GNU bug report logs - #54949
13.1.2; Indentation for matrix environments

Previous Next

To add a comment to this bug, you must first unarchive it, by sending
a message to control AT debbugs.gnu.org, with unarchive 54949 in the body.
You can then email your comments to 54949 AT debbugs.gnu.org in the normal way.

Toggle the display of automated, internal messages from the tracker.

Message #5 received at submit <at> debbugs.gnu.org (full text, mbox):

From: Robert Morelli <morelli <at> flux.utah.edu>
To: bug-auctex <at> gnu.org
Subject: 13.1.2; Indentation for matrix environments
Date: Fri, 15 Apr 2022 02:31:32 -0600

The only way I can get the indentation to work with matrix environments
like matrix, bmatrix, etc. is to start the line with \begin{matrix}. For
instance, putting:
Let $E = \begin{bmatrix}
1 & 0\\
0 & 1
\end{bmatrix}$ be the
identity matrix.
results in the indentation being offset for the remainder of the file.

Emacs : GNU Emacs 28.1 (build 1, x86_64-apple-darwin18.7.0, NS 
appkit-1671.60 Version 10.14.6 (Build 18G95))
of 2022-04-04
Package: 13.1.2

current state:
==============
State could not be dumped due to the following error:

(void-function ,)

You should still send this bug report.

-- 
Robert Morelli
Department of Mathematics
Salt Lake Community College

Message #8 received at 54949 <at> debbugs.gnu.org (full text, mbox):

From: Ikumi Keita <ikumi <at> ikumi.que.jp>
To: Robert Morelli <morelli <at> flux.utah.edu>
Cc: 54949 <at> debbugs.gnu.org
Subject: Re: bug#54949: 13.1.2; Indentation for matrix environments
Date: Fri, 15 Apr 2022 23:50:45 +0900

Hi Robert,

>>>>> Robert Morelli <morelli <at> flux.utah.edu> writes:
> The only way I can get the indentation to work with matrix environments
> like matrix, bmatrix, etc. is to start the line with \begin{matrix}. For
> instance, putting:
> Let $E = \begin{bmatrix}
> 1 & 0\\
> 0 & 1
> \end{bmatrix}$ be the
> identity matrix.
> results in the indentation being offset for the remainder of the file.

That's an effect of recent AUCTeX update, which indents matrix-like
environments by the same rule as tabular-like environments. Please
disable that behavior by customizing `LaTeX-indent-environment-list' to
remove entries of matrix-like environments.

Regards,
Ikumi Keita
#StandWithUkraine #StopWarInUkraine

Message #11 received at 54949 <at> debbugs.gnu.org (full text, mbox):

From: Robert Morelli <morelli <at> flux.utah.edu>
To: 54949 <at> debbugs.gnu.org
Subject: bug54949
Date: Fri, 15 Apr 2022 15:18:07 -0600

For me, this seems problematic.

1. It is common to use matrix environments in mathematical writing 
inline, rather than displayed.

2. It is quite unexpected that indentation for an environment persists 
to the end of the file, even after the environment has ended. This seems 
to break the expectation that indentation reflects the compositional 
structure of a document.

3. Asking an end user to modify a variable like 
LaTeX-indent-environment-list puts an unwanted burden on the user. There 
are potentially many ways to remove an element from a list. One way is 
through the customization interface. Like many Emacs users, I avoid the 
customization interface because I find it awkward to us, and it creates 
two potentially conflicting sources of modifying Emacs behavior if you 
also use an init file.

4. If the variable is programmatically changed in the init file, it in 
turn raises the question of how to do so. There are potentially many 
ways. Some might be safer than others, or easier. What is being removed 
is itself a list, and there are a number of such elements.

5. If I modify the variable in my init file, I potentially lose future 
improvements to AUCTeX's indentation algorithms. To avoid that, I would 
have to recheck with each update of AUCTeX what new features are in the 
default. Remember, AUCTeX is one of more than 100 packages I have 
installed in Emacs.

6. Like most people, I'm very busy. I don't have time to create a mini 
project out of investigating how to modify a single variable in a single 
one of the 100+ packages I use in my text editor, which is one of dozens 
of programs I use on my computer, which is only a part of my life. In 
the big scheme of things, it would be foolish of me to spend more time 
on LaTeX-indent-environment-list, so the simplest solution for me is 
probably to regress AUCTeX, if I can do that easily.


-- 
Robert Morelli
Department of Mathematics
Salt Lake Community College

Message #14 received at 54949 <at> debbugs.gnu.org (full text, mbox):

From: Ikumi Keita <ikumi <at> ikumi.que.jp>
To: Robert Morelli <morelli <at> flux.utah.edu>
Cc: 54949 <at> debbugs.gnu.org
Subject: Re: bug#54949: bug54949
Date: Sun, 17 Apr 2022 20:11:28 +0900

Hi Robert,

Thanks for your comments, I understand what you expect: "AUCTeX should
be smart enough about indentation." :-)

However, it is very difficult to find out the "correct" indentation in
all cases. Thus AUCTeX actually uses only limited set of heuristics. In
particular, it assumes that there are only whitespaces between
\begin{xxx} and its beginning of line; this avoids that the amount of
computation of indentation goes beyond practical degree of complexity.

