GNU bug report logs - #1948
confusion and bug in dabbrev.el

Previous Next

Full log

View this message in rfc822 format

From: Chong Yidong <cyd <at> stupidchicken.com>
To: Lars Lindberg <Lars.Lindberg <at> sypro.cap.se>
Cc: 1948 <at> debbugs.gnu.org, Peter Tury <tury.peter <at> gmail.com>
Subject: bug#1948: confusion and bug in dabbrev.el
Date: Mon, 19 Jan 2009 09:54:48 -0500

Hi Lars, could you take a look at this bug report for dabbrev?  Thanks.

Peter Tury <tury.peter <at> gmail.com> wrote:

> 1) in `dabbrev--substitute-expansion' (in dabbrev.el), `t' value of
> variable `dabbrev-eliminate-newlines' causes elimination of spaces
> (and tabs) as well. Thus its name and its functionality are not really
> compatible. In the best case it should have a better name
> (`dabbrev-eliminate-whitespace'??), but as a minimum, its
> documentation should be improved and mention elimination of tabs and
> spaces as well, I think.

> Moreover, contrary to its documentation again, it doesn't "converts
> them [=newlines] to spaces": instead it just removes them. So its
> documentation should be fixed here as well in my opinion.

> 2) However, the more serious problem is that it "forgets" such
> eliminations when it searches for `old' expansion in case of repeated
> calls of `dabbrev-expand' (i.e. when the check "(eq last-command
> this-command)" in this function results t) what results in weird text
> in some cases. To check this:

> * customize `dabbrev-abbrev-char-regexp' to be a period (.) Thus
> dabbrev-expand will replace whole lines.

> * type these two lines into a buffer (with double spaces between the
>   words!):
> one  tt
> one  ww

> * then type "on" in the next line and call `dabbrev-expand' two times.
> Your buffer should look like this at the end:
> one  tt
> one  ww
> one tt
> But instead you will get this:
> one  tt
> one tt
> one ww
> What seems to be a complete mess.

> As I see the key here is the search for `old' at the end of
> `dabbrev--substitute-expansion' (`old' is the first parameter of this
> function). If I replace the following original code:
> "    (if old
>      (save-excursion
>        (search-backward old))"
> to be
> "    (if old
>      ;; Convert whitespace to single spaces.
>      (progn
>      (if dabbrev-eliminate-newlines
>          (let ((pos
>                  (if (equal abbrev " ") 0 (length abbrev))))
>                        ;; If ABBREV is real, search after the end of it.
>                              ;; If ABBREV is space and we are copying
>                              successive words,
>                                    ;; search starting at the front.
>                                          (while (string-match "[\n \t]+"
>                                          old pos)
>                                              (setq pos (1+
>                                              (match-beginning 0)))
>                                                               (setq old
>                                                               (replace-match
>                                                               " " nil
>                                                               nil
>                                                               old)))))
>                                                               (save-excursion
>                                                                 (search-backward
>                                                                 old)))"
> my test (above) works well. Here I simply copy-pasted (...) original
> whitespace-elimination code from few lines above, so this is clearly
> not a nice solution: it just shows a way of fixing the broken
> functionality.

> (I "reported" this problem in gnu.emacs.help with the subject "strange
> dabbrev for whole lines and double spaces in cvs Emacs 23" on
> Jan.8.2009)

Previous Next

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