Both
tilly and
indigo have exactly the same result,
which is a fix of
twerq's including 1) my, and 2) the
test for $t which prevents infinite looping.
Here is a solution which is short, but not 'strict' compliant:
sub f{for(($s,$t)=@_;$_=$$t{(0,l,r)[$$t{d}cmp$s]};$t=$_){}$t}
Alas, although 4 characters lighter, it doesn't return 0 on misses properly. This does,
but is admittedly 3 characters heavy:
sub f{for(($s,$t)=@_;$_=$t&&(0,l,r)[$$t{d}cmp$s];){$t=$$t{$_}||0}$t}
Or, a tie, provided the tree has 0-value stubs as it does in
other examples:
sub f{for(($s,$t)=@_;$_=$t&&(0,l,r)[$$t{d}cmp$s];){$t=$$t{$_}}$t}
Fun test code below for a tree:
$|++;
$table = {
d => 'h',
l => {
d => 'd',
l => { d => 'b', l => { d => 'a' }, r => { d => 'c' } },
r => { d => 'f', l => { d => 'e' }, r => { d => 'g' } },
},
r => {
d => 'l',
l => { d => 'j', l => { d => 'i' }, r => { d => 'k' } },
r => { d => 'm', l => { d => 'l' }, r => { d => 'n' } },
}
};
foreach $y (qw[ a b c d e f g h i j k l m n o p q ])
{
$x = f($y,$table);
print "Answer for $y = ";
print $x," ", $$x{d},"\n";
}
value entry in the hash.