This PR was merged into the master branch.

Commits
-------

6a4fffeb Added a note about writing documentation

Discussion
----------

Added a note about writing documentation

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

by igorw at 2013-03-08T15:52:03Z

👍

This PR was squashed before being merged into the master branch (closes #553).

Commits
-------

7c38c9a8 Add services allowing for form type extensions and guessers

Discussion
----------

Add services allowing for form type extensions and guessers

I don't like the empty `FormTypeGuesserChain`, but `PreloadedExtension` requires one to be passed until symfony/symfony#5479 gets merged.

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

by fabpot at 2012-11-24T11:54:08Z

symfony/symfony#5479 has been merged now.

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

by davedevelopment at 2012-11-25T19:59:41Z

So, do we want this documenting at this point? I'm glad to have removed the dummy TypeGuesser, but without it, people who don't have a 'dev' version of symfony/form, will need the workaround in their userland code, which may cause headaches.

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

by fabpot at 2013-01-10T07:34:15Z

I think we need to switch to Symfony 2.2 in Silex now. What others think? @igorw

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

by davedevelopment at 2013-01-10T07:54:20Z

👍 from me, though while we're on the topic, are there any plans to tag a 1.0?

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

by igorw at 2013-01-10T13:52:38Z

@fabpot Is there any reason why you would want to enforce users to use an unstable version?

I am against it for the following reasons:

* Users should be able to get stable versions if the stable versions work correctly.
* Silex already allows you to get 2.2, simply by lowering the `minimum-stability` to `beta`.
* Our current recommended way to install silex is `1.0.*@dev` (except on the download page, because it was not updated correctly when we changed it). This means that once they update it will blow up (due to `minimum-stability`), which will result in a lot of frustration, and a lot of support requests.

IMO we should encourage our users to try 2.2 by lowering their `minimum-stability`, but not enforce it.

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

by stof at 2013-01-10T13:55:34Z

another solution would be to bump the requirement to 2.2 only after releasing Silex.

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

by fabpot at 2013-01-11T07:51:36Z

Silex has not reached 1.0 yet. So, switching to 2.2 seems logical to me. That being said, we can wait for 2.2 stable to be out if this is what you suggest.

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

by igorw at 2013-01-11T16:23:06Z

Bumping to 2.2 after it is stable sounds acceptable to me, but unless there is a specific change in 2.2 that silex *needs*, I still don't really see the point. Supporting 2.1+ is easy enough to do.

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

by davedevelopment at 2013-03-06T13:28:19Z

So the current status with this PR, is that adding a TypeExtension with anything less than 2.1.4 of the symfony components will break, unless the user also adds an FormTypeGuesserChain to the guessers array. That is, the test `testFormServiceProviderWillLoadTypeExtensions` will fail with symfony/form < 2.1.4.

If we want to merge this in I guess we can:

* Document the whole feature, with a note/gotcha for Symfony <2.1.4
* Bump the minimum Symfony component versions to 2.1.4
* Just merge it anyway

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

by stof at 2013-03-06T14:13:38Z

bumping to 2.1.4 as minimum is fine IMO. I don't see why someone would want to use 2.1.0 explicitly instead of using the bufix releases for 2.1

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

by fabpot at 2013-03-06T21:39:29Z

👍 for bumping to 2.1.4+

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

by igorw at 2013-03-06T22:08:37Z

In this particular case I'd say it's acceptable to bump to `2.1.4`.

In the future these kind of changes will need more consideration though, as they can cause BC issues when multiple branches of symfony are supported by silex.

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

by fabpot at 2013-03-08T12:05:04Z

When we said bumping to 2.1.4, I thought we were talking about bumping for just the component that has problems with earlier Symfony version, not all of them.

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

by davedevelopment at 2013-03-08T12:14:06Z

I did that because I wasn't sure how to approach it, I'll have a think about it, but I'll probably need some guidance.

I guess we may have to fall back on documenting that we have a dependency on 2.1.4 version of the Form Component, because it's not actually required? Maybe that should be in `suggest`? I don't think suggest enforces version requirements, so it either needs to be in the suggest message, or in the docs?

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

by igorw at 2013-03-08T12:22:59Z

It should be in both. Specifically, the form service provider docs.

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

by davedevelopment at 2013-03-08T13:01:26Z

Changed the requirements bump and documented, do we need to have something more specific/wordy in the docs?

Squashed a little as well, can squash further if it's desirable.

This PR was squashed before being merged into the master branch (closes #645).

Commits
-------

adc172b3 Add support for remember-me with RememberMeServiceProvider

Discussion
----------

Add support for remember-me with RememberMeServiceProvider

This is a refactoring of #464.

