[Webkit-unassigned] [Bug 148965] Split mixed font GlyphPage functionality to separate class
bugzilla-daemon at webkit.org
bugzilla-daemon at webkit.org
Tue Sep 8 14:47:36 PDT 2015
https://bugs.webkit.org/show_bug.cgi?id=148965
Myles C. Maxfield <mmaxfield at apple.com> changed:
What |Removed |Added
----------------------------------------------------------------------------
Attachment #260783|review? |review+
Flags| |
--- Comment #3 from Myles C. Maxfield <mmaxfield at apple.com> ---
Comment on attachment 260783
--> https://bugs.webkit.org/attachment.cgi?id=260783
patch
View in context: https://bugs.webkit.org/attachment.cgi?id=260783&action=review
I think this is fine.
> Source/WebCore/platform/graphics/FontCascadeFonts.cpp:56
> +void FontCascadeFonts::GlyphPageCacheEntry::setSingleFontPage(RefPtr<GlyphPage>&& page)
Shouldn't this take a Ref?
> Source/WebCore/platform/graphics/FontCascadeFonts.h:98
> + RefPtr<GlyphPage> m_singleFont;
> + std::unique_ptr<MixedFontGlyphPage> m_mixedFont;
Unclear why one is a RefPtr and the other one is a unique_ptr.
> Source/WebCore/platform/graphics/GlyphPage.h:74
> static unsigned indexForCharacter(UChar32 c) { return c % GlyphPage::size; }
Shouldn't this be private (maybe with a friend of MixedFontGlyphPage?) Seems to me that external people don't need to know this.
> Source/WebCore/platform/graphics/GlyphPage.h:83
> + return m_glyphs[indexForCharacter(c)];
Please add ASSERT_WITH_SECURITY_IMPLICATION() in all member functions which deal with this stuff
> Source/WebCore/platform/graphics/GlyphPage.h:94
> + void setGlyphDataForIndex(unsigned index, Glyph glyph)
Looks to me like this function is just setting the glyph, not the GlyphData. Now that you are separating mixed parts from the non-mixed parts, I think it makes sense to give them different APIs.
> Source/WebCore/platform/graphics/GlyphPage.h:121
> +class MixedFontGlyphPage {
Seems to me that this should get its own file.
> Source/WebCore/platform/graphics/GlyphPage.h:124
> + MixedFontGlyphPage(RefPtr<GlyphPage>&& initialPage)
Judging from the function body, shouldn't this take a regular lvalue reference instead of an rvalue reference?
--
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.webkit.org/pipermail/webkit-unassigned/attachments/20150908/f7a99901/attachment.html>
More information about the webkit-unassigned
mailing list