I think there have been some changes to 'charsub' in the latest beta to
allow for Unicode strings.
However, this has had some negative effects on code that used to work....
This is the kind of construct I was using:
'bB' charsub >0 { "1 (3 2)$ ;:'abcd efgh'
aBcd
aBcd
aBcd
which now in the L-beta gives:
'bB' charsub >0 { "1 (3 2)$ ;:'abcd efgh'
|index error: charsub
| (t {~f i.n{y)n}y
I think it still worked in the J601K beta, but I have no way of verifying
since I don't know how to download previous betas.
"J504B" charsub
+-+-+-----------------------------+
|4|:|'f t'=. ((#x.)$0 1)<@,&a./.x.|
| | |t {~ f i. y. |
+-+-+-----------------------------+
"J601L" charsub
┌─┬─┬───────────────────────┐
│4│:│'f t'=. |: _2 [\ x │
│ │ │n=. I. y e. f │
│ │ │(t {~ f i. n { y) n } y│
└─┴─┴───────────────────────┘
--
Gerald
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
