* 2.3:
[ci] Phpunit tests wont run if composer is installed in a wrapper
[ci] Add version tag in phpunit wrapper to trigger cache-reset on demand
fix race condition at mkdir (#16258)
[PropertyAccess] Test access to dynamic properties
[PropertyAccess] Fix dynamic property accessing.
Conflicts:
src/Symfony/Component/PropertyAccess/Tests/PropertyAccessorTest.php
This PR was merged into the 2.7 branch.
Discussion
----------
[Form] trigger deprecation warning when using empty_value
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #15908
| License | MIT
| Doc PR |
The `empty_value` option is deprecated in the `choice`, `date`, and
`time` form types. Therefore, a deprecation warning must be triggered
when the users configures a value for this option.
The `datetime` form type does not need to be updated as it passes
configured values to the `date` and `time` form types which trigger
deprecation warnings anyway.
Commits
-------
405d4a8 trigger deprecation warning when using empty_value
This PR was merged into the 2.7 branch.
Discussion
----------
[FrameworkBundle] JsonDescriptor - encode container params only once
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #16379
| License | MIT
| Doc PR |
Commits
-------
dd9d076 JsonDescriptor - encode container params only once
This PR was squashed before being merged into the 2.7 branch (closes#16463).
Discussion
----------
[PropertyAccess] Port of the performance optimization from 2.3
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #16179
| License | MIT
| Doc PR | n/a
Portage of #16294 in the 2.7 branch.
Commits
-------
aa4cc90 [PropertyAccess] Port of the performance optimization from 2.3
This PR was merged into the 2.7 branch.
Discussion
----------
[Serializer] GetSetNormalizer shouldn't set/get static methods
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | maybe
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | -
| License | MIT
| Doc PR | -
This PR fixes setting static getters and setters when serializing.
This is strictly speaking a BC break but in my opinion a needed one because the current behavior is unexpected.
Commits
-------
d8d4405 [Serializer] GetSetNormalizer shouldn't set/get static methods
This PR was merged into the 2.7 branch.
Discussion
----------
[Serializer] PropertyNormalizer shouldn't set static properties
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | -
| License | MIT
| Doc PR | -
This PR fixes a bug where static properties where set by the PropertyNormalizer.
Commits
-------
b15bdca [Serializer] PropertyNormalizer shouldn't set static properties
The `empty_value` option is deprecated in the `choice`, `date`, and
`time` form types. Therefore, a deprecation warning must be triggered
when the users configures a value for this option.
The `datetime` form type does not need to be updated as it passes
configured values to the `date` and `time` form types which trigger
deprecation warnings anyway.
This PR was merged into the 2.3 branch.
Discussion
----------
removed all @covers annotations
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | n/a
| License | MIT
| Doc PR | n/a
Some unit tests have a `@covers` PHPUnit annotations. Most of them were added a very long time ago, but since then, we did not use them anymore and the existing ones are not maintained (see #16413). So, I propose to remove them all.
Commits
-------
1e0af36 removed all @covers annotations
This PR was squashed before being merged into the 2.3 branch (closes#16294).
Discussion
----------
[PropertyAccess] Major performance improvement
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #16179
| License | MIT
| Doc PR | n/a
This PR improves performance of the PropertyAccess component of ~70%.
The two main changes are:
* caching the `PropertyPath` initialization
* caching the guessed access strategy
This is especially important for the `ObjectNormalizer` (Symfony Serializer) and the JSON-LD normalizer ([API Platform](https://api-platform.com)) because they use the `PropertyAccessor` class in large loops (ex: normalization of a list of entities).
Here is the Blackfire comparison: https://blackfire.io/profiles/compare/c42fd275-2b0c-4ce5-8bf3-84762054d31e/graph
The code of the benchmark I've used (with Symfony 2.3 as dependency):
```php
<?php
require 'vendor/autoload.php';
class Foo
{
private $baz;
public $bar;
public function getBaz()
{
return $this->baz;
}
public function setBaz($baz)
{
$this->baz = $baz;
}
}
use Symfony\Component\PropertyAccess\PropertyAccess;
$accessor = PropertyAccess::createPropertyAccessor();
$start = microtime(true);
for ($i = 0; $i < 10000; ++$i) {
$foo = new Foo();
$accessor->setValue($foo, 'bar', 'Lorem');
$accessor->setValue($foo, 'baz', 'Ipsum');
$accessor->getValue($foo, 'bar');
$accessor->getValue($foo, 'baz');
}
echo 'Time: '.(microtime(true) - $start).PHP_EOL;
```
This PR also adds an optional support for Doctrine cache to keep access information across requests and improve the overall application performance (even outside of loops).
