=head1 NAME
ModPerl::Util - Helper mod_perl Functions
=head1 Synopsis
use ModPerl::Util;
# e.g. PerlResponseHandler
$callback = ModPerl::Util::current_callback;
# exit w/o killing the interpreter
ModPerl::Util::exit();
# untaint a string (do not use it! see the doc)
ModPerl::Util::untaint($string);
# removes a stash (.so, %INC{$stash}, etc.) as best as it can
ModPerl::Util::unload_package($stash);
# current perl's address (0x92ac760 or 0x0 under non-threaded perl)
ModPerl::Util::current_perl_id();
=head1 Description
C<ModPerl::Util> provides mod_perl utilities API.
=head1 API
C<ModPerl::Util> provides the following functions and/or methods:
=head2 C<current_callback>
Returns the currently running callback name,
e.g. C<'PerlResponseHandler'>.
$callback = ModPerl::Util::current_callback();
=over 4
=item ret: C<$callback> ( string )
=item since: 2.0.00
=back
=head2 C<current_perl_id>
Return the memory address of the perl interpreter
$perl_id = ModPerl::Util::current_perl_id();
=over 4
=item ret: C<$perl_id> ( string )
Under threaded perl returns something like: C<0x92ac760>
Under non-thread perl returns C<0x0>
=item since: 2.0.00
=back
Mainly useful for debugging applications running under threaded-perl.
=head2 C<exit>
Terminate the request, but not the current process (or not the current
Perl interpreter with threaded mpms).
ModPerl::Util::exit($status);
=over 4
=item opt arg1: C<$status> ( integer )
The exit status, which as of this writing is ignored. (it's accepted
to be compatible with the core C<exit> function.)
=item ret: no return value
=item since: 2.0.00
=back
Normally you will use the plain C<exit()> in your code. You don't need
to use C<ModPerl::Util::exit> explicitly, since mod_perl overrides
C<exit()> by setting C<CORE::GLOBAL::exit> to
C<ModPerl::Util::exit>. Only if you redefine C<CORE::GLOBAL::exit>
once mod_perl is running, you may want to use this function.
The original C<exit()> is still available via C<CORE::exit()>.
C<ModPerl::Util::exit> is implemented as a special C<die()> call,
therefore if you call it inside C<eval BLOCK> or C<eval "STRING">,
while an exception is being thrown, it is caught by C<eval>. For
example:
exit;
print "Still running";
will not print anything. But:
eval {
exit;
}
print "Still running";
will print I<Still running>. So you either need to check whether L<the
exception|docs::2.0::api::APR::Error> is specific to C<exit> and call
C<exit()> again:
use ModPerl::Const -compile => 'EXIT';
eval {
exit;
}
exit if $@ && ref $@ eq 'APR::Error' && $@ == ModPerl::EXIT;
print "Still running";
or use C<CORE::exit()>:
eval {
CORE::exit;
}
print "Still running";
and nothing will be printed. The problem with the latter is the
current process (or a Perl Interpreter) will be killed; something that
you really want to avoid under mod_perl.
=head2 C<unload_package>
Unloads a stash from the current Perl interpreter in the safest way
possible.
ModPerl::Util::unload_package($stash);
=over 4
=item arg1: C<$stash> ( string )
The Perl stash to unload. e.g. C<MyApache2::MyData>.
=item ret: no return value
=item since: 2.0.00
=back
Unloading a Perl stash (package) is a complicated business. This
function tries very hard to do the right thing. After calling this
function, it should be safe to C<use()> a new version of the module
that loads the wiped package.
References to stash elements (functions, variables, etc.) taken from
outside the unloaded package will still be valid.
This function may wipe off things loaded by other modules, if the
latter have inserted things into the C<$stash> it was told to unload.
If a stash had a corresponding XS shared object (.so) loaded it will
be unloaded as well.
If the stash had a corresponding entry in C<%INC>, it will be removed
from there.
C<unload_package()> takes care to leave sub-stashes intact while
deleting the requested stash. So for example if C<CGI> and
C<CGI::Carp> are loaded, calling C<unload_package('CGI')> won't affect
C<CGI::Carp>.
=head2 C<untaint>
Untaint the variable, by turning its tainted SV flag off (used
internally).
ModPerl::Util::untaint($tainted_var);
=over 4
=item arg1: C<$tainted_var> (scalar)
=item ret: no return value
C<$tainted_var> is untainted.
=item since: 2.0.00
=back
Do not use this function unless you know what you are doing. To learn
how to properly untaint variables refer to the I<perlsec> manpage.
=head1 See Also
L<mod_perl 2.0 documentation|docs::2.0::index>.
=head1 Copyright
mod_perl 2.0 and its core modules are copyrighted under
The Apache Software License, Version 2.0.
=head1 Authors
L<The mod_perl development team and numerous
contributors|about::contributors::people>.
=cut
