Eric Wilhelm wrote:
>1. I think we should officially deprecate dir WRT Build.PL
Yes, that's sensible. We just can't remove support for "dir" immediately,
because it's currently the only correct way to specify that data.
>2. This code will put them both in META.yml if they both exist.
That's a weird situation that should never occur: it's invalid input
data under any version of the spec. I let it pass through and produce
equally-invalid output, which is what happens with any keys that aren't
allowed to exist and so on. I'd have done it differently if M::B were
validating the input, or otherwise taking measures to generate only
strictly valid output.
>3. why bother with the dereference?
Do you mean the copying of the hash:
$no_index = { %$no_index };
? This is to avoid modifying a hash that was supplied as an argument.
I think it's bad style to cause undocumented side effects on one's
parameters.
>We have no documentation coverage for this?
There's a lot in M::B that is not properly documented.
-zefram
