Thanks for that. I found my printer manual and toggled
> error printing. I get a simalr message:
> ERROR: invalidfont OFFENDING COMMAND: stringwidth
> STACK:
> (psmarker0)
matplotlib embeds the entire font file into the generated, and if on
your system some huge font is being found by the font matching
mechanism, it may be screwing up your printer driver. Try setting the
ps.useafm : True
option in your rc file and see if this helps.
Also, if you run with --verbose-debug it will give you information
about which font files are being loaded (and lots of other useless
junk too...
JDH
John Hunter writes:
> matplotlib embeds the entire font file into the generated, and if on
> your system some huge font is being found by the font matching
> mechanism, it may be screwing up your printer driver. Try setting the
>
> ps.useafm : True
>
> option in your rc file and see if this helps.
>
> Also, if you run with --verbose-debug it will give you information
> about which font files are being loaded (and lots of other useless
> junk too...
>
Thanks, John -- if I set (ps.useafm : True) and also
(text.usetex : False) then findfont locates a helvetca font and the
file prints fine:
(http://clouds.eos.ubc.ca/~phil/matplotlib_postscript/debug_notex.txt,
http://clouds.eos.ubc.ca/~phil/matplotlib_postscript/file1_notex.ps)
If I set (text.usetex : True) then (ps.useafm : True) is
overridden, dvips is invoked and I get the same erroneous postscript file as previously.
generated by dvips using cmr10:
(http://clouds.eos.ubc.ca/~phil/matplotlib_postscript/debug_withtex.txt,
http://clouds.eos.ubc.ca/~phil/matplotlib_postscript/file1.ps)
(Note that TeX/latex is definitely working on this machine).
best, Phil