It seems like there's several messages that have gone to this list
but never made it to my mailbox. I wondered why all the discussion
dried up. :)
I'm worried that there seems to have been something read into RFC 6
I didn't intend. RFC 6 is about making lexical variables the default
by making strict 'vars' on by default in Perl6. It explicitly says
that it is not addressing any of the other strict options and says
nothing about -w. While I always program with all of these options in
effect, I think it's very wise to keep all of these issues separate.
If you were around to read both versions of RFC 6, you know that my
position changed substantially between versions 1 and 2. In version
1, I proposed variables be made lexical by default by changing the
language such that any undeclared variable is lexically scoped instead
of dynamically scoped. Based on feedback I received, I decided that
this should be accomplished instead by requiring strict 'vars' to be
on by default and I relegated my former proposal to the "ALTERNATE
SOLUTION" section of RFC 6.
What I'd really like to do is just wave a magic wand and make all of
the heartache associated with the original Perl dynamic scoping to go
I'm very intrigued by the things Mark-Jason Dominus has said; if I
read him correctly, he supports my original proposal and has quite a
bit more information on making this work. I'd like to see this
material either be an additional RFC or incorporated into the
"ALTERNATE SOLUTION" section of RFC 6, but Mark-Jason said he won't
have the time, himself. Most of the "big names" I recognize in the
Perl world don't seem to be evident on this list other than MJD and
perhaps a couple of others, and I wonder if any of the rest of us is
really qualified to author such a document, but I'm hoping someone
qualified would show up and be interested. If not, I'll try to
incorporate this info into my RFC. Such a revision would have to
occur no sooner than this weekend.
I'll also mention that I intend for the next version of RFC 6 to add
the following two minor items:
* strict 'vars' will not be in effect by default for a script run with
the -e switch
* strict 'vars' can be disabled by a command-line switch (-z seems to
be the favorite)
I think that catches me up to date on everything I missed and
everything that's on my mind. I've resubscribed to the list.
Hopefully it'll work.