bug#954: indentation in latex-mode in emacs-23

bug#954: indentation in latex-mode in emacs-23

[Chirok Han]

Hi

I've been using gnu emacs 22 so far (happily) and recently I compiled
and installed cvs emacs-23 (updated today: 2008-09-10). I am very
happy with this new version. Thank you!

I do lots of latex work (and I don't use auctex). Up to emacs 22, the
line following \[ is not indented but now in emacs 23, the line after
\[ is indent. For example, in emacs 22, `indent-region' gives

\[
abc
\]

but now in emacs 23 it looks like

\[
  abc
  \]

This may be a new feature, but I would like the old style. Is there
any way I can get back the old behavior? I tried to read tex-mode.el
but it was too difficult for me.

Thank you.

ch

bug#954: indentation in latex-mode in emacs-23

[Stefan Monnier]

> but now in emacs 23 it looks like

> \[
>   abc
>   \]

This lookds like a bug in my indentation code which fails to take the \
into account.  I'll take a look at it,


        Stefan

bug#954: indentation in latex-mode in emacs-23

[martin rudalics]

 > but now in emacs 23 it looks like
 >
 > \[
 >   abc
 >   \]

I've checked in a fix for this.  Please retry.

Thanks for reporting, martin.

bug#954: indentation in latex-mode in emacs-23

[Chirok Han]

Thanks.

It is now
\[
  abc
\]

I would like
\[
abc
\]
but if that's what you intend.

Thanks again. Chirok


On Tue, Sep 23, 2008 at 3:33 PM, martin rudalics <rudalics@gmx.at> wrote:
>> but now in emacs 23 it looks like
>>
>> \[
>>   abc
>>   \]
>
> I've checked in a fix for this.  Please retry.
>
> Thanks for reporting, martin.
>

bug#954: indentation in latex-mode in emacs-23

[martin rudalics]

 > It is now
 > \[
 >   abc
 > \]
 >
 > I would like
 > \[
 > abc
 > \]
 > but if that's what you intend.

This part of the Emacs 23 behavior is due to that change:

2007-09-21  Glenn Morris  <rgm@gnu.org>

	(latex-forward-sexp-1, latex-backward-sexp-1): Doc fix.
	Handle escaped parens.

If you convinced us that

\[
abc
\]

is more correct LaTeX style than

\[
   abc
\]

we might reconsider this.

Leslie, what's the preferred way to indent this?

martin

bug#954: indentation in latex-mode in emacs-23

[Leslie Lamport]

Martin,

I have no opinion on how a LaTeX input file should be formatted.

Leslie

-----Original Message-----
From: martin rudalics [mailto:rudalics@gmx.at]
Sent: Tuesday, September 23, 2008 5:28 AM
To: Chirok Han
Cc: 954@emacsbugs.donarmstrong.com; Stefan Monnier; Leslie Lamport
Subject: Re: bug#954: indentation in latex-mode in emacs-23

 > It is now
 > \[
 >   abc
 > \]
 >
 > I would like
 > \[
 > abc
 > \]
 > but if that's what you intend.

This part of the Emacs 23 behavior is due to that change:

2007-09-21  Glenn Morris  <rgm@gnu.org>

        (latex-forward-sexp-1, latex-backward-sexp-1): Doc fix.
        Handle escaped parens.

If you convinced us that

\[
abc
\]

is more correct LaTeX style than

\[
   abc
\]

we might reconsider this.

Leslie, what's the preferred way to indent this?

martin

bug#954: indentation in latex-mode in emacs-23

[Chirok Han]

Hi Martin,

It's not something I can convince you about. And I don't think there
is any correct or incorrect formatting style. It's just from my 10
years experience. As I get older, it gets harder to accommodate new
things no matter how small they are....

Many thanks for your efforts. I like latex on emacs very much.

