This PR was merged into the 2.3 branch.
Discussion
----------
Fix performance (PHP5) and memory (PHP7) issues when using token_get_all
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #16868
| License | MIT
| Doc PR | -
Although it's not the case anymore on PHP 7, on PHP 5, `is_array` checks are much slower than `isset` checks.
Also from @peteward in #17384:
> New PHP7 memory manager will not release small buckets to OS automatically in cases exposed by `token_get_all()`. This function call addition specifically for PHP7 will reclaim this memory to keep the footprint down of long processe
> See above ticket and suggested actions by PHP internals team for long-running tasks (https://bugs.php.net/70098) - I think `cache:clear/warmup` on a heavy app justifies this.
> We're running on cloud-based hosting platforms under memory limitations (Platform.sh). When memory is exceeded we're into swap and the cache clearing process goes from seconds to minutes for the initial deployment, which really slows our development workflow and also causes holding page delays.
Commits
-------
e555aad Add gc_mem_caches() call for PHP7 after itoken_get_all() as new memory manager will not release small buckets to OS automatically
d1f72d8 Fix perf and mem issue when using token_get_all
This PR was merged into the 2.3 branch.
Discussion
----------
[ClassLoader] Use symfony/polyfill-apcu
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | -
| License | MIT
| Doc PR | -
Fixes APC usage on PHP 7 or when APCu bc-layer is disabled
Commits
-------
623595f [ClassLoader] Use symfony/polyfill-apcu
This PR was squashed before being merged into the 2.3 branch (closes#17326).
Discussion
----------
[Console] Display console application name even when no version set
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| License | MIT
When displaying help of an Application with a name but no version, it show "Console Tool" instead of the application name.
Commits
-------
61e810e [Console] Display console application name even when no version set
This PR was merged into the 2.3 branch.
Discussion
----------
[Yaml] Fix tests on PHP 7.0.2
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #1234
| License | MIT
| Doc PR | -
Related to https://bugs.php.net/66179 that has been fixed in 7.0.2
Commits
-------
902f6dd [Yaml] Fix tests on PHP 7.0.2
This PR was merged into the 2.3 branch.
Discussion
----------
[Serializer] Remove normalizer cache in Serializer class
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets |
| License | MIT
| Doc PR |
The serializer cache the normalizer/denormalizer to use based only on the class and format. But the supportsNormalization and supportDenormalization methods can decide based on the data passed leading to hard to find bugs when the serializer used a cached normalizer it shouldn't use.
Commits
-------
8566dc1 Remove normalizer cache in Serializer class
This PR was merged into the 2.3 branch.
Discussion
----------
[Yaml] recognize when a block scalar is left
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #17116
| License | MIT
| Doc PR |
The parser did not recognize when the block scalar was completely parsed
and thus treated following comments as they need to be kept leading to
parse errors on the following lines.
Commits
-------
3a22165 [Yaml] recognize when a block scalar is left
The parser did not recognize when the block scalar was completely parsed
and thus treated following comments as they need to be kept leading to
parse errors on the following lines.
This PR was merged into the 2.3 branch.
Discussion
----------
bug #14246 [Filesystem] dumpFile() non atomic
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #14246
| License | MIT
| Doc PR | none
Remove race condition where `dumpFile` has written the resulting file but has not given access yet.
Commits
-------
e4015d5 bug #14246 [Filesystem] dumpFile() non atomic
This PR was merged into the 2.3 branch.
Discussion
----------
CS: use nowdoc instead of heredoc
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | —
| License | MIT
| Doc PR | —
if this is accepted, we could add the fixer to symfony level of php-cs-fixer: https://github.com/FriendsOfPHP/PHP-CS-Fixer/pull/1580
Commits
-------
3dca549 use nowdoc instead of heredoc
This PR was merged into the 2.3 branch.
Discussion
----------
Do not use HttpKernel Extension when not needed
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | -
| License | MIT
| Doc PR | -
HttpKernel Extension only adds `addClassesToCompile`. So the class hierarchy should be slim if it's not used.
