Behdad Esfahbod

I checked as far back as 4.0.0; the bug is there. I don't think it's a regression.

Okay. Looks like the font uses operators (like `FontBBox`) in FontDicts, whereas those only belong to TopDict... FontTools also detects, and ignores, that: ``` ``` I think we should simply...

This does it: ```diff diff --git a/src/hb-cff-interp-dict-common.hh b/src/hb-cff-interp-dict-common.hh index a08b10b5f..fc128e4da 100644 --- a/src/hb-cff-interp-dict-common.hh +++ b/src/hb-cff-interp-dict-common.hh @@ -71,6 +71,25 @@ struct dict_opset_t : opset_t env.argStack.push_real (parse_bcd (env.str_ref)); break; + case OpCode_isFixedPitch:...

Or we can change the catch-all to not err in general. I think that's a better way to handle broken fonts.

> Or we can change the catch-all to not err in general. I think that's a better way to handle broken fonts. Using a hammer: ```diff diff --git a/src/hb-cff-interp-common.hh b/src/hb-cff-interp-common.hh...

I'm curious how those fonts were built.

@matthiasclasen can you update the test results? @drott where's your live test page? I can test all against this branch to make sure.

> > @drott where's your live test page? I can test all against this branch to make sure. > > https://roettsch.es/var_colrv1.html I tested these and all look the same before/after...

> We're failing to render any of the `sweep_varsweep` tests though. I'll investigate separately. We render them fine. It's just that the image is 18k pixels wide and my tools...

@matthiasclasen If you update the expected outputs, this is good to go.