Respect 'Propagate Anchors' custom parameter

[felipesanches]

I've tested this by running ttx_diff on notofonts/arabic/sources/NotoSansArabic.glyphspackage and got these improvements:

COMPARISON	Before	After
(mark / kern)	29.249%	Identical
GDEF	56.062%	Identical
GPOS	87.500%	Identical
ligcaret	34.043%	Identical

(issue #889)

[anthrotype]

thanks. Could you also add a test?

you could create a minimal test font in Glyphs.app with two glyphs, a simple one with an anchor, and a composite glyph using the latter as component. You also add the Propagate Anchors custom param set to false. Pass it through the resources/scripts/tidy_glyphs.py python script before committing to get rid of unwated stuff (timestamps etc.).

then in your test you would simply Font::load the test file and assert that the composite glyph gets no propagated anchors

[felipesanches]

@anthrotype, I have attempted to create the test, but I am still not getting good results. Can you help me understand what I'm doing wrong here?

[felipesanches]

I'm getting this panic:

thread 'propagate_anchors::tests::dont_propagate_anchors' panicked at glyphs-reader/src/propagate_anchors.rs:1038:9:
assertion failed: glyph.layers.first().unwrap().anchors.is_empty()

My reasoning here is that:

• I know that the glyph "Aacute" exists
• I know that it has a single layer, so .layers.first().unwrap() should be fine
• I expect it not to have anchors and since .anchors is a Vec<Anchor>, then it should be fine to use .is_empty()

But I must have missed something, because it panics... What do you think?

[cmyr]

the panic is just saying that the test failed, and it failed because apparently anchors isn't empty, so something unexpected is going on..

[cmyr]

okay the issue seems to be that the "Propagate Anchors" custom param is set on the master, not on the font?

[felipesanches]

okay the issue seems to be that the "Propagate Anchors" custom param is set on the master, not on the font?

Yep! Thanks! Just moved it and the check now passes.

[cmyr]

looks good, thanks!