Chirok


On Tue, Sep 23, 2008 at 9:28 PM, martin rudalics <rudalics@gmx.at> wrote:
>> It is now
>> \[
>>   abc
>> \]
>>
>> I would like
>> \[
>> abc
>> \]
>> but if that's what you intend.
>
> This part of the Emacs 23 behavior is due to that change:
>
> 2007-09-21  Glenn Morris  <rgm@gnu.org>
>
>        (latex-forward-sexp-1, latex-backward-sexp-1): Doc fix.
>        Handle escaped parens.
>
> If you convinced us that
>
> \[
> abc
> \]
>
> is more correct LaTeX style than
>
> \[
>  abc
> \]
>
> we might reconsider this.
>
> Leslie, what's the preferred way to indent this?
>
> martin
>

bug#954: indentation in latex-mode in emacs-23

[martin rudalics]

[-- Attachment #1: Type: text/plain, Size: 325 bytes --]

  > It's not something I can convince you about. And I don't think there
  > is any correct or incorrect formatting style. It's just from my 10
  > years experience. As I get older, it gets harder to accommodate new
  > things no matter how small they are....

... sounds convincing.  What about the attached patch?

martin


[-- Attachment #2: 954.diff --]
[-- Type: text/plain, Size: 11418 bytes --]

*** textmodes/tex-mode.el.~1.218.~	2008-09-23 10:17:14.531250000 +0200
--- textmodes/tex-mode.el	2008-09-24 09:57:17.203125000 +0200
***************
*** 1482,1487 ****
--- 1482,1489 ----
      (push-mark)
      (goto-char spot)))
  
+ (defvar latex-handle-escaped-parens t)
+ 
  ;; Don't think this one actually _needs_ (for the purposes of
  ;; tex-mode) to handle escaped parens.
  (defun latex-backward-sexp-1 ()
***************
*** 1489,1499 ****
    (let ((pos (point))
  	(forward-sexp-function))
      (backward-sexp 1)
!     (cond ((looking-at "\\\\\\(begin\\>\\|[[({]\\)")
  	   (signal 'scan-error
  		   (list "Containing expression ends prematurely"
  			 (point) (prog1 (point) (goto-char pos)))))
! 	  ((looking-at "\\\\\\([])}]\\)")
  	   (tex-last-unended-eparen (match-string 1)))
  	  ((eq (char-after) ?{)
  	   (let ((newpos (point)))
--- 1491,1505 ----
    (let ((pos (point))
  	(forward-sexp-function))
      (backward-sexp 1)
!     (cond ((looking-at
! 	    (if latex-handle-escaped-parens
! 		"\\\\\\(begin\\>\\|[[({]\\)"
! 	      "\\\\begin\\>"))
  	   (signal 'scan-error
  		   (list "Containing expression ends prematurely"
  			 (point) (prog1 (point) (goto-char pos)))))
! 	  ((and latex-handle-escaped-parens
! 		(looking-at "\\\\\\([])}]\\)"))
  	   (tex-last-unended-eparen (match-string 1)))
  	  ((eq (char-after) ?{)
  	   (let ((newpos (point)))
***************
*** 1528,1539 ****
  	(tex-next-unmatched-end))
         ;; A better way to handle this, \( .. \) etc, is probably to
         ;; temporarily change the syntax of the \ in \( to punctuation.
!        ((looking-back "\\\\[])}]")
  	(signal 'scan-error
  		(list "Containing expression ends prematurely"
  		      (- (point) 2) (prog1 (point)
  				      (goto-char pos)))))
!        ((looking-back "\\\\\\([({[]\\)")
  	(tex-next-unmatched-eparen (match-string 1)))
         (t (goto-char newpos))))))
  
--- 1534,1547 ----
  	(tex-next-unmatched-end))
         ;; A better way to handle this, \( .. \) etc, is probably to
         ;; temporarily change the syntax of the \ in \( to punctuation.
!        ((and latex-handle-escaped-parens
! 	     (looking-back "\\\\[])}]"))
  	(signal 'scan-error
  		(list "Containing expression ends prematurely"
  		      (- (point) 2) (prog1 (point)
  				      (goto-char pos)))))
!        ((and latex-handle-escaped-parens
! 	     (looking-back "\\\\\\([({[]\\)"))
  	(tex-next-unmatched-eparen (match-string 1)))
         (t (goto-char newpos))))))
  
