Better failure mode for sstruct

[simoncozens]

Currently if you give a field value - any field value - a number it isn't expecting for whatever reason (float when uint is expected, number out of range, etc.), you get this:

font["OS/2"].usWinDescent = -600
font.save("foo.ttf")

...

  File "/opt/homebrew/lib/python3.11/site-packages/fontTools/ttLib/tables/O_S_2f_2.py", line 173, in compile
    data = sstruct.pack(OS2_format_2, self)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/opt/homebrew/lib/python3.11/site-packages/fontTools/misc/sstruct.py", line 75, in pack
    data = struct.pack(*(formatstring,) + tuple(elements))
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
struct.error: argument out of range

Which at least leads you to the right table, but nothing more than that. With this PR, you get:

    data = sstruct.pack(OS2_format_2, self)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/simon/others-repos/fonttools/Lib/fontTools/misc/sstruct.py", line 82, in pack
    raise ValueError(
ValueError: Value -600 does not fit in format H for usWinDescent

[anthrotype]

i like this change, I'm only a bit wary if external code depends on sstruct.getformat and may break now that names is a dict and no longer a list..

[simoncozens]

The only code I've found on GitHub that isn't a vendoring of fontTools is Inter's "fontinfo.py". Amusingly the reason it is reaching for those names is to turn them into a dict...

Also we probably won't be able to write GPKG tables correctly any more, but I reckon we should just leave that and see if anyone notices. :-)

[anthrotype]

What even is a GPKG table?!

[simoncozens]

SING Glyphlet stuff. Adobe idea for embeddable single-glyph fonts, abandoned in 2009.