* 3.4:
Fix travis script
minor fix for wrong case
[HttpFoundation] Fix `getMaxFilesize`
[Cache] fix warning on PHP 7.4
[Console] fix warning on PHP 7.4
Don't add value of (default/static) objects to the signature
fix(yml): fix comment in milti line value
This PR was merged into the 3.4 branch.
Discussion
----------
[Debug] minor fix for wrong method name case
| Q | A
| ------------- | ---
| Branch? | 3.4
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | -
| License | MIT
| Doc PR | -
Minor replacement of wrong case for function name, needs to be fixed in 3.4 as hinten [here](https://github.com/symfony/symfony/pull/32812#discussion_r308815666).
Commits
-------
8b2d67b minor fix for wrong case
This PR was merged into the 3.4 branch.
Discussion
----------
[Yaml] fix comment in multi line value
| Q | A
| ------------- | ---
| Branch? | 4.3 and lower <!-- see below -->
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes <!-- please add some, will be required by reviewers -->
| Fixed tickets | #32667 <!-- #-prefixed issue number(s), if any -->
| License | MIT
| Doc PR | symfony/symfony-docs#... <!-- required for new features -->
<!--
Replace this notice by a short README for your feature/bugfix. This will help people
understand your PR and can be used as a start for the documentation.
Additionally (see https://symfony.com/roadmap):
- Bug fixes must be submitted against the lowest maintained branch where they apply
(lowest branches are regularly merged to upper ones so they get the fixes too).
- Features and deprecations must be submitted against branch 4.4.
- Legacy code removals go to the master branch.
-->
Comments after blank lines are read as value in yaml.
```
`$yaml = <<<YAML
parameters:
abc
# Comment
YAML;
var_dump(Symfony\Component\Yaml\Yaml::parse($yaml));`
```
**Result before fix:**
```
array(1) {
["parameters"]=>
string(13) "abc
# Comment"
}
```
**Result after fix:**
```
array(1) {
["parameters"]=>
string(3) "abc"
}
```
Commits
-------
dd945e375c fix(yml): fix comment in milti line value
This PR was squashed before being merged into the 3.4 branch (closes#32790).
Discussion
----------
[HttpFoundation] Fix `getMaxFilesize`
When checking for the maximum size of an uploaded file you can't just rely on `upload_max_filesize` since the request might also exceed `post_max_size`. Also discussed in contao/contao#498
| Q | A
| ------------- | ---
| Branch? | 3.4
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | -
| License | MIT
| Doc PR | -
Commits
-------
54107bac33 [HttpFoundation] Fix `getMaxFilesize`
This PR was merged into the 3.4 branch.
Discussion
----------
[Cache] fix warning on PHP 7.4
| Q | A
| ------------- | ---
| Branch? | 4.3
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | not all, still some missing I cant find
| Fixed tickets |
| License | MIT
| Doc PR |
Some fixes in tests for PHP 7.4
Commits
-------
3324505b51 [Cache] fix warning on PHP 7.4
This PR was submitted for the 4.3 branch but it was squashed and merged into the 3.4 branch instead (closes#32806).
Discussion
----------
[Console] fix warning on PHP 7.4
| Q | A
| ------------- | ---
| Branch? | 4.3
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| License | MIT
In php 7.4rc, accessing an array by key, sometimes causes an error.
Rarely this error occurs on php 7.2.
```bash
ErrorException: Trying to access array offset on value of type int
/app/vendor/symfony/console/Input/ArrayInput.php:135
/app/vendor/symfony/console/Input/Input.php:55
/app/vendor/symfony/console/Application.php:208
/app/vendor/symfony/console/Application.php:149
/app/vendor/laravel/framework/src/Illuminate/Console/Application.php:90
/app/vendor/laravel/framework/src/Illuminate/Console/Application.php:182
/app/vendor/laravel/framework/src/Illuminate/Foundation/Console/Kernel.php:275
/app/vendor/laravel/framework/src/Illuminate/Foundation/Testing/PendingCommand.php:136
/app/vendor/orchestra/testbench-core/src/Database/MigrateProcessor.php:73
/app/vendor/orchestra/testbench-core/src/Database/MigrateProcessor.php:44
/app/vendor/orchestra/testbench-core/src/Concerns/WithLoadMigrationsFrom.php:27
```
Commits
-------
5f451e6f70 [Console] fix warning on PHP 7.4
This PR was merged into the 3.4 branch.
