GNU bug report logs - #16160
[PATCH] define-derived-mode clobbers syntax tables

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 16160 in the body.
You can then email your comments to 16160 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: Daniel Colascione <dancol <at> dancol.org>
To: bug-emacs <bug-gnu-emacs <at> gnu.org>
Subject: [PATCH] define-derived-mode clobbers syntax tables
Date: Mon, 16 Dec 2013 03:12:43 -0800

Repro:

1) Visit shell script with comments. See comments fontified properly.
2) M-x find-library sh-script RET
3) M-x eval-buffer
4) Return to shell script buffer
5) C-x C-v RET
6) Observe that comments are no longer fontified as comments.

The define-derived-mode macro in sh-script.el emits code that clobbers 
sh-mode-syntax-table because we don't have a :syntax-table argument. I 
think the following patch changes the code to the expected behavior.

~/edev/trunk
$ bzr diff
=== modified file 'lisp/emacs-lisp/derived.el'
--- lisp/emacs-lisp/derived.el	2013-05-27 16:12:52 +0000
+++ lisp/emacs-lisp/derived.el	2013-12-16 11:09:41 +0000
@@ -206,11 +206,11 @@
        ,(if declare-syntax
 	    `(progn
 	       (unless (boundp ',syntax)
-		 (put ',syntax 'definition-name ',child))
-	       (defvar ,syntax (make-syntax-table))
-	       (unless (get ',syntax 'variable-documentation)
-		 (put ',syntax 'variable-documentation
-		      (purecopy ,(format "Syntax table for `%s'." child))))))
+		 (put ',syntax 'definition-name ',child)
+                 (defvar ,syntax (make-syntax-table))
+                 (unless (get ',syntax 'variable-documentation)
+                   (put ',syntax 'variable-documentation
+                        (purecopy ,(format "Syntax table for `%s'." 
child)))))))
        ,(if declare-abbrev
 	    `(progn
 	       (put ',abbrev 'definition-name ',child)


That is, we shouldn't touch the syntax table variable at all unless the 
variable is unbound.

Message #8 received at 16160 <at> debbugs.gnu.org (full text, mbox):

From: Stefan Monnier <monnier <at> IRO.UMontreal.CA>
To: Daniel Colascione <dancol <at> dancol.org>
Cc: 16160 <at> debbugs.gnu.org
Subject: Re: bug#16160: [PATCH] define-derived-mode clobbers syntax tables
Date: Mon, 16 Dec 2013 14:17:19 -0500

> 1) Visit shell script with comments. See comments fontified properly.
> 2) M-x find-library sh-script RET
> 3) M-x eval-buffer
> 4) Return to shell script buffer
> 5) C-x C-v RET
> 6) Observe that comments are no longer fontified as comments.

I can't reproduce it, starting from "emacs -Q".

> -		 (put ',syntax 'definition-name ',child))
> -	       (defvar ,syntax (make-syntax-table))
> -	       (unless (get ',syntax 'variable-documentation)
> -		 (put ',syntax 'variable-documentation
> -		      (purecopy ,(format "Syntax table for `%s'." child))))))
> +		 (put ',syntax 'definition-name ',child)
> +                 (defvar ,syntax (make-syntax-table))
> +                 (unless (get ',syntax 'variable-documentation)
> +                   (put ',syntax 'variable-documentation
> +                        (purecopy ,(format "Syntax table for `%s'." child)))))))
> That is, we shouldn't touch the syntax table variable at all unless the
> variable is unbound.

But that shouldn't make much difference since none of the code you
changed should affect the var's value when it is already bound.
I'm not necessarily opposed to the change (haven't dug enough to try and
remember why it's written this way), but I'd first like to understand
why it fixes the problem you see.


        Stefan

Message #11 received at 16160 <at> debbugs.gnu.org (full text, mbox):

From: Daniel Colascione <dancol <at> dancol.org>
To: Stefan Monnier <monnier <at> IRO.UMontreal.CA>
Cc: 16160 <at> debbugs.gnu.org
Subject: Re: bug#16160: [PATCH] define-derived-mode clobbers syntax tables
Date: Mon, 16 Dec 2013 18:17:55 -0800

On 12/16/2013 11:17 AM, Stefan Monnier wrote:
>> 1) Visit shell script with comments. See comments fontified properly.
>> 2) M-x find-library sh-script RET
>> 3) M-x eval-buffer
>> 4) Return to shell script buffer
>> 5) C-x C-v RET
>> 6) Observe that comments are no longer fontified as comments.
>
> I can't reproduce it, starting from "emacs -Q".

