[VARC] Variable Composites table

[behdad]

Based on harfbuzz/boring-expansion-spec#103 (comment)

[behdad]

This is mostly steady now. I'll go ahead and write a spec for it.

What is left to do is:

• Rip out the glyf1 VarComposite glyphs,
• Hook VARC up in ttGlyphSet so we can draw.

[behdad]

I should also somehow figure out how to do the VARC glyph loading lazy.

[behdad]

TODO:

• Make VARC table lazy-load glyphs.
• Subsetter
• Instancer
• scaleUpem
• tests

[behdad]

I can use some directions here. The multiVarStore and varStore share some similar code. Should I try to subclass them from a common class?

It will be more pronounced when adding subsetting to multiVarStore.

[behdad]

I can use some directions here. The multiVarStore and varStore share some similar code. Should I try to subclass them from a common class?

It will be more pronounced when adding subsetting to multiVarStore.

I managed to share some code by monkey-patching for now.

[behdad]

This is ready for review. Thanks in advance @anthrotype @justvanrossum

[behdad]

This is ready for review. Thanks in advance @anthrotype @justvanrossum

It's terse but here's something of a spec for this:
https://github.com/harfbuzz/boring-expansion-spec/blob/main/varc.md

[behdad]

I probably should move some of the new code from otTables to a side module.

[behdad]

HarfBuzz counterpart WIP: harfbuzz/harfbuzz#4578

Now can render fonts generated with this fonttools branch.

[behdad]

This is ready. Has tests too.

[anthrotype]

de/compiling TupleVariation with this new DELTAS_ARE_LONGS flag doesn't seem to be covered by tests

[behdad]

You caught me!

I'll see if I can make one.

[anthrotype]

You caught me!

codecov did! :)

[justvanrossum]

Suggest to apply this patch to make a var composite referencing its own GID work correctly with the pen protocol:

diff --git a/Lib/fontTools/ttLib/ttGlyphSet.py b/Lib/fontTools/ttLib/ttGlyphSet.py
index 8b276aa37..5d373cfb9 100644
--- a/Lib/fontTools/ttLib/ttGlyphSet.py
+++ b/Lib/fontTools/ttLib/ttGlyphSet.py
@@ -344,11 +344,17 @@ class _TTGlyphVARC(_TTGlyph):
             reset = comp.flags & VarComponentFlags.RESET_UNSPECIFIED_AXES
             with self.glyphSet.glyphSet.pushLocation(location, reset):
                 with self.glyphSet.pushLocation(location, reset):
-                    try:
-                        pen.addVarComponent(
-                            comp.glyphName, transform, self.glyphSet.rawLocation
-                        )
-                    except AttributeError:
+                    shouldDecompose = self.name == comp.glyphName
+
+                    if not shouldDecompose:
+                        try:
+                            pen.addVarComponent(
+                                comp.glyphName, transform, self.glyphSet.rawLocation
+                            )
+                        except AttributeError:
+                            shouldDecompose = True
+
+                    if shouldDecompose:
                         t = transform.toTransform()
                         compGlyphSet = (
                             self.glyphSet

Without this, it will emit a self-referencing component which (as far as I can see) can't be resolved in another way.

@behdad, I can't find a reference to the prescribed behavior in the proposal doc.

Others: what was new to me is that a VARC Composite may reference its own GID, which will then bypass VARC and go straight to glyf or CFF2. This is a handy way to mix outlines and (variable) components.

[behdad]

@justvanrossum Thanks. Can you push that change to the branch?

@behdad, I can't find a reference to the prescribed behavior in the proposal doc.

Here:

The component then is recursively loaded from the VARC table, if it is present in this table; otherwise, falls back to glyf/gvar, CFF2, or any other mechanism to get raw outlines. An exception to the recursive rule is that if a component refers to the glyphID of the current glyph, instead of recursing (which would result in infinite recursion), the glyph outline for the component is loaded directly instead, as if the glyphID had no VARC entry.

[justvanrossum]

Can you push that change to the branch?

Will do, thanks!