auctex-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [AUCTeX-devel] Style system Was: Inconsistency in loading/saving .el


From: David Kastrup
Subject: Re: [AUCTeX-devel] Style system Was: Inconsistency in loading/saving .el files
Date: Fri, 23 May 2008 20:04:09 +0200
User-agent: Gnus/5.13 (Gnus v5.13) Emacs/23.0.60 (gnu/linux)

Pierre Lorenzon <address@hidden> writes:

> When investigating the code to check how I could improve the
> style system I discover this apparent contradiction. 
>
> It would be better (in my opinion) or at least more comfortable
> if the TeX-update-style was called after all local variables
> (including file locale one and especially TeX-master) are
> evaluated and set. 

In AUCTeX 10, TeX-update-style was called when first calling C-c C-c or
when first opening a submenu in the LaTeX menu that needed the
information.

We did not find a robust way to ask the user questions when a menu was
first opened: at least some Emacs versions were uncomfortable with a
user prompt and/or file operations in a dynamic menu entry.

The current solution is in my opinion quite inferior to the previous one
(or rather: the way the previous one was supposed to work), but I have
been somewhat alone with that opinion, and since no working solution for
all supported Emacs versions appeared, the current behavior was done.

The main problem I see with it is that it inserts file master stuff when
first opening a non-existing file.  That means that the local variables
are set to the default TeX mode.  If one has a default TeX mode of
LaTeX, and wants to create a plain TeX file, one needs to edit the local
variable section, save the file, reload it.  A complete nuisance.

Also it is not possible to delay the parsing in a new file until having
entered a basic frame work (document class/usepackages and so on).  That
means one has to manually parse again.  Or be forced to enter all
information using C-c RET.


So if you have proposals to make how to change things, by all means go
ahead: at least in my book the current behavior is not cast in stone,
and if a better and/or more convenient solution can be found, I'll be
happy to have it included.  We need to be reasonable sure, however, that
we can get basically equivalent behavior on all supported platforms
(Emacs 21.4 or later, current versions of XEmacs 21.4 and 21.5).  If
getting the same behavior on XEmacs proves to be a showstopper, we might
consider not supporting it until we find a volunteer working on AUCTeX
and/or XEmacs as required to get it working.  And if Emacs 21.4 proves
too much of a burden, I am willing to be persuaded to let it go, too.
11.85 provides a reasonably complete environment for people having to
use older software.

I would not lightly do so, but backward compatibility at all costs
should not keep us from improving things.

-- 
David Kastrup, Kriemhildstr. 15, 44793 Bochum




reply via email to

[Prev in Thread] Current Thread [Next in Thread]