I can't repro it either now, with or without -Q. That was odd. I know 
what I saw, and I know that my patch resolved the problem in that 
instance of Emacs, but I have no idea how defvar was clobbering that 
variable. I was able to repro it over an dover again though. I can only 
conclude that Emacs was punishing me for my many sins.

If I see it again, I'll try to hook Emacs up to a debugger. In the 
meantime, let's avoid changing fundamental infrastructure before a 
feature freeze.

Message #14 received at 16160 <at> debbugs.gnu.org (full text, mbox):

From: Stefan Monnier <monnier <at> IRO.UMontreal.CA>
To: Daniel Colascione <dancol <at> dancol.org>
Cc: 16160 <at> debbugs.gnu.org
Subject: Re: bug#16160: [PATCH] define-derived-mode clobbers syntax tables
Date: Mon, 16 Dec 2013 21:36:28 -0500

> I can only conclude that Emacs was punishing me for my many sins.

I believe He may forgive you, if you fix a few bugs for penance.


        Stefan

Message #17 received at 16160 <at> debbugs.gnu.org (full text, mbox):

From: Daniel Colascione <dancol <at> dancol.org>
To: Stefan Monnier <monnier <at> IRO.UMontreal.CA>
Cc: 16160 <at> debbugs.gnu.org
Subject: Re: bug#16160: [PATCH] define-derived-mode clobbers syntax tables
Date: Fri, 20 Dec 2013 13:45:43 -0800

Actually...

On 12/16/2013 06:17 PM, Daniel Colascione wrote:
> On 12/16/2013 11:17 AM, Stefan Monnier wrote:
>>> 1) Visit shell script with comments. See comments fontified properly.
>>> 2) M-x find-library sh-script RET
>>> 3) M-x eval-buffer
>>> 4) Return to shell script buffer
>>> 5) C-x C-v RET
>>> 6) Observe that comments are no longer fontified as comments.

The repro steps are incorrect. Try these:

1) emacs -Q
2) visit a shell script
3) M-x find-library sh-script RET
4) C-s define-derived-mode RET
5) C-M-x
6) switch to your shell script buffer
7) C-x C-v
8) Observe that the shell script is fontified incorrectly; the syntax 
table is now clobbered.

Message #22 received at 16160 <at> debbugs.gnu.org (full text, mbox):

From: Lars Ingebrigtsen <larsi <at> gnus.org>
To: Daniel Colascione <dancol <at> dancol.org>
Cc: 16160 <at> debbugs.gnu.org, Stefan Monnier <monnier <at> IRO.UMontreal.CA>
Subject: Re: bug#16160: [PATCH] define-derived-mode clobbers syntax tables
Date: Tue, 01 Mar 2016 13:05:13 +1100

Daniel Colascione <dancol <at> dancol.org> writes:

> The repro steps are incorrect. Try these:
>
> 1) emacs -Q
> 2) visit a shell script
> 3) M-x find-library sh-script RET
> 4) C-s define-derived-mode RET
> 5) C-M-x
> 6) switch to your shell script buffer
> 7) C-x C-v
> 8) Observe that the shell script is fontified incorrectly; the syntax
> table is now clobbered.

I can confirm that this bug still exists.

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

Message #31 received at 16160-done <at> debbugs.gnu.org (full text, mbox):

From: Glenn Morris <rgm <at> gnu.org>
To: 16160-done <at> debbugs.gnu.org
Subject: Re: bug#16160: [PATCH] define-derived-mode clobbers syntax tables
Date: Mon, 12 Dec 2016 20:04:52 -0500

Version: 26.1

>> The repro steps are incorrect. Try these:
>>
>> 1) emacs -Q
>> 2) visit a shell script
>> 3) M-x find-library sh-script RET
>> 4) C-s define-derived-mode RET
>> 5) C-M-x
>> 6) switch to your shell script buffer
>> 7) C-x C-v
>> 8) Observe that the shell script is fontified incorrectly; the syntax
>> table is now clobbered.

Fixed in 8db7b65 along the suggested lines.

So AIUI the issue is that define-derived-mode expands to 

(defvar ,syntax (make-syntax-table))

Normally this has no effect if ,syntax is already bound.
But M-x eval-defun resets defvars to their default settings...
Hence if you selectively re-evaluate only the mode definition from
sh-script.el, you get the reported problem.

So I changed it to explicitly check if ,syntax is bound.

This will however now make things go wrong in the opposite way for
someone who has been playing around with customizing a syntax table that
was actually defined by define-derived-mode and wants to reset it by
re-evaluating the derived mode definition. But AFAICS there's no way to
fix both scenarios, and the second seems less likely (?) to me than the
one reported here.

Previous Next

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