I made a separate service provider to avoid adding more complexity to the SecurityServiceProvider.

This PR was merged into the master branch.

Commits
-------

eb941a8f proper prefixes for cookie options

Discussion
----------

All cookie-related options of NativeSessionStorage should be prefixed with `cookie_`

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

by indeyets at 2013-02-16T09:25:59Z

Separate issue is, that there are, actually, much more options available. Listing all of those wouldn't be useful for this doc, so I skipped it for now.

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

by igorw at 2013-02-16T15:33:22Z

The doc should probably link to http://php.net/session.configuration

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

by igorw at 2013-03-05T02:30:32Z

@indeyets what's the status on this? can you update your patch with the link?

This PR was merged into the master branch.

Commits
-------

e11b93e3 Force travis to use 2.1.*

Discussion
----------

Force travis to use 2.1.* now 2.2 has a stable release

This PR was merged into the master branch.

Commits
-------

36b879b3 Use addPrefix to avoid deprecated calls in >2.2.0

Discussion
----------

Use addPrefix to avoid deprecated calls in >2.2.0

I think this should take care of #642

This PR was merged into the master branch.

Commits
-------

fdd89b91 [Security] Create a distinct service for each authentication provider

Discussion
----------

[Security] Rename authentication_provider services to avoid name conflict

When using both anonymous and form authentication providers, they share the same service name whereas they use totally different classes.

Before : `security.authentication_provider.myfirewall`
After : `security.authentication_provider.myfirewall.dao` and `security.authentication_provider.myfirewall.anonymous`

The naming follows the recommendation for custom auth providers.
http://silex.sensiolabs.org/doc/providers/security.html#defining-a-custom-authentication-provider

This PR was merged into the master branch.

Commits
-------

6e8d404e Add symfony/property-access require for 2.2
337a92c4 Make travis pass on Symfony 2.2

Discussion
----------

Make travis pass on Symfony 2.2

Make travis actually update the lock file and install the 2.2 components.
Document deprecation of getClassMetadata in validator docs.

This gives an actual good version of symfony/form instead of some old beta.

Make travis actually update the lock file and install the 2.2 components.
Document deprecation of getClassMetadata in validator docs.

When using multiple both

This PR was squashed before being merged into the master branch (closes #636).

Commits
-------

abdc1fc7 Updated docs to include _method and enableHttpMethodParameterOverride()

Discussion
----------

Updated docs to include _method and enableHttpMethodParameterOverride()

Should be self explanatory. Small addition to the docs to explain about the _method form value and also to advise abut the explicit enableHttpMethodParameterOverride() call required with Symfony 2.2.0 and later.

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

by igorw at 2013-02-18T13:57:37Z

Please wrap the docs at 80 chars.

👍

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

by davedevelopment at 2013-02-18T14:02:08Z

👍

Only one thing, do we need to explicitly state to do this before the request is handled? No-brainer to us, but what about others?

Or should I say, how likely are people to try and do something like:

``` php

$app->put("/some/resource", function(Request $request) {
    Request::enableHttpMethodParameterOverride();
});

```

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

by robdudley at 2013-02-18T14:11:15Z

I've used a similar example to the reverse proxy tip which I hope would make the context clear. Not sure how else to word it without becoming unnecessarily verbose. Also I'm assuming that the setting is going to be rolled into core at some point as it has pretty big backwards compatibility implications.

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

by davedevelopment at 2013-02-18T14:14:00Z

:shipit:

This PR was merged into the master branch.

Commits
-------

7e554db2 improved the performance of the Twig loader by not registering a chain loader if not needed

Discussion
----------

improved the performance of the Twig loader by not registering a chain loader if not needed

When `twig.templates` is empty, there is no need to register the Twig array loader as it will never be used anyway.

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

by lyrixx at 2013-02-18T09:50:07Z

totally 👍 More over, I think this feature was not documented.

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

by GromNaN at 2013-02-18T10:04:32Z

Nice tip. How much is it more performant ?

@lyrixx it is documented here : http://silex.sensiolabs.org/doc/providers/twig.html#parameters

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

by fabpot at 2013-02-18T10:06:17Z

It's not really about how much but just keeping the array loader around is a waste of time.

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

by igorw at 2013-02-18T13:37:58Z

👎 This prevents extending the `twig.loader` service and also feels like a micro-optimization to me.

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

by fabpot at 2013-02-18T13:57:13Z

I've removed the optimization but I've kept the different order in registration as it makes more sense.

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

by igorw at 2013-02-18T13:59:55Z

👍

This PR was merged into the master branch.

