bug#11165: 24.0.95; c-indent-new-comment-line bug or feature?

bug#11165: 24.0.95; c-indent-new-comment-line bug or feature?

[Leo]

1. Emacs -q
2. M-x c-mode
3. customise the indentation for syntactic symbol c to 4
4. Enter /* and then M-j (c-indent-new-comment-line)

you should see point located at column 2 instead of 4, which is a wrong
indentation position for that line and unlike indent-new-comment-line or
newline-and-indent.

Is this a bug or feature? Thanks.

Leo

bug#11165: 24.0.95; c-indent-new-comment-line bug or feature?

[Alan Mackenzie]

> 1. Emacs -q
> 2. M-x c-mode
> 3. customise the indentation for syntactic symbol c to 4
> 4. Enter /* and then M-j (c-indent-new-comment-line)

I think you've missed out some steps here.  When I do exactly that, I do
M-x c-mode on the opening splash screen (which isn't sensible), which is
read-only, preventing step 4.

What I tried was, instead of your step 4, creating a new buffer with C-x
b asdf.c, then setting the mode with M-x c-mode.

> you should see point located at column 2 instead of 4, which is a wrong
> indentation position for that line and unlike indent-new-comment-line
> or newline-and-indent.

I do indeed see point at column 2, which is correct.  I think you might
have set the offset for "c" to 4 before creating the empty C buffer.
Indentation configuration is buffer local.

When I then repeat the C-c C-o c <ret> 4 <ret> I then get the second line
at offset 4, where it ought to be.

> Is this a bug or feature? Thanks.

Could you please indicate exactly what you did.  I suspect this is
neither bug nor feature.  :-)  It would be good to get this bug report
closed off.

Leo

-- 
Alan Mackenzie (Nuremberg, Germany).

bug#11165: 24.0.95; c-indent-new-comment-line bug or feature?

[Leo]

On 2012-04-17 18:53 +0800, Alan Mackenzie wrote:
> I think you've missed out some steps here.  When I do exactly that, I do
> M-x c-mode on the opening splash screen (which isn't sensible), which is
> read-only, preventing step 4.
>
> What I tried was, instead of your step 4, creating a new buffer with C-x
> b asdf.c, then setting the mode with M-x c-mode.

Sorry for the splash screen. What you did instead is correct.

> I do indeed see point at column 2, which is correct.  I think you might
> have set the offset for "c" to 4 before creating the empty C buffer.
> Indentation configuration is buffer local.
>
> When I then repeat the C-c C-o c <ret> 4 <ret> I then get the second line
> at offset 4, where it ought to be.

I customise "c" after entering C mode so the customisation is effective.
If you type M-j after /*, point is moved to the newline with 2 leading
spaces. Now if you type TAB, it will go to column 4. so M-j is
misbehaving i.e. it doesn't go to the right indentation.

In contrast, M-x indent-new-comment-line following /* will move point to
column 4 in the next line.

Leo

bug#11165: 24.0.95; c-indent-new-comment-line bug or feature?

[Alan Mackenzie]

Hello, Leo.

On Tue, Apr 17, 2012 at 09:57:52PM +0800, Leo wrote:
> On 2012-04-17 18:53 +0800, Alan Mackenzie wrote:
> > I think you've missed out some steps here.  When I do exactly that, I do
> > M-x c-mode on the opening splash screen (which isn't sensible), which is
> > read-only, preventing step 4.

> > What I tried was, instead of your step 4, creating a new buffer with C-x
> > b asdf.c, then setting the mode with M-x c-mode.

> Sorry for the splash screen. What you did instead is correct.

> > I do indeed see point at column 2, which is correct.  I think you might
> > have set the offset for "c" to 4 before creating the empty C buffer.
> > Indentation configuration is buffer local.

> > When I then repeat the C-c C-o c <ret> 4 <ret> I then get the second line
> > at offset 4, where it ought to be.

> I customise "c" after entering C mode so the customisation is effective.
> If you type M-j after /*, point is moved to the newline with 2 leading
> spaces. Now if you type TAB, it will go to column 4. so M-j is
> misbehaving i.e. it doesn't go to the right indentation.

> In contrast, M-x indent-new-comment-line following /* will move point to
> column 4 in the next line.

This is getting confusing.  I'm asking you at this point to read the
chapter "Config Basics" of the CC Mode manual.  About 36 lines down from
the top, it explains how setq'ing/customizing CC Mode's style variables
interacts with loading a buffer.

I don't think this is a bug.  When you've read that manual page, would
you please check exactly what you've done, and whether your buffer
indents as we both would like.

If it still seems wrong, could you give the exact description of how to
create the problem starting from emacs -Q (yes, I know this'll involve
some repetition).

Thanks!

> Leo

-- 
Alan Mackenzie (Nuremberg, Germany).

bug#11165: 24.0.95; c-indent-new-comment-line bug or feature?

[Leo]

[-- Attachment #1: Type: text/plain, Size: 565 bytes --]

On 2012-05-16 00:50 +0800, Alan Mackenzie wrote:
> I don't think this is a bug.  When you've read that manual page, would
> you please check exactly what you've done, and whether your buffer
> indents as we both would like.

