This removes yet another special case that should not exist. Now, all
listeners/subscribers registration are done in one place.

If you want to disable the exception_handler, you now should call the
disable() method on it. You can still unset it but be careful to do it
as early as possible (which is anyway always the best idea).

Removed the SilexEvents class. Using the standard Symfony events allows
us to be more flexible (the priorities are the same as the one from
Symfony, so you can hook at any place into the HTTP handling).

Updated documentation to make it accurate with the current behavior, and
extended it to explain the flexibility the developer has.

It is indeed a good thing to do but for consistency, it must be done for
all events. As this is not possible now, I prefer to revert it for now.

This PR was merged into the master branch.

Commits
-------

5c5dae19 Flipped condition
9f4a122f Log a little more information about exceptions

Discussion
----------

Log a little more information about exceptions

This takes a little from Symfony's ExceptionListener, logging a little more appropriate information. Without this code, 404's etc are logged with a blank message.

---------------------------------------------------------------------------

by MarcinKleczek at 2012-10-29T11:53:41Z

+1

This PR was merged into the master branch.

Commits
-------

33d802ec updated FormTrait documentation

Discussion
----------

Updated doc/providers/form.rst

Hello,

Only one line changed. FormTrait adds ```function form($data, $options)``` to \Silex\Application, not ```form($name, $data, $options)``` like createBuilder; this example in docs was throwing exception.

(this is my first pull request and now I see I should have done it in separate branch; sorry. I hope I didn't broke anything)

---------------------------------------------------------------------------

by igorw at 2012-10-29T02:25:44Z

👍

form function takes only 2 parameters: $data and $options

This PR was merged into the master branch.

Commits
-------

26c50d4a Grammar tweak
44c11ca2 Aren't parameters set after the provider is registered ?

Discussion
----------

Aren't parameters set after the provider is registered ?

fabpot/Silex@9766faf0616a67cd0ff7437950d5453deebac9c4

---------------------------------------------------------------------------

by igorw at 2012-10-01T14:19:19Z

👍

This PR was merged into the master branch.

Commits
-------

03bccc83 Add a `SerializerServiceProvider`.

Discussion
----------

Add a SerializerServiceProvider

Usage
-----

```php
<?php

use Silex\Application;
use Silex\Provider\SerializerServiceProvider;
use Symfony\Component\HttpFoundation\Response;

$app = new Application();

$app->register(new SerializerServiceProvider);

// only accept content types supported by the serializer via the assert method.
$app->get("/pages/{id}.{_format}", function ($id) use ($app) {
    // assume a page_repository service exists that returns Page objects. The
    // object returned has getters and setters exposing the state.
    $page = $app['page_repository']->find($id);
    $format = $app['request']->getFormat();

    if (!$page instanceof Page) {
        $this->abort("No page found for id: $id");
    }

    return new Response($app['serializer']->serialize($page, $format), 200, array(
        "Content-Type" => $app['request']->getMimeType($format)
    ));
})->assert("_format", "xml|json")
  ->assert("id", "\d+");
```

---------------------------------------------------------------------------

by GromNaN at 2012-06-28T14:29:50Z

You can create this ServiceProvider in its own repository and put a link on the wiki :
https://github.com/fabpot/Silex/wiki/Third-Party-ServiceProviders

---------------------------------------------------------------------------

by fabpot at 2012-06-28T14:38:42Z

I think it makes sense to have this in core.

---------------------------------------------------------------------------

by marijn at 2012-06-28T14:41:03Z

Ok. Haven't tested in "in the wild" yet (working on that as we speak). Any pointers, concerns or other comments are more than welcome 😄

Should I add more documentation or is this enough?

---------------------------------------------------------------------------

by igorw at 2012-06-28T14:47:03Z

@fabpot I agree. 👍

Please add some more tests.

---------------------------------------------------------------------------

by marijn at 2012-06-28T14:50:13Z

@igorw in regards to the tests: what would you like to see added? I figured we should only test if the serializer is configured properly, not if it actually serializes.

---------------------------------------------------------------------------

by igorw at 2012-06-28T15:11:06Z

You're right, no need to test serialization. Looks good. 👍

---------------------------------------------------------------------------

by marijn at 2012-06-28T16:02:20Z

Somehow I have trouble running the full test suite. The swift mailer configuration doesn't seem to work. Is this a known issue?

---------------------------------------------------------------------------

by GromNaN at 2012-06-28T20:33:14Z

A new trait can also be added with the method `serialize()`.

---------------------------------------------------------------------------

by marijn at 2012-06-28T21:34:58Z

At the moment I haven't got a working version of PHP 5.4 so I cannot test traits.

---------------------------------------------------------------------------

by marijn at 2012-06-29T19:01:21Z

It seems to me that adding a `trait` for the serializer doesn't really make any sense: the API for the `Serializer` component is pretty extensive, it would add a lot of methods to the `Application`. What do you think?

---------------------------------------------------------------------------

by stof at 2012-06-30T10:32:07Z

@marijn the trait would cover the simple use of the serializer, i.e. a ``serialize()`` method. It is not intended to cover all possible use cases.

---------------------------------------------------------------------------

by fabpot at 2012-06-30T18:31:32Z

Can you squash your commits before I merge? Thanks.

---------------------------------------------------------------------------

by marijn at 2012-07-10T10:58:58Z

@fabpot this has been squashed. I think it's mergeable 😄

---------------------------------------------------------------------------

by marijn at 2012-08-17T13:39:17Z

Have I missed something here? To the best of my knowledge, this is mergeable. Please let me know if there is anything I still need to do 😄

---------------------------------------------------------------------------

by alanbem at 2012-10-27T20:39:02Z

What is status of this PR? Is it going to be merged?

