Re: [Devel] cmap in webdings

[Dirck Blaskey]

David,

Thanks for the feedback;
I'm going to go back to the customer and
try to find out where this font came from.

I think there may be an approach that
could correctly fix the cmap after loading,
but, as you say, doctoring the cmap isn't an
option with ttcmap0.c.

If I come up with anything, I'll let you know.

d


----- Original Message -----
From: "David Turner" <address@hidden>
To: <address@hidden>
Sent: Tuesday, May 21, 2002 5:59 PM
Subject: Re: [Devel] cmap in webdings


> Hello Dirck,
>
> Dirck Blaskey a écrit :
> >
> > I have a pdf file with an embedded True Type font, webdings,
> > that has some oddities with its cmap, so FT_Get_Char_Index
> > wasn't working.
> >
> In other words, the embedded font is buggy... Humm...
>
> > The cmap doesn't explicitly cover the character ranges,
> > most of the entries have startCount == endCount.
> >
> > Here is a patch to Freetype 2.0.9 sfnt/ttcmap.c;
> > The patch should work with the current development snapshot.
> >
> > In theory, there might be a problem with the patch,
> > when looking for characters that _really_ aren't there.
> > Maybe an additional check for startCount == endCount?
> > Is there any safe way to be sure?
> >
>
> I'm sorry, but I'm going to reject this patch for the moment.
> I have studied it, the charmap you describe as well as the
> original webdings.ttf font. I simply cannot think of any
> reasonable way to ensure reliable support for both correct
> fonts per se the TrueType/OpenType specs and the broken one
> you're using.
>
> If you can come up with a scheme that would ensure reliable
> charmap indexing in all cases, I'd be happy to incorporate
> it within FreeType. Your patch unfortunately doesn't provide
> enough guarantees about the correctness of the returned
> function.
>
> That's not to say that a solution is impossible, simply that
> I couldn't come up with something satisfying myself.
>
>
> Best Regards,
>
> - David Turner
> - The FreeType Project  (www.freetype.org)
>
> PS: By the way, if you intend to perform some doctoring, please
>     be aware that the code in "src/sfnt/ttcmap.c" is being
>     replaced by the one in "src/sfnt/ttcmap0.c" (found in 2.1.0)
>
>     this one directly parses the table's raw bytes, and doctoring
>     is not directly allowed there, since we may be accessing ROM
>     or memory-mapped files in certain cases...
>
> _______________________________________________
> Devel mailing list
> address@hidden
> http://www.freetype.org/mailman/listinfo/devel