Yue Wang wrote:
>> Would it be acceptable to have
>>
>> fontforge.open("file", false)
>>
>> which makes to_table return four zeroes in the character boundingboxes?
>>
> no. they should be get from the glyf (xMin, yMin, xMax, yMax), hmtx,
> vmtx, vorg, etc. table directly.
CFF based fonts (Adobe Song) do not have the glyf (boundingbox)
information, which is why I have to do the expensive calculation to
begin with.
So, the question remains the same: would it be ok to have a correct
vwidth (cf width in horizontal fonts) and tsidebearing (analogous to
the already existing lsidebearing), but no boundingbox? In other
words, is the glyph boundingbox information relevant in the general
case?
The exact boundingbox is relevant for latin fonts because the glyph
depth and height are important for the line height calculations,
because the horizontal baseline lives somewhere in the middle of
the glyphs.
Best wishes,
Taco
