Re^2: Globs and globals

[Eyck]

This advice is factually incorrect.

GLOB is a glob, and local doesen't help here ( besides providing false sense of security )

my $socket=\*S1;
print "Orig:",$socket,"\n";
{
local(*S1, *S2);
my $s=\*S1;
print "Local: ",$s,"\n";
}

Orig:GLOB(0x8160be4)
Local: GLOB(0x8160be4)

[BrentDax]

Odd--you're right, though I can't figure out why. The docs seem to say that you get a new glob when you localize:

The construct local *name; creates a whole new symbol table entry for the glob "name" in the current package. That means that all variables in its glob slot ($name, @name, %name, &name, and the "name" filehandle) are dynamically reset.

In any case, the following does work correctly:

open(S1, "/dev/null");
$a=*S1{IO};
{
    local(*S1);
    open(S1, "/dev/null");
    $b=*S1{IO};
}
print "$a $b\n";

IO::Handle=IO(0x101536f8) IO::Handle=IO(0x1013cdec)

[Eyck]

Interesting, your code works, although I would expect it not to. I haven't seen this excerpt in documentation about local and globs so it's a bit of surprise for me, thanks.

There must be something about dual nature of globs - whatever experiment you devise it proves your theory, even if it contradicts other experiments.

[BrentDax]

I think it may be that local *glob localizes the contents of *glob, not the glob itself. The docs could be read to support that.

Or maybe it's just that the superposition-like nature of globs means that their behavior is governed by the Heisenburg Uncertainty Principle. Yeah, that's it...