Alban Browaeys authored Oct 16, 2017 and Akira TAGOH committed Oct 17, 2017

Remove leftover references to run-test271.sh.

Alexander Kanavin authored Oct 11, 2017 and Akira TAGOH committed Oct 12, 2017

It's also freed after bail1, so no need to do it here.

Bastien Nocera authored Oct 05, 2017 and Akira TAGOH committed Oct 05, 2017

Prefer the system provided emoji fonts on systems which provide one,
such as Windows, MacOS and Android, even if the Emoji One or Emoji Two
fonts are installed.

This also allows free software OSes such as GNOME to prefer the Emoji
One font, which is not used in other OSes and therefore has a unique
brand identity, by installing them and only them by default.

Users can use more capable fonts while Emoji One and Emoji Two catch up
by installing a font otherwise already used by another system, such as
Google's freely redistributable Noto Emoji font.

https://bugzilla.redhat.com/show_bug.cgi?id=1496761

2.8.0 had a bad bug with loading 'avar' table.  Let's update requirement and cleanup
fifteen years of ifdefs!

https://lists.freedesktop.org/archives/fontconfig/2017-September/006048.html

Makes VotoSerifGX scanning another 40% faster...  Down to 36ms now.

VotoSerifGX has over 500 named instances, which means it also has over a thousand
name table entries.  So we were looking for names for over 500 pattern, looking for
some thirty different name-ids, and using linear search across the 1000 entries!

Makes scanning VotoSerifGX three times faster.  The rest is probably the lang
matching, which can also be shared across named-instances.  Upcoming.

There's no "all other platforms", there was just ISO left.
Hardcode it in.

If it's good, it's good for everything!

This didn't give me the speedup I was hoping for, though I do get around 15% for VotoSerifGX.

For variable-font named-instances we want to reuse the same charset and redo the spacing.

https://bugs.freedesktop.org/show_bug.cgi?id=102141

This reverts commit dc56ff80.

Makes scanning of Voto (over 500 named instaces) twice faster.

Next, avoid charset / lang recalculation for each of those.

This reverts commit 57764e3a36449da25bb829c34cb08c54e9e5de90.

For regular font pattern we don't look into fvar, so it doesn't make sense to
get non-variation from it either.

Meh.

This can be used for FC_VARIABLE=FcDontCare for example, to opt into getting
variable fonts for clients that support using them.

Oops.

Old logic was really bad.  If you requested weight=102 and got a medium
font (weight=100), it would still enable emboldening...

Adjust it to only embolden if request was >= bold and font was <= regular.

Has two distinctions from FcCompareRange():
1. As best value, it returns query pattern size, even if it's out of font range,
2. Implements semi-closed interval, as that's what OS/2 v5 table defines

Blanks are the new black, err, dead!

If font claims to support range [100,900], and request is for [250], then
return [250] in "rendered" pattern.  Previously was returning [100,900].

This is desirable for varfonts weight and width, but probably not for size.
Will roll back size to return request size always, for non-empty ranges.