bug#15126: 24.3; kmacro-end-and-call-macro repeats macro multiple times following isearch-forward

[Simon Katz]

If I have a buffer with some text that is repeated several times and

do the following:

  - C-s <some text>  (isearch-forward)   to search for the repeated text

  - C-x (   (kmacro-start-macro)

  - Enter some new text

  - C-s C-s   to search again

  - C-x e  (kmacro-end-and-call-macro)

I get different behaviour in my old and new versions of Emacs.

In Emacs 24.2.1 and all previous versions I've used,

kmacro-end-and-call-macro calls the macro once more, and I can go

through the file making changes one at a time, and I can stop when I

want.

In Emacs 24.3.1, kmacro-end-and-call-macro calls the macro repeatedly

making changes to all the repeated text up to the end of the file.

Interestingly, if I define a similar keyboard macro that uses

next-error rather than isearch-forward I get my changes one at a time.

Is this some bizarre new behaviour for isearch-forward? Can I get the

old behaviour back?

Here's a precise recipe:

  /Applications/Emacs-24-3-1.app/Contents/MacOS/Emacs -Q

  Then at the start of the *scratch* buffer:

  C-s   e r   C-x (   x x x x   C-s C-s   C-x e

  Emacs makes multiple changes for the  C-x e.

In GNU Emacs 24.3.1 (x86_64-apple-darwin, NS apple-appkit-1038.36)

 of 2013-03-13 on bob.porkrind.org

Windowing system distributor `Apple', version 10.3.1138

Configured using:

 `configure '--host=x86_64-apple-darwin' '--build=i686-apple-darwin'

 '--with-ns' 'build_alias=i686-apple-darwin'

 'host_alias=x86_64-apple-darwin' 'CC=gcc -mmacosx-version-min=10.7

 -isystem

 /Users/david/Xcode-10.7_4.5.2/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.7.sdk/usr/include/

 -F/Users/david/Xcode-10.7_4.5.2/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.7.sdk/System/Library/Frameworks''

Important settings:

  value of $LANG: en_GB.UTF-8

  locale-coding-system: utf-8-unix

  default enable-multibyte-characters: t

Major mode: Lisp Interaction

Minor modes in effect:

  tooltip-mode: t

  mouse-wheel-mode: t

  tool-bar-mode: t

  menu-bar-mode: t

  file-name-shadow-mode: t

  global-font-lock-mode: t

  font-lock-mode: t

  blink-cursor-mode: t

  auto-composition-mode: t

  auto-encryption-mode: t

  auto-compression-mode: t

  line-number-mode: t

  transient-mark-mode: t

Recent input:

s-x M-x r e p <tab> o r t <tab> <return>

Recent messages:

For information about GNU Emacs and the GNU system, type C-h C-a.

kill-region: The mark is not set now, so there is no region

Making completion list...

Load-path shadows:

None found.

Features:

(shadow sort gnus-util mail-extr emacsbug message format-spec rfc822 mml

mml-sec mm-decode mm-bodies mm-encode mail-parse rfc2231 mailabbrev

gmm-utils mailheader sendmail rfc2047 rfc2045 ietf-drums mm-util

mail-prsvr mail-utils help-mode easymenu time-date tooltip ediff-hook

vc-hooks lisp-float-type mwheel ns-win tool-bar dnd fontset image

regexp-opt fringe tabulated-list newcomment lisp-mode register page

menu-bar rfn-eshadow timer select scroll-bar mouse jit-lock font-lock

syntax facemenu font-core frame cham georgian utf-8-lang misc-lang

vietnamese tibetan thai tai-viet lao korean japanese hebrew greek

romanian slovak czech european ethiopic indian cyrillic chinese

case-table epa-hook jka-cmpr-hook help simple abbrev minibuffer loaddefs

button faces cus-face macroexp files text-properties overlay sha1 md5

base64 format env code-pages mule custom widget hashtable-print-readable

backquote make-network-process ns multi-tty emacs)

[Drew Adams]

> Here's a precise recipe:
>   /Applications/Emacs-24-3-1.app/Contents/MacOS/Emacs -Q
>   Then at the start of the *scratch* buffer:
>   C-s   e r   C-x (   x x x x   C-s C-s   C-x e
>   Emacs makes multiple changes for the  C-x e.

Confirmed in MS Windows builds.
Emacs 23.4 through Emacs 24.2 were OK.
Emacs 24.3 introduced the regression.

