Re: A new golf "Unfragment" has started

Hello All!

As usual, playing another round of golf turns my mind to another round of
questions...

My solution below works for all test cases:
-lp0a
for$=(@F){@;=grep{$*=$_;grep$*eq$_,@;?@;:@*}@*=map{$_.$=,$=.$_}@F}$_=pop@I
thought, however, that I could save a stroke by replacing the first grep
statement with a map:

-lp0a
for$=(@F){@;=map{$*=$_;grep$*eq$_,@;?@;:@*}@*=map{$_.$=,$=.$_}@F}$_=pop@
This solution (though seeming to work at first) fails on test 81 (seen at
the bottom of this post).

Maybe it actually passes, but I didn't have the patience to wait for it... I
noticed Juho mentioned something similar about test 81.

Why does this substitution (grep to map) make such a huge difference in
running time? Both statements should return the same array, right?

Many thanks for any thoughts.
-Riley (o0lit3)

Test case 81:
0000000
00
000000
000
000
000000
0000
00000
0000
00000
0000000
00
00000000
0
0
00000000