Front page | perl.perl5.porters |
Postings from April 2012
On Sun Dec 26 17:35:35 2004, ysth wrote:
> On Mon, Dec 27, 2004 at 01:27:27AM +0000, Ton Hospel <perl5-
> porters@ton.iguana.be> wrote:
> > In article <20041226230112.GA3552@e_n.org>,
> > Yitzchak Scott-Thoennes <sthoenna@efn.org> writes:
> > > That's insufficient for overloading (which isn't handled like
> regular
> > > magic). I'd suggest just going ahead and doing an SvPV.
> > >
> > mm, didn't know that, that makes some of my XS code incomplete then.
> > But SvPV seems apporpiate enough here.
>
> The rule is, you can only check the UTF8 flag *after* calling SvPV
> (at least for 5.8.1 and later - before that only a direct stringify
> like "$x" would preserve the UTF8 flag). Same thing with stringified
> Regexp's that contain utf8 literals.
>
Running this on 5.14.2 and blead, I get:
$ perl -wlT 33186.pl
0
Can't locate object method "FETCH" via package "main" at 33186.pl line 10.
So utf8::is_utf8() was fixed, but tainted() was not. I tried with both
the XS and PP versions and got the same result.
For the XS version, the issue is in the core itself, since sv_tainted()
isn't calling get magic before checking for taintedness. So I just went
and added a SvGETMAGIC(sv); to sv_taint(). That solves this ticket (and
one TODO) but breaks one test in t/op/taint.t:
{
# Bug ID 20010730.010
my $i = 0;
sub Tie::TIESCALAR {
my $class = shift;
my $arg = shift;
bless \$arg => $class;
}
sub Tie::FETCH {
$i ++;
${$_ [0]}
}
package main;
my $bar = "The Big Bright Green Pleasure Machine";
taint_these $bar;
tie my ($foo), Tie => $bar;
my $baz = $foo;
ok $i == 1;
}
Because now FETCH gets called twice, so $i ends up as 2. Unfortunately I
can't find the bug report that the test references, and I'm already way
out of my depth here, so this is as far as I can go -- Could someone
else take a look?
---
via perlbug: queue: perl5 status: open
https://rt.perl.org:443/rt3/Ticket/Display.html?id=33186
Thread Previous