Simon Cozens

In fact, I don't need a way to trigger it, we *explicitly test* for the axis being added twice. https://github.com/googlefonts/axisregistry/blob/3db2ff673bb85c8ecee55f9e38ce601e8bc2e2f4/tests/data/OpenSans-Italic%5Bwdth%2Cwght%5D_STAT.ttx#L15-L24

Maybe an add_tag method on the Set?

Using paintcompiler on Bitcount: FontTools main: 2:19.07 (peak memory usage 4.8G) `cache-color-builds`: 54.015 (peak memory usage 236M) `cache-color-builds-2`: 2:41.66 (peak memory usage 5G) Probably better to do nothing than to...

Don't know. The TTX is different between the master build and my build, but visually it's the same. I'm measuring the memory footprint with "top", which I know isn't completely...

Will do, although the actual "error" case is trivial: ``` if ( !name.endsWith(".otf") && !name.endsWith(".ttf") && !name.endsWith(".ttc") && !name.endsWith(".otc") ) { setShaking(true); ```

Confirmed, this should skip discrete axes. (I think.)

OK, this is a bigger issue - the builder doesn't support DS5 designspaces with multiple variable fonts, because fontmake doesn't know which one we are trying to build and we...

I removed the old wheels because they were corrupted. I uploaded a new version on Friday with non-corrupt wheels.

This stuff (any kind of cubic-cubic intersection logic) is horribly complicated and full of floating-point imprecision and edge cases and I've basically hit the limit of my knowledge here. (Actually...

This is what I have been saying for the past few years to anyone who would listen: these kinds of problems are inherent in a "distributed" build workflow. Using a...