On Wed, 2002-09-11 at 19:11, [EMAIL PROTECTED] wrote:
> Hello everybody,
>
> that's a few month i'm following this mailing list. i dont like all the changes
> that are being made to perl
Imagine how I felt when Perl 4 came out, with all that silly binary I/O
stuff! I had no idea what Larry had been smoking! :-)
> This is the part i like about perl 5. Really. I like it, it seems when you're
> doing I/O on a device, and nobody knows about the features. I like it! You're
> writting a driver an nobody suspects what's you're about.
If no one can determine what your code does, it's bad code. That's one
of the reasons that I don't particularly like the implicit semicolon,
though I do very much like the idea of solving the eval{}; vs for{}
discrepancy.
Here's how I would solve the problem:
If the last argument to a subroutine is a code value (e.g. C<&block>)
then the subroutine invocation does not require a semicolon. Thus:
eval {...} / 2
would be an error in exactly the same way that
if 0 {...} / 2
would be, but the common
eval {...};
would be ok, even though it's got a null statement.
Remember, all's fair if you pre-declare.
--
Aaron Sherman <[EMAIL PROTECTED]>
http://www.ajs.com/~ajs