Discussion
----------
Don't add object-value of static properties in the signature of container metadata-cache
| Q | A
| ------------- | ---
| Branch? | 3.4
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | No ticket created, see below
| License | MIT
Running bin/console cache:clear can take a very long time and/or uses explosive amounts of memory if large protected/public static objects are available in classes for which metadata will be generated during the 'container dump'.
I.e. in \Symfony\Component\HttpKernel\Kernel::dumpContainer eventually a call to \Symfony\Component\Config\ResourceCheckerConfigCache::write is done. That writes a serialized version of the $metadata.
In that serialize a call to \Symfony\Component\Config\Resource\ReflectionClassResource::generateSignature for each service will eventually do a print_r on the class's (default) properties.
Class properties can also be (public/protected) static. PHP Reflection doesn't differentiate between those in that context. And _those_ static objects can also have non-scalar values, including objects.
If such an object is very large, for instance a service that references the Container (with many services, some also referencing the container, etc), the print_r will run into very deep recursions. I.e. it will either take "forever" or use so much memory the OOM killer is activated (it got to 25GB...).
This scenario only triggers when a cache existed prior to running cache:clear, i.e. when 'isFresh' is called. When first removing the cache dir (or running cache:clear with --no-warmup) and than running cache:clear does not trigger the behavior, apparently the hash is created in a saver time in that scenario.
This sample project offers code that will prime a static property into a service (the trick is having a command with such a class-reference/service):
https://github.com/arjenm/symfony-service-recursion
Add a var_dump to the `$defaults = $class->getDefaultProperties();` in \Symfony\Component\Config\Resource\ReflectionClassResource::generateSignature to see it in effect.
The first cache:clear will show a few null's; the second run - without removing the cache - will show a container instead of null.
Its still a tiny container, so it'll not trigger serious issues.
This PR prevents an object "default value" for a property from being fed to 'print_r'. Since normally an object-property should be null on a "fresh" look at a non-instantiated class it should be save to use (i.e. no BC).
Commits
-------
a80e56c460 Don't add value of (default/static) objects to the signature
This PR was merged into the 4.4 branch.
Discussion
----------
Ensure $request->hasSession() is always checked before calling getSession()
| Q | A
| ------------- | ---
| Branch? | 4.4
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | -
| License | MIT
| Doc PR | -
Commits
-------
7b2c326719 Ensure $request->hasSession() is always checked before calling getSession()
This PR was merged into the 4.3 branch.
Discussion
----------
Recompile container when translations directory changes
| Q | A
| ------------- | ---
| Branch? | 4.3
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #32706
| License | MIT
| Doc PR | N/A
The list of translation resources is cached by the container, so when adding a new translation file, the container needs to be recompiled otherwise the translator won't know about the new file.
Commits
-------
7f2e7e2e9a Recompile container when translations directory changes
This PR was merged into the 4.3 branch.
Discussion
----------
[DependencyInjection] Fix bindings and tagged_locator
| Q | A
| ------------- | ---
| Branch? | 4.3
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes <!-- please add some, will be required by reviewers -->
| Fixed tickets | <!-- #-prefixed issue number(s), if any -->
| License | MIT
| Doc PR |
Working `tagged_locator`
```yaml
App\XXX:
arguments:
$taggedLocator: !tagged_locator { tag: 'my_tag', index_by: 'my_key' }
```
Not working `tagged_locator`
```yaml
App\XXX:
bind:
$taggedLocator: !tagged_locator { tag: 'my_tag', index_by: 'my_key' }
```
Currently `ResolveBindingsPass` is executed after `ServiceLocatorTagPass`, which produces empty service locators when using `bind`.
I'm proposing to change the optimization passes order, so that this issue is solved.
I'm not confident with the impact, so let's discuss about it.
