* 2.4:
  fixed CS
  Fix "absolute path" when we look to the cache directory

…colas-grekas)

This PR was merged into the 2.4-dev branch.

Discussion
----------

[Debug] cleanup interfaces before 2.5-final

| Q             | A
| ------------- | ---
| Bug fix?      | yes
| New feature?  | no
| BC breaks?    | no
| Deprecations? | no
| Tests pass?   | yes
| Fixed tickets | none
| License       | MIT
| Doc PR        | none

This PR is targeted at cleaning up interfaces before 2.5 final:
- ExceptionHandlerInterface has never been released in a stable Symfony, lets drop it, not deprecate it,
- generalize a little bit how fatal errors are handled and make them take the same path as uncaught exceptions,
- enhance handling of out of memory situations.

Commits
-------

e3255bf [Debug] better ouf of memory error handling
dfa8ff8 [Debug] cleanup interfaces before 2.5-final

… of ArrayAccess implementations

…ng non-existing indices of ArrayAccess implementations (webmozart)

This PR was merged into the 2.5-dev branch.

Discussion
----------

[2.5][PropertyAccess] Fixed getValue() when accessing non-existing indices of ArrayAccess implementations

| Q             | A
| ------------- | ---
| Bug fix?      | yes
| New feature?  | no
| BC breaks?    | no
| Deprecations? | no
| Tests pass?   | yes
| Fixed tickets | -
| License       | MIT
| Doc PR        | -

Previously, when the following code was executed:

```php
$object = new ImplOfArrayAccess();

$propertyAccessor->setValue($object, '[index]', 'Value');
```

and that index did not exist, a fatal error would be generated because `array_keys()` was executed on `$object`. This error is fixed now.

Commits
-------

fef698e [PropertyAccess] Fixed getValue() when accessing non-existing indices of ArrayAccess implementations

* 2.4:
  Fixed the XPath filtering to have the same behavior than Symfony 2.4
  [DomCrawler] Fixed filterXPath() chaining
  [DomCrawler] Added more tests for the XPath filtering
  [PropertyAccess] Fixed inverted logic
  [HttpKernel] fixed file uploads in functional tests when no file was selected
  [PropertyAccess] fixed CS
  Fixed test cases failing when the Intl extension is not installed
  Fixed the Travis build to avoid exiting too early
  [PropertyAccess] Fixed getValue() when accessing non-existing indices of ArrayAccess implementations
  [PropertyAccess] Refactored PropertyAccessorCollectionTest
  [PropertyAccess] Refactored PropertyAccessorTest

… "kB")

…ebmozart)

This PR was merged into the 2.5-dev branch.

Discussion
----------

Made use of "kB" vs. "KiB" consistent

| Q             | A
| ------------- | ---
| Bug fix?      | no
| New feature?  | no
| BC breaks?    | no
| Deprecations? | no
| Tests pass?   | yes
| Fixed tickets | -
| License       | MIT
| Doc PR        | -

Continuation of symfony#10661.

This PR makes the usage of "kB" and "KiB" consistent across the project. The notations equal the internationally recognized ones:

Short form | Long form | Value in Bytes
--- | --- | ---
B | bytes | 1
kB | kilobytes | 1000
KiB | kibibytes | 1024
MB | megabytes | 1000000
MiB | mebibytes | 1048576
GB | gigabytes | 1000000000
GiB | gibibytes | 1073741824

The reason for differentiating between the two is that several users got confused with the current mix (see symfony#10648, symfony#10917, symfony#10661).

FileValidator, UploadedFile and the ProgressBar helper were changed accordingly.

Follow-up feature request: symfony#10962

Commits
-------

e4c6da5 [Validator] Improved to-string conversion of the file size/size limit
bbe1045 [Validator][Console][HttpFoundation] Use "KiB" everywhere (instead of "kB")

…of Process::setInput (romainneutron)

This PR was merged into the 2.3-dev branch.

