[email protected] skrev:
On Tue 25th May, 2010 at 12:25, Lars Hellström seems to have written:

\installfont{clmcsc8t10}{%
 ec-lmcsc10 encoding txtfdmns,lmcsc8ttl10,%
 \metrics\unsetint{acccapheight},newlatin%
}{t1-clm}{T1}{clm}{m}{sc}{}

That's wonderful - thank you.

I altered my driver file to do this with all of the fonts (though
without the line breaks). I ran the file and the output looks normal
except that I have a dvi in addition to what I'd expect. When this
happens, I can usually figure out why but not this time. The file
consists of 15 pages of a single line repeated - "0a1,16". Checking the
files in my working directory, grep tells me this is found only in the
dvi itself.

The main ecoding file, t1-clm.etx, and the driver file itself I thought
to be prime suspects. The encoding file seems clean, though. (A shorter
test driver file using the same encoding does not produce dvi output.)
And I can't see anything in the driver file which looks suspicious - it
does not even contain the number 16, never mind the whole line. The
other custom encoding files I use are unaltered from the last time I
produced this package so I'm assuming I can rule them out.

What else should I be checking/looking for? I think I must be missing
something obvious if it is repeated enough in my source to fill almost
15 pages!

It could be a bug which hasn't surfaced before because something is being used in an unusual way; the "encoding txtfdmns" trick only occurred to me as I was answering your mail.

One thing you could try is to add the command \nullfont near the top of the driver file (after \begin{document} if you're working under LaTeX). This will cause the log file to contain warnings about missing characters rather than a .dvi file to be generated, and should at least narrow down what file is being input when these lines are being produced.

Lars Hellström


Reply via email to