basic support for delayed writes/async with docs

[catagits/Catalyst-Runtime.git] / lib / Catalyst / Upgrading.pod

=head1 NAME

Catalyst::Upgrading - Instructions for upgrading to the latest Catalyst

=head1 Upgrading to Catalyst 5.90040

This version of L<Catalyst> offers some support for using L<AnyEvent> and
L<IO::Async> event loops in your application.  In order to achieve this goal
we needed to make some changes to the way the we finalize the HTTP response
such that sloppy code that closed over $c and leaked memory will no longer
work in some manner.  For example you might accidently have:

    $c->stash(my_model => sub { $c->model->find(shift) });

If you have old code that leaks memory in this way but otherwise seemed to
work, it will no longer complete the response properly.

If you don't want to fix your code, you can force the old behavior with the
global configuration key C<aggressively_close_writer_on_finalize_body>.  This
of course will still leave you with a leaky application and you lose the new
event loop support, but your application will go back to completing its
response output.  For example:

    package MyApp::Web;

    use Moose;
    use Catalyst;

    __PACKAGE__->config(
      name => 'MyApp::Web',
      enable_catalyst_header => 1,
      disable_component_resolution_regex_fallback => 1,
      aggressively_close_writer_on_finalize_body => 1,
    );

    __PACKAGE__->setup;

See L<Catalyst::Component::ContextClosure> for help on how to close over the
context safely, should you need to do this.  See L<CatalystX::LeakChecker>
and L<Catalyst::Controller::LeakTracker> for help if you want to solve your
memory leak issues.

=head1 Upgrading to Catalyst 5.9

The major change is that L<Plack>, a toolkit for using the L<PSGI>
specification, now replaces most of the subclasses of L<Catalyst::Engine>. If
you are using one of the standard subclasses of L<Catalyst::Engine> this
should be a straightforward upgrade for you. It was a design goal for
this release to preserve as much backwards compatibility as possible.
However, since L<Plack> is different from L<Catalyst::Engine>, it is
possible that differences exist for edge cases. Therefore, we recommend
that care be taken with this upgrade and that testing should be greater
than would be the case with a minor point update. Please inform the
Catalyst developers of any problems so that we can fix them and
incorporate tests.

It is highly recommended that you become familiar with the L<Plack> ecosystem
and documentation. Being able to take advantage of L<Plack> development and
middleware is a major bonus to this upgrade. Documentation about how to
take advantage of L<Plack::Middleware> by writing your own C<< .psgi >> file
is contained in L<Catalyst::PSGI>.

If you have created a custom subclass of L<Catalyst:Engine>, you will
need to convert it to be a subclass of L<Plack::Handler>.

If you are using the L<Plack> engine, L<Catalyst::Engine::PSGI>, this new
release supersedes that code.

If you are using a subclass of L<Catalyst::Engine> that is aimed at
nonstandard or internal/testing uses, such as
L<Catalyst::Engine::Embeddable>, you should still be able to continue
using that engine.

Advice for specific subclasses of L<Catalyst::Engine> follows:

=head2 Upgrading the FastCGI Engine

No upgrade is needed if your myapp_fastcgi.pl script is already upgraded
to use L<Catalyst::Script::FastCGI>.

=head2 Upgrading the mod_perl / Apache Engines

The engines that are built upon the various iterations of mod_perl,
L<Catalyst::Engine::Apache::MP13> (for mod_perl 1, and Apache 1.x) and
L<Catalyst::Engine::Apache2::MP20> (for mod_perl 2, and Apache 2.x),
should be seamless upgrades and will work using using L<Plack::Handler::Apache1>
or L<Plack::Handler::Apache2> as required.

L<Catalyst::Engine::Apache2::MP19>, however, is no longer supported, as
Plack does not support mod_perl version 1.99. This is unlikely to be a
problem for anyone, as 1.99 was a brief beta-test release for mod_perl
2, and all users of mod_perl 1.99 are encouraged to upgrade to a
supported release of Apache 2 and mod_perl 2.

=head2 Upgrading the HTTP Engine

The default development server that comes with the L<Catalyst> distribution
should continue to work as expected with no changes as long as your C<myapp_server>
script is upgraded to use L<Catalyst::Script::HTTP>.

=head2 Upgrading the CGI Engine

If you were using L<Catalyst::Engine::CGI> there is no upgrade needed if your
myapp_cgi.pl script is already upgraded to use L<Catalyst::Script::CGI>.