>>>>> Robert Morelli <morelli <at> flux.utah.edu> writes:
> 2. It is quite unexpected that indentation for an environment persists
> to the end of the file, even after the environment has ended. This
> seems to break the expectation that indentation reflects the
> compositional structure of a document.

Yeah. But it would be rather inefficient to re-calculate the indent
every time that an environment ends, which needs to find out the
corresponding \begin{xxx} skipping back over all potential inner nested
\begin-\end pairs, especially when indenting region with wide span.
(...but I began to feel that that idea isn't much bad. Maybe I try to
implement such behavior when I have spare time.)

> 1. It is common to use matrix environments in mathematical writing
> inline, rather than displayed.

Well, it isn't the matter of inline v.s. displayed style. LaTeX typesets
the following inline math expression without any problems:
----------------------------------------------------------------------
Let $E =
\begin{bmatrix}
  1 & 0\\
  0 & 1
\end{bmatrix}
$ be the identity matrix.
----------------------------------------------------------------------
Hence I expect you won't suffer from the heuristics that \begin{xxx}
stands on its own dedicated line as long as you enter environments by
C-c C-e in AUCTeX, because C-c C-e always put \begin{xxx} and \end{xxx}
on their own dedicated lines (except possible comment prefix).

If I understand correctly, this issue is only problematic for users who
frequently receive LaTeX documents written by other persons and compose
those into one large document, adapting the components accordingly. I
suppose you are in such a position, perhaps an editor of proceedings of
some math conferences?

Actually, reports about similar shortcoming in AUCTeX recurred recently:
https://debbugs.gnu.org/cgi/bugreport.cgi?bug=28405#41
https://lists.gnu.org/r/auctex-devel/2022-02/msg00014.html
We haven't come up with a satisfactory solution for manuscripts which
don't conform to AUCTeX heuristics with respect to indentation.

> 3. Asking an end user to modify a variable like
> LaTeX-indent-environment-list puts an unwanted burden on the user.
> There are potentially many ways to remove an element from a list. One
> way is through the customization interface. Like many Emacs users, I
> avoid the customization interface because I find it awkward to us,

Sorry, what I had in mind was that awkward method. It's really awkward
since the default value contains dozens of matrix-like environments.
Currently I don't have good idea what to do about that tediousness. :-(

> 6. Like most people, I'm very busy. I don't have time to create a mini
> project out of investigating how to modify a single variable in a
> single one of the 100+ packages I use in my text editor, which is one
> of dozens of programs I use on my computer, which is only a part of my
> life. In the big scheme of things, it would be foolish of me to spend
> more time on LaTeX-indent-environment-list, so the simplest solution
> for me is probably to regress AUCTeX, if I can do that easily.

> 5. If I modify the variable in my init file, I potentially lose future
> improvements to AUCTeX's indentation algorithms.

Then you are expecting to have improvements to AUCTeX. The update in
`LaTeX-indent-environment-list' is one of such improvements, which
aligns "&" sign vertically in suitable environments. So your
inconvenience is someone's benefit. It's inevitable to have sometimes
such unfortunate conflicts as long as you are expecting to experience
improvements to AUCTeX.

Regards,
Ikumi Keita
#StandWithUkraine #StopWarInUkraine

Message #17 received at 54949 <at> debbugs.gnu.org (full text, mbox):

From: Robert Morelli <morelli <at> flux.utah.edu>
To: 54949 <at> debbugs.gnu.org
Subject: bug#54949
Date: Mon, 18 Apr 2022 00:24:34 -0600

[Message part 1 (text/plain, inline)]

Thank you for you prompt replies.

For the reasons I've already stated, I've decided my best course at this 
point is to revert to an earlier version of AUCTeX, (perhaps the last 
12.x version?), and otherwise freeze all Emacs packages at their current 
versions, and revert Emacs to 27.x. Reverting Emacs has been easy, but I 
don't know how to get back to AUCTeX 12.x. If you can suggest how I can 
do this, I would be grateful.

If I may further comment, I do not agree with backwards incompatible 
changes in the behavior of Emacs or its packages. It is my opinion that 
Emacs development from this point forward should transition to legacy 
mode, with bug fixes and adaptations to OS upgrades, but with no further 
changes in behavior. New development should be diverted to using 
cross-editor technologies like language server protocol, and widely used 
non-proprietary programming languages like Javascript.

