Re: [AUCTeX-devel] Problems with tex-site.el.in

[David Kastrup]

Reiner Steib <address@hidden> writes:

> I'm not sure if the current code in CVS is supposed to build at all
> outside David's machine.

Not even there.  I have not yet started building.  Up to now, I have
just been getting the configure process to a level where I think it is
more or less supposed to work (supposed to!)

> Anyhow, here we go...
>
> With current AUCTeX CVS and Emacs 21.3, I get:
>
> $ ./autogen.sh
> $ /configure --prefix=/usr/local \
>   --with-auto-dir=/usr/local/share/emacs/site-lisp/auctex/auto \
>   --with-lispdir=/usr/local/share/emacs/site-lisp
> $ make
> [...]
> While compiling toplevel forms in file [...]/auctex/tex.el:
>   !! Wrong type argument ((listp 65))
>
> With "emacs-cvs -Q -l lpath.el" I get after `M-: (require 'tex-site)
> RET':
>
> --8<---------------cut here---------------start------------->8---
> Debugger entered--Lisp error: (wrong-type-argument listp 65)
>   car(65)
>   (list (quote const) (car x))
>   (lambda (x) (list (quote const) (car x)))(65)
>   mapcar((lambda (x) (list (quote const) (car x))) "Alist of built-in TeX 
> modes and their load files.")
>   (cons (quote set) (mapcar (lambda ... ...) TeX-mode-alist))
>   (custom-declare-variable (quote TeX-modes) (quote (mapcar ... 
> TeX-mode-alist)) "List of modes provided by AUCTeX.\n\nThis variable can't be 
> set normally, use customize for it." :type (cons (quote set) (mapcar ... 
> TeX-mode-alist)) :set (quote TeX-modes-set) :initialize (lambda (var value) 
> (custom-initialize-reset var value) (dolist ... ...)))
>   (defcustom TeX-modes (mapcar (quote car) TeX-mode-alist) "List of modes 
> provided by AUCTeX.\n\nThis variable can't be set normally, use customize for 
> it." :type (cons (quote set) (mapcar ... TeX-mode-alist)) :set (quote 
> TeX-modes-set) :initialize (lambda (var value) (custom-initialize-reset var 
> value) (dolist ... ...)))
>   eval-buffer(#<buffer  *load*> nil "tex-site" nil t)
>   ;;; Reading at buffer position 3616
>   
> load-with-code-conversion("/home/dept/ste/src/links/auctex/auctex/tex-site.el"
>  "tex-site" nil t)
>   require(tex-site)
>   eval((require (quote tex-site)))
>   eval-expression((require (quote tex-site)) nil)
>   call-interactively(eval-expression)
> --8<---------------cut here---------------end--------------->8---
>
> I found out that this is due to the bogus defcustom of
> `TeX-mode-alist' (value and doc-string exchanged):
>
> ,----[ C-h v TeX-mode-alist RET ]
> | TeX-mode-alist's value is 
> | "Alist of built-in TeX modes and their load files."
> | 
> | ((tex-mode . tex-mode) (plain-tex-mode . tex-mode) (texinfo-mode .
> | texinfo-mode) (latex-mode . tex-mode) (doctex-mode . tex-mode))
> `----
>
> --8<---------------cut here---------------start------------->8---
> --- tex-site.el.in    28 Apr 2005 12:53:34 +0200      1.21
> +++ tex-site.el.in    28 Apr 2005 18:55:13 +0200      
> @@ -65,12 +65,12 @@
>    :type 'directory)
>  
>  (defconst TeX-mode-alist
> -  "Alist of built-in TeX modes and their load files."
>    '((tex-mode . tex-mode)
>      (plain-tex-mode . tex-mode)
>      (texinfo-mode . texinfo-mode)
>      (latex-mode . tex-mode)
> -    (doctex-mode . tex-mode)))
> +    (doctex-mode . tex-mode))
> +  "Alist of built-in TeX modes and their load files.")
>  
>  (defun TeX-modes-set (var value &optional update)
>    "Set VAR (which should be `TeX-modes') to VALUE.
> --8<---------------cut here---------------end--------------->8---
>
> The next problem is, that `doctex-mode' is not defined in Emacs < 22
> (or XEmacs most probably).   We should probably check this:
>
> --8<---------------cut here---------------start------------->8---
> @@ -85,7 +85,8 @@
>      (while list
>        (setq elt (car (pop list)))
>        (when (or update (null (get elt 'tex-saved)))
> -     (put elt 'tex-saved (symbol-function elt)))
> +     (when (fboundp elt)
> +       (put elt 'tex-saved (symbol-function elt))))
>        (fset elt
>           (if (memq elt value)
>               (intern (concat "TeX-" (symbol-name elt)))
> --8<---------------cut here---------------end--------------->8---
>
> I will check in these patches now.  I hope this is correct.
>
> Next one: Starting Emacs 21.3 and opening a LaTeX file, I get:
>
> ,----
> | File mode specification error: (void-function tex-mode)
> `----
>
> and `C-h f tex-mode RET' gives:
>
> ,----
> | describe-function: Symbol's function definition is void: TeX-tex-mode
> `----
>
> I don't know how to fix this.  Probably something is still wrong in
> `TeX-mode-alist' and/or `TeX-modes-set'.

Uh, nothing unusual here.  This is all done by autoloads, and that was
the next thing I wanted to get into the Makefile.

I _knew_ I had forgotten something pretty important...

Hey, I want to give Ralf a good reason to call me criminally insane,
ok?  It makes all people feel warm and fuzzy knowing that they could
not possibly be held responsible for the project management's
decisions.  Everybody needs to say "told you so" once in a while.

Ok, I'll try cleaning up the mess until it builds and works on my
machine.  I don't think I'll have the time and energy left for letting
the release machinery roll before I leave.

But I'll have Internet access in Poland, and though the mail servers
of GNU are down this weekend, the CVS servers can still be
used....  If you need to reach me by mail, try my forename dot
lastname at arcor.de.  It does not look like more people than Reiner
and Ralf would be actively involved right now.

Anyway: just to give you an impression just _how_ mad I can be if I
want something: I am singing in a pretty reputable choir (we compete
at state level) and we just had our last rehearsal before our
Berlin/Greifswald tournee (which is right after BachoTeX in Poland).
We have a pretty large program, and the last piece is Monteverdi's
1651 mass.  To some consternation of the other choir members ("You are
standing in the wrong place!"  "Trust me."), I switched from a rather
high tenor voice (which is for this reason shared between high tenors
and low altos) into the first alto (having asked the conductor for
permission to try this two weeks ago, but we did not rehearse
Monteverdi until today, and then only partially).  Hey, I made it.  I
was pretty unhappy with the previous arrangement, because it was
merely half mad, and constantly having to mix and/or switch registers
between normal and crazy was a nuisance and frustrating.

So don't be surprised if at times I opt for the full madness...

-- 
David Kastrup, Kriemhildstr. 15, 44793 Bochum