[tex-k] bug in dvips 5.86d in partial font downloading

Tom Kacvinsky tjk@ams.org
Tue, 16 Oct 2001 09:05:49 -0400 (EDT)


Sounds like a known bug in dvips.  Let me guess -- you are trying to use one
font "container" in the resulting PS file, but the font that is in the font
"container" is re-encoded multiple times.  If this is the case, OOPSIE!  You
have just tripped the bug in dvips.

Best bet:  do not subset fonts when re-encoding them multiple times.

Tom

On Tue, 16 Oct 2001, Vladimir Volovich wrote:

> Hi,
>
> when i process the attached dvi file with dvips 5.86d, i get:
>
> $ dvips test -o
> This is dvips(k) 5.86d Copyright 1999 Radical Eye Software (www.radicaleye.com)
> 'DVIcopy 1.5 output from TeX output 2001.10.15:1931' -> test.ps
> <texc.pro><t1.enc><ts1.enc><texps.pro>. <sfrm1000.pfb>
> Warning: module writet1 of dvips (file sfrm1000.pfb): character 89 is mapped to .notdef
>
> Warning: module writet1 of dvips (file sfrm1000.pfb): character 97 is mapped to .notdef
>
> Warning: module writet1 of dvips (file sfrm1000.pfb): character 101 is mapped to .notdef
>
> Warning: module writet1 of dvips (file sfrm1000.pfb): character 114 is mapped to .notdef
> [1]
>
> I have the cm-super fonts installed (see
> CTAN:fonts/ps-type1/cm-super), and psfonts.map uses those fonts.
>
> The file test.dvi was obtained by processing the attached test.tex and
> then running dvicopy on that file (to remove the possible virtual font
> issues).
>
> I think that this is a bug in dvips because the same file test.tex is
> processed by pdftex 0.14e without errors (and gived correct PDF file).
>
> Best,
> v.
>