[CLOSED] Win Descent

Post your bug reports here. Include information that helps us to understand and reproduce the bug.
Post Reply
MRoberts
Posts: 5
Joined: Wed Jun 30, 2021 5:13 pm

[CLOSED] Win Descent

Post by MRoberts »

For all 5 versions of the OS/2 table, the usWinDescent is an unsigned 16 bit (uint16), so it will always be a positive number. But FontCreator is displaying the Win Descent as a negative number. The descender value in the hhea table is a signed 16 bit (FWORD), so it can be negative, and FontCreator is displaying it correctly.

The OpenType specification appendix titled, "Recommendations for OpenType Fonts" states:
internal leading = usWinAscent + usWinDescent - unitsPerEm
If the usWinDescent were negative, then this expression would be incorrect.

I found that the Win Descent has shown as a negative number going back to at least FontCreator v6. But I suspect this is just a display issue. A positive number is probably being stored in the OS/2 table; otherwise, fonts would not be rendering correctly. Just out of curiosity, I checked another font editor and it showed a positive number for the Win Descent and a negative number for the Descender. Both had the same magnitude.
Erwin Denissen
Moderator
Moderator
Posts: 11108
Joined: Fri Oct 04, 2002 12:41 am
Location: Bilthoven, The Netherlands
Contact:

Re: Win Descent

Post by Erwin Denissen »

It is not a bug, but a feature :D

FontCreator indeed shows WinDescent as -1*usWinDescent. The other descender values also show negative values, so this is more in line with them. People who do not read the specs prefer the current presentation.
Erwin Denissen
High-Logic
Proven Font Technology
MRoberts
Posts: 5
Joined: Wed Jun 30, 2021 5:13 pm

Re: [CLOSED] Win Descent

Post by MRoberts »

I thought that was probably the case. You may want to consider including it as a note in the user guide.
Post Reply