> >We already have a trailing if statement, so this argument is edgy, at
> >best.
>
> Your brain/eye can look one statement. It's evil and wrong to make
> it look infinite statements forward. How can you pretend those are
> the same?
Well, I think trailing statements already have only limited usefulness.
Consider these examples:
die "Bad stuff: $?" if (system "mkdir -p /usr/bin");
@data = get_some_data(\*FILE, @args) unless (@data);
Blech, personally. I only like trailing statements when I'm decently
sure they're going to be true. So, a trailing {}if can be useful in some
situations:
$tmpdir = generate_tmpdir;
{ mkdir $tmpdir;
chown $uid, $gid, $tmpdir;
chmod 0700, $tmpdir; } unless -d $tmpdir;
This reads better than a leading if. Why? Because we're pretty certain
that the $tmpdir is not going to exist. So putting the statements first
- which will probably get executed - makes sense, and reduces clutter.
It is still easily readable, too, and not far at all from what a
trailing statement is used for currently.
*NOTE*: Just for the record, I would never advocate trailing if's on
long conditions, like the one you demonstrated. It's garbage, I agree.
But that's a style issue.
Anyways, I think this should be RFC'ed. It's not a revolutionary idea
that's going to destroy Perl. If Larry decides he still doesn't like it,
it doesn't make it in. No biggie.
> >Granted, the do {} is misleading. A do {} block is always executed, at
> >least once.
>
> No, it's not.
>
> do { .... } if 0;
>
> never executes. Nay, not even once.
I know. I agree. If you read my email, that's why I said:
Nathan Wiger wrote:
>
>We should drop the do {} if we have a trailing condition:
>
> {
> stuff;
> otherstuff;
> morestuff;
> } if $x > y;
