In article <[EMAIL PROTECTED]>, Tassilo Von Parseval wrote:
[...]
> If you from then on referred to elements of the hash with something like
>
> $score_ref->{ key };
>
> then this would be it. It depends on what you want. By dereferencing the
> whole data-structure you're essentially creating a copy. The downside of
> that is that it needs more memory and is slightly less efficient.
> However, that way your subroutine wont have side-effects. If you
> directly work on the references, you'd change the hashes in your main
> program (only on read-access, of course). That itself can be a valid
> design decision. So it's up to you what you want.
Thanks, I know this question is very similar to one I posed not so long ago,
but some concepts percolate in more slowly than others ("deep-watering" as
they say in California).
>> my $sort_hi if ($opt{h});
Ummm... a "digital artifact"? :-)
> Ah, I should have read on. Either the unless or the else branch is the
> last thing in your subroutine and so the last statement they evaluate is
> returned. Hmmh, no. I wouldn't do it like that. It requires too much of
> brain-power. ;-) What I often do in such a case is return early from a
> sub and thus doing away with an else block. Like this:
>
> sub sort_results {
>
> my %scores = %{ (shift) };
> my %opt = %{ (shift) };
>
> if (%opt) {
> return sort {
> compare($scores{ $a }, $scores{ $b }, $opt{ h })
> } keys %scores;
> }
>
> return sort keys %scores;
>
> }
>
> First of all, I turned the logic around. What previously was the else
> part is now the if part. I use an explicit return there. So the second
> return statement is only reached when %opt was false.
That makes good sense. I had originally added returns in there just to make
it clearer, but your way looks even smoother.
[...]
>
> The rest looks fine to me. However - and so turning back to your
> original question - for this program you are probably best off by
> directly working with the references. sort_results() would then look
> like:
Thanks!
--
Kevin Pfeiffer
International University Bremen
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
