Re: Would like MakeTeXPK to write into texmf/fonts/public/cm/pk/..

[Christopher Lott <lott@informatik.uni-kl.de>]

Hi,

Thank you all very much for your responses to my complaints about
installing chunks of TeX.  Sorry to be so slow in responding.  Our 
son was just born last week, so things have been a little chaotic on
the home front!

I would first like to ask about teTeX.  This sounds like the solution
to frustrating installs.  Is this available from a CTAN archive, or
only commercially?  No matter what the answer, may I suggest making
extremely clear references to this package somewhere where a person
such as myself might find this notice before blithely ftp'ing
lib.tar.gz, base.tar.gz, packages.tar.gz etc etc.?

Sorry too that I was so unspecific about package versions.  Usually
I'm better at this.  I should have written that I had just pulled
everything from CTAN a few days earlier, so I think I have the very
latest versions.  I have the LaTeX distribution from 21 Nov 95,
kpathsea-2.6, version 18f of xdvik, version 5.58f of dvipsk, a
psnfss2e.tex file dated (internally) 95/11/21, etc etc.  I can dig 
out all the numbers if it would help.

What did psnfss2e.tex mention that was misleading..  I'm trying to
reconstruct my notes.  Basically it's an extremely concise document
that assumes you know exactly what you're doing, which in my case was
certainly not true.  So keep that in mind when reading this list:

 + v5.58f of dvipsk includes a big psfonts.map file that seems to
   have everything supplied with the psnfss2e package.  
 + The .fd files for adobe as available on CTAN are stored in
   subdirectories named "tex".   Directories from other foundries are
   set up differently.
 + Might mention that directories texmf/fonts/* are (as set up in the
   supplied texmf.cnf) file are not searched by LaTeX, but they ARE
   searched by dvips. 
 + Maybe I'm on drugs, but I only made the .fd files available to
   LaTeX2e, not the .tfm files, and it works fine.  What will not work?
 + A good reference to "Karl Berry's font naming documentation" would
   have been very helpful, e.g., a ftp address.
 + A little more explanation about the decision to/not to use Cork
   encoding (I still have no clue what that is) would have been great.
 + I could barely get test0.tex working, test1.tex was hopeless.
   Maybe this is where the missing .tfm files bit me?  Should test1
   have worked?

I wrote:
 >> 2. Fontname seems to be essentially undocumented.  I found that pk
You replied:
 >it has excellent documentation!

WRT fontname: I was trying to say that I could not find any
description of why I might need "fontname" in any of the
documentation, especially kpathsea.texi.  Try for yourself:
'grep fontname kpathsea.texi'.  I get "@cindex fontnames, unlimited length"
that's it.  Once I pulled the fontname package, sure I found docs with
it.  The real leap was guessing that "fontname" is a package and
ftp'ing off to dante in search of it.  I got lucky there.

I'm really quite uncertain about the possible bug I reported WRT
prioritizing directory seeks when using the ls-R database.  Some more
info: we only have latex2e, and it's installed as "latex".  We do not
have latex209 in any shape or form, except for compatability mode.
My texmf.cnf is the original that's supplied with whatever package I
ftp'd.  (Sorry, I'm confused here.)  Maybe I should name the installed
binary latex2e???  I was trying to keep things simple.  What's the
usual rule here?  And I don't have the double-slash in TEXINPUTS.

I hope that I've answered all the salient points.  Thanks again for
all your time in responding.  I'll be happy to supply more info as
needed.

chris...
--
Christopher Lott    lott@informatik.uni-kl.de    +49 (631) 205-3342, Fax -3331
Address: FB Informatik-Bau 57, Universitaet KL, 67653 Kaiserslautern, Germany
World-wide web: http://uomo.informatik.uni-kl.de:2080/Personalia/cml/