Those few of us who are still using Emacs are not eager to invest time 
adapting to new and potentially problematic behaviors, especially since 
almost all of us are probably already transitioning to other editors.

If I may suggest, I would be happy to have the functionality of AUCTeX 
available in alternative editors. I think this would be welcomed by a 
much larger group of potential users than Emacs will likely ever again have.

-- 
Robert Morelli
Department of Mathematics
Salt Lake Community College

[Message part 2 (text/html, inline)]

Message #20 received at 54949 <at> debbugs.gnu.org (full text, mbox):

From: Ikumi Keita <ikumi <at> ikumi.que.jp>
To: Robert Morelli <morelli <at> flux.utah.edu>
Cc: 54949 <at> debbugs.gnu.org
Subject: Re: bug#54949:
Date: Mon, 18 Apr 2022 21:44:04 +0900

Hi Robert,

>>>>> Robert Morelli <morelli <at> flux.utah.edu> writes:
> For the reasons I've already stated, I've decided my best course at
> this point is to revert to an earlier version of AUCTeX, (perhaps the
> last 12.x version?), and otherwise freeze all Emacs packages at their
> current versions, and revert Emacs to 27.x. Reverting Emacs has been
> easy, but I don't know how to get back to AUCTeX 12.x. If you can
> suggest how I can do this, I would be grateful.

I suppose the following prescreption would work:
1. Delete the current AUCTeX by M-x package-delete RET auctex-13.1.2 RET
2. Download one of the package files from
       https://elpa.gnu.org/packages/auctex.html
   and decompress it to obtain tar file.
3. Restart emacs session and specify the tar file as the argument for
   M-x package-install-file RET

Please let us know if it doesn't.

> If I may further comment, I do not agree with backwards incompatible
> changes in the behavior of Emacs or its packages. It is my opinion
> that Emacs development from this point forward should transition to
> legacy mode, with bug fixes and adaptations to OS upgrades, but with
> no further changes in behavior.

There is another LaTeX mode called YaTeX (Yet Another TeX mode for
Emacs) available from MELPA, documents of which are available at
https://www.yatex.org/ . I'd recommend to taste it because YaTeX tries
very hard to keep backward compatibility; for example, it still works
on emacs 19.

Regards,
Ikumi Keita
#StandWithUkraine #StopWarInUkraine

Message #25 received at 54949-done <at> debbugs.gnu.org (full text, mbox):

From: Arash Esbati <arash <at> gnu.org>
To: Robert Morelli <morelli <at> flux.utah.edu>
Cc: Keita Ikumi <ikumi <at> ikumi.que.jp>, 54949-done <at> debbugs.gnu.org
Subject: Re: bug#54949:
Date: Tue, 19 Apr 2022 14:01:40 +0200

Robert Morelli <morelli <at> flux.utah.edu> writes:

> Reverting Emacs has been easy, but I don't know how to get back to
> AUCTeX 12.x. If you can suggest how I can do this, I would be
> grateful.

The recipe provided in the other response should work, it is also
described here[1].

> It is my opinion that Emacs development from this point forward should
> transition to legacy mode, with bug fixes and adaptations to OS
> upgrades, but with no further changes in behavior.

I suggest you take this comment to emacs-devel <at> gnu.org if you want to
discuss it seriously with the Emacs maintainers.

> New development should be diverted to using cross-editor technologies
> like language server protocol,

You can use LSP servers with Emacs through one of these clients:

 • https://github.com/joaotavora/eglot
 • https://github.com/emacs-lsp/lsp-mode

> Those few of us who are still using Emacs are not eager to invest time
> adapting to new and potentially problematic behaviors, especially
> since almost all of us are probably already transitioning to other
> editors.
>
> If I may suggest, I would be happy to have the functionality of AUCTeX
> available in alternative editors. I think this would be welcomed by a
> much larger group of potential users than Emacs will likely ever again
> have.

As always, this means that someone(™) has to do the job.  AUCTeX's
source code is available, if that someone wants to get an idea how it's
done in elisp.

I don't see we're discussing a bug, hence I'm closing this report.
Please feel free to re-open if a bug discussion will happen.

Best, Arash

Footnotes:
[1]  https://www.emacswiki.org/emacs/ELPA#h5o-10

Message #32 received at 54949 <at> debbugs.gnu.org (full text, mbox):

From: Ikumi Keita <ikumi <at> ikumi.que.jp>
To: Robert Morelli <morelli <at> flux.utah.edu>
Cc: 54949 <at> debbugs.gnu.org
Subject: Re: bug#54949: 13.1.2; Indentation for matrix environments
Date: Thu, 14 Sep 2023 14:47:25 +0900

Hi Robert,

I think the latest commit to the AUCTeX git repository addresses this
issue.  Please give it a try if you are still interested in this topic.

Regards,
Ikumi Keita
#StandWithUkraine #StopWarInUkraine

Previous Next

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