Your message dated Tue, 28 Jul 2009 21:39:55 +0300
with message-id <[email protected]>
and subject line Re: Bug#195138: Probably fixed
has caused the Debian Bug report #195138,
regarding taint checks having weird effects on performance
to be marked as done.
This means that you claim that the problem has been dealt with.
If this is not the case it is now your responsibility to reopen the
Bug report if necessary, and/or fix the problem forthwith.
(NB: If you are a system administrator and have no idea what this
message is talking about, this may indicate a serious mail system
misconfiguration somewhere. Please contact [email protected]
immediately.)
--
195138: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=195138
Debian Bug Tracking System
Contact [email protected] with problems
--- Begin Message ---
Package: perl
Version: 5.8.0-17
Here's a little script:
#!/usr/bin/perl -T
use strict;
use warnings;
open IN, "Contents-i386";
foreach my $line (<IN>)
{
if ($line =~ m,^(\S+)\t,)
{
}
}
close IN;
The input is the Contents file for unstable.
Here's it running:
asuffi...@cyclone:~/tmp$ time ./scan.pl
real 0m14.630s
user 0m13.950s
sys 0m0.530s
Now, edit the script and *remove* the -T, thusly disabling taint
checks.
asuffi...@cyclone:~/tmp$ time ./scan.pl
real 1m28.588s
user 1m26.810s
sys 0m0.520s
This would imply that taint checks make code run several times faster.
If I run this under woody, with perl 5.6.1, the effect is
reversed. That does not give me a warm fuzzy feeling. What is going on
here? Taint checks superficially appear to be working...
--
.''`. ** Debian GNU/Linux ** | Andrew Suffield
: :' : http://www.debian.org/ | Dept. of Computing,
`. `' | Imperial College,
`- -><- | London, UK
pgplQwgiWyiBc.pgp
Description: PGP signature
--- End Message ---
--- Begin Message ---
On Sun, Jul 26, 2009 at 07:31:40PM +0200, Chris Butler wrote:
> I've just had a go at running the testcase in this bug:
>
> chr...@bacon:~$ time perl scan.pl
>
> real 0m10.085s
> user 0m9.177s
> sys 0m0.700s
> chr...@bacon:~$ time perl -T scan.pl
>
> real 0m13.949s
> user 0m12.377s
> sys 0m0.956s
>
>
> From the looks of those results, I think whatever was causing this problem
> is probably fixed by now. This bug can probably be closed.
Thanks, closing. I see similarly sane numbers on 5.8.8 too FWIW.
--
Niko Tyni [email protected]
--- End Message ---