Commits
-------
4978e19 Do not use HttpKernel Extension when not needed
This PR was merged into the 2.3 branch.
Discussion
----------
[Yaml] fix indented line handling in folded blocks
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #9105
| License | MIT
| Doc PR |
Commits
-------
756834c [Yaml] fix indented line handling in folded blocks
This PR was merged into the 2.3 branch.
Discussion
----------
[Config] Fix array sort on normalization in edge case
| 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
-------
eca41a8 [Config] Fix array sort on normalization in edge case
This PR was merged into the 2.3 branch.
Discussion
----------
improve BrowserKit test coverage p1
Hi,
This PR improves BrowserKit test coverage.
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| License | MIT
I rebased previous PR https://github.com/symfony/symfony/pull/16549 to 2.3
Commits
-------
0261b48 improve BrowserKit test coverage p1
This PR was merged into the 2.3 branch.
Discussion
----------
[Process] More robustness and deterministic tests
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #16988, #15617
| License | MIT
| Doc PR | -
Commits
-------
d1a178a [Process] More robustness and deterministic tests
This PR was merged into the 2.3 branch.
Discussion
----------
[Security] skip bcrypt tests on incompatible platforms
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #17055
| License | MIT
| Doc PR |
Not all PHP versions before 5.3.7 have backported fixes that make it
possible to use `password_hash()` function. Therefore, we have to skip
tests on not supported platforms.
Commits
-------
2a6fa7b use requires annotation
65eb188 skip bcrypt tests on incompatible platforms
Not all PHP versions before 5.3.7 have backported fixes that make it
possible to use `password_hash()` function. Therefore, we have to skip
tests on not supported platforms.
The test introduced in #16797 used the `assertFilePermissions()` method
to test for the expected result. This worked quite well for the PR as it
was submitted for the `master` branch. However, the tests now fail on
2.3 as the `FilesystemTestCase` class which contains this method was
introduced with Symfony 2.4.
This PR was merged into the 2.3 branch.
Discussion
----------
[Console] Avoid extra blank lines when rendering exceptions
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | -
| License | MIT
| Doc PR | -
Don't know if it can be considered as a bug fix as it is the behavior since a while (https://github.com/symfony/symfony/pull/9055), but I let you see by yourself the old vs new rendering:
Commits
-------
2e158df [Console] Avoid extra blank lines when rendering exceptions
This PR was merged into the 2.3 branch.
Discussion
----------
[Security] Verify if a password encoded with bcrypt is no longer than 72 characters
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #17047
| License | MIT
| Doc PR | -
From the [password_hash() docs](http://php.net/password_hash):
> Caution Using the PASSWORD_BCRYPT as the algorithm, will result in the password parameter being truncated to a maximum length of 72 characters.
Commits
-------
0a496e7 [Security] Enable bcrypt validation and result length tests on all PHP versions
5c30266 [Security] Verify if a password encoded with bcrypt is no longer than 72 characters
This PR was squashed before being merged into the 2.3 branch (closes#16959).
Discussion
----------
[Form] fix#15544 when a collection type attribute "required" is false, "prototype" should too
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #15544
| License | MIT
| Doc PR | -
Commits
-------
b4b5d63 [Form] fix#15544 when a collection type attribute "required" is false, "prototype" should too
This PR was merged into the 2.3 branch.
Discussion
----------
[Yaml] do not remove "comments" in scalar blocks
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #12478, #15857
| License | MIT
| Doc PR |
Inside scalar blocks, lines starting with a `#` character must be
treated like every other strings and must not be ignored as comments.
Commits
-------
3c72fcc [Yaml] do not remove "comments" in scalar blocks
This PR was merged into the 2.3 branch.
