File dvips/config/japanese.map is invalid as it does not specify .pfb,
or .pfa, files from which these fonts are supposed to come (not mentioning
encoding but maybe this is ok). Compare with other map files in the same
This deficiency is causing later spurious wailings from pdftex and
pdflatex and maybe from other programs as well.
Can you tell me how I can see these wailings? Do you have a test case?
Oh, I see what is happening. After one added some new font maps, or one
changed font configuration options (turning on support for CM in
Type1 format, for example in 'updmap') then one is supposed to rerun
/usr/share/texmf/dvips/config/updmap to generate new 'psfonts.map'.
It looks like that although japanese.map was added, and also got listed
on 'extra_modules' list in updmap, nobody bothered to execute updmap.
Therefore old maps were not updated and and you do not see anything
wrong. I was playing with 'type1_default' and was hit by that.
BTW - I checked that 'cmpskit' (cf. #31957) does not add a single pfb file
to what is in the current tetex distribution and only some buggy scripts.
The only useful piece of that package seems to be 'cmsample.tex'.
Turn on 'type1_default=true', rerun 'updmap', and do 'pdftex cmsample.tex'
just for fun of it. This should also show complaints if you kept
japanese.map in updmap configuration.
Regardless - japanese.map simply does not conform to a required format.
Is this the same as bug #17333? It seems related.
http://www.io.com/~kazushi/unix/ptex.html#psfontsmap says this:
Changing a psfonts.map for printer having no Japanese fonts
Edit /usr/local/share/texmf/dvips/psfonts.map. Add "<`min" and "<`goth" like
rml Ryumin-Light-H <`min
rmlv Ryumin-Light-V <`min
gbm GothicBBB-Medium-H <`goth
gbmv GothicBBB-Medium-V <`goth
In addition, I describe meanings of name. "rml" is a font name in DVI file.
"Ryumin-Light-H" is a font name in PS file. "min is a font name in VFlib.
Does this sound like the right solution? (Bug #17333 is a separate issue I
> Does this sound like the right solution?
This for sure would make these map entries acceptable in that sense
that they would contain a minimal number of required elements. The
last one is a name of a file from which glyphs description will be
read. How "'min" is suppposed to be interpreted in the context
I am not sure and even less what ' means here. Usually you see there
names like "lbfa.pfb" or "xyline10.pfa".
This fix is in tetex-1.0.7-16, anyhow.