* bug#28003: 26.0.50; Auto fill broken in Message mode
@ 2017-08-07 19:56 Stephen Berman
2017-08-09 3:33 ` Katsumi Yamaoka
2017-08-27 16:55 ` bug#28003: done Tom Tromey
0 siblings, 2 replies; 11+ messages in thread
From: Stephen Berman @ 2017-08-07 19:56 UTC (permalink / raw)
To: 28003
This change:
commit 9b463fa8648b7baed95a44f4317cb7402fd8bf1c
Author: Tom Tromey <tom@tromey.com>
Date: Sat Aug 5 18:30:52 2017 -0600
Respect comment-auto-fill-only-comments
Respect comment-auto-fill-only-comments when auto-filling and a
comment syntax is defined.
* lisp/newcomment.el (comment-indent-new-line): Do not check
comment-auto-fill-only-comments.
* lisp/simple.el (internal-auto-fill): New defun.
* src/cmds.c (internal_self_insert): Call Qinternal_auto_fill, not
auto_fill_function.
(syms_of_cmds): Define Qinternal_auto_fill.
broke auto filling in Message mode; the breakage is due specifically to
this change in comment-indent-new-line:
diff --git a/lisp/newcomment.el b/lisp/newcomment.el
index 8772b52376..e3ee4dfab1 100644
--- a/lisp/newcomment.el
+++ b/lisp/newcomment.el
@@ -1382,10 +1382,9 @@ comment-indent-new-line
(interactive)
(comment-normalize-vars t)
(let (compos comin)
- ;; If we are not inside a comment and we only auto-fill comments,
- ;; don't do anything (unless no comment syntax is defined).
+ ;; If we are not inside a comment don't do anything (unless no
+ ;; comment syntax is defined).
(unless (and comment-start
- comment-auto-fill-only-comments
(not (called-interactively-p 'interactive))
(not (save-excursion
(prog1 (setq compos (comment-beginning))
To reproduce:
0. emacs -Q
1. Type `C-x m' to get a Message mode buffer, with a fill-column of 70.
2. Insert at point-max this text (all in one unbroken line):
This is a test. This is a test. This is a test. This is a test. This is a test.
3. Point is now in column 79; type SPC.
=> The line is not broken and filled.
4. Delete the last SPC, type `C-a > SPC C-e SPC'.
=> This line is now broken and filled.
Message mode sets "> " and comment-auto-fill-only-comments to nil by
default, but since the above change eliminated checking the latter,
after step 3 the unless condition evaluates to t, so there's no
auto-filling; while after step 4, (comment-beginning) is non-nil, so the
unless condition is nil and auto-filling occurs.
In GNU Emacs 26.0.50 (build 2, x86_64-pc-linux-gnu, GTK+ Version 3.22.8)
of 2017-08-07 built on rosalinde
Repository revision: 14ea76af5f3596d48747c2437006f6e1abcb67a7
Windowing system distributor 'The X.Org Foundation', version 11.0.11901000
^ permalink raw reply related [flat|nested] 11+ messages in thread* bug#28003: 26.0.50; Auto fill broken in Message mode
2017-08-07 19:56 bug#28003: 26.0.50; Auto fill broken in Message mode Stephen Berman
@ 2017-08-09 3:33 ` Katsumi Yamaoka
2017-08-09 8:58 ` Stephen Berman
2017-08-27 16:55 ` bug#28003: done Tom Tromey
1 sibling, 1 reply; 11+ messages in thread
From: Katsumi Yamaoka @ 2017-08-09 3:33 UTC (permalink / raw)
To: Stephen Berman; +Cc: 28003
On Mon, 07 Aug 2017 21:56:42 +0200, Stephen Berman wrote:
> This change:
[...]
> broke auto filling in Message mode; the breakage is due specifically to
> this change in comment-indent-new-line:
This will probably be no more than a workaround, though.
--- message.el~ 2017-08-08 22:08:28.694840400 +0000
+++ message.el 2017-08-09 03:30:33.407170200 +0000
@@ -3451,7 +3451,7 @@
(defun message-do-auto-fill ()
"Like `do-auto-fill', but don't fill in message header."
(unless (message-point-in-header-p)
- (do-auto-fill)))
+ (let ((comment-start nil)) (do-auto-fill))))
(defun message-insert-signature (&optional force)
"Insert a signature. See documentation for variable `message-signature'."
^ permalink raw reply [flat|nested] 11+ messages in thread
* bug#28003: 26.0.50; Auto fill broken in Message mode
2017-08-09 3:33 ` Katsumi Yamaoka
@ 2017-08-09 8:58 ` Stephen Berman
2017-08-09 12:37 ` Tom Tromey
0 siblings, 1 reply; 11+ messages in thread
From: Stephen Berman @ 2017-08-09 8:58 UTC (permalink / raw)
To: Katsumi Yamaoka; +Cc: Tom Tromey, 28003
On Wed, 09 Aug 2017 12:33:01 +0900 Katsumi Yamaoka <yamaoka@jpl.org> wrote:
> On Mon, 07 Aug 2017 21:56:42 +0200, Stephen Berman wrote:
>> This change:
> [...]
