Avis, Ed wrote:
> >There are no array slices in your examples. In a [] subscript,
> >expressions are evaluated in scalar context.
>
> Well, not in the case of
>
> @a[0, 1, 2]
>
> It is tempting for the programmer to try the same thing with a listref:
>
> $r->[0, 1, 2]
>
> although in fact the correct way to do it is
>
> @$r[0, 1, 2]
>
> A warning for the second case would be a good idea, since it's a little
> counterintuitive that the square brackets behave differently in the two
> cases.
Maybe, but such a warning can only be produced by dirty heuristics (like
the C<print ("foo")> warning, and I'm slightly opposed to
heuristic-driven warnings. Sometimes they are wrong, and sometimes they
don't occur where they would have been useful.
> perlop says:
>
> >If either operand of scalar ".." is a constant expression, that operand
> >is considered true if it is equal ("==") to the current input line num-
> >ber (the $. variable).
>
> So somehow it must be comparing 0 and 1 to $., which in this example
> program is always undef since we haven't read a line from input.
>
> I'll remember to be careful when using .. for ranges in case I get the
> wrong operator by accident.
Right. So, no bug. (Although the incrementation of the return value of
.. is not expected/documented, but a side-effect of the implementation).
--
Fix in six / Fix in six / And we'll take over Unix
