[Devel] Re: Font transforms and autohinter...

[Tom Kacvinsky]

Well, it looks like one cannot use a FontMatrix whose FT transform component
is not invertible.  I just checked with some Adobe interpreters (A Xerox printer
and Distiller), and they do not like this FontMatrix: [.001 .001 .001 .001 0 0].

But looking at the PLRM and/or Type 1 spec., it is not clear that these matrices
need to be invertible.  Hmmm...

Tom

On Thu, 21 Dec 2000, Tom Kacvinsky wrote:

> I have a concern here.  The function ah_hinter_load (as I have recently 
> patched
> it) calculates the inverse of the font transformation matrix.  This is not 
> good,
> because it makes the assumptiont hat the matrix is invertible.  For instance, 
> it
> is possible to have a Type 1 FontMatrix that looks like
>
>     [.001 .001 .001 .001 0 0].
>
> The matrix used by FT for font transforms then becomes
>
>     [.001 .001 .001 .001].
>
> This is not an invertible matrix!
>
> What say you, esteemed members of the FT devel list?
>