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