- When using the `InputInterface` mock with interactive option, asking a
ChoiceQuestion without specifying max attempts nor a default value led
to an infinite loop when invalid values are submitted. In case there is
something wrong in tests, it wasn't obvious to understand what's
happening, as the process was constantly waiting for inputs.
- `testAmbiguousChoiceFromChoicelist` test was wrong as the
autocompleter filled the value and the test was never triggered (+max
attempt issue).
This PR was merged into the 2.7 branch.
Discussion
----------
[2.7][DoctrineBridge] change Logger interface in tests to use Psr\\Lo…
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets |
| License | MIT
| Doc PR |
Commits
-------
5e8864d [2.7][DoctrineBridge] change Logger interface in tests to use Psr\\Log\\LoggerInterface instead of Symfony\\Component\\HttpKernel\\Log\\LoggerInterface"
This PR was merged into the 2.3 branch.
Discussion
----------
[travis] Don't use the cache
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | -
| License | MIT
| Doc PR | -
The composer cache breaks per components builds (deps=low/high)
Commits
-------
103c0df [travis] Don't use the cache
* 2.6:
[travis] Use container-based infrastructure
[PropertyAccess] Fix setting public property on a class having a magic getter
ContainerInterface: unused exception dropped
Conflicts:
.travis.yml
This PR was merged into the 2.3 branch.
Discussion
----------
[travis] Use container-based infrastructure
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | -
| License | MIT
| Doc PR | -
Let's see if this works...
Commits
-------
2aea3aa [travis] Use container-based infrastructure
This PR was merged into the 2.7 branch.
Discussion
----------
[HttpKernel] use ConfigCache::getPath() method when it exists
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets |
| License | MIT
| Doc PR |
This way, the deprecated `__toString()` method of the `ConfigCache`
class will not be called.
Commits
-------
ab70632 [HttpKernel] use ConfigCache::getPath() method when it exists
This PR was merged into the 2.6 branch.
Discussion
----------
[PropertyAccess] Fix setting public property on a class having a magic getter
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | N/A
| License | MIT
| Doc PR | N/A
When using PropertyAccessor with an object having both a public property and a magic getter, and one wants to update this property, PropertyAccessor may lose the property reference.
This occurs when the public property value is a hash:
```php
class Foo
{
/**
* Example: $this->someProperty['foo']['bar'] = 'baz'
* @var array
*/
public $someProperty;
public function __get($name)
{
// ...
}
}
$obj = new Foo();
$obj->someProperty = ['foo' => ['bar' => 'some_value']];
$propertyAccessor->setValue($obj, 'someProperty[foo][bar]', 'another_value');
echo $obj->someProperty['foo']['bar'];
// Before this patch: 'some_value' => fail
// After this patch: 'another_value' => correct
```
Furthermore, public properties are always used before `__get()` by PHP.
This bug is visible since v2.6.5 as d733a887 changed the way
`setValue()` works.
Commits
-------
8b8feff [PropertyAccess] Fix setting public property on a class having a magic getter
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | N/A
| License | MIT
| Doc PR | N/A
When using PropertyAccessor with an object having both a public property
and a magic getter, and one wants to update this property,
PropertyAccessor may lose the property reference.
This occurs when the public property value is a hash:
```php
class Foo
{
/**
* Example: $this->someProperty['foo']['bar'] = 'baz'
* @var array
*/
public $someProperty;
public function __get($name)
{
// ...
}
}
$obj = new Foo();
$obj->someProperty = ['foo' => ['bar' => 'some_value']];
$propertyAccessor->setValue($obj, 'someProperty[foo][bar]', 'another_value');
echo $obj->someProperty['foo']['bar'];
// Before this patch: 'some_value' => fail
// After this patch: 'another_value' => correct
```
Furthermore, public properties are always used before `__get()` by PHP.
This bug is visible since v2.6.5 as d733a887 changed the way
`setValue()` works.
* 2.6: (21 commits)
bumped Symfony version to 2.6.8
updated VERSION for 2.6.7
updated CHANGELOG for 2.6.7
bumped Symfony version to 2.3.29
updated VERSION for 2.3.28
update CONTRIBUTORS for 2.3.28
updated CHANGELOG for 2.3.28
[Debug] Fixed ClassNotFoundFatalErrorHandlerTest
[SecurityBundle] use access decision constants in config
[SecurityBundle] use session auth constants in config
PhpDoc fix in AbstractRememberMeServices
[Filesystem] Simplified an if statement
[SecurityBundle] Use Enum Nodes Instead Of Scalar
[Debug 2.3] Fix test for PHP7
[HttpKernel] Check if "symfony/proxy-manager-bridge" package is installed
[Translation] simplify getMessages.
[Framework][Translation] added test for debug command.
Run tests on hhvm instead of hhvm-nightly
Use HTTPS in README and some other fixes
add more entropy to generated classnames
...
Conflicts:
.travis.yml
src/Symfony/Component/HttpKernel/Kernel.php
* 2.3:
bumped Symfony version to 2.3.29
updated VERSION for 2.3.28
update CONTRIBUTORS for 2.3.28
updated CHANGELOG for 2.3.28
PhpDoc fix in AbstractRememberMeServices
Conflicts:
src/Symfony/Component/HttpKernel/Kernel.php
This PR was merged into the 2.6 branch.
Discussion
----------
[Debug] Fixed ClassNotFoundFatalErrorHandlerTest
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #14573
| License | MIT
| Doc PR |
Since `testCannotRedeclareClass` is skipped on case-sensitive filesystems, it won't have any effect on the travis build, but fixes the issue described in #14573 for e.g. OS X.
Commits
-------
f7def8c [Debug] Fixed ClassNotFoundFatalErrorHandlerTest
This PR was merged into the 2.3 branch.
Discussion
----------
[SecurityBundle] Use Enum Nodes Instead Of Scalar
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | ~
| License | MIT
| Doc PR | ~
Commits
-------
ce7fb04 [SecurityBundle] Use Enum Nodes Instead Of Scalar
This PR was submitted for the master branch but it was merged into the 2.3 branch instead (closes#14567).
Discussion
----------
[Filesystem] Simplify if statement
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | none
| License | MIT
| Doc PR | none
Simplified the `if` statement if `Symfony\Component\Filesystem\Filesystem::isAbsolutePath`
Commits
-------
4d974ced [Filesystem] Simplified an if statement
This PR was merged into the 2.3 branch.
Discussion
----------
[HttpKernel] Check if "symfony/proxy-manager-bridge" package is installed
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | -
| License | MIT
| Doc PR | -
Commits
-------
43cc877 [HttpKernel] Check if "symfony/proxy-manager-bridge" package is installed
* 2.3:
[Debug 2.3] Fix test for PHP7
Run tests on hhvm instead of hhvm-nightly
Use HTTPS in README and some other fixes
add more entropy to generated classnames
Conflicts:
src/Symfony/Component/Debug/Tests/ErrorHandlerTest.php