Discussion
----------

[Process] Deprecate Process::setStdin in favor of Process::setInput

| Q             | A
| ------------- | ---
| Bug fix?      | no
| New feature?  | no
| BC breaks?    | no
| Deprecations? | yes
| Tests pass?   | yes
| Fixed tickets | n/a
| License       | MIT

From the `Process` point of view, what we pass is an *input*, as well as we retrieve *output* and *error output*.
As we use `getOutput` and `getErrorOutput` we should use `setInput` and let the underlying `ProcessPipes` deal with the actual `STDIN`.

By the way, `ProcessBuilder` already has `setInput` method and no `setStdin` method

Commits
-------

53b9d73 [Process] Deprecate Process::setStdin in favor of Process::setInput

* 2.4:
  [WIP][Finder] Fix wrong implementation on sortable callback comparator
  ommited space
  [Validator] Fixed StaticMethodLoader on systems that don't have E_STRICT enabled by default
  Include file path in exception
  [Process] Add validation on Process input

Conflicts:
	src/Symfony/Component/Process/Process.php
	src/Symfony/Component/Process/ProcessBuilder.php
	src/Symfony/Component/Process/Tests/ProcessBuilderTest.php

* 2.4:
  Revert "bug symfony#10908 [HttpFoundation] implement session locking for PDO (Tobion)"
  bumped Symfony version to 2.3.15
  updated VERSION for 2.3.14
  update CONTRIBUTORS for 2.3.14
  updated CHANGELOG for 2.3.14

…2.5 deprecation (romainneutron)

This PR was merged into the 2.3-dev branch.

Discussion
----------

[Process] Fix conflicts between latest 2.3 fix and 2.5 deprecation

| Q             | A
| ------------- | ---
| Bug fix?      | yes
| New feature?  | no
| BC breaks?    | no
| Deprecations? | no
| Tests pass?   | yes
| Fixed tickets | n/a
| License       | MIT

As reported in symfony#10932 (comment), it's been introduced when merging 2.3 in master

Commits
-------

3454d60 [Process] Fix conflicts between latest 2.3 fix and 2.5 deprecation

…Stdin and ProcessBuilder::setInput

…tring for Process::setStdin and ProcessBuilder::setInput (romainneutron)

This PR was merged into the 2.4-dev branch.

Discussion
----------

[Process] Deprecate using values that are not string for Process::setStdin and ProcessBuilder::setInput

| Q             | A
| ------------- | ---
| Bug fix?      | no
| New feature?  | no
| BC breaks?    | no
| Deprecations? | yes
| Tests pass?   | yes
| Fixed tickets | n/a
| License       | MIT

This deprecates passing a `Process` input any value that is not a strict string. This needs symfony#10929 to be merged.
I don't know if the use of `trigger_error` is correct or should be removed.

Commits
-------

9887b83 [Process] Deprecate using values that are not string for Process::setStdin and ProcessBuilder::setInput

…e paths (nicolas-grekas)

This PR was merged into the 2.5 branch.

Discussion
----------

[Debug] throw even in stacking mode to preserve code paths

| Q             | A
| ------------- | ---
| Bug fix?      | yes
| New feature?  | no
| BC breaks?    | no if throwing a parent class is accepted
| Deprecations? | no
| Tests pass?   | yes
| Fixed tickets | none
| License       | MIT
| Doc PR        | none

Throwing \ErrorException instead of ContextErrorException is a better work around for https://bugs.php.net/42098

Commits
-------

d06b206 [Debug] throw even in stacking mode to preserve code paths

…ekas)

This PR was merged into the 2.5 branch.