>> broke auto filling in Message mode; the breakage is due specifically to
>> this change in comment-indent-new-line:
>
> This will probably be no more than a workaround, though.
>
> --- message.el~ 2017-08-08 22:08:28.694840400 +0000
> +++ message.el 2017-08-09 03:30:33.407170200 +0000
> @@ -3451,7 +3451,7 @@
> (defun message-do-auto-fill ()
> "Like `do-auto-fill', but don't fill in message header."
> (unless (message-point-in-header-p)
> - (do-auto-fill)))
> + (let ((comment-start nil)) (do-auto-fill))))
>
> (defun message-insert-signature (&optional force)
> "Insert a signature. See documentation for variable `message-signature'."
This change doesn't fix the problem. I don't know what the motivation
for the change that caused this problem was, so I don't know how to fix
it; cc'ing Tom Tromey, who made the change.
Steve Berman
^ permalink raw reply [flat|nested] 11+ messages in thread
* bug#28003: 26.0.50; Auto fill broken in Message mode
2017-08-09 8:58 ` Stephen Berman
@ 2017-08-09 12:37 ` Tom Tromey
2017-08-09 13:04 ` Stephen Berman
0 siblings, 1 reply; 11+ messages in thread
From: Tom Tromey @ 2017-08-09 12:37 UTC (permalink / raw)
To: Stephen Berman; +Cc: Katsumi Yamaoka, Tom Tromey, 28003
>>>>> "Stephen" == Stephen Berman <stephen.berman@gmx.net> writes:
Stephen> On Wed, 09 Aug 2017 12:33:01 +0900 Katsumi Yamaoka <yamaoka@jpl.org> wrote:
>> On Mon, 07 Aug 2017 21:56:42 +0200, Stephen Berman wrote:
>>> This change:
>> [...]
>>> broke auto filling in Message mode; the breakage is due specifically to
>>> this change in comment-indent-new-line:
>>
>> This will probably be no more than a workaround, though.
>>
>> --- message.el~ 2017-08-08 22:08:28.694840400 +0000
>> +++ message.el 2017-08-09 03:30:33.407170200 +0000
>> @@ -3451,7 +3451,7 @@
>> (defun message-do-auto-fill ()
>> "Like `do-auto-fill', but don't fill in message header."
>> (unless (message-point-in-header-p)
>> - (do-auto-fill)))
>> + (let ((comment-start nil)) (do-auto-fill))))
>>
>> (defun message-insert-signature (&optional force)
>> "Insert a signature. See documentation for variable `message-signature'."
Stephen> This change doesn't fix the problem. I don't know what the motivation
Stephen> for the change that caused this problem was, so I don't know how to fix
Stephen> it; cc'ing Tom Tromey, who made the change.
I've been discussing this off-list with another person who ran into this
problem, but I haven't managed to reproduce it myself.
I suspect the newcomment.el change is incorrect and should be reverted.
Could you try that? The other changes should, I think, remain in place though.
The basic idea of my patch is to make comment-auto-fill-only-comments
work properly. Without it, setting comment-auto-fill-only-comments and
enabling auto-fill in js-mode will cause auto-filling even when typing a
long line of code, which is clearly incorrect.
thanks,
Tom
^ permalink raw reply [flat|nested] 11+ messages in thread
* bug#28003: 26.0.50; Auto fill broken in Message mode
2017-08-09 12:37 ` Tom Tromey
@ 2017-08-09 13:04 ` Stephen Berman
2017-08-09 14:54 ` Stephen Berman
0 siblings, 1 reply; 11+ messages in thread
From: Stephen Berman @ 2017-08-09 13:04 UTC (permalink / raw)
To: Tom Tromey; +Cc: Katsumi Yamaoka, Stephen Berman, 28003
On Wed, 09 Aug 2017 06:37:42 -0600 Tom Tromey <tom@tromey.com> wrote:
>>>>>> "Stephen" == Stephen Berman <stephen.berman@gmx.net> writes:
>
> Stephen> On Wed, 09 Aug 2017 12:33:01 +0900 Katsumi Yamaoka <yamaoka@jpl.org> wrote:
>>> On Mon, 07 Aug 2017 21:56:42 +0200, Stephen Berman wrote:
>>>> This change:
>>> [...]
>>>> broke auto filling in Message mode; the breakage is due specifically to
>>>> this change in comment-indent-new-line:
>>>
>>> This will probably be no more than a workaround, though.
