Re: [ft] Incorrect y position glyph render

[Edward Lam]

Hi Werner,

Thanks for the reply! I was using left to right. I'm new to fonts so your 
response lead me resolving my issue. It was "wrong" because I was erroneously 
comparing between different versions of the font, where the '*' glyph had 
changed vertical positioning.

This was after trying out harfbuzz to no avail, then dumping out to ttx, trying 
to make sense of the GPOS/GSUB tables, learning about the Adobe Glyph List so 
that I then realized to search for "asterisk", and then finally stumbling into 
the logged github issue where it was decided to change the vertical positioning 
of the glyph. :(

-Edward

On 21/10/2016 2:05 AM, Werner LEMBERG wrote:
> What do you mean with `vertically'?  Is your typesetting direction
> from top do bottom?  Or left to right?
>
> > The sample text I'm rendering is '1+4*8' and the '*' character in
> > ftstring isn't being rendered higher than the '+' like it should
> > (eg. on Windows using WordPad).  This results in it being hard to
> > tell the difference between the plus and multiply operators.
>
> For left-to-right typesetting, FreeType simply takes the metrics
> directly from the font.  However, it doesn't apply higher-level
> OpenType tables like `GSUB' and `GPOS'.  Right now I'm too lazy to
> check, but it is possible that the font in question either contains
> GSUB data to replace the glyphs with special variants or GPOS data to
> move glyphs vertically if `+' or `*' is positioned between digits.
>
> Either have a look into the font (ttx, your preferred font editor,
> ...) or try the demo programs of HarfBuzz to check whether you get
> better results.
>
> [BTW, you should also check whether WordPad really uses the glyphs of
>  the font for mathematical stuff.]
>
> For top-to-bottom typesetting, there are known problems with the CFF
> format since FreeType doesn't currently support the `VORG' table.
>
>
>     Werner