[Gs-code-review] A fix for #517474 "Problem with hyphen in PDF using Minion"

[Helge Blischke]

L. Peter Deutsch wrote:
> 
> This is a reasonable patch, but I'm disturbed that we have to keep piling
> one patch on another to address this issue.
> 
> I think there are two basic problems here.
> 
> 1) It seems that many PDF-generating applications are violating the PDF
>    specification, which says that the Widths must match the widths in the
>    underlying font (i.e., are only a shortcut to help the viewer), and are
>    instead either using Widths like Metrics (to override the widths in the
>    font) or putting garbage in them (like the 0 values that this patch
>    addresses).
> 
> 2) PostScript's Metrics facility overrides metrics on the basis of glyph
>    name, not character code, so we have to resort to hacks to make it work
>    for PDF, where Widths is indexed by character code.
> 
> We can't do anything about #1.  But we could do something about #2.  We
> could add a new (non-standard) .Widths array to the fonts that now can have
> a Metrics dictionary, indexed by character code, that the C code would
> access directly.  That would get rid of these piled-up patches, and would
> add very little code (in fact, it might not net add code at all, because of
> the code that we could remove from the PDF interpreter).
> 
> What do you think about this idea?
> 
I'm not very fond of this idea. Why not address the widths via the
font's _actual_ encoding
vector?

Helge

-- 
H.Blischke at srz-berlin.de
H.Blischke at srz-berlin.com
H.Blischke at acm.org