Discussion
----------

[Debug] preserve modified error level

| Q             | A
| ------------- | ---
| Bug fix?      | yes
| New feature?  | no
| BC breaks?    | no
| Deprecations? | no
| Tests pass?   | yes
| Fixed tickets | none
| License       | MIT
| Doc PR        | none

Minor edge case, but still a bug fix.
Replaces symfony#10978

Commits
-------

e40b717 [Debug] preserve modified error level

* 2.4:
  [TwigBridge][Trans]set %count% only on transChoice
  [DomCrawler] Fixed a forgotten case of complex XPath queries
  bumped Symfony version to 2.4.6
  updated VERSION for 2.4.5
  updated CHANGELOG for 2.4.5
  bumped Symfony version to 2.3.16
  updated VERSION for 2.3.15
  updated CHANGELOG for 2.3.15

Conflicts:
	src/Symfony/Component/HttpKernel/Kernel.php

* 2.5:
  [TwigBridge][Trans]set %count% only on transChoice
  [Debug] throw even in stacking mode to preserve code paths
  [Debug] preserve modified error level
  [DomCrawler] Fixed a forgotten case of complex XPath queries
  bumped Symfony version to 2.4.6
  updated VERSION for 2.4.5
  updated CHANGELOG for 2.4.5
  bumped Symfony version to 2.3.16
  updated VERSION for 2.3.15
  updated CHANGELOG for 2.3.15

* 2.4:
  Revert "bug symfony#10894 [HttpKernel] removed absolute paths from the generated container (fabpot)"
  Revert "bug symfony#10937 [HttpKernel] Fix "absolute path" when we look to the cache directory (BenoitLeveque)"
  Revert "fixed CS"
  Revert "bug symfony#10979 Make rootPath part of regex greedy (artursvonda)"
  Revert "[HttpKernel] simplified some tests"
  [HttpKernel] simplified some tests
  Make rootPath part of regex greedy

* 2.5:
  Revert "bug symfony#10894 [HttpKernel] removed absolute paths from the generated container (fabpot)"
  Revert "bug symfony#10937 [HttpKernel] Fix "absolute path" when we look to the cache directory (BenoitLeveque)"
  Revert "fixed CS"
  Revert "bug symfony#10979 Make rootPath part of regex greedy (artursvonda)"
  Revert "[HttpKernel] simplified some tests"
  [HttpKernel] simplified some tests
  Make rootPath part of regex greedy

…for PDO (Tobion)"

This reverts commit 8c71454, reversing
changes made to 735e9a4.

* 2.5:
  Revert "bug symfony#10908 [HttpFoundation] implement session locking for PDO (Tobion)"

…er:debug command (xabbuh)

This PR was merged into the 2.6-dev branch.

Discussion
----------

remove the scope from the output of the container:debug command

| Q             | A
| ------------- | ---
| Bug fix?      | no
| New feature?  | yes
| BC breaks?    | no
| Deprecations? | no
| Tests pass?   | yes
| Fixed tickets | symfony#11302
| License       | MIT
| Doc PR        |

Commits
-------

f56b731 remove the scope from the output of the container:debug command

…_SERVER['KERNEL_DIR'] before invoking getPhpUnitXmlDir()

…xception when specifying $_SERVER['KERNEL_DIR'] (iteman)

This PR was submitted for the master branch but it was merged into the 2.5 branch instead (closes symfony#11370).

Discussion
----------

[FrameworkBundle] avoid raising unexpected RuntimeException when specifying $_SERVER['KERNEL_DIR']

| Q             | A
| ------------- | ---
| Bug fix?      | yes
| New feature?  | no
| BC breaks?    | no
| Deprecations? | no
| Tests pass?   | yes
| Fixed tickets | -
| License       | MIT
| Doc PR        | -

Since v2.5.0 (exactly e778cf1), all tests that use WebTestCase are to be errors "RuntimeException: Unable to guess the Kernel directory." even though $_SERVER['KERNEL_DIR'] is specified. This has been preventing another test runner (e.g. [MakeGood](https://github.com/piece/makegood)) from running tests for a Symfony application without overriding KernelTestCase::getPhpUnitXmlDir() as follows.

