Torsten wasn't changing the value pointed to by $ctx, he was changing
$ctx - therefore it looks like pnotes actually stores a reference to the
scalar you give it, instead of the actual scalar itself as the docs
imply. So I'd say it's a doc bug, but I'd prefer pnotes to store the
actual scalar instead.
Torsten - for a workaround try not re-using $ctx until it's gone out of
scope, like this:
$r->content_type('text/plain');
{
my $tmp={one=>1,two=>2};
$r->pnotes('x',$tmp);
}
{
my $tmp='nothing';
print Dumper($r->pnotes('x'))."\n------------\n";
}
print Dumper($r->pnotes('x')); # prints the same hash
Clint Edwards wrote:
Torsten,
Yes this is expected, $ctx contains a reference to an anonymous hash,
and when you store the value of $ctx in your pnotes, you are merely
storing another reference, therefore whenever the values pointed to by
$ctx change, you effectively change what is seen in pnotes.