I have read the manual before and I just read it again. Here is a
detailed recipe to reproduce:

1. Emacs -q
2. Make a new buffer named BUG and enable c-mode in it
3. C-c C-o and customise syntactic 'c' to 4 in BUG buffer
4. Now typing in the BUG buffer slash star /*, then press M-j
5. type TAB to indent

In step 4 the buffer looks like:

[-- Attachment #2: bug1.png --]
[-- Type: image/png, Size: 8322 bytes --]

[-- Attachment #3: Type: text/plain, Size: 34 bytes --]


In step 5 the buffer looks like:

[-- Attachment #4: bug2.png --]
[-- Type: image/png, Size: 8301 bytes --]

[-- Attachment #5: Type: text/plain, Size: 73 bytes --]


The bug is M-j (c-indent-new-comment-line) indents to the wrong column.

bug#11165: 24.0.95; c-indent-new-comment-line bug or feature?

[Alan Mackenzie]

Hi, Leo.

On Thu, May 17, 2012 at 03:19:56PM +0800, Leo wrote:
> On 2012-05-16 00:50 +0800, Alan Mackenzie wrote:
> > I don't think this is a bug.  When you've read that manual page, would
> > you please check exactly what you've done, and whether your buffer
> > indents as we both would like.

> I have read the manual before and I just read it again. Here is a
> detailed recipe to reproduce:

> 1. Emacs -q
> 2. Make a new buffer named BUG and enable c-mode in it
> 3. C-c C-o and customise syntactic 'c' to 4 in BUG buffer
> 4. Now typing in the BUG buffer slash star /*, then press M-j
> 5. type TAB to indent

Thanks for this clarification.

[ .... ]

> The bug is M-j (c-indent-new-comment-line) indents to the wrong column.

Yes, indeed.  In c-indent-new-comment-line there are some lines of code
which prevent point ending up further to the right than its starting
point.  This was to fix a bug in which, after M-; M-j, point ended up
too far to the right.  The fix was erroneous.

This code was committed on 2007-02-15, so it doesn't count as a
regression in the sense of the Emacs 24 release.  So even though the fix
shouldn't be too difficult, it isn't the most urgent thing at the
moment.  I'll get it fixed in CC Mode standalone and the Emacs trunk
sometime.

In the meantime, a workaround is to type <tab> after each M-j,
irritating though this may be.

-- 
Alan Mackenzie (Nuremberg, Germany).

bug#11165: 24.0.95; c-indent-new-comment-line bug or feature?

[Alan Mackenzie]

Hello, Leo.

On Wed, Apr 04, 2012 at 02:30:32 +0800, Leo wrote:
> 1. Emacs -q
> 2. M-x c-mode
> 3. customise the indentation for syntactic symbol c to 4
> 4. Enter /* and then M-j (c-indent-new-comment-line)

> you should see point located at column 2 instead of 4, which is a wrong
> indentation position for that line and unlike indent-new-comment-line
> or newline-and-indent.

> Is this a bug or feature? Thanks.

It was a bug.  Now, at long last, it has been fixed, specifically in
commit 25ed447b7bec3af66cf0322239cfabbaf71bef26 of today, "CC Mode: Fix
positioning of point whilst inserting comments without non-ws".

The fix is also in the head of the default branch of stand alone CC Mode
at Sourceforge.

I am closing the bug with this post.  Sorry it was ignored for so long.

> Leo

-- 
Alan Mackenzie (Nuremberg, Germany).

bug#11165: 24.0.95; c-indent-new-comment-line bug or feature?

[Leo Liu]

On 2019-10-25 20:34 +0000, Alan Mackenzie wrote:
> It was a bug.  Now, at long last, it has been fixed, specifically in
> commit 25ed447b7bec3af66cf0322239cfabbaf71bef26 of today, "CC Mode: Fix
> positioning of point whilst inserting comments without non-ws".
>
> The fix is also in the head of the default branch of stand alone CC Mode
> at Sourceforge.

Excellent thanks for fixing it.

> I am closing the bug with this post. Sorry it was ignored for so long.

No worries. I am glad it is fixed ;)

Leo

bug#11165: 24.0.95; c-indent-new-comment-line bug or feature?

[Glenn Morris]

Alan Mackenzie wrote:

> commit 25ed447b7bec3af66cf0322239cfabbaf71bef26 of today, "CC Mode: Fix
> positioning of point whilst inserting comments without non-ws".

This change causes some test failures, ref eg
https://hydra.nixos.org/build/104351524

bug#11165: 24.0.95; c-indent-new-comment-line bug or feature?

[Alan Mackenzie]

Hello, Daniel and Glenn.

On Sat, Oct 26, 2019 at 13:08:30 -0400, Glenn Morris wrote:
> Alan Mackenzie wrote:

> > commit 25ed447b7bec3af66cf0322239cfabbaf71bef26 of today, "CC Mode: Fix
> > positioning of point whilst inserting comments without non-ws".

