Support new AFDKO variable layout syntax #2228

simoncozens · 2021-03-18T20:36:05Z

Implements the variable scalar and conditional feature format proposed in this AFDKO discussion:

pos A B' <0 (wght=200:-100 wght=900:-150 wght=900,wdth=150:-120) 0 0> C' D;

davelab6 · 2021-03-20T21:05:17Z

https://twitter.com/simoncozens/status/1373175391133655041 has a visual demo :)

simoncozens · 2021-05-24T14:38:37Z

OK, something odd seems to have happen to main since I started writing this, because it passed tests until I rebased. What is happening is that when builder_test does:

        self.expect_ttx(font, self.getpath("%s.ttx" % name))
        # Check that:
        # 1) tables do compile (only G* tables as long as we have a mock font)
        # 2) dumping after save-reload yields the same TTX dump as before
        for tag in ('GDEF', 'GSUB', 'GPOS'):
            if tag in font:
                data = font[tag].compile(font)
                font[tag].decompile(data, font)
        self.expect_ttx(font, self.getpath("%s.ttx" % name))

the first TTX is correct, but after compile/decompile roundtripping, additional comments for FeatureVariationCount, ConditionCount and LookupCount have been added to the TTX:

--- /Users/simon/others-repos/fonttools/Tests/feaLib/data/variable_conditionset.ttx
+++ /var/folders/js/26vkgd8s5x15c5kwbhq2c2jw0000gn/T/tmpp1m8cmke/tmp1.ttx
@@ -39,10 +39,8 @@
     </LookupList>
     <FeatureVariations>
       <Version value="0x00010000"/>
-      <!-- FeatureVariationCount=1 -->
       <FeatureVariationRecord index="0">
         <ConditionSet>
-          <!-- ConditionCount=1 -->
           <ConditionTable index="0" Format="1">
             <AxisIndex value="0"/>
             <FilterRangeMinValue value="0.625"/>
@@ -54,7 +52,6 @@
           <SubstitutionRecord index="0">
             <FeatureIndex value="0"/>
             <Feature>
-              <!-- LookupCount=1 -->
               <LookupListIndex index="0" value="0"/>
             </Feature>
           </SubstitutionRecord>

Obviously I can't keep both tests happy. I'm not sure what to do about this.

anthrotype · 2021-05-24T14:56:17Z

the meaning of that tests is, if you're building those FeatureVariations and Feature records, you need to make sure the *Count fields are set at build time, and not rely on the compiler which will (re)set those for you; that way dumping to TTX straight after constructing the table objects will give the same output as building followed by compile.
The expected variable_conditionset.ttx should have those count fields set, but from that diff it looks like it has not.

anthrotype · 2021-05-24T14:57:17Z

it passed tests until I rebased

I'm not exactly sure what would have changed in the meantime

simoncozens · 2021-05-24T15:21:54Z

Got it, thank you.

ctrlcctrlv · 2021-10-22T02:36:40Z

Please rebase again when you get a chance @simoncozens. I need a commit later in history.

…ase"

Previously otlLib was generating None if the values themselves were empty even if the value format was non-empty. This happened to work for compiling to binary since the compiler handles Value=None. But this was confusing varLib.merger module (as in when building variable fonts from such otlLib-built master GSUB/GPOS tables, without roundtripping to OTF/TTF binary first), because in varLib.merger, a None means "this master doesn't provide that info; skip it"; whereas in a PairPos table a None as generated by otlLib simply meant "all values are zero", which is different from "this master doesn't provide this value". This fixes that, such that ufo2ft can build variable-font without saving masters to binary. Part of googlefonts/ufo2ft#486

…ubst The format values for those are automatically handled in postRead/preWrite to choose optimal format. As such, don't write them in XML. Reduces noise. fonttools#2236 (comment)

This reverts commit fb0c60c.