Commits
-------
284dc75 [PropertyAccess] Major performance improvement
* 2.3:
added the new Composer exclude-from-classmap option
fix expected argument type docblock
Set back libxml settings after testings.
fixed Twig deprecation notices
This PR was merged into the 2.3 branch.
Discussion
----------
added the new Composer exclude-from-classmap option
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | n/a
| License | MIT
| Doc PR | n/a
Commits
-------
65bef75 added the new Composer exclude-from-classmap option
This PR was merged into the 2.3 branch.
Discussion
----------
[HttpFoundation] fix expected argument type docblock
| Q | A
| ------------- | ---
| Fixed tickets |
| License | MIT
Commits
-------
1c01ebc fix expected argument type docblock
Calling the parent constructor of the mocked `Mongo` class tries to
connect to a local MongoDB server which fails in case no local server
was configured.
Similarly, when the parent constructor of the mocked `MongoCollection`
class is called it performs checks on the passed arguments which fails
again when a connection was not established successfully before.
* 2.3:
added missing quotes in YAML files
[HttpKernel] Add `@group time-sensitive` on some transient tests
[DoctrineBridge] Fix issue which prevent the profiler to explain a query
Use mb_detect_encoding with $strict = true
don't allow to install the split Security packages
bumped Symfony version to 2.3.35
updated VERSION for 2.3.34
update CONTRIBUTORS for 2.3.34
updated CHANGELOG for 2.3.34
This PR was merged into the 2.3 branch.
Discussion
----------
Use mb_detect_encoding with $strict = true
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | -
| License | MIT
| Doc PR | -
Otherwise, UTF-8 can be returned for non-UTF8 strings...
See e.g. https://3v4l.org/oMMnX
Commits
-------
e6c89f1 Use mb_detect_encoding with $strict = true
This PR was merged into the 2.3 branch.
Discussion
----------
[Security] don't allow to install the split Security packages
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #16134
| License | MIT
| Doc PR |
Currently, you would be able to install the Security component fromm
Symfony 2.3 together with one of the split packages from a higher
Symfony vesion like this:
```json
{
"require": {
"symfony/symfony": "2.3.*",
"symfony/security-core": "~2.7"
}
}
```
However, you will end up with classes being present twice.
This must be reverted after merging up in the `2.7` branch.
Commits
-------
0d14064 don't allow to install the split Security packages
Currently, you would be able to install the Security component fromm
Symfony 2.3 together with one of the split packages from a higher
Symfony vesion like this:
```json
{
"require": {
"symfony/symfony": "2.3.*",
"symfony/security-core": "~2.7"
}
}
```
However, you will end up with classes being present twice.
This must be reverted after merging up in the `2.7` branch.
* 2.3:
fixed YAML files missing quotes when a string starts with @
[Routing] mark internal classes
[Translation][Csv file] remove unnecessary statements, for better readability.
[Form] remove validation of FormRegistry::getType as FormRegistry::hasType does not validate either
This PR was merged into the 2.3 branch.
Discussion
----------
fixed YAML files missing quotes when a string starts with @
| 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
Commits
-------
e36fea8 fixed YAML files missing quotes when a string starts with @
This PR was merged into the 2.3 branch.
Discussion
----------
[Translation][Csv loader] remove unnecessary statements, for better readability.
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Fixed tickets | ~
| Tests pass? | yes
| License | MIT
Commits
-------
47b8c3e [Translation][Csv file] remove unnecessary statements, for better readability.
This PR was merged into the 2.3 branch.
Discussion
----------
[Form] remove type check in FormRegistry::getType
| Q | A
| ------------- | ---
| Bug fix? | yes (consistency)
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | -
| License | MIT
| Doc PR | -
remove validation of `FormRegistry::getType` as `FormRegistry::hasType` does not validate either. So `hasType` currently triggers a PHP warning with a wrong argument.
also developers do not work with the registry directly anyway but through the factory. and the factory already validates the value. So this validation is useless in reality.