=head2 Upgrading Catalyst::Engine::HTTP::Prefork

If you were using L<Catalyst::Engine::HTTP::Prefork> then L<Starman>
is automatically loaded. You should (at least) change your C<Makefile.PL>
to depend on Starman.

You can regenerate your C<myapp_server.pl> script with C<catalyst.pl>
and implement a C<MyApp::Script::Server> class that looks like this:

    package MyApp::Script::Server;
    use Moose;
    use namespace::autoclean;

    extends 'CatalystX::Script::Server::Starman';

    1;

This takes advantage of the new script system, and will add a number of
options to the standard server script as extra options are added by
Starman.

More information about these options can be seen at
L<CatalystX::Script::Server::Starman/SYNOPSIS>.

An alternate route to implement this functionality is to write a simple .psgi
file for your application, and then use the L<plackup> utility to start the
server.

=head2 Upgrading the PSGI Engine

If you were using L<Catalyst::Engine::PSGI>, this new release supersedes
this engine in supporting L<Plack>. By default the Engine is now always
L<Plack>. As a result, you can remove the dependency on
L<Catalyst::Engine::PSGI> in your C<Makefile.PL>.

Applications that were using L<Catalyst::Engine::PSGI>
previously should entirely continue to work in this release with no changes.

However, if you have an C<app.psgi> script, then you no longer need to
specify the PSGI engine. Instead, the L<Catalyst> application class now
has a new method C<psgi_app> which returns a L<PSGI> compatible coderef
which you can wrap in the middleware of your choice.

Catalyst will use the .psgi for your application if it is located in the C<home>
directory of the application.

For example, if you were using L<Catalyst::Engine::PSGI> in the past, you will
have written (or generated) a C<script/myapp.psgi> file similar to this one:

    use Plack::Builder;
    use MyCatalytApp;

    MyCatalystApp->setup_engine('PSGI');

    builder {
        enable ... # enable your desired middleware
        sub { MyCatalystApp->run(@_) };
    };

Instead, you now say:

    use Plack::Builder;
    use MyCatalystApp;

    builder {
        enable ... #enable your desired middleware
        MyCatalystApp->psgi_app;
    };

In the simplest case:

    MyCatalystApp->setup_engine('PSGI');
    my $app = sub { MyCatalystApp->run(@_) }

becomes

    my $app = MyCatalystApp->psgi_app(@_);

B<NOT>:

    my $app = sub { MyCatalystApp->psgi_app(@_) };
    # If you make ^^ this mistake, your app won't work, and will confuse the hell out of you!

You can now move C<< script/myapp.psgi >> to C<< myapp.psgi >>, and the built-in
Catalyst scripts and your test suite will start using your .psgi file.

B<NOTE:> If you rename your .psgi file without these modifications, then
any tests run via L<Catalyst::Test> will not be compatible with the new
release, and will result in the development server starting, rather than
the expected test running.

B<NOTE:> If you are directly accessing C<< $c->req->env >> to get the PSGI
environment then this accessor is moved to C<< $c->engine->env >>,
you will need to update your code.

=head2 Engines which are known to be broken

The following engines B<DO NOT> work as of Catalyst version 5.9. The
core team will be happy to work with the developers and/or users of
these engines to help them port to the new Plack/Engine system, but for
now, applications which are currently using these engines B<WILL NOT>
run without modification to the engine code.

=over

=item Catalyst::Engine::Wx

=item Catalyst::Engine::Zeus

=item Catalyst::Engine::JobQueue::POE

=item Catalyst::Engine::XMPP2

=item Catalyst::Engine::SCGI

=back

=head2 Engines with unknown status

The following engines are untested or have unknown compatibility.
Reports are highly encouraged:

=over

=item Catalyst::Engine::Mojo

=item Catalyst::Engine::Server (marked as Deprecated)

=item Catalyst::Engine::HTTP::POE (marked as Deprecated)

=back

=head2 Plack functionality

See L<Catalyst::PSGI>.

=head2 Tests in 5.9

Tests should generally work the same in Catalyst 5.9, but there are
some differences.

Previously, if using L<Catalyst::Test> and doing local requests (against
a local server), if the application threw an exception then this
exception propagated into the test.