Discussion
----------
[HttpFoundation] Added the ability of using BinaryFileResponse with stream wrappers
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #12990#13187
| License | MIT
| Doc PR | ~
Commits
-------
1da3d61 [HttpFoundation] Added the ability of mapping stream wrapper protocols when using X-Sendfile
dd129b7 [HttpFoundation] Add a test case for using BinaryFileResponse with stream wrappers
* Fixes edge case on windows where PHP does not generate a PHP Warning but instead returns a wrong result https://bugs.php.net/bug.php?id=71103
* Improved error reporting on `unlink` used in `remove()`
This PR was merged into the 2.3 branch.
Discussion
----------
Clean useless deprecation silencing
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | -
| License | MIT
| Doc PR | -
We added these when we triggered unsilenced notices. But we silence them now.
Commits
-------
9953550 Clean useless deprecation silencing
This PR was merged into the 2.3 branch.
Discussion
----------
[Process] Enhance compatiblity with --enable-sigchild
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #16888
| License | MIT
| Doc PR | -
This is complete rewrite of the fallback `--enable-sigchild` handling in the Process class.
It removes most of the differences between this and a non-sigchild-enabled php.
Which means the test suite doesn't need anymore to be replayed 3 times (which is how I started this PR, looking for a way to test this component in less time).
I validated this with a locally compiled php, sigchild-enabled. Green.
Changes affect only this special-mode php.
Ping @romainneutron and @Seldaek (original writer of the sigchild support)
Submitted on 2.3 as bugfix, which it is to me.
Commits
-------
e7cc4aa [Process] Enhance compatiblity with --enable-sigchild
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
Conflicts:
src/Symfony/Component/Validator/Resources/translations/validators.lb.xlf
This PR was merged into the 2.3 branch.
Discussion
----------
[Process] Fix stopping a process on Windows
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #15617
| License | MIT
| Doc PR | -
Commits
-------
80fb51c [Process] Fix stopping a process on Windows
This PR was merged into the 2.3 branch.
Discussion
----------
CS: general fixes
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | ?
| Fixed tickets | N/A
| License | MIT
| Doc PR | N/A
Commits
-------
d3f671e CS: general fixes
This PR was squashed before being merged into the 2.3 branch (closes#16799).
Discussion
----------
Improve error message for undefined DIC aliases
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets |
| License | MIT
| Doc PR | n/a
Additionally, I think the arguments need to be the other way round (`templating` is an alias pointing to `templating.engine.twig`).
Commits
-------
9195cd3 Improve error message for undefined DIC aliases
This PR was merged into the 2.3 branch.
Discussion
----------
CS: remove unneeded parentheses around control statements
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | ?
| Fixed tickets | N/A
| License | MIT
| Doc PR | N/A
Commits
-------
45d250d CS: remove unneeded parentheses around control statements
This PR was squashed before being merged into the 2.3 branch (closes#16312).
Discussion
----------
[HttpKernel] clearstatcache() so the Cache sees when a .lck file has been released
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #15813
| License | MIT
| Doc PR | n/a
I've been trying to debug #15813 and modified the Store in a way to keep unique request IDs in the .lck file. That way, I was hoping to find out which request is blocking and/or if the request is actually still running.
It turned out that `is_file()` would claim that a lock file still exists, but a subsequent attempt to read the information from that file returned "file not found" errors.
So, my assumption is that the `is_file()` result is based on the fstat cache and wrong once a process has seen the lock file.
@jakzal said in https://github.com/symfony/symfony/issues/15813#issuecomment-149013691 that `unlink()`ing the lock file should clear the statcache, but I doubt this is true across PHP processes.
Commits
-------
982710f [HttpKernel] clearstatcache() so the Cache sees when a .lck file has been released
This PR was merged into the 2.3 branch.
Discussion
----------
Add missing exclusions from phpunit.xml.dist
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | -
| License | MIT
| Doc PR | -
Commits
-------
d4880c4 Add missing exclusions from phpunit.xml.dist
This PR was squashed before being merged into the 2.3 branch (closes#16510).
Discussion
----------
[2.3][Process] fix Proccess run with pts enabled
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #12643
| License | MIT
| Doc PR |
Commits
-------
9cf90fb [2.3][Process] fix Proccess run with pts enabled
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
