> Why does this work? (same result as your proposals)
There is no clear rule to distinguish between code-block and hash-ref!!!
The parser uses DWIM heuristics to guess if it looks more like a hash-ref or not.
(Furthermore "w/o look-ahead")
So in complicated cases you need to give hints for what you mean.
This ambiguity should be one reason why Perl6 switched from {} to <> to indicate hashes.
> UPDATE: Guess: Because there is some statement in there that needs to be executed rather than a static string only?
well it's a perfect hash-ref!
DB<112> $hr={ "$_"."=s" => \$opt{$_} }
=> { "=s" => \undef }
But I think the used heuristic must be quite simple, something like
(bareword || variable) (comma || fatcomma) (expr)
So "$_"."=s" is already to complex to fit this rule of thumb.
Cheers Rolf
( addicted to the Perl Programming Language)
|