[subset] zero glyf and loca memory before writing so pads are clean. Test just the part of maxp we care about

[subset] sketch out dropping in-glyf instructions. Sometimes yields differnet size glyphs than fonttools, possibly due to padding not being corrected

Fixes https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=6022

New API:
- hb_set_previous()
- hb_set_previous_range()

To signify that unlike rest of library, returns nullptr on failure.

[subset] Remove test-subset, testing is planned to be done at the table level here with test/subset covering the complete subsetting operation.

I broke this in c31fcf4c