Eric wrote:
>I've never heard of the non-unique sparse elements bug. Do you have a way to
>produce this on demand?
Sure. In j.exe (not jconsole.exe, nor j.jar) for any J6 beta (not J5), run the
following sentences. I've elided the output, but for obvious reasons, you
should not suppress it.
(9!:12 (, ' ; ' &, )&": 9!:14) '' NB. Windows ; J6
6 j601/beta/2006-05-20/22:30
wd'qwd' NB. j.exe (not jconsole.exe nor j.jar)
jwin32
boxdraw_j_ 0 NB. Have to have unicode boxes; ASCII just
causes 'too much data' error.
9!:37 ] 0 1024 300 0 NB. Long, wide output.
<"> i. 1024 300 NB. Big boxed display
...
<"> i. 1024 300 NB. And again
...
wdhandler'' NB. Calling wdhandler'' directly only
results in 'too much data'
|domain error: wd
| wdq=: wd'q'
NB. Instead, press ESC, or F12, or mouse-click on the little [x],
etc.
NB. Basically, do anything that causes the a windows message loop
event.
|non-unique sparse elements: wd
| wdq=: wd'q'
wdhandler'' NB. After its first appearance, any call to
wd 'q' fails with the sparse error
|non-unique sparse elements: wd
| wdq=: wd'q'
11!:0 'q'
|non-unique sparse elements
| 11!:0'q'
My speculation is that (A) you're using a sparse array to store the wd data,
to leverage the fact that spaces are common, and (B) the error message is due
to the split personality of unicode characters. What I mean by (B) is that
they look like one character, but they're treated like several:
#'?'
3
If that character doesn't come through email correctly, it is a copy-and-pasted
u:16 (the top-left corner box drawing character). But do not think:
#u:16
1
is equivalent. It is not; you must literally copy and paste the result of
u:16 , put quotes around it, and put a # in front. (the character is really
2 u: u: 226 148 140 ). This has been discussed previously on the Forums.
I'll dig up a link if you ask.
Anyway, since the bug requires boxdraw 0 and since sometimes single unicode
characters are really multiple characters, I speculate this message is just:
($.i.4) { 226 ,. i. 4
|non-unique sparse elements
| ($.i.4) {226,.i.4
in disguise (contrast with examples in postscript).
-Dan
PS:
( i.4) { i.4
0 1 2 3
( i.4) { 0 ,. i.4
0 0
0 1
0 2
0 3
( i.4) { 226 ,. i.4
16 0
16 1
16 2
16 3
($. i.4) { i.4
1 | 1
2 | 2
3 | 3
($. i.4) { 0 ,. i.4
1 | 0 1
2 | 0 2
3 | 0 3
(I am not saying the actual value 226 is special or relevant.)
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