>>>
>>> --- message.el~ 2017-08-08 22:08:28.694840400 +0000
>>> +++ message.el 2017-08-09 03:30:33.407170200 +0000
>>> @@ -3451,7 +3451,7 @@
>>> (defun message-do-auto-fill ()
>>> "Like `do-auto-fill', but don't fill in message header."
>>> (unless (message-point-in-header-p)
>>> - (do-auto-fill)))
>>> + (let ((comment-start nil)) (do-auto-fill))))
>>>
>>> (defun message-insert-signature (&optional force)
>>> "Insert a signature. See documentation for variable `message-signature'."
>
> Stephen> This change doesn't fix the problem. I don't know what the motivation
> Stephen> for the change that caused this problem was, so I don't know how to fix
> Stephen> it; cc'ing Tom Tromey, who made the change.
>
> I've been discussing this off-list with another person who ran into this
> problem, but I haven't managed to reproduce it myself.
Strange; you did try the recipe of my OP? That shows the problem for me
every time (with emacs -Q and well as with my initializations). But see
below...
> I suspect the newcomment.el change is incorrect and should be reverted.
> Could you try that? The other changes should, I think, remain in place though.
Yes, restoring the variable comment-auto-fill-only-comments in
comment-indent-new-line restores auto-filling in Message mode.
> The basic idea of my patch is to make comment-auto-fill-only-comments
> work properly. Without it, setting comment-auto-fill-only-comments and
> enabling auto-fill in js-mode will cause auto-filling even when typing a
> long line of code, which is clearly incorrect.
I haven't looked at js-mode, but it seems strange that setting
comment-auto-fill-only-comments to t there does auto-filling of
non-commented lines...
This prompted me to look closer at Katsumi Yamaoka's patch above, and it
turns out that message-do-auto-fill is not being called, although it is
the value of normal-auto-fill-function in the Message mode buffer I'm
typing in right now; instead, the default value do-auto-fill is called.
When you try my recipe is message-do-auto-fill being used? That may
explain why you can't reproduce the problem. I have no idea why
message-do-auto-fill isn't being called here...
Steve Berman
^ permalink raw reply [flat|nested] 11+ messages in thread
* bug#28003: 26.0.50; Auto fill broken in Message mode
2017-08-09 13:04 ` Stephen Berman
@ 2017-08-09 14:54 ` Stephen Berman
2017-08-09 15:41 ` Stephen Berman
0 siblings, 1 reply; 11+ messages in thread
From: Stephen Berman @ 2017-08-09 14:54 UTC (permalink / raw)
To: Tom Tromey; +Cc: Katsumi Yamaoka, 28003
On Wed, 09 Aug 2017 15:04:56 +0200 Stephen Berman <stephen.berman@gmx.net> wrote:
> This prompted me to look closer at Katsumi Yamaoka's patch above, and it
> turns out that message-do-auto-fill is not being called, although it is
> the value of normal-auto-fill-function in the Message mode buffer I'm
> typing in right now; instead, the default value do-auto-fill is called.
> When you try my recipe is message-do-auto-fill being used? That may
> explain why you can't reproduce the problem. I have no idea why
> message-do-auto-fill isn't being called here...
It's not being called because internal-auto-fill calls do-auto-fill
instead of funcalling auto-fill-function. But changing it to do the
latter makes no difference, since message-do-auto-fill also calls
do-auto-fill, unless point is on a message header. (So it's still a
mystery to me why you can't reproduce the problem...)
Steve Berman
^ permalink raw reply [flat|nested] 11+ messages in thread
* bug#28003: 26.0.50; Auto fill broken in Message mode
2017-08-09 14:54 ` Stephen Berman
@ 2017-08-09 15:41 ` Stephen Berman
2017-08-09 22:08 ` Tom Tromey
0 siblings, 1 reply; 11+ messages in thread
From: Stephen Berman @ 2017-08-09 15:41 UTC (permalink / raw)
To: Stephen Berman; +Cc: Katsumi Yamaoka, Tom Tromey, 28003
On Wed, 09 Aug 2017 16:54:27 +0200 Stephen Berman <stephen.berman@gmx.net> wrote:
> On Wed, 09 Aug 2017 15:04:56 +0200 Stephen Berman <stephen.berman@gmx.net> wrote:
>
>> This prompted me to look closer at Katsumi Yamaoka's patch above, and it
>> turns out that message-do-auto-fill is not being called, although it is
>> the value of normal-auto-fill-function in the Message mode buffer I'm
>> typing in right now; instead, the default value do-auto-fill is called.
>> When you try my recipe is message-do-auto-fill being used? That may
>> explain why you can't reproduce the problem. I have no idea why
>> message-do-auto-fill isn't being called here...
