I decided to time smoutput to see if there were any speed issues
as the session got large.
I also wanted to have a good base line idea of its speed.
In the process of developing this test methodology, I managed
to "hang" the j session. By this, I mean that J was using
100% of the cpu and was not updating the display.
Here's the code I was using:
t=:":i. 3 60
l=:":(32+40?96){a.
g=:6!:2 bind 'smoutput l' ` (6!:2 bind 'smoutput t')
times=: g/\bind (i. 40)/\i.40
When I ran this code, the edit window was partially
obscuring the J session window. I let this run for
a while, then clicked on the J session window to
bring it to the front. After it came to the front,
J ceased the session window updating (while using 100%
cpu). (Before I brought the J window to the front, my cpu
load was 50% in the system kernel -- this dropped to
nearly zero when the J session window came forward.)
jbreak was sufficient to regain control of the session.
I then created a new J session and opened up the script.
This time, the editor window seemed to completely obscure
the J canvas (I was using the default window locations).
Behavior seemed similar, except that it took a while
for the window to come forward. (Before the window
came forward, my cpu load was maybe 10% in the system
kernel -- this dropped to nearly zero when the J
session window came forward.)
In both J sessions, I am unable to exit J using the
close widget or the File->Exit menu option. When I try,
I get an error:
|domain error: wd
| wdq=: wd'q'
FYI,
--
Raul
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