Commits
-------
d37b9e6 [Form] remove validation of FormRegistry::getType as FormRegistry::hasType does not validate either
* 2.3:
[DoctrineBridge] Fix required guess of boolean fields
[DI] don't use array_map to resolve services
Remove dead code in the PropertyPath constructor
[Process] Inherit env vars by default in PhpProcess
[HttpFoundation] Fixes /0 subnet handling in IpUtils
[Form] Simplify DateTimeToStringTransformer Avoid unneeded catch and re-throw of the same exception.
[HttpKernel] Remove a duplicate test for the EsiFragmentRenderer
Conflicts:
src/Symfony/Component/Process/Process.php
src/Symfony/Component/PropertyAccess/Tests/PropertyAccessorCollectionTest.php
src/Symfony/Component/PropertyAccess/Tests/PropertyAccessorTest.php
This PR was merged into the 2.3 branch.
Discussion
----------
[Process] Inherit env vars by default in PhpProcess
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | -
| License | MIT
| Doc PR | -
This is the cause of our failures on Windows, where the SYSTEMROOT env var is mandatory for mcrypt_create_iv to work.
I don't know why the browserkit client is run with no env inheritance and this looks like a bug.
Same for PhpProcess emptying the env by default, this looks like a bug, esp. since the parent `Process` class defaults to inheriting the env.
Tests are not broken by this change.
Commits
-------
ab8cc29 [Process] Inherit env vars by default in PhpProcess
This PR was merged into the 2.3 branch.
Discussion
----------
Remove dead code in the PropertyPath constructor
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | n/a
| License | MIT
| Doc PR | n/a
Custom singulars have been removed from the component before merging it in Symfony, but the code parsing them was only removed partially.
Commits
-------
ad4d0eb Remove dead code in the PropertyPath constructor
This PR was squashed before being merged into the 2.3 branch (closes#16177).
Discussion
----------
[HttpFoundation] Fixes /0 subnet handling in IpUtils
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #16055
| License | MIT
| Doc PR | Not needed
Fixes bug #16055. For IP addresses with CIDR subnet length 0, the IP address must be valid - IPs with subnet masks greater than zero are implicitly validated due to the use of `ip2long` and `substr_compare` (although it's not particularly robust - there could be some future work to improve this here).
Commits
-------
d9ac571 [HttpFoundation] Fixes /0 subnet handling in IpUtils
This PR was merged into the 2.3 branch.
Discussion
----------
[HttpKernel] Remove a duplicate test for the EsiFragmentRenderer
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | -
| License | MIT
| Doc PR | -
Since the request was made a required argument to the `render()` method in #6829, this [test became a duplicate](23f51450bd/src/Symfony/Component/HttpKernel/Tests/RenderingStrategy/EsiRenderingStrategyTest.php (L28-L38)) of `testRenderFallbackToInlineStrategyIfEsiNotSupported()`.
Commits
-------
44d57a3 [HttpKernel] Remove a duplicate test for the EsiFragmentRenderer
Since the request was made a required argument to the `render()` method in #6829, this test became a duplicate of `testRenderFallbackToInlineStrategyIfEsiNotSupported()`.
* 2.3:
[Routing] use constants in tests
[Validator] Allow an empty path in a URL with only a fragment or a query
[HttpFoundation] Fix some typo in the Request doc
fixed CS
Added separated handling of root paths
This PR was merged into the 2.7 branch.
Discussion
----------
[TwigBundle] Fix Twig cache is not properly warmed
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #15982
| License | MIT
| Doc PR | -
Alternative to #15034
Commits
-------
e704ee4 [TwigBundle] Fix Twig cache is not properly warmed
This PR was merged into the 2.3 branch.
Discussion
----------
[filesystem] makeRelativePath does not work correctly from root
| Q | A
| ------------- | ---
| Bug fix? | yes/no
| New feature? | yes/no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #1234
| License | MIT
| Doc PR | #14066, #14067
When using `makeRelativePath`, it returns an incorrect path when trying to fetch an entry from the root:
$fs->makePathRelative('/foo/bar/baz', '/');
Actual result:
../foo/bar/baz
Expected result:
foo/bar/baz
As we have specified an absolute path, there is no point on having an `..` added. It works, because a root directory has a `..` which points to itself, but it could result in issues when the relative path is actually prefixed or concatted.
Commits
-------
791b124 fixed CS
7bb394e Added separated handling of root paths