***************
*** 2568,2670 ****
  	    (indent-line-to indent)
  	  (save-excursion (indent-line-to indent)))))))
  
  (defun latex-find-indent (&optional virtual)
    "Find the proper indentation of text after point.
  VIRTUAL if non-nil indicates that we're only trying to find the indentation
    in order to determine the indentation of something else.
  There might be text before point."
!   (save-excursion
!     (skip-chars-forward " \t")
!     (or
!      ;; Stick the first line at column 0.
!      (and (= (point-min) (line-beginning-position)) 0)
!      ;; Trust the current indentation, if such info is applicable.
!      (and virtual (save-excursion (skip-chars-backward " \t&") (bolp))
! 	  (current-column))
!      ;; Stick verbatim environments to the left margin.
!      (and (looking-at "\\\\\\(begin\\|end\\) *{\\([^\n}]+\\)")
! 	  (member (match-string 2) tex-verbatim-environments)
! 	  0)
!      ;; Put leading close-paren where the matching open paren would be.
!      (let (escaped)
!        (and (or (eq (latex-syntax-after) ?\))
! 		;; Try to handle escaped close parens but keep original
! 		;; position if it doesn't work out.
! 		(setq escaped (looking-at "\\\\\\([])}]\\)")))
! 	    (ignore-errors
! 	     (save-excursion
! 	       (when escaped
! 		 (goto-char (match-beginning 1)))
! 	       (latex-skip-close-parens)
  	       (latex-backward-sexp-1)
! 	       (latex-find-indent 'virtual)))))
!      ;; Default (maybe an argument)
!      (let ((pos (point))
! 	   ;; Outdent \item if necessary.
! 	   (indent (if (looking-at tex-indent-item-re) (- tex-indent-item) 0))
! 	   up-list-pos)
!        ;; Find the previous point which determines our current indentation.
!        (condition-case err
! 	   (progn
! 	     (latex-backward-sexp-1)
! 	     (while (> (current-column) (current-indentation))
! 	       (latex-backward-sexp-1)))
! 	 (scan-error
! 	  (setq up-list-pos (nth 2 err))))
!        (cond
! 	((= (point-min) pos) 0) ; We're really just indenting the first line.
! 	((integerp up-list-pos)
! 	 ;; Have to indent relative to the open-paren.
! 	 (goto-char up-list-pos)
! 	 (if (and (not tex-indent-allhanging)
! 		  (save-excursion
! 		    ;; Make sure we're an argument to a macro and
! 		    ;; that the macro is at the beginning of a line.
! 		    (condition-case nil
! 			(progn
! 			  (while (eq (char-syntax (char-after)) ?\()
! 			    (forward-sexp -1))
! 			  (and (eq (char-syntax (char-after)) ?/)
! 			       (progn (skip-chars-backward " \t&")
! 				      (bolp))))
! 		      (scan-error nil)))
! 		  (> pos (progn (latex-down-list)
! 				(forward-comment (point-max))
! 				(point))))
! 		 ;; Align with the first element after the open-paren.
! 	     (current-column)
! 	   ;; We're the first element after a hanging brace.
  	   (goto-char up-list-pos)
! 	   (+ (if (and (looking-at "\\\\begin *{\\([^\n}]+\\)")
! 		       (member (match-string 1)
! 			       latex-noindent-environments))
! 		  0 tex-indent-basic)
! 	      indent (latex-find-indent 'virtual))))
! 	;; We're now at the "beginning" of a line.
! 	((not (and (not virtual) (eq (char-after) ?\\)))
! 	 ;; Nothing particular here: just keep the same indentation.
! 	 (+ indent (current-column)))
! 	;; We're now looking at a macro call.
! 	((looking-at tex-indent-item-re)
! 	 ;; Indenting relative to an item, have to re-add the outdenting.
! 	 (+ indent (current-column) tex-indent-item))
! 	(t
! 	 (let ((col (current-column)))
! 	   (if (or (not (eq (char-syntax (or (char-after pos) ?\s)) ?\())
! 		   ;; Can't be an arg if there's an empty line inbetween.
! 		   (save-excursion (re-search-forward "^[ \t]*$" pos t)))
! 	       ;; If the first char was not an open-paren, there's
! 	       ;; a risk that this is really not an argument to the
! 	       ;; macro at all.
! 	       (+ indent col)
! 	     (forward-sexp 1)
! 	     (if (< (line-end-position)
! 		    (save-excursion (forward-comment (point-max))
! 				    (point)))
! 		 ;; we're indenting the first argument.
! 		 (min (current-column) (+ tex-indent-arg col))
! 	       (skip-syntax-forward " ")
! 	       (current-column))))))))))
  ;;; DocTeX support
  