This behavior has been removed, and now a 500 response will be returned
to the test. This change standardizes behavior, so that local test
requests behave similarly to remote requests.

=head2 Regex dispatch type is deprecated.

The Regex dispatchtype (L<Catalyst::DispatchType::Regex>) has been deprecated.

You are encouraged to move your application to Chained dispatch (L<Catalyst::DispatchType::Chained>).

If you cannot do so, please add a dependency to Catalyst::DispatchType::Regex to your application's
Makefile.PL

=head1 Upgrading to Catalyst 5.80

Most applications and plugins should run unaltered on Catalyst 5.80.

However, a lot of refactoring work has taken place, and several changes have
been made which could cause incompatibilities. If your application or plugin
is using deprecated code, or relying on side effects, then you could have
issues upgrading to this release.

Most issues found with existing components have been easy to
solve. This document provides a complete description of behavior changes
which may cause compatibility issues, and of new Catalyst warnings which
might be unclear.

If you think you have found an upgrade-related issue which is not covered in
this document, please email the Catalyst list to discuss the problem.

=head1 Moose features

=head2 Application class roles

You can only apply method modifiers after the application's C<< ->setup >>
method has been called. This means that modifiers will not work with methods
run during the call to C<< ->setup >>.

See L<Catalyst::Manual::ExtendingCatalyst> for more information about using
L<Moose> in your applications.

=head2 Controller actions in Moose roles

You can use L<MooseX::MethodAttributes::Role> if you want to declare actions
inside Moose roles.

=head2 Using Moose in Components

The correct way to use Moose in a component in a both forward and backwards
compatible way is:

    package TestApp::Controller::Root;
    use Moose;
    BEGIN { extends 'Catalyst::Component' }; # Or ::Controller, or whatever

See L<Components which inherit from Moose::Object before Catalyst::Component>.

=head1 Known backwards compatibility breakages

=head2 Applications in a single file

Applications must be in their own file, and loaded at compile time. This
issue generally only affects the tests of CPAN distributions. Your
application will fail if you try to define an application inline in a
block, and use plugins which supply a C< new > method, then use that
application latter in tests within the same file.

This is due to the fact that Catalyst is inlining a new method on your
application class allowing it to be compatible with Moose. The method
used to do this changed in 5.80004 to avoid the possibility of reporting
an 'Unknown Error' if your application failed to compile.

=head2 Issues with Class::C3

Catalyst 5.80 uses the L<Algorithm::C3> method dispatch order. This is
built into Perl 5.10, and comes via L<Class::C3> for Perl 5.8. This
replaces L<NEXT> with L<Class::C3::Adopt::NEXT>, forcing all components
to resolve methods using C3, rather than the unpredictable dispatch
order of L<NEXT>.

This issue manifests itself by your application failing to start due to an
error message about having a non-linear @ISA.

The Catalyst plugin most often causing this is
L<Catalyst::Plugin::Session::Store::FastMmap> - if you are using this
plugin and see issues, then please upgrade your plugins, as it has been
fixed. Note that Makefile.PL in the distribution will warn about known
incompatible components.

This issue can, however, be found in your own application - the only solution is
to go through each base class of the class the error was reported against, until
you identify the ones in conflict, and resolve them.

To be able to generate a linear @ISA, the list of superclasses for each
class must be resolvable using the C3 algorithm. Unfortunately, when
superclasses are being used as mixins (to add functionality used in your class),
and with multiple inheritance, it is easy to get this wrong.

Most common is the case of:

    package Component1; # Note, this is the common case
    use base qw/Class::Accessor::Fast Class::Data::Inheritable/;

    package Component2; # Accidentally saying it this way causes a failure
    use base qw/Class::Data::Inheritable Class::Accessor::Fast/;

    package GoesBang;
    use base qw/Component1 Component2/;

Any situation like this will cause your application to fail to start.

For additional documentation about this issue, and how to resolve it, see
L<Class::C3::Adopt::NEXT>.

=head2 Components which inherit from Moose::Object before Catalyst::Component

Moose components which say:

    package TestApp::Controller::Example;
    use Moose;
    extends qw/Moose::Object Catalyst::Component/;

to use the constructor provided by Moose, while working (if you do some hacks
with the C< BUILDARGS > method), will not work with Catalyst 5.80 as
C<Catalyst::Component> inherits from C<Moose::Object>, and so C< @ISA > fails
to linearize.

