Re: CC Mode 5.31.4 (Java/l); Incorrent behaviour of uncomment-region in java-mode

[Alan M. <ac...@mu...>]

Hi, Guillaume!

On Tue, Oct 14, 2008 at 07:26:41PM +1000, Guillaume Salagnac wrote:
> Hi,

> I have a strange problem with uncomment-region eating my lines of
> code. I can reproduce it deterministically by:

> - starting emacs (with -Q)
> - changing to java-mode (in the scratch buffer, for example)
> - typing a two line c-style comment :

> /* first line
> second line */

> - selecting all the buffer (or just the two lines)
> - M-x uncomment region

> I end up with the following text

> first lin

> (which is kinda scary, and has already caused me problems with lines
> of code silentely disappearing, not causing a compile-time error but a
> strange run-time behaviour instead. but I'm getting off-topic :-)

> BTW, this is emacs 22.1.1 (CC Mode version 5.31.4), I get the same
> behaviour with emacs 23.0.60.1 (CC Mode version 5.31.5), but it works
> fine in emacs 21 (CC Mode version 5.28). I'm on ubuntu hardy.


> Is this a known bug, or am I doing something wrong ?

Yes, it was indeed a known bug: A `save-excursion' had been omitted from
the function `comment-enter-backward' in ..../lisp/newcomment.el.

I would recommend you to upgrade your emacs to 22.3.  You can download
the tarball from <http://ftp.gnu.org/pub/gnu/emacs/emacs-22.3.tar.gz>.
It is easy indeed to build, especially on a system as common as Ubuntu.
Or you might be able to download it as a Ubuntu style package from
somewhere, which would be even easier.

Alternatively, if you really don't want to upgrade your emacs, apply this
patch to ...../lisp/newcomment.el, and byte-compile it, replacing the
existing newcomment.elc.  If you need any help applying the patch or
byte-compiling, feel free to send me a personal Email.

*** /usr/local/share/emacs/22.1/lisp/newcomment.el	2007-05-14 14:56:29.000000000 +0000
--- /home/acm/emacs/emacs/lisp/newcomment.el	2008-05-29 18:19:55.000000000 +0000
***************
*** 486,506 ****
        ;; comment-end = ""
        (progn (backward-char) (skip-syntax-backward " "))
      (cond
!      ((save-restriction
!         (narrow-to-region (line-beginning-position) (point))
!         (goto-char (point-min))
!         (re-search-forward (concat comment-end-skip "\\'") nil t))
        (goto-char (match-beginning 0)))
!      ;; comment-end-skip not found.  Maybe we're at EOB which implicitly
!      ;; closes the comment.
!      ((eobp) (skip-syntax-backward " "))
!      (t
!       ;; else comment-end-skip was not found probably because it was not
!       ;; set right.  Since \\s> should catch the single-char case, we'll
!       ;; blindly assume we're at the end of a two-char comment-end.
        (backward-char 2)
        (skip-chars-backward (string (char-after)))
!       (skip-syntax-backward " ")))))
  
  ;;;;
  ;;;; Commands
--- 490,519 ----
        ;; comment-end = ""
        (progn (backward-char) (skip-syntax-backward " "))
      (cond
!      ((save-excursion
!         (save-restriction
!           (narrow-to-region (line-beginning-position) (point))
!           (goto-char (point-min))
!           (re-search-forward (concat comment-end-skip "\\'") nil t)))
        (goto-char (match-beginning 0)))
!      ;; comment-end-skip not found probably because it was not set
!      ;; right.  Since \\s> should catch the single-char case, let's
!      ;; check that we're looking at a two-char comment ender.
!      ((not (or (<= (- (point-max) (line-beginning-position)) 1)
!                (zerop (logand (car (syntax-after (- (point) 1)))
!                               ;; Here we take advantage of the fact that
!                               ;; the syntax class " " is encoded to 0,
!                               ;; so "  4" gives us just the 4 bit.
!                               (car (string-to-syntax "  4"))))
!                (zerop (logand (car (syntax-after (- (point) 2)))
!                               (car (string-to-syntax "  3"))))))
        (backward-char 2)
        (skip-chars-backward (string (char-after)))
!       (skip-syntax-backward " "))
!      ;; No clue what's going on: maybe we're really not right after the
!      ;; end of a comment.  Maybe we're at the "end" because of EOB rather
!      ;; than because of a marker.
!      (t (skip-syntax-backward " ")))))
  
  ;;;;
  ;;;; Commands




> Cheers,
> -Guillaume

-- 
Alan Mackenzie (Nuremberg, Germany).