This fixes a potential problem with Carp::longmess
firing somewhere deeper in the stack. This is not a known
problem at this time, but something I noticed while chasing
something else.
sub do_sock_stream { # via wq_io_do, for big requests
my ($self, $len) = @_;
sub do_sock_stream { # via wq_io_do, for big requests
my ($self, $len) = @_;
- recv_and_run($self, delete $self->{0}, $len, 1);
+ recv_and_run($self, my $s2 = delete $self->{0}, $len, 1);