The correct way to use Moose in a component in a both forward and backwards
compatible way is:

    package TestApp::Controller::Root;
    use Moose;
    BEGIN { extends 'Catalyst::Component' }; # Or ::Controller, or whatever

Note that the C< extends > declaration needs to occur in a begin block for
L<attributes> to operate correctly.

This way you do not inherit directly from C<Moose::Object>
yourself. Having components which do not inherit their constructor from
C<Catalyst::Component> is B<unsupported>, and has never been recommended,
therefore you're on your own if you're using this technique. You'll need
to detect the version of Catalyst your application is running, and deal
with it appropriately.

You also don't get the L<Moose::Object> constructor, and therefore attribute
initialization will not work as normally expected. If you want to use Moose
attributes, then they need to be made lazy to correctly initialize.

Note that this only applies if your component needs to maintain component
backwards compatibility for Catalyst versions before 5.71001 - in 5.71001
attributes work as expected, and the BUILD method is called normally
(although BUILDARGS is not).

If you depend on Catalyst 5.8, then B<all> Moose features work as expected.

You will also see this issue if you do the following:

    package TestApp::Controller::Example;
    use Moose;
    use base 'Catalyst::Controller';

as C< use base > appends to @ISA.

=head3 use Moose in MyApp

Similar to the above, this will also fail:

    package MyApp;
    use Moose;
    use Catalyst qw/
      ConfigLoader
    /;
    __PACKAGE__->setup;

If you need to use Moose in your application class (e.g. for method modifiers
etc.) then the correct technique is:

    package MyApp;
    use Moose;
    use Catalyst;

    extends 'Catalyst';

    __PACKAGE__->config( name => 'MyApp' );
    __PACKAGE__->setup(qw/
        ConfigLoader
    /);

=head2 Anonymous closures installed directly into the symbol table

If you have any code which installs anonymous subroutine references directly
into the symbol table, you may encounter breakages. The simplest solution is
to use L<Sub::Name> to name the subroutine. Example:

    # Original code, likely to break:
    my $full_method_name = join('::', $package_name, $method_name);
    *$full_method_name = sub { ... };

    # Fixed Code
    use Sub::Name 'subname';
    my $full_method_name = join('::',$package_name, $method_name);
    *$full_method_name = subname $full_method_name, sub { ... };

Additionally, you can take advantage of Catalyst's use of L<Class::MOP> and
install the closure using the appropriate metaclass. Example:

    use Class::MOP;
    my $metaclass = Moose::Meta::Class->initialize($package_name);
    $metaclass->add_method($method_name => sub { ... });

=head2 Hooking into application setup

To execute code during application start-up, the following snippet in MyApp.pm
used to work:

    sub setup {
        my ($class, @args) = @_;
        $class->NEXT::setup(@args);
        ... # things to do after the actual setup
    }

With Catalyst 5.80 this won't work anymore, because Catalyst no longer
uses NEXT.pm for method resolution. The functionality was only ever
originally operational as L<NEXT> remembers what methods have already
been called, and will not call them again.

Using this now causes infinite recursion between MyApp::setup and
Catalyst::setup, due to other backwards compatibility issues related to how
plugin setup works. Moose method modifiers like C<< before|after|around setup
=> sub { ... }; >> also will not operate correctly on the setup method.

The right way to do it is this:

    after setup_finalize => sub {
        ... # things to do after the actual setup
    };

The setup_finalize hook was introduced as a way to avoid this issue.

=head2 Components with a new method which returns false

Previously, if you had a component which inherited from Catalyst::COMPONENT,
but overrode the new method to return false, then your class's configuration
would be blessed into a hash on your behalf, and this would be returned from
the COMPONENT method.

This behavior makes no sense, and so has been removed. Implementing your own
C< new > method in components is B<highly> discouraged. Instead, you should
inherit the new method from Catalyst::Component, and use Moose's BUILD
functionality and/or Moose attributes to perform any construction work
necessary for your class.

=head2 __PACKAGE__->mk_accessor('meta');

Won't work due to a limitation of L<Moose>. This is currently being fixed
inside Moose.

=head2 Class::Data::Inheritable side effects

Previously, writing to a class data accessor would copy the accessor method
down into your package.

This behavior has been removed. While the class data is still stored
per-class, it is stored on the metaclass of the class defining the accessor.

