Apache2::PerlSections - write Apache configuration files in Perl

  <Perl>
  @PerlModule = qw(Mail::Send Devel::Peek);

  #run the server as whoever starts it
  $User  = getpwuid(>) || >;
  $Group = getgrgid()) || );

  $ServerAdmin = $User;

  </Perl>

With CW<Perl>...CW</Perl> sections, it is possible to configure your server entirely in Perl.

CW<Perl> sections can contain any and as much Perl code as you wish. These sections are compiled into a special package whose symbol table mod_perl can then walk and grind the names and values of Perl variables/structures through the Apache core configuration gears.

Block sections such as CW<Location>..CW</Location> are represented in a CW%Location hash, e.g.:

  <Perl>
  $Location{"/~dougm/"} = {
    AuthUserFile   => '/tmp/htpasswd',
    AuthType       => 'Basic',
    AuthName       => 'test',
    DirectoryIndex => [qw(index.html index.htm)],
    Limit          => {
        "GET POST"    => {
            require => 'user dougm',
        }
    },
  };
  </Perl>

If an Apache directive can take two or three arguments you may push strings (the lowest number of arguments will be shifted off the CW@list) or use an array reference to handle any number greater than the minimum for that directive:

  push @Redirect, "/foo", "http://www.foo.com/";

  push @Redirect, "/imdb", "http://www.imdb.com/";

  push @Redirect, [qw(temp "/here" "http://www.there.com")];

Other section counterparts include CW%VirtualHost, CW%Directory and CW%Files.

To pass all environment variables to the children with a single configuration directive, rather than listing each one via CWPassEnv or CWPerlPassEnv, a CW<Perl> section could read in a file and:

  push @PerlPassEnv, [$key => $val];

or

  Apache2->httpd_conf("PerlPassEnv $key $val");

These are somewhat simple examples, but they should give you the basic idea. You can mix in any Perl code you desire. See eg/httpd.conf.pl and eg/perl_sections.txt in the mod_perl distribution for more examples.

Assume that you have a cluster of machines with similar configurations and only small distinctions between them: ideally you would want to maintain a single configuration file, but because the configurations aren't exactly the same (e.g. the CWServerName directive) it's not quite that simple.

CW<Perl> sections come to rescue. Now you have a single configuration file and the full power of Perl to tweak the local configuration. For example to solve the problem of the CWServerName directive you might have this CW<Perl> section:

  <Perl>
  $ServerName = `hostname`;
  </Perl>

For example if you want to allow personal directories on all machines except the ones whose names start with secure:

  <Perl>
  $ServerName = `hostname`;
  if ($ServerName !~ /^secure/) {
      $UserDir = "public.html";
  }
  else {
      $UserDir = "DISABLED";
  }
  </Perl>

  push @PerlConfig, 'Alias /foo /bar';

Or: CW$PerlConfig .= Alias /foo /bar\n;

See also CW$r->add_config

There are a few variables that can be set to change the default behaviour of CW<Perl> sections. Each CW<Perl> section is evaluated in its unique namespace, by default residing in a sub-namespace of CWApache2::ReadConfig::, therefore any local variables will end up in that namespace. For example if a CW<Perl> section happened to be in file /tmp/httpd.conf starting on line 20, the namespace: CWApache2::ReadConfig::tmp::httpd_conf::line_20 will be used. Now if it had:

  <Perl>
    $foo     = 5;
    my $bar  = 6;
    $My::tar = 7;
  </Perl>

The local global variable CW$foo becomes CW$Apache2::ReadConfig::tmp::httpd_conf::line_20::foo, the other variable remain where they are.

By default, the namespace in which CW<Perl> sections are evaluated is cleared after each block closes. In our example nuking CW$Apache2::ReadConfig::tmp::httpd_conf::line_20::foo, leaving the rest untouched.

By setting CW$Apache2::PerlSections::Save to a true value, the content of those namespaces will be preserved and will be available for inspection by CWApache2::Status and CWApache2::PerlSections->dump In our example CW$Apache2::ReadConfig::tmp::httpd_conf::line_20::foo will still be accessible from other perl code, after the CW<Perl> section was parsed.