…hods Fixes fonttools#2319

See adobe-type-tools/afdko#153 (comment)

… to do with them

…okup ID) See fonttools#2316

simoncozens · 2021-10-22T07:46:47Z

Ack, messed this up. Will reopen.

simoncozens · 2021-10-22T07:49:28Z

-> #2432

simoncozens mentioned this pull request May 20, 2021

varLib.featureVars needs rethinking #2316

Open

simoncozens changed the title ~~Support variable scalars in feature file~~ Support new AFDKO variable layout syntax May 24, 2021

simoncozens mentioned this pull request May 24, 2021

[spec] Variable FEA Syntax adobe-type-tools/afdko#1350

Open

2 tasks

simoncozens marked this pull request as ready for review May 24, 2021 14:24

simoncozens force-pushed the fealib-variable-scalar branch from db7bbdb to 7a69fb1 Compare May 24, 2021 14:34

simoncozens and others added 19 commits October 22, 2021 08:35

Shorten exception names

da39a79

super doesn’t magically call its own methods

6227a73

Improve error message

207083a

More tests

b3266ca

Use kwargs and explicit stack

8ba64df

Make CFF merge errors inherit differently

1ca6ee8

Remove pass if we have a docstring

914b664

[py23] change deprecation warning from "next release" to "future rele…

9e0cf4c

…ase"

[subset] Call VariationStore.subset_varidxes() as method

97fb234

Update test data

f348db3

Don't write Format for Coverage, ClassDef, Single/Alternate/LigatureS…

edc4bba

…ubst The format values for those are automatically handled in postRead/preWrite to choose optimal format. As such, don't write them in XML. Reduces noise. fonttools#2236 (comment)

Revert "[varLib.merge] Fix tests"

c825def

This reverts commit fb0c60c.

[varLib.merger] Adjust for removal of Coverage / ClassDef .Format

cd4dc79

[Tests] Fix most tests expectations for Coverage/ClassDef .Format change

af81665

[Tests] Fix more tests expectations for Coverage/ClassDef .Format change

f8fdfca

[Tests] Fix most tests expectations for SingleSubst .Format change

f1f7816

[Tests] Fix most tests expectations for LigatureSubst .Format change

63e8dd4

[Tests] Fix most test expectations for AlternateSubst .Format removal

6ebbdf7

anthrotype and others added 25 commits October 22, 2021 08:35

mention release --minor, --major and --sign options

789a11b

otTables: ignore Format attribute in split.*Subst overflow-fixing met…

490e9a8

…hods Fixes fonttools#2319

otTables_test: mixed tabs and spaces

870b767

Update changelog [skip ci]

4f96b4b

Release 4.24.3

01c87a2

Bump version: 4.24.3 → 4.24.4.dev0

5b2fb0b

Parse variable scalars in feature files.

ec816ed

See adobe-type-tools/afdko#153 (comment)

Build variable scalars for anchors/value records

2e86b2c

Variable scalar tests

2e55709

Support variable scalars in pos A V (...);

1216eb2

Format bug

16afa35

Useful property

b38bf4a

Add support for the conditionset statement

9a3fbac

Parse variation blocks

4165bb2

Round-trip FEA correctly

259204e

Store the conditional features in a dictionary until we work out what…

301c17d

… to do with them

Rearrange featureVars so we can do *really* raw feature builds (by lo…

f99b6fb

…okup ID) See fonttools#2316

Oops, leftover when testing something else

1f758f2

A more generic interface, taking either GSUB or GPOS table

f7589ca

Normalize condition sets when storing them

635c29f

feature_variations_ is a better name than conditionalFeatures_

dcf985d

Support building feature variations using the "condition" statement.

c58ddae

Put count fields in ttx

3c19ae4

Set count fields when building feature variations

778a7ca

Allow float values in condition sets.

4639015

simoncozens force-pushed the fealib-variable-scalar branch from b880ae1 to 4639015 Compare October 22, 2021 07:35

simoncozens closed this Oct 22, 2021

simoncozens deleted the fealib-variable-scalar branch October 22, 2021 07:49

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Support new AFDKO variable layout syntax #2228

Support new AFDKO variable layout syntax #2228

simoncozens commented Mar 18, 2021 •

edited

davelab6 commented Mar 20, 2021

simoncozens commented May 24, 2021

anthrotype commented May 24, 2021

anthrotype commented May 24, 2021

simoncozens commented May 24, 2021

ctrlcctrlv commented Oct 22, 2021

simoncozens commented Oct 22, 2021

simoncozens commented Oct 22, 2021

Support new AFDKO variable layout syntax #2228

Support new AFDKO variable layout syntax #2228

Conversation

simoncozens commented Mar 18, 2021 • edited

davelab6 commented Mar 20, 2021

simoncozens commented May 24, 2021

anthrotype commented May 24, 2021

anthrotype commented May 24, 2021

simoncozens commented May 24, 2021

ctrlcctrlv commented Oct 22, 2021

simoncozens commented Oct 22, 2021

simoncozens commented Oct 22, 2021

simoncozens commented Mar 18, 2021 •

edited