The London Perl and Raku Workshop takes place on 26th Oct 2024. If your company depends on Perl, please consider sponsoring and/or attending.

NAME

Plack::Middleware::Rewrite - mod_rewrite for Plack

SYNOPSIS

# in app.psgi
use Plack::Builder;

builder {
    enable 'Rewrite', request => sub {
        s{^/here(?=/|$)}{/there};

        return [303]
            if s{^/foo/?$}{/bar/}
            or s{^/baz/?$}{/quux/};

        return [301, [ Location => 'http://example.org/' ], []]
            if m{^/example/?$};

        return [201] if $_ eq '/favicon.ico';

        return [503] if -e '/path/to/app/maintenance.lock';

        return [200, [qw(Content-Type text/plain)], ['You found it!']]
            if $_ eq '/easter-egg';
    },
    response => sub {
        $_->status( 303 )
            if $_->status eq 201 and $_->get( 'Location' );

        $_->set( 'Content-Type', 'application/xhtml+xml' )
            if ( $_[0]{'HTTP_ACCEPT'} || '' ) =~ m{application/xhtml\+xml(?!\s*;\s*q=0)};
    };
    $app;
};

DESCRIPTION

This middleware provides a convenient way to modify requests in flight in Plack apps. Rewrite rules are simply written in Perl, which means everything that can be done with mod_rewrite can be done with this middleware much more intuitively (if in syntactically wordier ways). Its primary purpose is rewriting paths, but almost anything is possible very easily.

CONFIGURATION OPTIONS

request

Takes a reference to a function that will be called in scalar context for each request. On call, $_ will be aliased to PATH_INFO, so that you can easily use regexp matches and subtitutions to examine and modify it. The PSGI environment will be passed to the function as its first and only argument.

The function may return three kinds of valid value:

A plain scalar

Ignored. The value will be thrown away and any path rewriting (or any other modifications of the PSGI environment) will take effect during the current request cycle, invisibly to the client user agent.

An array reference

A PSGI array response to return immediately without invoking the wrapped PSGI application.

The array may have fewer than 3 elements, in which case it will be filled to 3 elements by pushing the default values: an empty body array, empty headers array, and a 303 status code.

If the Location header is missing from a redirect response (i.e. one with 3xx status code), it will be filled in automatically from the value left in PATH_INFO by your callback. (Note that this only allows you to redirect to URLs with the same hostname. To redirect the client to a different host, you will have to supply a Location header manually.)

A code reference

A PSGI application which will be called to process the request. This prevents the wrapped application from being called.

Any other kind of reference

Error. An exception will be thrown.

response

Takes a reference to a function that will be called after the request has been processed and the response is ready to be returned.

On call, $_ will be aliased to a specially extended a Plack::Util::headers object for the response, for convenient alteration of headers. The extension is a status method, which allows you to inspect and modify the response status code.

Just as in "request", the PSGI environment is passed as first and only argument.

Any return value from this function will be ignored unless it is a code reference. In that case it will be used to filter the response body, as documented in "RESPONSE CALLBACK" in Plack::Middleware:

    return sub {
        my $chunk = shift;
        return unless defined $chunk;
        $chunk =~ s/Foo/Bar/g;
        return $chunk;
    };

    The callback takes one argument $chunk and your callback is expected to return the updated chunk. If the given $chunk is undef, it means the stream has reached the end, so your callback should also return undef, or return the final chunk and return undef when called next time.

LEGACY INTERFACE

The old interface uses a single attribute, rules, instead of the request and response pair, with a more complex set of return values, containing an ambiguity. It is also less expressive than the new interface.

The old interface is documented here for the purposes of maintaining old code; its use in new code is discouraged. In the far future it may get removed entirely, and in the meantime it will not gain new features.

The return value of the rules callback is interpreted as follows:

An array reference (with at least one element)

A regular PSGI response, except that you may omit either or both the headers and body elements. You may not omit the status.

A scalar value that looks like an HTTP status

Like returning a reference to a one-element.

Beware: every subroutine in Perl has a return value, even if you do not return anything explicitly. To avoid ambiguities you must return one-element arrays instead of plain values and use an explicit return at the end of your rules:

return [201] if $_ eq '/favicon.ico';
s{^/here(?=/|$)}{/there};
return;
A code reference

Equivalent to the "response" callback in the new interface, with the same arguments and return values.

Any other kind of value

Internal rewrite.

Porting from the old to the new interface

There are two major incompatibilities between the interfaces:

  1. You can no longer return status codes as plain scalars, as in return 301. You must now return [301] (which you could before, but didn't have to).

  2. Rewriting the response is no longer done by returning a sub. Instead you must use the response attribute.

    This may be inconvient if the function was closing over variables from the rules callback; in that case you now have to explicitly pass that state from one callback to the other through the environment hash. However, such code is rare, and in all other cases your code will be more readable under the new interface.

AUTHOR

Aristotle Pagaltzis <pagaltzis@gmx.de>

COPYRIGHT AND LICENSE

This software is copyright (c) 2022 by Aristotle Pagaltzis.

This is free software; you can redistribute it and/or modify it under the same terms as the Perl 5 programming language system itself.