Taint::Util - Test for and flip the taint flag without regex matches or eval

    #!/usr/bin/env perl -T
    use Taint::Util;

    # eek!
    untaint $ENV{PATH};

    # $sv now tainted under taint mode (-T)
    taint(my $sv = "hlagh");

    # Untaint $sv again
    untaint $sv if tainted $sv;

Wraps perl's internal routines for checking and setting the taint flag and thus does not rely on regular expressions for untainting or odd tricks involving eval and kill for checking whether data is tainted, instead it checks and flips a flag on the scalar in-place.

Returns a boolean indicating whether a scalar is tainted. Always false when not under taint mode.

Taints or untaints given values, arrays will be flattened and their elements tainted, likewise with the values of hashes (keys can't be tainted, see perlsec). Returns no value (which evaluates to false).

    untaint(%ENV);                  # Untaints the environment
    taint(my @hlagh = qw(a o e u)); # elements of @hlagh now tainted

References (being scalars) can also be tainted, a stringified reference reference raises an error where a tainted scalar would:

    taint(my $ar = \@hlagh);
    system echo => $ar;      # err: Insecure dependency in system

This feature is used by perl internally to taint the blessed object qr// stringifies to.

    taint(my $str = "oh noes");
    my $re = qr/$str/;
    system echo => $re;      # err: Insecure dependency in system

This does not mean that tainted blessed objects with overloaded stringification via overload need return a tainted object since those objects may return a non-tainted scalar when stringified (see t/usage.t for an example). The internal handling of qr// however ensures that this holds true.

File handles can also be tainted, but this is pretty useless as the handle itself and not lines retrieved from it will be tainted, see the next section for details.

    taint(*DATA);    # *DATA tainted
    my $ln = <DATA>; # $ln not tainted

Since this module is a low level interface that directly exposes the internal SvTAINTED* functions it also presents new and exciting ways for shooting yourself in the foot.

Tainting in Perl was always meant to be used for potentially hostile external data passed to the program. Perl is passed a soup of strings from the outside; it never receives any complex datatypes directly.

For instance, you might get tainted hash keys in %ENV or tainted strings from *STDIN, but you'll never get a tainted Hash reference or a tainted subroutine. Internally, the perl compiler sets the taint flag on external data in a select few functions mainly having to do with IO and string operations. For example, the ucfirst function will manually set a tainted flag on its newly created string depending on whether the original was tainted or not.

However, since Taint::Util is exposing some of perl's guts, things get more complex. Internally, tainting is implemented via perl's MAGIC facility, which allows you to attach attach magic to any scalar, but since perl doesn't liberally taint scalars it's there to back you up if you do.

You can taint(*DATA) and tainted(*DATA) will subsequently be true but if you read from the filehandle via <DATA> you'll get untainted data back. As you might have guessed this is completely useless.

The test file t/usage.t highlights some of these edge cases.

Back in the real world, the only reason tainting makes sense is because perl will back you up when you use it, e.g. it will slap your hand if you try to pass a tainted value to system().

If you taint references, perl doesn't offer that protection, because it doesn't know anything about tainted references since it would never create one. The things that do work like the stringification of taint($t = []) (i.e. ARRAY(0x11a5d48)) being tainted only work incidentally.

But I'm not going to stop you. By all means, have at it! Just don't expect it to do anything more useful than warming up your computer.

See RT #53988 for the bug that inspired this section.

Exports tainted, taint and untaint by default. Individual functions can be exported by specifying them in the use list, to export none use ().

I wrote this when implementing re::engine::Plugin so that someone writing a custom regex engine with it wouldn't have to rely on perl regexps for untainting capture variables, which would be a bit odd.

perlsec

Ævar Arnfjörð Bjarmason <avar@cpan.org>

Copyright 2007-2010 Ævar Arnfjörð Bjarmason.

This program is free software; you can redistribute it and/or modify it under the same terms as Perl itself.