learning .fea syntax from decompiled FontForge structures is tricky
OTMaster is another program I wouldn't recommend for learning how to encode features. — Its decompiled source maintains maximal faithfulness to the underlying structure of the font at the expense of readability. There are situations where this is highly desirable, but learning coding isn't one of them.
FF, which I find buggy (and sometimes it even crashes).
True. However, it'a free software with only a few developers and economic resources. So, it likes me in any case a meritorious undertaking.
I myself identified and reported in github a bug related to the automatic creation of aalt, which causes FF freezes.
On December 26th I received the following statement from a developer in FF users.list (so Linux users can hope in a better release):
There was a long-blocking list of issues that finally has fixes in
review. We expect a release before the end of the year. Please note any
other low-hanging fruit (open crashes with backtraces) here so that I
can pack them in.
FF is a precious gem of the open source type scene, it is like an unpolished diamond, its structure so brittle that polishing might break it I'll have a look on the changelist when the new release rolls in, that might be interesting.
Thanks to your generous help I finally almost succeeded. I created the glyph <quoterighr.fr> with less bearing, a Single Substitution Lookup and a Lookup calt Chaining Substitution with latn{FRA ,ITA }. It likes so (in the form FontForge exports .fes files):
1. With Libreoffice I obtain the desired results; if I change the languages, quoteright.fr is not used in English and it's in Italian language. With Xetex settings are ignored, even if I put in my .tex file <RawFeature=+calt>; infact there is no calt if I ask so:
<div>samiel@darkstar:~/work$ otfinfo -f SamielPro.otf
</div><div>aalt Access All Alternates</div><div>
c2sc Small Capitals From Capitals</div><div>
cpsp Capital Spacing</div><div>
dnom Denominators</div><div>
frac Fractions</div><div>
kern Kerning</div><div>
liga Standard Ligatures</div><div>
lnum Lining Figures</div><div>
numr Numerators</div><div>
onum Oldstyle Figures</div><div>
pnum Proportional Figures</div><div>
smcp Small Capitals</div><div>
sups Superscript</div><div>
tnum Tabular Figures</div><div>
zero Slashed Zero</div>
All works with Xetex if and only if I set calt languages to general default, i.e. <DFLT{dflt} grek{dflt} latn{dflt}>
2. When quoteright is not an apostrophe and I don't need narrower spacing? Have I to set calt in a different way, inserting a third condition, I mean letter + quoteright + letter?
Comments
http://typedrawers.com/discussion/2988/numbers-glyphs-in-aalt#latest
OTMaster is another program I wouldn't recommend for learning how to encode features. — Its decompiled source maintains maximal faithfulness to the underlying structure of the font at the expense of readability. There are situations where this is highly desirable, but learning coding isn't one of them.
I myself identified and reported in github a bug related to the automatic creation of aalt, which causes FF freezes.
On December 26th I received the following statement from a developer in FF users.list (so Linux users can hope in a better release):
I created the glyph <quoterighr.fr> with less bearing, a Single Substitution Lookup and a Lookup calt Chaining Substitution with latn{FRA ,ITA }.
It likes so (in the form FontForge exports .fes files):
<div>lookup caltContextualAlternateslookup15 { </div><div> lookupflag 0;</div><div> @cc0_match_1 = [\A \B \C \D \E \F \G \H \I \J \K \L \M \N \O \P \Q \R \S \T \U \V \W \X \Y \Z \a \b \c \d \e \f \g \h \i \j </div><div> \k \l \m \n \o \p \q \r \s \t \u \v \w \x \y ];</div><div> @cc0_match_2 = [\quoteright ];</div><div> @cc0_back_1 = [\A \B \C \D \E \F \G \H \I \J \K \L \M \N \O \P \Q \R \S \T \U \V \W \X \Y \Z \a \b \c \d \e \f \g \h \i \j </div><div> \k \l \m \n \o \p \q \r \s \t \u \v \w \x \y ];</div><div> @cc0_back_2 = [\quoteright ];</div><div> @cc0_ahead_1 = [\A \B \C \D \E \F \G \H \I \J \K \L \M \N \O \P \Q \R \S \T \U \V \W \X \Y \Z \a \b \c \d \e \f \g \h \i \j </div><div> \k \l \m \n \o \p \q \r \s \t \u \v \w \x \y ];</div><div> @cc0_ahead_2 = [\quoteright ];</div><div> sub @cc0_back_1 @cc0_match_2' by [\quoteright.fr ] ;</div><div> } caltContextualAlternateslookup15;</div>In the specific case, I created it by Classes and not by Coverage. I acted well, or is this last solution more appropriate?
However: Problems
1.
With Libreoffice I obtain the desired results; if I change the languages, quoteright.fr is not used in English and it's in Italian language. With Xetex settings are ignored, even if I put in my .tex file <RawFeature=+calt>; infact there is no calt if I ask so:
All works with Xetex if and only if I set calt languages to general default, i.e. <DFLT{dflt} grek{dflt} latn{dflt}>
2.
When quoteright is not an apostrophe and I don't need narrower spacing? Have I to set calt in a different way, inserting a third condition, I mean letter + quoteright + letter?
https://tex.stackexchange.com/questions/467943/xetex-support-for-otf-features .
So remain two questions:
better calt by classes o by coverage?
and the above question no. 2
thanx