---------------------------------------------------------------------------

by marijn at 2012-10-27T21:14:12Z

I have no clue, @alanbem. If you're in need of a serializer for Silex you can either merge this in your own fork of Silex or use the [`jms/serializer-service-provider`][1] that I released.

[1]: https://github.com/pink-tie/JMSSerializerServiceProvider

---------------------------------------------------------------------------

by igorw at 2012-10-27T23:15:53Z

@fabpot this looks good to me.

This PR was merged into the master branch.

Commits
-------

5e83c757 Update logout_path with pattern info per #423

Discussion
----------

Update logout_path with pattern info per #423

Because logout_path must exist inside the existing pattern, the example
in the docs would not work by default.

This PR was merged into the master branch.

Commits
-------

49cde048 Change 'security.authentication_listener.form._proto' definition to the same signature all other _protos have.

Discussion
----------

Change 'security.authentication_listener.form._proto' definition to the ...

...same signature all other _protos have.

This is a fix to remove the unussed (third) $class parameter from the 'security.authentication_listener.form._proto' definition and allow to customize the class via $options['listener_class'] instead.

Fixes https://github.com/fabpot/Silex/issues/459

---------------------------------------------------------------------------

by sli-systems at 2012-08-29T23:20:15Z

Any feedback on this? I'd love to get this sorted so if there are issues let me know and I'll address them.

---------------------------------------------------------------------------

by sli-systems at 2012-09-11T20:55:26Z

bump again - feedback either way would be appreciated.

---------------------------------------------------------------------------

by sli-systems at 2012-09-30T21:54:04Z

bump again (again) - it would be helpful to know if there is interest in this - it seems pointless to keep rebasing without any way of knowing if it is worthwhile...

---------------------------------------------------------------------------

by sli-systems at 2012-10-23T21:54:12Z

Just did another rebase - how long does it usually take for a one line PR to get accepted or rejected? Sorry, but I do not know what else to do other than bumping this thread every now and then...

If this PR is not good, please just reject it so I do not have to waste any more time on it.

Thanks, mano

This PR was merged into the master branch.

Commits
-------

ac9cdded Update doc/providers/twig.rst

Discussion
----------

Update doc/providers/twig.rst

corrected the grammatical possessiveness of the word services 

---------------------------------------------------------------------------

by igorw at 2012-10-21T18:20:54Z

👍

corrected the grammatical possessiveness of the services 

Because logout_path must exist inside the existing pattern, the example
in the docs would not work by default.

Change 'security.authentication_listener.form._proto' definition to the same signature all other _protos have.

This is a fix to remove the unussed (third) $class parameter from the 'security.authentication_listener.form._proto' definition and allow to customize the class via $options['listener_class'] instead.

This PR was merged into the master branch.

Commits
-------

f23f1589 [Doc] Added more sample to url_url_generator provider

Discussion
----------

[Doc] Added more sample to url_url_generator provider

fixed #508

fixed #508

This PR was merged into the master branch.

Commits
-------

be648301 Takes advantage of fluent interface on Response object
fac3649a Updated to use setTtl

Discussion
----------

Updated http cache docs to use setTtl

Previous version set Cache-Control s-maxage headers explicitly which is
not enough to trigger ESI caching in the reverse proxy. Using setTtl is
enough to make this work correctly.

---------------------------------------------------------------------------

by igorw at 2012-09-30T13:23:43Z

👍

Previous version set Cache-Control s-maxage headers explicitly which is
not enough to trigger ESI caching in the reverse proxy. Using setTtl is
enough to make this work correctly.

Commits
-------

6ca59f3b removed composer download

Discussion
----------

removed composer download

Commits
-------

eb5d0a48 Fixed missing line before list
07b1246b Fixed null from string to vartype
b5fb45f6 Added configuration example for external server

Discussion
----------

Added configuration example for external server

Just added an example of how to configure an external smtp server for newbie users which don't know where to set it up.

---------------------------------------------------------------------------

by alex88 at 2012-09-18T17:59:13Z

Sorry for the errors, is it ok now?

Just added an example of how to configure an external smtp server for newbie users which don't know where to set it up.

Commits
-------

b0dfe797 Removes non-lib directories from archives

Discussion
----------

Removes non-lib directories from archives

Removes the followind directories:
* vendor/swiftmailer/swiftmailer/test-suite
* vendor/twig/twig/test
* vendor/twig/twig/ext

Removes the followind directories:
vendor/swiftmailer/swiftmailer/test-suite
vendor/twig/twig/test
vendor/twig/twig/ext

Commits
-------

a51a9a14 Make the event dispatcher class configurable.

Discussion
----------

Make the event dispatcher class configurable.

<del>After 81bafcc6</del> There is no easy way to modify event dispatcher class.

This PR allows @davedevelopment to use a customized class like this #426.

---------------------------------------------------------------------------

by igorw at 2012-07-18T12:36:52Z

+1

---------------------------------------------------------------------------

by davedevelopment at 2012-07-18T12:41:07Z

+1, I'll settle for this :)

---------------------------------------------------------------------------

by stof at 2012-07-18T16:23:00Z

To be clear, the commit you linked is a revert of a change committed by mistake in the previous commit just a few minutes earlier.

---------------------------------------------------------------------------

by GromNaN at 2012-07-18T20:59:12Z

@stof indeed.