[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Devel] mailing list archive; patch submission
From: |
Dirck Blaskey |
Subject: |
Re: [Devel] mailing list archive; patch submission |
Date: |
Thu, 28 Jun 2001 13:37:33 -0700 |
Tom,
Here are some more details;
for CID/CFF, in cffload.c:
in CFF_Load_Font, CFF_Load_FD_Select uses charstrings_index.count,
not dict->cid_count (from Adobe's 5167.CFF.pdf spec), so
the charstrings index has to be loaded first, and then I set
dict->cid_count = font->charstrings_index.count before
calling CFF_Load_FD_Select. Setting cid_count probably isn't cool,
because the value may be needed later, but for now it fixed this call
and the num_glyphs thing in cffobjs.c.
The next problem has to do with encodings.
The font encoding is a literal array of DBCS character values,
so CFF_Load_Encoding is pretty much irrelevent, and the functionality
has to be replaced with something I haven't come up with yet.
Basically, given a DBCS character, it has to search the array
and return the index. I don't think I want to preallocate the whole
DBCS mapping of 16000 or so characters, though it's certainly doable,
if somewhat extravagant.
I'm not sure what's going to come up next after that. -
I also have some problems up in the PDF application to deal with.
I suppose I may be out on a limb here - if I get these issues sorted
out, is this likely to work, or is this beyond the bounds of what it
will do?
...
for Type42 (TrueType) in ttload.c
TT_Load_SFNT_Header, around:
/* now, check the values of `num_tables', `seach_range', etc. */
my type42 font has a bogus entry_selector specwise,
it is equal to num_tables before it is shifted. I check for
this and skip the error.
TT_Load_Names, around
/* The name table is required so indicate failure. */
there is no name table in the font, returned no error
TT_Load_CMap, around
error = TT_Err_CMap_Table_Missing;
also ignored this error.
Postscript has an encoding external to the font, so the internal
names table and cmap aren't needed, and are literally missing.
I would be happy to provide the postscript and pdf or the
extracted fonts if you would like to take a look, or anything
else, just let me know.
d
----- Original Message -----
From: "Tom Kacvinsky" <address@hidden>
To: "FreeType Development" <address@hidden>
Sent: Thursday, June 28, 2001 1:06 PM
Subject: Re: [Devel] mailing list archive; patch submission
> Hi,
>
> >
> > *) I have patches to sfnt/ttload.c to handle some
> > 'incomplete' type42 fonts in postscript, and i'm working on
> > patches to cff/cffload.c (and probably other) to get
> > Type0/CID/CFF Japanese fonts to work in a pdf application.
> >
>
> This is the list where you send patches.
>
> I am interested in your patches to the cffload.c. What exactly are you
> adding to make this "work" with CID fonts in CFF format?
>
> > Is there a standard person to send patches to?
> > Any known fixes/work I might be replication?
>
> I don't know of anyone working on a Type 42 driver. I wouldn't know if
> there is replication in the CFF stuff until you tell me what you plan on
> doing (if it involes Unicode stuff, yes, there is duplication)
>
> Tom
>
>
> _______________________________________________
> Devel mailing list
> address@hidden
> http://www.freetype.org/mailman/listinfo/devel
>