[cairo] Rounding bug in cairo_ft_font.c

[Keith Packard]

Around 0 o'clock on Jan 9, Carl Worth wrote:

> There are obviously some bugs in the exact metrics, (notice some of
> the glyphs extend beyond the boxes).

Most importantly, the metrics are obviously different under different CTM 
environments.  This can't be OK.  Making that work was what I did in my 
patch.

Note that the interface between the top and bottom of cairo is the source 
of the current bugs; it has the font driver attempt to return things in 
'user space', but there isn't any user space down there, only pixel space 
and font space.

With my patch in place, CTM has no effect on metrics.  And, I use the font 
matrix to scale the font metrics; I think the only change here is that 
Owen believes we should just use the font matrix to transform the font 
metrics.  I think I can do that; then we'll have a more reasonable 
comparison.

> This seems quite sane to me. Do we actually want something different?

I don't know.  PostScript returns just the font-space metrics and lets the 
application sort out the font matrix effect on metrics.  Owen (and I) 
think that this is a bad plan.  I think I agree with Owen that the only 
sensible thing to do is return a user space axis-aligned bounding 
rectangle that covers the glyphs.  Optimizing this bounding rectangle to 
minimally bound the glyphs seems optional.

-keith


-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 228 bytes
Desc: not available
Url : http://lists.freedesktop.org/archives/cairo/attachments/20050109/3357b76c/attachment.pgp