The bootstrap file for testing in the app directory:

```php
<?php
...
$_SERVER['KERNEL_DIR'] = __DIR__;
require_once dirname(__DIR__) . '/var/bootstrap.php.cache';
```

console output:

```console
PHPUnit 4.1.3 by Sebastian Bergmann.

EE

Acme\DemoBundle\Tests\Controller\DemoController
 [ ] Index
 [ ] Secure section

Time: 146 ms, Memory: 9.50Mb

There were 2 errors:

1) Acme\DemoBundle\Tests\Controller\DemoControllerTest::testIndex
RuntimeException: Unable to guess the Kernel directory.

/path/to/symfony-application/vendor/symfony/symfony/src/Symfony/Bundle/FrameworkBundle/Test/KernelTestCase.php:56
/path/to/symfony-application/vendor/symfony/symfony/src/Symfony/Bundle/FrameworkBundle/Test/KernelTestCase.php:103
/path/to/symfony-application/vendor/symfony/symfony/src/Symfony/Bundle/FrameworkBundle/Test/KernelTestCase.php:156
/path/to/symfony-application/vendor/symfony/symfony/src/Symfony/Bundle/FrameworkBundle/Test/KernelTestCase.php:137
/path/to/symfony-application/vendor/symfony/symfony/src/Symfony/Bundle/FrameworkBundle/Test/WebTestCase.php:33
/path/to/symfony-application/src/Acme/DemoBundle/Tests/Controller/DemoControllerTest.php:11
/path/to/eclipse/plugins/com.piece_framework.makegood.stagehandtestrunner_3.1.0.v201407050319/resources/php/vendor/piece/stagehand-testrunner/src/Runner/PHPUnitRunner.php:81
/path/to/eclipse/plugins/com.piece_framework.makegood.stagehandtestrunner_3.1.0.v201407050319/resources/php/vendor/piece/stagehand-testrunner/src/Process/TestRunner.php:100
/path/to/eclipse/plugins/com.piece_framework.makegood.stagehandtestrunner_3.1.0.v201407050319/resources/php/vendor/piece/stagehand-testrunner/src/CLI/TestRunnerApplication/Command/PluginCommand.php:149
/path/to/symfony-application/vendor/symfony/symfony/src/Symfony/Component/Console/Command/Command.php:252
/path/to/symfony-application/vendor/symfony/symfony/src/Symfony/Component/Console/Application.php:887
/path/to/symfony-application/vendor/symfony/symfony/src/Symfony/Component/Console/Application.php:193
/path/to/symfony-application/vendor/symfony/symfony/src/Symfony/Component/Console/Application.php:124
...
```

Commits
-------

6f58674 [FrameworkBundle] changed KernelTestCase::getKernelClass() to check $_SERVER['KERNEL_DIR'] before invoking getPhpUnitXmlDir()

…webmozart)

This PR was merged into the 2.5 branch.

Discussion
----------

[Validator] Fixed memory leak in ValidatorBuilder

| Q             | A
| ------------- | ---
| Bug fix?      | yes
| New feature?  | no
| BC breaks?    | no
| Deprecations? | no
| Tests pass?   | yes
| Fixed tickets | symfony#11236
| License       | MIT
| Doc PR        | -

In 23534ca, the following code was introduced in `ValidatorBuilder::getValidator()`:

```php
AnnotationRegistry::registerLoader(function ($class) {
    if (0 === strpos($class, __NAMESPACE__.'\\Constraints\\')) {
        $file = str_replace(__NAMESPACE__.'\\Constraints\\', __DIR__.'/Constraints/', $class).'.php';

        if (is_file($file)) {
            require_once $file;

            return true;
        }
    }

    return false;
});
```

`AnnotationRegistry::registerLoader()` stores all loaders in a global array. Every time that `getValidator()` is called, a new closure is put onto the loader stack, referencing `$this` (i.e. the ValidatorBuilder) and consequently preventing the ValidatorBuilder (and its references) from being garbage-collected.

The call to `registerLoader()` did not exist in 2.4 and I can't find a reason why it should be there. All tests are green without that code. I suppose I used it for debugging and then forgot to remove it again, so I'm removing it here.

Commits
-------

283387a [Validator] Fixed memory leak in ValidatorBuilder

…or (jskvara)

This PR was merged into the 2.6-dev branch.

