[Devel] Re: ttmkfdir problem

[Eric Seppanen]

(I'm switching over to devel from freetype, so please excuse the excessive 
quoting; I'm just including the background info)

I wrote:

>> I've been having problems with ttmkfdir not generating valid family 
>> names.  I have a (Windows) Truetype font called Concielian Jet, freely 
>> available from
>>
>> http://www.1001fonts.com/font_details.html?font_id=1077
>>
>> and this font works correctly on Windows boxes.
>>
>> However, on my Redhat machines (both 7.2 which distributes Freetype 
>> 2.0.3 and 7.3 which distributes Freetype 2.0.9), I am having great 
>> difficulty using this font, because ttmkfdir, which is run 
>> automatically by the xfs startup script, is generating an empty family 
>> name for this font:
>> 
>> address@hidden eds]$ ttmkfdir 
>> 6
>> conc21.ttf -macromedia--medium-r-normal--0-0-0-0-p-0-ascii-0
>> conc21.ttf -macromedia--medium-r-normal--0-0-0-0-p-0-fcd8859-15
>> conc21.ttf -macromedia--medium-r-normal--0-0-0-0-p-0-iso8859-1
>> conc21.ttf -macromedia--medium-r-normal--0-0-0-0-p-0-iso8859-15
>> conc21.ttf -macromedia--medium-r-normal--0-0-0-0-p-0-iso8859-2
>> conc21.ttf -macromedia--medium-r-normal--0-0-0-0-p-0-iso8859-9
>>
>> Note the two dashes between macromedia and medium; there's supposed to 
>> be a family name there.  This makes the font inaccessible from apps, 
>> particularly Gimp. <snip>


On Fri, May 17, 2002 at 12:21:23AM -0400, James H. Cloos Jr. wrote:
> I gave mkfontscale a try on that font.  It also generated an xlfd w/o
> a family name.  So I used fonttools to dump it to xml.  It looks like
> the font is buggy.  But ttmkfdir and mkfontscale probably ought to be
> able to deal better than they do.
> 
> The problem is that the font's name table has empty entries for the 8
> name IDs for platform 0 encoding 0.  ttmkfdir and mkfontscale must
> grab the first nameID="1" in the table.  They should probably check
> whether that entry is "" and skip to the next platform/encoding
> if it is, until one has a length > 0.

Thanks for the help Jim.

I'd like to fix this problem; can anyone point me at the correct bit of 
code to look at?  Was anyone already planning on looking at this?