  (defun doctex-font-lock-^^A ()
--- 2576,2690 ----
  	    (indent-line-to indent)
  	  (save-excursion (indent-line-to indent)))))))
  
+ (defcustom latex-indent-within-escaped-parens nil
+   "Non-nil means add extra indent to text within escaped parens.
+ When this is non-nil, text within matching pairs of escaped
+ parens is indented at the column following the open paren.  The
+ default value does not add any extra indent thus providing the
+ behavior of Emacs 22 and earlier."
+   :type 'boolean
+   :group 'tex
+   :version "23.1")
+ 
  (defun latex-find-indent (&optional virtual)
    "Find the proper indentation of text after point.
  VIRTUAL if non-nil indicates that we're only trying to find the indentation
    in order to determine the indentation of something else.
  There might be text before point."
!   (let ((latex-handle-escaped-parens latex-indent-within-escaped-parens))
!     (save-excursion
!       (skip-chars-forward " \t")
!       (or
!        ;; Stick the first line at column 0.
!        (and (= (point-min) (line-beginning-position)) 0)
!        ;; Trust the current indentation, if such info is applicable.
!        (and virtual (save-excursion (skip-chars-backward " \t&") (bolp))
! 	    (current-column))
!        ;; Stick verbatim environments to the left margin.
!        (and (looking-at "\\\\\\(begin\\|end\\) *{\\([^\n}]+\\)")
! 	    (member (match-string 2) tex-verbatim-environments)
! 	    0)
!        ;; Put leading close-paren where the matching open paren would be.
!        (let (escaped)
! 	 (and (or (eq (latex-syntax-after) ?\))
! 		  ;; Try to handle escaped close parens but keep
! 		  ;; original position if it doesn't work out.
! 		  (and latex-handle-escaped-parens
! 		       (setq escaped (looking-at "\\\\\\([])}]\\)"))))
! 	      (ignore-errors
! 	       (save-excursion
! 		 (when escaped
! 		   (goto-char (match-beginning 1)))
! 		 (latex-skip-close-parens)
! 		 (latex-backward-sexp-1)
! 		 (latex-find-indent 'virtual)))))
!        ;; Default (maybe an argument)
!        (let ((pos (point))
! 	     ;; Outdent \item if necessary.
! 	     (indent (if (looking-at tex-indent-item-re) (- tex-indent-item) 0))
! 	     up-list-pos)
! 	 ;; Find the previous point which determines our current indentation.
! 	 (condition-case err
! 	     (progn
  	       (latex-backward-sexp-1)
! 	       (while (> (current-column) (current-indentation))
! 		 (latex-backward-sexp-1)))
! 	   (scan-error
! 	    (setq up-list-pos (nth 2 err))))
! 	 (cond
! 	  ((= (point-min) pos) 0) ; We're really just indenting the first line.
! 	  ((integerp up-list-pos)
! 	   ;; Have to indent relative to the open-paren.
  	   (goto-char up-list-pos)
! 	   (if (and (not tex-indent-allhanging)
! 		    (save-excursion
! 		      ;; Make sure we're an argument to a macro and
! 		      ;; that the macro is at the beginning of a line.
! 		      (condition-case nil
! 			  (progn
! 			    (while (eq (char-syntax (char-after)) ?\()
! 			      (forward-sexp -1))
! 			    (and (eq (char-syntax (char-after)) ?/)
! 				 (progn (skip-chars-backward " \t&")
! 					(bolp))))
! 			(scan-error nil)))
! 		    (> pos (progn (latex-down-list)
! 				  (forward-comment (point-max))
! 				  (point))))
! 	       ;; Align with the first element after the open-paren.
! 	       (current-column)
! 	     ;; We're the first element after a hanging brace.
! 	     (goto-char up-list-pos)
! 	     (+ (if (and (looking-at "\\\\begin *{\\([^\n}]+\\)")
! 			 (member (match-string 1)
! 				 latex-noindent-environments))
! 		    0 tex-indent-basic)
! 		indent (latex-find-indent 'virtual))))
! 	  ;; We're now at the "beginning" of a line.
! 	  ((not (and (not virtual) (eq (char-after) ?\\)))
! 	   ;; Nothing particular here: just keep the same indentation.
! 	   (+ indent (current-column)))
! 	  ;; We're now looking at a macro call.
! 	  ((looking-at tex-indent-item-re)
! 	   ;; Indenting relative to an item, have to re-add the outdenting.
! 	   (+ indent (current-column) tex-indent-item))
! 	  (t
! 	   (let ((col (current-column)))
! 	     (if (or (not (eq (char-syntax (or (char-after pos) ?\s)) ?\())
! 		     ;; Can't be an arg if there's an empty line inbetween.
! 		     (save-excursion (re-search-forward "^[ \t]*$" pos t)))
! 		 ;; If the first char was not an open-paren, there's
! 		 ;; a risk that this is really not an argument to the
! 		 ;; macro at all.
! 		 (+ indent col)
! 	       (forward-sexp 1)
! 	       (if (< (line-end-position)
! 		      (save-excursion (forward-comment (point-max))
! 				      (point)))
! 		   ;; we're indenting the first argument.
! 		   (min (current-column) (+ tex-indent-arg col))
! 		 (skip-syntax-forward " ")
! 		 (current-column)))))))))))
  ;;; DocTeX support
  