Discussion
----------

[FrameworkBundle] Improved controller class error

| Q             | A
| ------------- | ---
| Bug fix?      | no
| New feature?  | yes
| BC breaks?    | no
| Deprecations? | no
| Tests pass?   | yes
| Fixed tickets | symfony#11304
| License       | MIT
| Doc PR        |

Commits
-------

da41eb1 [FrameworkBundle] improved controller name parse error message

* 2.4:
  [Translator] Use quote to surround invalid locale
  Optimize assertLocale regexp
  [ExpressionLanguage] Fixed an issue with # characters in double quoted string literals
  Add some tweaks to the pt_BR translations
  [Validator] Backported symfony#11410 to 2.3: Object initializers are called only once per object
  [Translator][FrameworkBundle] Added @ to the list of allowed chars in Translator
  [Process] Reduce I/O load on Windows platform
  [Form] Check if IntlDateFormatter constructor returned a valid object before using it

Conflicts:
	src/Symfony/Component/Validator/Tests/ValidationVisitorTest.php

* 2.5:
  [Translator] Use quote to surround invalid locale
  [Validator] Fixed memory leak in ValidatorBuilder
  [FrameworkBundle] changed KernelTestCase::getKernelClass() to check $_SERVER['KERNEL_DIR'] before invoking getPhpUnitXmlDir()
  Optimize assertLocale regexp
  [ExpressionLanguage] Fixed an issue with # characters in double quoted string literals
  [Validator] Fixed object initializers in 2.5 version of the Validator
  Add some tweaks to the pt_BR translations
  [Validator] Backported symfony#11410 to 2.3: Object initializers are called only once per object
  [Translator][FrameworkBundle] Added @ to the list of allowed chars in Translator
  [Process] Reduce I/O load on Windows platform
  [Form] Check if IntlDateFormatter constructor returned a valid object before using it

Conflicts:
	src/Symfony/Component/Process/ProcessPipes.php

… now accepts secure/httpOnly (adrienbrault)

This PR was merged into the 2.6-dev branch.

Discussion
----------

[HttpFoundation] ResponseHeaderBag::clearCookie now accepts secure/httpOnly

| Q             | A
| ------------- | ---
| Bug fix?      | no
| New feature?  | yes
| BC breaks?    | no
| Deprecations? | no
| Tests pass?   | yes
| Fixed tickets | -
| License       | MIT
| Doc PR        | -

Commits
-------

83f4ac5 [HttpFoundation] ResponseHeaderBag::clearCookie now accepts secure/httpOnly

…ion (xabbuh)

This PR was merged into the 2.6-dev branch.

Discussion
----------

server:run command: provide more error information

The server:run command didn't provide many information when the executed command exited unexpectedly. Now, the process' exit code is passed through and an error message is displayed.

| Q             | A
| ------------- | ---
| Bug fix?      | no
| New feature?  | yes
| BC breaks?    | no
| Deprecations? | no
| Tests pass?   | yes
| Fixed tickets |
| License       | MIT
| Doc PR        |

This is, for example, useful, if you provide an invalid ip address/hostname on which the server couldn't listen.

Commits
-------

5ba40bf server:run command: provide more error information

This PR was merged into the 2.6-dev branch.

Discussion
----------

[DomCrawler] Added node name getter

[DomCrawler] Added node name getter

Based on the use case came up in a SO question (http://stackoverflow.com/questions/22564465/getting-the-tag-name-of-a-element-crawled-with-domcrawler-in-php) I propose a new helper function to get the name of the active node

| Q             | A
| ------------- | ---
| Bug fix?      | no
| New feature?  | yes
| BC breaks?    | no
| Deprecations? | no
| Tests pass?   | yes
| Fixed tickets |
| License       | MIT
| Doc PR        | symfony/symfony-docs#4039

(Follow up of symfony#11351)

Commits
-------

2fee576 [DomCrawler] Added node name getter

…ced xml (xmlns) (mattvick)

This PR was merged into the 2.6-dev branch.

