Re: [ft-devel] Type42 handling

[Huver]

Hi,

I'm referring to the fact that the Type42 driver, as in version
2.3.5, loads the TT font; but does not treat it like one.  That
is, the "PostScript-ish" interpretation of "the font" gets in
the way.

For one, FontName in Type42 doesn't necessarily always match the TT
name.  For another, Type42 handler does not support "kerning info"
query; this makes it useless for text composition applications.

A T42 font file should be viewed as a "user convenince" file: one
need not have a TT font exist in two different font files, in order
to use it (on screen) and also print it to PS printers.  But to FT
clients, it should appear exactly the same as a TT font -- becaues
that's what the font really is.

It is my question/suggestion that FT rid the PostScript style
interpretation on T42 font, and treat it exactly as a TT font, in
all aspects of operations as FT provides.  FT may want to preserve
some property flag to remind itself and FT lib clients, that the
particular font origin format is T42, and the T42 handler may even
provide a way for clients to request "the TT font data" -- i.e.,
get the decoded "sfnts" data so that a client can subset the TT font
if need to -- this is optional of course.

But please, do not restrict FT operations on a T42 font to what Adeobe
thinks its PS RIP should be -- a PS RIP doesn't ever compose text, so
it can't care any less about kerning info, but FT lib clients almost
always do.


-huver  address@hidden