Re: state vars: init with handle?

On Wed, Sep 19, 2007 at 04:53:57PM +0200, Rafael
Garcia-Suarez wrote:
> On 19/09/2007, Abigail <abigailabigail.be> wrote:
> > On Wed, Sep 19, 2007 at 02:31:53PM +0200, Rafael
Garcia-Suarez wrote:
> > >
> > > Well, given that open() is one of the very
few built-ins that modify
> > > its argument (with sysopen() -- any
others?),
> >
> > substr, push, pop, shift, unshift, splice, vec,
chomp, chop, undef, tie,
> > bless.
> 
> >From this list, I see only tie() which has an
interest. The other most
> of the time operate on a variable that has already been
used. But
> C<tie state $foo> makes some sense.

Mmm, I think that it's going to be moderately tricky to
implement.
And I was trying to get all the blead stuff done to get back
to 5.8.9

Is the list longer than tie, open, sysopen?

Nicholas Clark

On 19/09/2007, Nicholas Clark <nickccl4.org> wrote:
> Mmm, I think that it's going to be moderately tricky to
implement.

Likewise.

> And I was trying to get all the blead stuff done to get
back to 5.8.9

You know, that can wait for another release. We could even
forbid that
for now, like state foo = qw(bar). There are more pressing
things.

> Is the list longer than tie, open, sysopen?

Doesn't seem so.