This is a framework based around the Model-View-Controller pattern and modeled after Ruby on Rails 1.2. It is compatible with PHP version 5.1.4 and later. Documentation can be found at: http://php-framework.readthedocs.org.
The repository shares the same layout as an application built with the framework. This is mostly because of historical reasons.
You can use ./script/createapp appname
to generate an application
directory in the current directory. Adjust the path to the createapp
script to create the application in a different place.
If you are familiar with Rails, using the framework should be rather straightfoward at this point.
The framework expects its index.php
file is located in the DocumentRoot
of
an Apache VirtualHost
and that mod_rewrite
is enabled. If you aren't familiar
with how to set this up, the framework is probably not for you. We haven't
made any attempt to support running it out of a subdirectory or on shared hosting.
One noticeable difference from Rails is that files are named with uppercase
as in PostsController.php
rather than posts_controller.php
. This is
largely because our libraries use PEAR-like conventions, where the uppercase
naming style is used. Browse the repository directories under app/
to see
the naming of controllers, models, and views. Similarly, we use camelCase
names for methods (e.g. respondTo
rather than respond_to
).
You can generate stubs with ./script/generate
. You'll probably want to
start with ./script/generate model post
, which will create a Post
model
file with associated migration and test files. Use ./script/generate
with no arguments for help.
Tasks such as db:migrate
are run as ./script/task db:migrate
. For a
list of tasks, use ./script/task -T
. If you copied the Rakefile
and have
Rake installed, you can call the tasks using the rake
command.
Applications built with the framework use class naming conventions similar to
Rails, e.g. classes named Post
and PostsController
. However, the
framework itself is built entirely with PEAR-style naming conventions and
does not pollute the global space.
Classes required by the framework are placed in the vendor/
directory,
which is a simple PEAR-style directory. Any framework classes are
prefixed with Mad_
(Mike and Derek) and placed under vendor/Mad/
.
There are some libraries from other projects included in the vendor/
directory as well that are dependencies. We have no interest in building
a monolith and try to share our code with other projects as we can. Most
of the dependencies in vendor/
started out as Mad_
classes that found
homes elsewhere. We hope in the future that the framework will continue
to shrink in this way.
The repository has the same layout as an application. The framework tests
are run the same way as an application's tests would be. Change to the
test/
directory and run phpunit AllTests
.
Before you can run the tests, you need to create a database and configure
the connection in database.yml
. You also need to build the tests database
using the file db/tests/madmodel_test.sql
.
Most of the development was done around the time of Rails 1.2. As such, Rails developers will notice many of the important additions since Rails 2.0 are not implemented.
The best way to contribute patches is to fork the repository on GitHub and then send us a pull request with your changes.
It's unlikely that we'll accept patches that deviate from the "Rails way". For example, a patch that implements support for Smarty templates will not be accepted. It's best to make your own fork for those kinds of changes.
If you implement a useful Rails feature that we have not yet implemented, it's more likely your patch will be accepted. We will also accept patches that improve the implementation of existing features.
We will not accept patches unless they use the same coding standards and include reasonable test coverage. Please refrain from sending us huge patches. Incremental improvements are best.
We would like to thank Chuck Hagenbuch of the Horde Project and Sebastian Bergmann of PHPUnit for helping us with maintenance by integrating some of our code into their projects.
We would also like to thank all the folks who have provided patches or support in other ways.