⎕R oddity - [resolved]

[Tomas Gustafsson]

      ('-' 'ef'⎕R'¯' 'x')'-ab cd ef'
¯xab cd ¯x                               <- I expected ¯ab cd x
     ('-' 'ef'⎕R'¯' 'xx')'-ab cd ef'
¯ab cd xx
     ('-' 'ef'⎕R'¯' 'xxx')'-ab cd ef'
¯ab cd xxx

      ('-' 'cd'⎕R'¯' 'x')'-ab cd ef'
¯xab ¯x ef                               <- I expected ¯ab x ef
     ('-' 'cd'⎕R'¯' 'xx')'-ab cd ef'
¯ab xx ef
     ('-' 'cd'⎕R'¯' 'xxx')'-ab cd ef'
¯ab xxx ef

Doesn't seem fully right, imho? x's and -'s appear at unexpected locations in the single-character cases. Anyone who could explain the behaviour for me?

      #.DISP #.APLVersion
┌───────┬────────────┬─┬───────────┐
│Windows│14.0.23160.0│W│Development│
└───────┴────────────┴─┴───────────┘

[Phil Last]

First off

      '¯' 'x' ≡ '¯x'

I guess if ⍺⍺ is a nested list and ⍵⍵ is a simple string then the string applies to all members of ⍺⍺. Normally you'd expect item conformability so that you'd need to enclose ⍵⍵ if you wanted it to apply to each ⍺⍺. But of course they aren't function arguments, they're operands. Nor is ⎕R scalar.

You could ensure what you expect by

      ('-' 'ef'⎕R(,¨'¯' 'x'))...

[Tomas Gustafsson]

Ah, good point Phil. I tried

      ⎕R ,¨ ... )

but not

      ⎕R (,¨ ...) )

Thx!