auctex-devel
[Top][All Lists]
Advanced

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

[AUCTeX-devel] Re: preview-latex 0.9.1 and AUCTeX 11.whatever


From: David Kastrup
Subject: [AUCTeX-devel] Re: preview-latex 0.9.1 and AUCTeX 11.whatever
Date: Fri, 01 Apr 2005 12:35:32 +0200
User-agent: Gnus/5.11 (Gnus v5.11) Emacs/22.0.50 (gnu/linux)

Jan-Åke Larsson <address@hidden> writes:

> David Kastrup wrote:
>> Try to see whether you can find the current code (I checked in
>> something that apparently works, but without the shadowing check)
>> tolerable. 
>
> It does seem usable. It is just that I've been thinking things like:
> "Why do we use 'sed' when we've had emacs up and running ~four times
> already" and "Why are we shell-escaping lisp code all the time when
> we could just fire up emacs -batch -l foo"

If we didn't, then Windows users could not complain about having to
install MSYS, and various Unix users would have nothing to complain
about what breaks their multitude of shells.  We'd be all out of a job
that is fun and rewarding.

>> I did not after all switch the order of checks depending
>> on whether we have a default or explicit prefix.
>
> No, no, it seems fine. I cannot say I found any errors on inspecting
> the code, but it feels more and more opaque to me, so no guarantees.

Well, the basic macro goes through three nested loops: general
directory tree, path tail, and list of directories, and finds the
first from the list of directories that one given tail and is in one
of the given directory trees (after running AC_FULL_EXPAND on the tree
name).  If it finds such a directory, it replaces the directory tree
with its unexpanded version and returns it.  This means that
relocating installation (which requires unexpanded '$prefix' and
similar in its paths) without DESTDIR will _not_ work in prefixes that
are autodetected relative to the binary: those are not made
prefix-relative, but since the binary-related locations are checked
last, any path that _could_ be made prefix-relative should have been
covered by the non-binary relative patterns, anyhow.

> Attached is a suggestion which would prefer (tetex-3) installs in
> texmf-local.
>
> /JÅ
> Index: aclocal.m4
> ===================================================================
> RCS file: /cvsroot/auctex/auctex/preview/aclocal.m4,v
> retrieving revision 1.47
> diff -r1.47 aclocal.m4
> 247c247,248
> <   [['${datadir}/texmf' "${texprefix}"]],
> ---
>>   [['${datadir}/texmf-local' '${datadir}/texmf' \ 
>>     "${texprefix}/texmf-local" "${texprefix}/texmf" "${texprefix}"]],

Oh rats, did not think of that.  While that seems like a generally
sensible idea, do we have a chance that this will ever do the right
thing in any situation?  Seems like barking up the wrong directory
tree:

If datadir is /usr/local/share, then /usr/local/share/texmf-local is
not likely to exist (the right place probably being
/usr/share/texmf-local), and that if datadir is /usr/share, then we
would want to install in /usr/share/texmf in the first place, as we
are then performing a system default instead of a site-wide
installation?

What is teTeX's take on the Linux file system standard in that regard?

-- 
David Kastrup, Kriemhildstr. 15, 44793 Bochum




reply via email to

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