11:14 a.m.
The following message is a courtesy copy of an article
that has been posted to comp.emacs.xemacs as well.
Julian Bradfield <jcb(a)inf.ed.ac.uk> writes:
Vin Shelton <acs(a)xemacs.org> writes:
>>>> I'm still using 21.4 (patch 17), and from time to time I run up
>>>> against the mind-numbing slowness of simple operations when wrapped in
>>>> a keyboard macro.
> Could you please give an example of the 'mind-numbing slowness' that
> you are talking about? Please describe a scenario (started from
> 'xemacs -vanilla' that displays the slowness you are describing.
xemacs -vanilla /tmp/procmail.log.4
C-x (
C-x r SP a
C-s newspam RET
C-r sf: RET
C-SP
C-x r j a
C-x C-x
C-w
C-s newspam RET
C-a
C-n
C-x )
Then each C-x e takes about twelve seconds of (elapsed and CPU)
time to execute.
/tmp/procmail.log.4 is a file of about 10MB; the typical distance
between successive occurrence of newspam is less than a hundred lines.
The file size is important - if I just take the first few thousand
lines, it's not a problem. It also occurs with xemacs -vanilla -nw .
I see what you mean. This seems to have something to do with
interactive search: if you change C-s to M-x search-forward and C-r to
M-x search-backward then your example runs very quickly for me. Do
you find that to be true?
I have forwarded this to the xemacs-beta list, perhaps someone there
will have some other thoughts.
Regards,
Vin
11:33 a.m.
New subject: very slow keyboard macros
Ar an t-aonú lá déag de mí Eanair, scríobh Vin Shelton:
I see what you mean. This seems to have something to do with
interactive search: if you change C-s to M-x search-forward and C-r to
M-x search-backward then your example runs very quickly for me. Do
you find that to be true?
If so, that seems a normal consequence of how keyboard macros are
implemented, and this patch (untested, against 21.4) will address it in the
case of isearch.
--- isearch-mode.el~ 2006-01-11 12:30:45.674397209 +0100
+++ isearch-mode.el 2006-01-11 12:31:06.335343912 +0100
＠＠ -707,7 +707,7 ＠＠
search and `search-nonincremental-instead' is non-nil, do a
nonincremental search instead via `isearch-edit-string'."
(interactive)
- (if (and search-nonincremental-instead
+ (if (and (or search-nonincremental-instead executing-kbd-macro)
(= 0 (length isearch-string)))
(let ((isearch-nonincremental t)
;; Highlighting only gets in the way of nonincremental
--
I AM IN JAIL AND ALLOWED SEND ONLY ONE CABLE SINCE WAS ARRESTED WHILE
MEASURING FIFTEEN FOOT WALL OUTSIDE PALACE AND HAVE JUST FINISHED COUNTING
THIRTY EIGHT THOUSAND FIVE HUNDERED TWENTY TWO NAMES WHOS WHO IN MIDEAST.
3:56 p.m.
New subject: very slow keyboard macros
> The file size is important - if I just take the first few
thousand
> lines, it's not a problem. It also occurs with xemacs -vanilla -nw .
Vin> I see what you mean. This seems to have something to do with
Vin> interactive search: if you change C-s to M-x search-forward and C-r
Vin> to M-x search-backward then your example runs very quickly for me.
I routinely disable isearch highlighting when running keyboard macros
especially if the file is large or the macro is run repeatedly. (Aidan's
patch seems related to that.) Also, if the macro modifies the file, turning
off auto-save for large files (or at increasing the auto-save interval)
also helps.
--
Skip Montanaro
http://www.musi-cal.com/
skip(a)pobox.com
4:31 p.m.
New subject: very slow keyboard macros
On 1/11/06, skip(a)pobox.com <skip(a)pobox.com> wrote:
>> The file size is important - if I just take the first few thousand
>> lines, it's not a problem. It also occurs with xemacs -vanilla -nw .
Vin> I see what you mean. This seems to have something to do with
Vin> interactive search: if you change C-s to M-x search-forward and C-r
Vin> to M-x search-backward then your example runs very quickly for me.
I routinely disable isearch highlighting when running keyboard macros
especially if the file is large or the macro is run repeatedly. (Aidan's
patch seems related to that.) Also, if the macro modifies the file, turning
off auto-save for large files (or at increasing the auto-save interval)
also helps.
--
Skip Montanaro
http://www.musi-cal.com/
skip(a)pobox.com
Skip,
Thanks for pointing this out. I've copied Julian on this follow-up.
- Vin
4:37 p.m.
New subject: very slow keyboard macros
Thanks for that, and I've seen Aidan's patch, which I'll probably use.
However, it remains the case that there's something deeply sick if
executing something as a keyboard macro increases the execution time
from neglible to 12 seconds! If I weren't too busy to do it, I'd
track it down for use as an example in classes of unexpected things
that happen when programming.
6686
days inactive
6686
days old
4 comments
4 participants
participants (4)
-
Aidan Kehoe -
Julian Bradfield -
Skip Montanaro -
Vin Shelton