  (defun doctex-font-lock-^^A ()

bug#954: indentation in latex-mode in emacs-23

[Glenn Morris]

martin rudalics wrote:

>  > It's not something I can convince you about. And I don't think there
>  > is any correct or incorrect formatting style. It's just from my 10
>  > years experience. As I get older, it gets harder to accommodate new
>  > things no matter how small they are....
>
> ... sounds convincing.  What about the attached patch?

Sorry to disagree, but "change bad" is not a convincing argument to me.

I prefer the current form because I think that

\begin{math}
  a = b
\end{math}

and

\(
  a = b
\)

should be formatted the same. A comment in the original bug that
motivated this change seems more convincing to me:

http://lists.gnu.org/archive/html/bug-gnu-emacs/2007-09/msg00045.html

    The correct solution would require that \(, \), \[, and \] are
    treated the same as \begin{math}, \end{math}, \begin{displaymath},
    and \end{displaymath}. (The latter are defined in terms of the
    former in latex.ltx.)

But I have to admit that: i) I use AUCTeX; and ii) I don't use \( \),
so take all that with a large grain of salt. :)

So please just check that whatever change is installed does not bring
back the original tex-validate-region problem:

http://lists.gnu.org/archive/html/bug-gnu-emacs/2007-09/msg00038.html

bug#954: indentation in latex-mode in emacs-23

[martin rudalics]

 > I prefer the current form because I think that
 >
 > \begin{math}
 >   a = b
 > \end{math}
 >
 > and
 >
 > \(
 >   a = b
 > \)
 >
 > should be formatted the same.

