Re: LaTeX conversion

[Michael Clarkson <clarkson@malte.ensmp.fr>]

   Date: Mon, 25 Nov 91 15:17:56 EST
   From: Rob_MacLachlan@LISP-PMAX1.SLISP.CS.CMU.EDU


   So how is is going?  

Pretty good.  I had to beef up LaTeXinfo to give it some more things, and
while I was at it, I figured I'd incorporate the latexest TeXinfo stuff,
which meant a new version of the manual, (loop until (hell-freezes-over)).
I just put the new version of LaTeXinfo out for beta test today; I want it
shaken down for a week by others before I give it to you, so that you can
really work with it if you need to extend/change things.

Take a look on isatis.ensmp.fr by anon ftp for cmu-user.info.Z.
Try it out with xinfo, or GNU Emacs info, and see what it's like.
It's not finished, but it's close.

   I thought I'd make clear that formatting like CLtL2 is
   only a general guideline; it is more important to have the documentation
   online via info than to have a maximally pretty hardcopy.

Yes, the hold up has been getting a new version of LaTeXinfo out.  I'm
going to let you do the LaTeX hacking on the defuns and such.

1)	I'm not sure that the examples file you sent me corresponds
to the tex code.

2)	Steele uses LaTeX to do fancy parsing of arguments.

3)	In most places you use a 

\defun{format-universal-time}[ \package{extensions}
       \args{\var{dest universal-time}}
       \keys{:timezone}
       \morekeys{:style :date-first}
       \yetmorekeys{:print-seconds ...}]

\defunx{format-decoded-time}[ \package{extensions}
      	\args{\var{dest seconds minutes hours day month year} \&key{} ...}]

type of format.  It has lots of information in there, which makes it easy
to make the Emacs code to handle the arguments.  If you take Steele's way,
you end up having to write the fancy parsing in Emacs code as well.

I kept with the 

\defun{format-universal-time}[...]
\defunx{format-decoded-time}[...]
\enddefun

style of writing things, which is different from normal LaTeXinfo.
What I've done is written a LaTeXinfo style called clisp, which has this,
plus the \mopt stuff, plus a few other things.  Then I defined another
LaTeXinfo style called cmulisp, which uses this, and has the \package...
and \Hemlock stuff in.  I'll try and write a page documenting this stuff,
but I've spent way too much time on it as it is, so it might not be much.

I'd appreciate anything you can do in LaTeX for the defuns; I'm using the
TeX code from TeXinfo and I don't like it.  You can safely pitch it and
just keep the Emacs code for the defuns.

The emacs code accepts

\defun{format-universal-time}[...]
\defunx{format-decoded-time}[...]
\enddefun

or

\begin{defun}{format-universal-time}[...]
\end{defun}

or 

\begin{defun}{format-universal-time}{...}
\end{defun}


The big trick for going from the Scribe to the LaTeX info, is the overuse
in the scribe documents of a font for many purposes.  So the scribe to
latexinfo converter that I wrote, uses Emacs elisp regexps to figure
out if @i{foo} is a metasyntactic variable or a title of a book.
(You'll see what I mean if you read the LaTeXinfo or TeXinfo manuals.)
So, I've made some decisions based on patterns that I spotted, but there
will still be some work to be done going through and patching up my wrong
assumptions.  It will print fine in the manual, but it may be off in Info.

Let me have a few days to shake down the new latexinfo, and I`ll give
you it and the LaTeXinfo'ed manual.

Mike