[nomi...@gmail.com]

On Sunday, 18 August 2013 22:39:41 UTC+1, Simon Katz wrote:
>
> Here's a precise recipe:
>
>
>   /Applications/Emacs-24-3-1.app/Contents/MacOS/Emacs -Q
>
>
>
>   Then at the start of the *scratch* buffer:
>
>
>   C-s   e r   C-x (   x x x x   C-s C-s   C-x e
>
>
>   Emacs makes multiple changes for the  C-x e.

Some more detail:

/Applications/Emacs-24-3-1.app/Contents/MacOS/Emacs -Q
Then at the start of the *scratch* buffer:

C-s   e r   C-x (   x x x x   C-s C-s   C-x ) F4 F4
This is as it should be -- one at a time.

But:

/Applications/Emacs-24-3-1.app/Contents/MacOS/Emacs -Q
Then at the start of the *scratch* buffer:
C-s   e r   C-x (   x x x x   C-s C-s   C-x e

(The original report -- multiple changes for the line above.)
Return to the start of the buffer.
F4
Multiple changes are made.

So F4 is ok in the first case but not in the second.

[Simon Katz]

[Juri Linkov]

> In Emacs 24.2.1 and all previous versions I've used,
> kmacro-end-and-call-macro calls the macro once more, and I can go
> through the file making changes one at a time, and I can stop when I
> want.
>
> In Emacs 24.3.1, kmacro-end-and-call-macro calls the macro repeatedly
> making changes to all the repeated text up to the end of the file.

I can't find a change that caused this regression.
At least, I see that it is not caused by the fix for bug#4894
that added `store-kbd-macro-event' to `isearch-other-meta-char'.
But removing `store-kbd-macro-event' from `isearch-unread'
that is called from `isearch-other-meta-char' fixes this bug.
Perhaps this is a wrong fix and I don't know how to fix this properly.
`isearch-other-meta-char' currently is full of crap that should be
rewritten and reduced to a more manageable size.

[Juri Linkov]

> In Emacs 24.2.1 and all previous versions I've used,
> kmacro-end-and-call-macro calls the macro once more, and I can go
> through the file making changes one at a time, and I can stop when I
> want.
>
> In Emacs 24.3.1, kmacro-end-and-call-macro calls the macro repeatedly
> making changes to all the repeated text up to the end of the file.

Actually this is an old bug uncovered by another unrelated change.
Emacs 24.2.1 stored the key `C-x' in the keyboard macro.
Emacs 24.3.1 now stores the whole key sequence `C-x e'.

Storing `C-x' had no effect, but also made no sense.
It's clear that neither `C-x' nor `C-x e' should be stored.
So not storing `C-x e' will not execute the same macro repeatedly.

Currently I see only one way to fix this properly:

=== modified file 'lisp/isearch.el'
--- lisp/isearch.el 2013-08-08 23:59:14 +0000
+++ lisp/isearch.el 2013-08-24 09:27:30 +0000
@@ -2386,6 +2386,13 @@ (defun isearch-other-meta-char (&optiona
(isearch-unread-key-sequence keylist)
(setq main-event (car unread-command-events))

+ ;; Don't store special commands in the keyboard macro.
+ (let (overriding-terminal-local-map)
+ (when (memq (key-binding key)
+ '(kmacro-start-macro
+ kmacro-end-macro kmacro-end-and-call-macro))
+ (cancel-kbd-macro-events)))
+
;; If we got a mouse click event, that event contains the
;; window clicked on. maybe it was read with the buffer
;; it was clicked on. If so, that buffer, not the current one,

[Juri Linkov]

>> In Emacs 24.3.1, kmacro-end-and-call-macro calls the macro repeatedly
>> making changes to all the repeated text up to the end of the file.
>

> It's clear that neither `C-x' nor `C-x e' should be stored.
> So not storing `C-x e' will not execute the same macro repeatedly.

This case is fixed now. Thanks for the bug report.

There is another case where instead of `C-x e'
typing more verbosely `M-x kmacro-end-and-call-macro RET'
will store this sequence in the macro.

But it's not a isearch specific problem as it can be reproduced
without isearch by the test case:

M-x kmacro-start-macro RET x x x x
M-x kmacro-end-and-call-macro RET

It doesn't stop after the first macro execution
because it stores M-x and the command in the macro:

Macro:
4*x ;; self-insert-command
M-x ;; execute-extended-command
kmacro-end-and-call-macro ;; self-insert-command * 25

This is a minor problem since nobody reported it
because there are more convenient keys to do the same
with `C-x ( x x x x C-x e' (or F3/F4).

[Simon Katz]

Great, and thanks for the additional info.