Therefore anything relying on the side effect of the accessor being copied down
will be broken.

The following test demonstrates the problem:

    {
        package BaseClass;
        use base qw/Class::Data::Inheritable/;
        __PACKAGE__->mk_classdata('foo');
    }

    {
        package Child;
        use base qw/BaseClass/;
    }

    BaseClass->foo('base class');
    Child->foo('sub class');

    use Test::More;
    isnt(BaseClass->can('foo'), Child->can('foo'));

=head2 Extending Catalyst::Request or other classes in an ad hoc manner using mk_accessors

Previously, it was possible to add additional accessors to Catalyst::Request
(or other classes) by calling the mk_accessors class method.

This is no longer supported - users should make a subclass of the class whose
behavior they would like to change, rather than globally polluting the
Catalyst objects.

=head2 Confused multiple inheritance with Catalyst::Component::COMPONENT

Previously, Catalyst's COMPONENT method would delegate to the method on
the right hand side, which could then delegate back again with
NEXT. This is poor practice, and in addition, makes no sense with C3
method dispatch order, and is therefore no longer supported.

If a COMPONENT method is detected in the inheritance hierarchy to the right
hand side of Catalyst::Component::COMPONENT, then the following warning
message will be emitted:

    There is a COMPONENT method resolving after Catalyst::Component
    in ${next_package}.

The correct fix is to re-arrange your class's inheritance hierarchy so that the
COMPONENT method you would like to inherit is the first (left-hand most)
COMPONENT method in your @ISA.

=head2 Development server relying on environment variables

Previously, the development server would allow propagation of system
environment variables into the request environment, this has changed with the
adoption of Plack. You can use L<Plack::Middleware::ForceEnv> to achieve the
same effect.

=head1 WARNINGS

=head2 Actions in your application class

Having actions in your application class will now emit a warning at application
startup as this is deprecated. It is highly recommended that these actions are moved
into a MyApp::Controller::Root (as demonstrated by the scaffold application
generated by catalyst.pl). 

This warning, also affects tests. You should move actions in your test,
creating a myTest::Controller::Root, like the following example:

    package MyTest::Controller::Root;

    use strict;
    use warnings;

    use parent 'Catalyst::Controller';

    __PACKAGE__->config(namespace => '');

    sub action : Local {
        my ( $self, $c ) = @_;
        $c->do_something; 
    }

    1;

=head2 ::[MVC]:: naming scheme

Having packages called MyApp::[MVC]::XX is deprecated and can no longer be generated
by catalyst.pl

This is still supported, but it is recommended that you rename your application
components to Model/View/Controller.

A warning will be issued at application startup if the ::[MVC]:: naming scheme is
in use.

=head2 Catalyst::Base

Any code using L<Catalyst::Base> will now emit a warning; this
module will be removed in a future release.

=head2 Methods in Catalyst::Dispatcher

The following methods in Catalyst::Dispatcher are implementation
details, which may change in the 5.8X release series, and therefore their use
is highly deprecated.

=over

=item tree

=item dispatch_types

=item registered_dispatch_types

=item method_action_class

=item action_hash

=item container_hash

=back

The first time one of these methods is called, a warning will be emitted:

    Class $class is calling the deprecated method Catalyst::Dispatcher::$public_method_name,
    this will be removed in Catalyst 5.9

You should B<NEVER> be calling any of these methods from application code.

Plugin authors and maintainers whose plugins currently call these methods
should change to using the public API, or, if you do not feel the public API
adequately supports your use case, please email the development list to
discuss what API features you need so that you can be appropriately supported.

=head2 Class files with names that don't correspond to the packages they define

In this version of Catalyst, if a component is loaded from disk, but no
symbols are defined in that component's name space after it is loaded, this
warning will be issued:

    require $class was successful but the package is not defined.

This is to protect against confusing bugs caused by mistyping package names,
and will become a fatal error in a future version.

Please note that 'inner packages' (via L<Devel::InnerPackage>) are still fully
supported; this warning is only issued when component file naming does not map
to B<any> of the packages defined within that component.

=head2 $c->plugin method

Calling the plugin method is deprecated, and calling it at run time is B<highly
deprecated>.

Instead you are recommended to use L<Catalyst::Model::Adaptor> or similar to
compose the functionality you need outside of the main application name space.

Calling the plugin method will not be supported past Catalyst 5.81.

=cut