Discussion
----------

Added support for encoding and decoding namespaced xml (xmlns)

Q | A
------------ | -------------
Bug fix? | no
New feature? | yes
BC breaks? | no
Deprecations? | no
Tests pass? | yes
Fixed tickets | symfony#10625
License | MIT

Complete this PR : symfony#9156

Commits
-------

9d6b20c Added support for encoding and decoding namespaced xml (xmlns)

…MeListener (lstrojny)

This PR was merged into the 2.6-dev branch.

Discussion
----------

[Security] Allow exception bubbling in RememberMeListener

- Allow optional exception bubbling so that the exception listener has a chance to handle those exceptions

#### While at it
- Test for dispatching the InteractiveLogin event
- Smaller cleanups in the test

| Q             | A
| ------------- | ---
| Bug fix?      | no
| New feature?  | yes
| BC breaks?    | no
| Deprecations? | no
| Tests pass?   | ye
| Fixed tickets | n.A.
| License       | MIT
| Doc PR        | n.A.

Commits
-------

fcb7f74 Allow exception bubbling in RememberMeListener

…Crunch)

This PR was merged into the 2.6-dev branch.

Discussion
----------

[HttpKernel][FrameworkBundle] SSI support

| Q             | A
| ------------- | ---
| Bug fix?      | No
| New feature?  | Yes
| BC breaks?    | No
| Deprecations? | No
| Tests pass?   | Yes
| Fixed tickets | symfony#9419 (, symfony#10684)
| License       | MIT

It does not support comments, or alternative URIs, or "continue" in case of errors. Maybe I can workaround that, but I've decided to left it out for this PR. Especially as far as I can see a "alternative URIs"-hack would _always_ lead to two requests, even if it's not needed.

Commits
-------

06cea08 SSI support

…(aferrandini)

This PR was merged into the 2.6-dev branch.

Discussion
----------

[DX] New service to simplify password encoding

| Q             | A
| ------------- | ---
| Bug fix?      | no
| New feature?  | yes
| BC breaks?    | no
| Deprecations? | no
| Tests pass?   | yes
| Fixed tickets | symfony#11299
| License       | MIT
| Doc PR | symfony/symfony-docs#3995

This new service siplifies the way to encode a password. Just get the `security.password_encoder` service and encode the `User` password.

```php
$encoded = $this->container->get('security.password_encoder')
    ->encodePassword($user, $plainPassword);

$user->setPassword($encoded);
```

Commits
-------

7bc190a New service to simplify password encoding

…on (text descriptor) (Maxime Douailin)

This PR was merged into the 2.6-dev branch.

Discussion
----------

[DX] Added line feed to container alias description (text descriptor)

| Q             | A
| ------------- | ---
| Bug fix?      | yes
| New feature?  | no
| BC breaks?    | no
| Deprecations? | no
| Tests pass?   | yes
| Fixed tickets | -
| License       | MIT
| Doc PR        | N/A

This PR just adds a line feed at the end of the alias service description delivered by the ContainerDebugCommand. Tests are passing because expected outputs are trimmed before comparison (which IMO is not a really good thing, but hey...)

Commits
-------

13c011c Added line feed to container alias description (text descriptor)

…tion message (rybakit)

This PR was merged into the 2.6-dev branch.

Discussion
----------

[Security] Add missing dot at the end of an exception message

| Q             | A
| ------------- | ---
| Fixed tickets | -
| License       | MIT

Commits
-------

8e1f978 Add missing dot at the end of an exception message

As @excelwebzone pointed out in fb9dc6a, `catchExceptions` is the 6th
argument of the RememberMeListener constructor.

…bbuh)

This PR was merged into the 2.6-dev branch.

Discussion
----------

[SecurityBundle] fix constructor argument index

| Q             | A
| ------------- | ---
| Bug fix?      | yes
| New feature?  | no
| BC breaks?    | no
| Deprecations? | no
| Tests pass?   | yes
| Fixed tickets | symfony#10793
| License       | MIT
| Doc PR        |

As @excelwebzone pointed out in fb9dc6a, `catchExceptions` is the 6th
argument of the RememberMeListener constructor.

Commits
-------

a32646e [SecurityBundle] fix constructor argument index

…ormatter