I think so too.  On the other hand, indenting habits are persistent.
And, according to Leslie, there doesn't seem to exist a convention for
indentation of LaTeX source.

 > A comment in the original bug that
 > motivated this change seems more convincing to me:
 >
 > http://lists.gnu.org/archive/html/bug-gnu-emacs/2007-09/msg00045.html
 >
 >     The correct solution would require that \(, \), \[, and \] are
 >     treated the same as \begin{math}, \end{math}, \begin{displaymath},
 >     and \end{displaymath}. (The latter are defined in terms of the
 >     former in latex.ltx.)
 >
 > But I have to admit that: i) I use AUCTeX; and ii) I don't use \( \),
 > so take all that with a large grain of salt. :)

I don't know - how does AUCTeX indent this?

 > So please just check that whatever change is installed does not bring
 > back the original tex-validate-region problem:
 >
 > http://lists.gnu.org/archive/html/bug-gnu-emacs/2007-09/msg00038.html

Sure.  I strongly hope that `tex-validate-region' doesn't care about
indentation.

martin

bug#954: indentation in latex-mode in emacs-23

[Stefan Monnier]

>> > It's not something I can convince you about. And I don't think there
>> > is any correct or incorrect formatting style. It's just from my 10
>> > years experience. As I get older, it gets harder to accommodate new
>> > things no matter how small they are....
>> 
>> ... sounds convincing.  What about the attached patch?

> Sorry to disagree, but "change bad" is not a convincing argument to me.

> I prefer the current form because I think that

> \begin{math}
>   a = b
> \end{math}

> and

> \(
>   a = b
> \)

> should be formatted the same. A comment in the original bug that

I don't think there's a right answer.  But at least if the above two
alternative forms indent differently, that makes it possible for the
user to choose the indentation (indirectly) by choosing which form
to use.
So I'd argue that the two should *not* for indented identically.


        Stefan


PS: The indentation in tex-mode (for which I largely plead guilty)
usually indents a good bit more than what I've seen used in most LaTeX
files.  I tend to think of LaTeX as a programming language, and most of
what I write is LaTeX is about programming languages, so I like to
indent it more than people who think of it as "text with markup".

bug#954: indentation in latex-mode in emacs-23

[Chirok Han]

Hi Glenn and Martin,

I have no objections to Glenn's argument. And also many thanks to
Martin; I can now figure out which part to modify for
customization.... :)

Chirok

PS. Just one correction. I think Glenn meant
\[
  abc
\]
instead of
\(
  abc
\)

On Thu, Sep 25, 2008 at 1:48 AM, Glenn Morris <rgm@gnu.org> wrote:
> martin rudalics wrote:
>
>>  > It's not something I can convince you about. And I don't think there
>>  > is any correct or incorrect formatting style. It's just from my 10
>>  > years experience. As I get older, it gets harder to accommodate new
>>  > things no matter how small they are....
>>
>> ... sounds convincing.  What about the attached patch?
>
> Sorry to disagree, but "change bad" is not a convincing argument to me.
>
> I prefer the current form because I think that
>
> \begin{math}
>  a = b
> \end{math}
>
> and
>
> \(
>  a = b
> \)
>
> should be formatted the same. A comment in the original bug that
> motivated this change seems more convincing to me:
>
> http://lists.gnu.org/archive/html/bug-gnu-emacs/2007-09/msg00045.html
>
>    The correct solution would require that \(, \), \[, and \] are
>    treated the same as \begin{math}, \end{math}, \begin{displaymath},
>    and \end{displaymath}. (The latter are defined in terms of the
>    former in latex.ltx.)
>
> But I have to admit that: i) I use AUCTeX; and ii) I don't use \( \),
> so take all that with a large grain of salt. :)
>
> So please just check that whatever change is installed does not bring
> back the original tex-validate-region problem:
>
> http://lists.gnu.org/archive/html/bug-gnu-emacs/2007-09/msg00038.html
>