Short reply: Yes! Nice!
Longer reply: See below...
[EMAIL PROTECTED] wrote:
>
> I think this is a better generalization:
>
> >> source repoke
> repoke: func [
> "sets value within nested blocks"
> block [block! hash! function! object!]
> indexes [block!]
> value [any-type!]
> ][
> indexes: reduce indexes
> while [1 < length? indexes] [
> block: pick :block first indexes
> indexes: next indexes
> ]
> poke block first indexes value
> ]
>
...
>
> >> source repick
> repick: func [
> "picks repeatedly"
> block [any-block! any-function! object!]
> indexes [block!]
> ][
> foreach i reduce indexes [block: pick :block i]
> :block
> ]
>
...
> >Try the following...
> >
> > >> set2d: func [b [block!] d1 [integer!] d2 [integer!] x] [
> > [ poke pick b d1 d2 x
> > [ ]
> > >> set2d a 3 2 "Goodbye!"
> >
> >.... and to generalize ...
> >
> > >> set3d: func [
> > [ b [block!] d1 [integer!] d2 [integer!] d3 [integer!] x][
> > [ poke pick pick b d1 d2 d3 x
> > [ ]
There's no question that Eric's single (pair of) function(s) is
much more elegant than the special purpose hack I tossed off the
top of my head, and I certainly prefer it as a standard item in
my toolkt.
I would only add that if I were writing code which did LOTS of
flogging of 2-D arrays, I'd probably code and use the special-
case 'set2d and 'get2d versions purely for performance reasons.
-jn-
