auctex-devel
[Top][All Lists]
Advanced

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

Re: [AUCTeX-devel] multi-file support


From: David Kastrup
Subject: Re: [AUCTeX-devel] multi-file support
Date: Fri, 06 May 2005 19:48:01 +0200
User-agent: Gnus/5.11 (Gnus v5.11) Emacs/22.0.50 (gnu/linux)

Ralf Angeli <address@hidden> writes:

> While I dislike the solution used until AUCTeX 11.14 for the reasons
> explained earlier, I wanted to have a look at it nevertheless both
> as a possible basis for a better implementation and for working with
> an alternative approach which would better support the 'shared case.
> And while this feels to me like laying a gun in the hand of somebody
> else, you can find a patch reverting the implementation to 11.14
> attached to this mail.

AFAIR, 11.14 was quite more broken than 11.13.  Although 11.13 was
broken already.

> Some remarks.  As mentioned earlier, the dynamic parts of the menus
> are now implemented with menu filters.  Contrary to what we
> discussed earlier, the filter functions will be called as soon as
> `easy-menu-add' is executed which means _very_ early during mode
> setup.  So after applying the patch, loading a LaTeX file won't work
> correctly anymore.

Sigh.  We are going in circles.  You object to what I consider the
best user interface by pointing to the broken implementations.  It was
exactly because we were not able to implement this cleanly that the
current behavior was chosen as a workaround.  As long as we can't
implement the desired behavior without everything breaking, there is
no sense in changing the code, obviously.

But that does not make the behavior undesirable.  It would definitely
be preferable to have the old behavior _as_ _intended_, at least as an
_option_.  One can then still talk about the most sensible default to
choose.

But it is equally clear that we won't make a behavior a default that
we can't keep from producing errors, and it would be unwise to even
provide it as an option.

> While playing with it, I saw the strangest errors, like ending up in
> fundamental mode or getting a backtrace about trying to use a menu
> from within a menu-entry.

Yes, those were the kind of things that made us switch schemes.  The
way you stated you had reimplemented the stuff, it sounded like it
might have become possible to avoid this kind of stuff.  If this is
not the case, too bad.  Then we'll have to close the case again until
somebody can come up with something that works.

> I am not sure if it is an Emacs bug that the menu filters are called
> during menu creation.

Almost sounds like it.

> Maybe somebody else can shed some light on this.

Perhaps emacs-devel?

-- 
David Kastrup, Kriemhildstr. 15, 44793 Bochum




reply via email to

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