>
> It's not being called because internal-auto-fill calls do-auto-fill
> instead of funcalling auto-fill-function. But changing it to do the
> latter makes no difference, since message-do-auto-fill also calls
> do-auto-fill, unless point is on a message header. (So it's still a
> mystery to me why you can't reproduce the problem...)
Well, I'm a bit slow today, but making internal-auto-fill funcall
auto-fill-function together with Yamaoka-san's patch does restore
auto-filling to non-comment lines in Message mode; though I can't judge
whether that's the right fix.
Steve Berman
^ permalink raw reply [flat|nested] 11+ messages in thread
* bug#28003: 26.0.50; Auto fill broken in Message mode
2017-08-09 15:41 ` Stephen Berman
@ 2017-08-09 22:08 ` Tom Tromey
2017-08-09 23:06 ` Katsumi Yamaoka
0 siblings, 1 reply; 11+ messages in thread
From: Tom Tromey @ 2017-08-09 22:08 UTC (permalink / raw)
To: Stephen Berman; +Cc: Katsumi Yamaoka, Tom Tromey, 28003
>>>>> "Stephen" == Stephen Berman <stephen.berman@gmx.net> writes:
Stephen> Well, I'm a bit slow today, but making internal-auto-fill funcall
Stephen> auto-fill-function together with Yamaoka-san's patch does restore
Stephen> auto-filling to non-comment lines in Message mode; though I can't judge
Stephen> whether that's the right fix.
Yes, I think so. I found a way to reproduce the bug (thanks to Yuri
D'Elia) and I am going to push the fix now.
Tom
^ permalink raw reply [flat|nested] 11+ messages in thread
* bug#28003: 26.0.50; Auto fill broken in Message mode
2017-08-09 22:08 ` Tom Tromey
@ 2017-08-09 23:06 ` Katsumi Yamaoka
2017-08-10 8:07 ` Stephen Berman
0 siblings, 1 reply; 11+ messages in thread
From: Katsumi Yamaoka @ 2017-08-09 23:06 UTC (permalink / raw)
To: Tom Tromey; +Cc: Stephen Berman, 28003
On Wed, 09 Aug 2017 16:08:31 -0600, Tom Tromey wrote:
> Yes, I think so. I found a way to reproduce the bug (thanks to Yuri
> D'Elia) and I am going to push the fix now.
Works in the message mode as before. Thank you!
^ permalink raw reply [flat|nested] 11+ messages in thread
* bug#28003: 26.0.50; Auto fill broken in Message mode
2017-08-09 23:06 ` Katsumi Yamaoka
@ 2017-08-10 8:07 ` Stephen Berman
0 siblings, 0 replies; 11+ messages in thread
From: Stephen Berman @ 2017-08-10 8:07 UTC (permalink / raw)
To: Katsumi Yamaoka; +Cc: Tom Tromey, 28003
On Thu, 10 Aug 2017 08:06:23 +0900 Katsumi Yamaoka <yamaoka@jpl.org> wrote:
> On Wed, 09 Aug 2017 16:08:31 -0600, Tom Tromey wrote:
>> Yes, I think so. I found a way to reproduce the bug (thanks to Yuri
>> D'Elia) and I am going to push the fix now.
>
> Works in the message mode as before. Thank you!
Works again for me, too; thanks.
Steve Berman
^ permalink raw reply [flat|nested] 11+ messages in thread
* bug#28003: done
2017-08-07 19:56 bug#28003: 26.0.50; Auto fill broken in Message mode Stephen Berman
2017-08-09 3:33 ` Katsumi Yamaoka
@ 2017-08-27 16:55 ` Tom Tromey
1 sibling, 0 replies; 11+ messages in thread
From: Tom Tromey @ 2017-08-27 16:55 UTC (permalink / raw)
To: 28003-done
The fix was checked in a while ago.
Tom
^ permalink raw reply [flat|nested] 11+ messages in thread
end of thread, other threads:[~2017-08-27 16:55 UTC | newest]
Thread overview: 11+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2017-08-07 19:56 bug#28003: 26.0.50; Auto fill broken in Message mode Stephen Berman
2017-08-09 3:33 ` Katsumi Yamaoka
2017-08-09 8:58 ` Stephen Berman
2017-08-09 12:37 ` Tom Tromey
2017-08-09 13:04 ` Stephen Berman
2017-08-09 14:54 ` Stephen Berman
2017-08-09 15:41 ` Stephen Berman
2017-08-09 22:08 ` Tom Tromey
2017-08-09 23:06 ` Katsumi Yamaoka
2017-08-10 8:07 ` Stephen Berman
2017-08-27 16:55 ` bug#28003: done Tom Tromey
Code repositories for project(s) associated with this public inbox
https://git.savannah.gnu.org/cgit/emacs.git
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).