625596>
Created attachment 449233 [details] Reduction
(The attached reduction needs the custom font from https://cmex-30.github.io/Bopomofo_on_Web/testpage/index.html)
For some reason we're getting two different Font objects in FontCascade::drawGlyphBuffer() which point to the same CTFont
m_font.glyphDataForCharacter() is reporting different Font objects for those characters. They point to the same CTFont though.
I see what's happening. The FontCascade::isCJKIdeographOrSymbol() check in FontCascadeFonts::glyphDataForVariant() is returning true for the modifiers but false for the regular characters.
Created attachment 449287 [details] Patch
Comment on attachment 449287 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=449287&action=review > Source/WebCore/ChangeLog:16 > + No new tests because this requires a custom font to test. We need to do something about this; maybe make some more custom fonts for the WebKit test suite. That situation is high risk for the project, I think. I’d even like to see this kind of thing tested in WPT.
Created attachment 450730 [details] Needs attribution for font creator
Created attachment 450845 [details] Patch for committing
Committed r289101 (246798@trunk): <https://commits.webkit.org/246798@trunk>