Commits
-------

fbffc2b0 Enable logger and stopwatch for Doctrine queries

Discussion
----------

Add logger and stopwatch for Doctrine queries

Add SQL queries to the logs using the DbalLogger of the DoctrineBridge.

In addition, if the WebprofilerServiceProvider is registered (stopwatch service set), the timeline show doctrine queries.

![Profiler](https://f.cloud.github.com/assets/400034/152414/877dc396-75c4-11e2-96df-6dde93c7d86e.png)

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

by DerManoMann at 2013-02-13T10:37:42Z

👍

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

by fabpot at 2013-02-13T10:39:34Z

I'm wondering if the stopwatch service should be moved to Silex core.

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

by lyrixx at 2013-02-13T11:05:00Z

Totally 👍 for moving it to silex.

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

by GromNaN at 2013-02-13T11:16:20Z

Is there a use case for the "stopwatch" outside of the webprofiler ?

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

by lyrixx at 2013-02-13T11:23:44Z

@GromNaN Yes, it you want to bench some code.

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

by henrikbjorn at 2013-02-13T11:35:28Z

@fabpot please not it will only add cruft. i think it is a niche thing to use the Stopwatch component.

This PR was merged into the master branch.

Commits
-------

829c8085 [Application] fix typo

Discussion
----------

[Application] fix typo

This PR was merged into the master branch.

Commits
-------

ef85340f Fixed brace and spaces
6ec11533 Changed Assert\MinLength to Assert\Length in doc

Discussion
----------

Removed Assert\minLength in doc

Changed to Assert\Length

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

by igorw at 2013-02-05T00:10:22Z

For the record, this change is due to Assert\MinLength and MaxLength being deprecated in 2.1, in favour of Length.

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

by igorw at 2013-02-05T00:16:47Z

👍

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

by fabpot at 2013-02-07T11:22:31Z

As we still support 2.1 and 2.2, I would suggest to add a note to indicate the name of the constraint in 2.1.

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

by igorw at 2013-02-11T14:23:51Z

@fabpot Length is already available in 2.1. There is no reason not to use it already. This PR looks good to me.

This PR was merged into the master branch.

Commits
-------

e055ed3e [docs] Add code examples for PUT and DELETE routes

Discussion
----------

[docs] Add code examples for PUT and DELETE routes

This PR was merged into the master branch.

Commits
-------

1e2655f4 Remove the "show" verb from example routes names

Discussion
----------

[docs] Remove the "show" verb from example routes names

This changes all examples from /blog/show/{id} to /blog/{id}. This makes
more sense in terms of HTTP, since the HTTP methods are the verbs and
the URL represents a resource.

This changes all examples from /blog/show/{id} to /blog/{id}. This makes
more sense in terms of HTTP, since the HTTP methods are the verbs and
the URL represents a resource.

This PR was merged into the master branch.

Commits
-------

3a5d992c Replace individual test suite component checks with global one

Discussion
----------

Replace individual test suite component checks with global one

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

by davedevelopment at 2013-02-09T19:43:06Z

👍 nice to see code disappearing

This PR was merged into the master branch.

Commits
-------

8e11cfa5 [Tests] Make use of the @requires annotation for PHP version
1e0dd622 Fix Doctrine tests

Discussion
----------

[Test] Improvements

see https://github.com/fabpot/Silex/pull/620

This PR was merged into the master branch.

Commits
-------

46e6e283 [Doc] Fix references to Symfony 2.0
9eb3b9b2 [Doc] Document Application::sendFile()
570df9d4 implement feedback
734d5738 [Application] Add the ability to send files

Discussion
----------

[Application] Add the ability to send files

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

by GromNaN at 2013-02-06T17:10:24Z

Nice feature 👍

As the `->json` method creates a `JsonResponse`, for consistency the method that create a `BinaryFileResponse` should be named `->binaryFile`.

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

by vicb at 2013-02-06T17:14:43Z

I deeply love consistency but I really can't be consistent with a method named `json` !

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

by vicb at 2013-02-06T18:01:40Z

@igorw will do thanks.

i have also warneed f&j about the test result, subtree split ?

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

by vicb at 2013-02-07T10:29:07Z

@igorw thanks for the feedback, all should be fixed.

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

by fabpot at 2013-02-07T11:26:50Z

Can you add a note in the documentation about this new feature?

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

by vicb at 2013-02-07T12:15:10Z

@fabpot, the doc has been updated

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

by vicb at 2013-02-07T12:23:10Z

@fabpot did you get my email regarding the issue with the tests (sub-stree split ?).

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

by igorw at 2013-02-07T15:45:09Z

Looks good to me. 👍