> This change causes some test failures, ref eg
> https://hydra.nixos.org/build/104351524

The failures were in js-mode (.../lisp/progmodes/js.el).

The reason for the failures was js-mode failing completely to initialise
a set of CC Mode "language variables", instead manually setting the few
that have been used in the past.

The lastest fix to CC Mode used another language variable, one that had
not been explicitly set by js-mode.  Thus js-mode's filling failed.

I think the best way to fix this is, rather than adding an ad hoc setq
for this variable, initialising a full set of language variables for
js-mode (based on Java Mode).  The following patch does this.  It passes
the Emacs test suite completely.

Daniel, what do you think?



diff --git a/lisp/progmodes/js.el b/lisp/progmodes/js.el
index 599923dd27..5f0913470f 100644
--- a/lisp/progmodes/js.el
+++ b/lisp/progmodes/js.el
@@ -46,6 +46,9 @@
 ;;; Code:
 
 (require 'cc-mode)
+(eval-when-compile
+  (require 'cc-langs)
+  (require 'cc-fonts))
 (require 'newcomment)
 (require 'imenu)
 (require 'moz nil t)
@@ -4529,12 +4532,22 @@ js-jsx--detect-after-change
         (when (js-jsx--detect-and-enable 'arbitrarily)
           (remove-hook 'after-change-functions #'js-jsx--detect-after-change t))))))
 
+;; Ensure all CC Mode "lang variables" are set to valid values.
+;; js-mode, however, currently uses only those needed for filling.
+(eval-and-compile
+  (c-add-language 'js-mode 'java-mode))
+
+(c-lang-defconst c-paragraph-start
+  js-mode "\\(@[[:alpha:]]+\\>\\|$\\)")
+
 ;;; Main Function
 
 ;;;###autoload
 (define-derived-mode js-mode prog-mode "JavaScript"
   "Major mode for editing JavaScript."
   :group 'js
+  ;; Ensure all CC Mode "lang variables" are set to valid values.
+  (c-init-language-vars js-mode)
   (setq-local indent-line-function #'js-indent-line)
   (setq-local beginning-of-defun-function #'js-beginning-of-defun)
   (setq-local end-of-defun-function #'js-end-of-defun)
@@ -4576,16 +4589,9 @@ js-mode
   (setq imenu-create-index-function #'js--imenu-create-index)
 
   ;; for filling, pretend we're cc-mode
-  (setq c-comment-prefix-regexp "//+\\|\\**"
-        c-paragraph-start "\\(@[[:alpha:]]+\\>\\|$\\)"
-        c-paragraph-separate "$"
-        c-block-comment-prefix "* "
-        c-line-comment-starter "//"
-        c-comment-start-regexp "/[*/]\\|\\s!")
+  (c-init-language-vars js-mode)
   (setq-local comment-line-break-function #'c-indent-new-comment-line)
-  (setq-local c-block-comment-start-regexp "/\\*")
   (setq-local comment-multi-line t)
-
   (setq-local electric-indent-chars
 	      (append "{}():;," electric-indent-chars)) ;FIXME: js2-mode adds "[]*".
   (setq-local electric-layout-rules
@@ -4599,6 +4605,13 @@ js-mode
     (make-local-variable 'paragraph-ignore-fill-prefix)
     (make-local-variable 'adaptive-fill-mode)
     (make-local-variable 'adaptive-fill-regexp)
+    ;; While the full CC Mode style system is not yet in use, set the
+    ;; pertinent style variables manually.
+    (c-initialize-builtin-style)
+    (let ((style (cc-choose-style-for-mode 'js-mode c-default-style)))
+      (c-set-style style))
+    (setq c-block-comment-prefix "* "
+          c-comment-prefix-regexp "//+\\|\\**")
     (c-setup-paragraph-variables))
 
   ;; Important to fontify the whole buffer syntactically! If we don't,



-- 
Alan Mackenzie (Nuremberg, Germany).

bug#11165: 24.0.95; c-indent-new-comment-line bug or feature?

[Glenn Morris]

> I think the best way to fix this is, rather than adding an ad hoc setq
> for this variable, initialising a full set of language variables for
> js-mode (based on Java Mode).  The following patch does this.  It passes
> the Emacs test suite completely.

It's inconvenient to have the test suite fail for an extended period of
time, so, in the absence of any other comments, could I encourage you to
apply this? TIA.

bug#11165: 24.0.95; c-indent-new-comment-line bug or feature?

[Alan Mackenzie]

Hello, Glenn.

On Wed, Nov 06, 2019 at 11:43:49 -0500, Glenn Morris wrote:

> > I think the best way to fix this is, rather than adding an ad hoc setq
> > for this variable, initialising a full set of language variables for
> > js-mode (based on Java Mode).  The following patch does this.  It passes
> > the Emacs test suite completely.

> It's inconvenient to have the test suite fail for an extended period of
> time, so, in the absence of any other comments, could I encourage you to
> apply this? TIA.

Yes, it's been 10 days with no comment from Daniel.  I've thus committed
the patch.

Thanks for the reminder.

-- 
Alan Mackenzie (Nuremberg, Germany).