That's a good idea & I will do it if it doesn't create problems. I
already detect u@> and u&.> but I don't think I detect dyadic u&> .
Henry Rich
On 10/19/2018 7:45 PM, 'Pascal Jasmin' via Programming wrote:
lots of boxing overhead removed in 807
https://code.jsoftware.com/wiki/System/ReleaseNotes/J807
it would be nice if for <;. and dyadic <\ <\. (first 1 priority),
if special code removed the boxing overhead in
(x u&> <;.) and (x u&> dyadicparam <;. ]) where x is arbitrary/optional left
fork item(s)
The use case for these improvements is that the verb argument to ;. \ \. are
monads, but the operation on the data can use another parameter, and if so a
boxing intermediate step is needed.
The case against special coding dyadic <\ is that ]\ is usually possible followed by
u"1 on result, and maybe dyadic \. isn't used enough?
----------------------------------------------------------------------
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