Re: [Devel] CFF Cleanup (was bug in BBox_Conic_Check?)

[David Turner]

Hi Eric,

> Since you are soliciting input... I don't think the CFF Unicode charmap or
> Adobe specific FT_CharMaps will be of use to me in the immediate future, but
> I suspect it could become useful further down the road.
> 
Good, I won't investigate further this field for now :-)

> As far as CFF goes, I'm currently most interested in the charset and
> encoding tables getting loaded by the driver. I'm using the patches Wes
> submitted, and Tom is continuing that work. It will be nice when it finds
> its way into CVS, as I'm currently maintaining it on the side and merging it
> in whenever I pick up a new library version.
> 
I'm currently tagging the CVS tree to prepare the 2.0.1 release. It
only contains the latest fixes from Tom. I'll commit changes I've made
to the CFF driver yesterday a bit later.

Basically, it seems that reading the Charset table is important to
synthetize a Unicode charmap for a pure CFF font, as well as to be
able to support the "seac" trick that comes from Acrobat-produced
CFF fonts.

I have no idea wether the "Encoding" table is going to be really useful
if we don't intend to use it for Adobe-specific charmaps. Please let me
know if you're still interested in it, as I'll forget about it otherwise..

> The font transform support will also be helpful. You may have seen some of
> the list traffic about the glyphs being flipped in Distiller produced CFF
> fonts. I've been using a variation of the previously disabled call to
> FT_Outline_Transform() in T2_Load_Glyph() to "right" them. It seems to serve
> my purposes fine, but I don't know if it is really the best way to handle
> things.
> 
The correct way is the one used in the T1 driver, and I've started modifying
the CFF driver to handle this. It will need some testing though :-)

Thanks a lot for your comments, they're really appreciated..

- David