This method will dump out all the configuration variables mod_perl will be feeding to the apache config gears. The output is suitable to read back in via CWeval.

  my $dump = Apache2::PerlSections->dump;

For example:

  <Perl>

  $Apache2::PerlSections::Save = 1;

  $Listen = 8529;

  $Location{"/perl"} = {
     SetHandler => "perl-script",
     PerlHandler => "ModPerl::Registry",
     Options => "ExecCGI",
  };

  @DirectoryIndex = qw(index.htm index.html);

  $VirtualHost{"www.foo.com"} = {
     DocumentRoot => "/tmp/docs",
     ErrorLog => "/dev/null",
     Location => {
       "/" => {
         Allowoverride => 'All',
         Order => 'deny,allow',
         Deny  => 'from all',
         Allow => 'from foo.com',
       },
     },
  };
  </Perl>

  <Perl>
  print Apache2::PerlSections->dump;
  </Perl>

This will print something like this:

  $Listen = 8529;

  @DirectoryIndex = (
    'index.htm',
    'index.html'
  );

  $Location{'/perl'} = (
      PerlHandler => 'Apache2::Registry',
      SetHandler => 'perl-script',
      Options => 'ExecCGI'
  );

  $VirtualHost{'www.foo.com'} = (
      Location => {
        '/' => {
          Deny => 'from all',
          Order => 'deny,allow',
          Allow => 'from foo.com',
          Allowoverride => 'All'
        }
      },
      DocumentRoot => '/tmp/docs',
      ErrorLog => '/dev/null'
  );

  1;
  __END__

It is important to put the call to CWdump in it's own CW<Perl> section, otherwise the content of the current CW<Perl> section will not be dumped. This method will call the CWdump method, writing the output to a file, suitable to be pulled in via CWrequire or CWdo.

  Apache2::PerlSections->store($filename);

ret: no return value

mod_perl 2.0 now introduces the same general concept of handlers to CW<Perl> sections. Apache2::PerlSections simply being the default handler for them.

To specify a different handler for a given perl section, an extra handler argument must be given to the section:

  <Perl handler="My::PerlSection::Handler" somearg="test1">
    $foo = 1;
    $bar = 2;
  </Perl>

And in My/PerlSection/Handler.pm:

  sub My::Handler::handler : handler {
      my ($self, $parms, $args) = @_;
      #do your thing!
  }

So, when that given CW<Perl> block in encountered, the code within will first be evaluated, then the handler routine will be invoked with 3 arguments: self-explanatory CW$parms is specific for the current Container, for example, you might want to call CW$parms->server() to get the current server. the table object of the section arguments. The 2 guaranteed ones will be:

  $args->{'handler'} = 'My::PerlSection::Handler';
  $args->{'package'} = 'Apache2::ReadConfig';

At this point, it's up to the handler routing to inspect the namespace of the CW$args->{'package'} and chooses what to do.

The most likely thing to do is to feed configuration data back into apache. To do that, use Apache2::Server->add_config(directive), for example:

  $parms->server->add_config("Alias /foo /bar");

Would create a new alias. The source code of CWApache2::PerlSections is a good place to look for a practical example. If the CW<Perl> sections include no code requiring a running mod_perl, it is possible to check those from the command line. But the following trick should be used:

  # file: httpd.conf
  <Perl>
  #!perl

  # ... code here ...

  __END__
  </Perl>

Now you can run:

  % perl -c httpd.conf

httpd-2.0.47 had a bug in the configuration parser which caused the startup failure with the following error:

  Starting httpd:
  Syntax error on line ... of /etc/httpd/conf/httpd.conf:
  <Perl> directive missing closing '>'     [FAILED]

This has been fixed in httpd-2.0.48. If you can't upgrade to this or a higher version, please add a space before the closing '>' of the opening tag as a workaround. So if you had:

  <Perl>
  # some code
  </Perl>

change it to be:

  <Perl >
  # some code
  </Perl>

On encountering a one-line <Perl> block, httpd's configuration parser will cause a startup failure with an error similar to this one:

  Starting httpd:
  Syntax error on line ... of /etc/httpd/conf/httpd.conf:
  <Perl>use> was not closed.

If you have written a simple one-line <Perl> section like this one :

  <Perl>use Apache::DBI;</Perl>

change it to be:

   <Perl>
   use Apache::DBI;
   </Perl>

This is caused by a limitation of httpd's configuration parser and is not likely to be changed to allow one-line block like the example above. Use multi-line blocks instead.

mod_perl 2.0 documentation.

mod_perl 2.0 and its core modules are copyrighted under The Apache Software License, Version 2.0.

The mod_perl development team and numerous contributors.