looks ok.
Bill H.
On 2018-12-14 6:32 p.m., Henry Rich wrote:
Fixed for next release. Good report.
The problem was in the new result-assembly code, and afflicted only
monad /\., as in f/\. y .
If the result of an application of f had shape that was not the shape of
an item of y, memory was corrupted. No workaround that I can see.
Henry Rich
BTW, when you say you think 0x40 looks like a space you're showing your
EBCDIC upbringing!
hr
On 12/14/2018 2:23 PM, Devon McCormick wrote:
I can reproduce it in 8.07 but not 8.06:
load 'plot'
d=: 4 : '''pensize 3'' plot ;/ |: 2{."1 x points/\. 0,~ |.
($~(4*x)*#) y'
points=: 1 : '((4*m)|[+{:@]) ,~ (2&{.)@] + (3 :
''|:(,:y|.])(,-)}:y-|i: y''
m) {~ (4*m)|[+{:@]'
90 d 7 13 270 9 29 290
90 d 7 13 270 9 29 240
Sometimes it fails on the second-to-last line but, so far, by the last
one
5 out of 5 times.
On Fri, Dec 14, 2018 at 11:56 AM Raul Miller <[email protected]>
wrote:
On Fri, Dec 14, 2018 at 11:43 AM Henry Rich <[email protected]>
wrote:
Can you turn this into a sequence that I can use to reproduce the
error?
With these definitions:
d=: 4 : '''pensize 3'' plot ;/ |: 2{."1 x points/\. 0,~ |.
($~(4*x)*#)
y'
points=: 1 : '((4*m)|[+{:@]) ,~ (2&{.)@] + (3 :
''|:(,:y|.])(,-)}:y-|i: y''m) {~ (4*m)|[+{:@]'
(two lines)
90 d 7 13 270 9 29 270
doesn't cause that error for me, either.
So one question is whether the crash is repeatable (as opposed to
being the consequence of some externally imposed glitch)?
Thanks,
--
Raul
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
---
This email has been checked for viruses by AVG.
https://www.avg.com
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
