[OS X TeX] xelatex fontmap error

[cfrees at imapmail.org]

On Fri 28th May, 2010 at 00:49, Peter Dyballa seems to have written:

>
> Am 27.05.2010 um 22:53 schrieb cfrees at imapmail.org:
>
>> So how can I figure out where the problematic lines are coming from?
>
>
> /usr/local/texlive/2009/texmf-dist/fonts/map/dvips/avantgar/uag.map is the 
> standard MAP file fragment for the AvantGarde clone font. If it's as clean as 
> you cite, then it could come from some private or local MAP file fragment (I 
> have half a dozen of them, one for the fonts of my PostScript 3 printer). 
> XeTeX's xdvipdfmx uses the pdftex.map MAP file. With TeX Live 2009 a first 
> default variant is installed and sometimes updated: 
> /usr/local/texlive/2009/texmf/fonts/map/pdftex/updmap/pdftex.map. The 
> system's working copy is a symbolic link: 
> /usr/local/texlive/2009/texmf-var/fonts/map/pdftex/updmap/pdftex.map (you can 
> use it like the file itself). One command you could try is:
>
> 	wc -l 
> /usr/local/texlive/2009/texmf/fonts/map/pdftex/updmap/pdftex.map 
> /usr/local/texlive/2009/texmf-var/fonts/map/pdftex/updmap/pdftex.map
>
> It will tell the number of lines the files have. Then you could modify the 
> command like this:
>
> 	grep uagd8a 
> /usr/local/texlive/2009/texmf/fonts/map/pdftex/updmap/pdftex.map 
> /usr/local/texlive/2009/texmf-var/fonts/map/pdftex/updmap/pdftex.map
>
> This will tell you the uagd8a font mappings used in both MAP files. It will 
> not reveal from where the faulty line comes. For this you would need to 
> iterate over all enabled MAP file fragments. This command could perform the 
> search:
>
> 	find /usr/local/texlive/2009/texmf-dist/fonts/map -type f -name 
> "*.map" -exec grep uagd8a {} /dev/null \;

Yes. I was rather afraid that might be the answer.

> I presume that no culprit will be found. Probably the updmap-sys script had 
> introduced the fault. This could be checked by running:
>
> 	sudo /usr/local/texlive/2009/bin/universal-darwin/updmap-sys 
> --disable uag.map
> 	sudo sh -x /usr/local/texlive/2009/bin/universal-darwin/updmap-sys 
> --enable Map=uag.map
>
> The latter command will (hopefully) show in detail how the shell scripts 
> performs it's job. You will see a lot of similar blocks of commands or 
> statements when all the MAP fragments are processed. Somewhere there the 
> error happened. Or later, when the first created output files are prepared 
> for actual use and modified for other convertors. It's very likely that the 
> same error will *not* happen again. When you now, after updmap-sys has 
> finished, invoke again the first two command, starting with 'wc -l' resp. 
> 'grep uagd8a', you should see very similar results and no faulty line. If 
> you'll still see that faulty line, then we'll really have a problem that 
> can't be solved that easily. It would need examination of all your local and 
> private MAP file fragments. For example by sending me privately an archive of 
> all these files... (which I would check in nine or ten hours)

That is very generous (if foolhardy) but thankfully unnecessary.

XeTeX apparently dislikes some map files which pdfTeX is quite happy
about. These include gentium.map (which should not have been enabled
through updmap anyway) and, in my local tree, 8r-urw-urwgothic.map,
ec-urw-urwgothic.map, texansi-urw-urwgothic.map and sil-charis.map.
Serves me right for experimenting... (At least I didn't *create* any of
these, even if I did install most of them...)

Anyway, thanks _very_ much for your help and Herb's. None of this
solves the problem with Venturis but then I never really expected it to.

Incidentally, uag.map is not included in the default updmap.cfg in my
install of TL. Is that correct?

Thanks again,
cfr