GNU bug report logs - #53749
29.0.50; [PATCH] Xref backend for TeX buffers

Previous Next

Full log

Message #334 received at 53749 <at> debbugs.gnu.org (full text, mbox):

From: David Fussner <dfussner <at> googlemail.com>
To: Mattias Engdegård <mattias.engdegard <at> gmail.com>
Cc: 53749 <at> debbugs.gnu.org, Ikumi Keita <ikumi <at> ikumi.que.jp>,
 Dmitry Gutov <dmitry <at> gutov.dev>, Arash Esbati <arash <at> gnu.org>,
 Stefan Kangas <stefankangas <at> gmail.com>, Tassilo Horn <tsdh <at> gnu.org>,
 Eli Zaretskii <eliz <at> gnu.org>, Stefan Monnier <monnier <at> iro.umontreal.ca>
Subject: Re: bug#53749: 29.0.50; [PATCH] Xref backend for TeX buffers
Date: Mon, 16 Sep 2024 09:33:48 +0100

Hi Mattias,

Thanks for this, and of course you are quite right on both counts:

> but it also depends on what you want that `tex-thingatpt-exclude-chars` variable to be

I believe it's time I abandoned this piece of idiocy, already flagged
by several commentators before you. The basic motivation was to allow
using the new xref code in TeX files that used non-standard escape and
grouping characters. Both etags and AUCTeX address this possibility,
but the latter does it right by having three separate variables:
TeX-esc, TeX-grop, and TeX-grcl, so that these can be part of the
whole regular expression apparatus for syntax highlighting and
everything else. I had in the back of my mind to implement something
similar in tex-mode, but the half-measure of
`tex-thingatpt-exclude-chars` is the wrong way to go.

I propose, therefore, to eliminate this variable from tex-mode and
just hard code the standard escape and grouping characters for the
time being, as in the rest of tex-mode.el, then to submit later a
patch to implement something similar to what AUCTeX does, and one
interoperable with AUCTeX, also.

> This is not a correct way to build a regexp;

Given the above changes, would you and your electronic servant be OK
with code like the following?

(re-search-forward "[][\\{}\"*`'#=&()%,|$[:cntrl:][:blank:]]")

(I still, I'm ashamed to confess, find the traditional regexp syntax
easier to read than rx notation, but could overcome this if you think
rx is a better fit here.)

What do you think?

Best,  David.

On Sat, 14 Sept 2024 at 18:26, Mattias Engdegård
<mattias.engdegard <at> gmail.com> wrote:
>
> Thanks for the contribution, David! My electronic servant complained about some regexps:
>
> 3881    (re-search-backward (concat "[]["
> 3882                                (mapconcat #'regexp-quote
> 3883                                           (mapcar #'char-to-string
> 3884                                                   tex-thingatpt-exclude-chars))
> 3885                                "\"*`'#=&()%,|$[:cntrl:][:blank:]]"))
>
> This is not a correct way to build a regexp; `regexp-quote` can only be used to quote strings that should match literally, not characters inside [...], where backslashes have no escaping power.
>
> There are various ways of doing this properly. I would suggest something like
>
>   (rx-to-string `(or (in "\"#$%&'()*,=[]`|" cntrl blank)
>                      ,@tex-thingatpt-exclude-chars)
>                 t))
>
> but it also depends on what you want that `tex-thingatpt-exclude-chars` variable to be. Should it be a user option (defcustom)? The variable's doc string is a wall of text that basically says that it can be set to whatever you want but things will stop working so you'd better not.
>

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.