-
Notifications
You must be signed in to change notification settings - Fork 26
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[skrifa] autohint bug fixes to match FT #1145
Conversation
This is a largish collection of small changes that bring the autohinter into a state where we match with FreeType for at least the Google Fonts and Noto collections. Also updates fauntlet to use the new hinting API.
cc @drott |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I can see the hours that went into finding those differences, thanks for bringing this in alignment with FT. LGTM.
@@ -13,6 +15,20 @@ pub fn compare_glyphs( | |||
// Don't run on bitmap fonts (yet) | |||
return true; | |||
} | |||
if let Some(Hinting::Auto(_)) = options.hinting { | |||
// This font is a pathological case for autohinting due to the |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Perhaps add an issue reference?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Will do. I want to get some actual metrics from this font so we can potentially add limits to prevent this performance cliff. Beyond a certain level of geometric complexity, hinting is likely to produce useless results anyway.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Yes, similarly we might want to disable it for COLRv1 fonts, but I can do that on the client side.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Sgtm.
Filed #1147 and added a link in the comment
// Note: this is actually critical for matching FreeType in cases where | ||
// we have more than one edge with the same fpos. When this happens, | ||
// linear and binary searches can produce different results. | ||
// See <https://gitlab.freedesktop.org/freetype/freetype/-/blob/57617782464411201ce7bbc93b086c1b4d7d84a5/src/autofit/afhints.c#L1489> |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
🤯
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Took me a while to find this one :)
if group != ScriptGroup::Default { | ||
// Divergent CJK behavior | ||
// See <https://gitlab.freedesktop.org/freetype/freetype/-/blob/57617782464411201ce7bbc93b086c1b4d7d84a5/src/autofit/afcjk.c#L1544> | ||
if dist < 54 { |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I am aware these are coming from the source, but perhaps in a separate, follow-up CL, any chance we could attempt to give this bouquet of magic numbers some names, or add some comments?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I’ll do the best I can. Updated #1141 to include this.
This is a largish collection of small changes that bring the autohinter into a state where we match with FreeType for at least the Google Fonts and Noto collections.
Also updates fauntlet to use the new hinting API.
Builds on #1144 which should land first.