Commits
-------
bf4c713ad7 Fix bindings and tagged_locator
This PR was merged into the 3.4 branch.
Discussion
----------
[DI] Fix dumping Doctrine-like service graphs (bis)
| Q | A
| ------------- | ---
| Branch? | 3.4
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #30091, #29637
| License | MIT
| Doc PR | -
Dumping the container while accounting for circular references is hard :)
Commits
-------
a37f3e0807 [DI] Fix dumping Doctrine-like service graphs (bis)
ee491444f4 Failing test case for complex near-circular situation + lazy
This PR was merged into the 3.4 branch.
Discussion
----------
[HttpKernel] do not stopwatch sections when profiler is disabled
| Q | A
| ------------- | ---
| Branch? | 4.3
| Bug fix? | yes
| New feature? | no <!-- please update src/**/CHANGELOG.md files -->
| BC breaks? | no <!-- see https://symfony.com/bc -->
| Deprecations? | no <!-- please update UPGRADE-*.md and src/**/CHANGELOG.md files -->
| Tests pass? | yes <!-- please add some, will be required by reviewers -->
| Fixed tickets |
| License | MIT
| Doc PR |
the toolbar and profiler panel disable to profiler which then does not set the X-Debug-Token. so when the header does not exist, do not call the stopwatch methods with `null` which violates the contract and does not make sense. found with #32242
Commits
-------
8718cd1b15 [HttpKernel] do not stopwatch sections when profiler is disabled
This PR was merged into the 4.4 branch.
Discussion
----------
[Form][DX] derive default timezone from reference_date option when possible
| Q | A
| ------------- | ---
| Branch? | 4.4
| Bug fix? | no
| New feature? | yes
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | -
| License | MIT
| Doc PR | -
Currently, when `reference_date` and `model_timezone` are provided but are not the same, we get a validation exception:
9216cb75ac/src/Symfony/Component/Form/Extension/Core/Type/TimeType.php (L48-L50)
But that happen even if the `model_timezone` is `null` !
I propose to relax this behavior by passing the time zone of the configured `reference_date` to this `model_timezone` (by default), thus avoiding annoyances.
Commits
-------
a62feea4ad [DX] derive default timezone from reference_date option when possible
This PR was merged into the 3.4 branch.
Discussion
----------
[Stopwatch] fix some phpdocs
| Q | A
| ------------- | ---
| Branch? | 3.4
| Bug fix? | no
| New feature? | no <!-- please update src/**/CHANGELOG.md files -->
| BC breaks? | no <!-- see https://symfony.com/bc -->
| Deprecations? | no <!-- please update UPGRADE-*.md and src/**/CHANGELOG.md files -->
| Tests pass? | yes <!-- please add some, will be required by reviewers -->
| Fixed tickets |
| License | MIT
| Doc PR |
Fix phpdocs found in #32242
Commits
-------
66dc9069aa [Stopwatch] fix some phpdocs
This PR was merged into the 4.4 branch.
Discussion
----------
[Lock] Add return type to the new LockFactory
| Q | A
| ------------- | ---
| Branch? | 4.4
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | -
| License | MIT
| Doc PR | -
Let's leverage renaming to add a return type.
Commits
-------
489f36daa2 [Lock] Add return type to the new LockFactory
the toolbar and profiler panel disable to profiler which then does not set the X-Debug-Token. so when the header does not exist, do not call the stopwatch methods with null which violates the contract and does not make sense
This PR was squashed before being merged into the 4.3 branch (closes#32641).
Discussion
----------
[Messenger] Retrieve table default options from the SchemaManager
| Q | A
| ------------- | ---
| Branch? | 4.3
| Bug fix? | yes
| New feature? | no <!-- please update src/**/CHANGELOG.md files -->
| BC breaks? | no <!-- see https://symfony.com/bc -->
| Deprecations? | no <!-- please update UPGRADE-*.md and src/**/CHANGELOG.md files -->
| Tests pass? | yes <!-- please add some, will be required by reviewers -->
| Fixed tickets | #32321 <!-- #-prefixed issue number(s), if any -->
| License | MIT
This PR modify the Connection of the Doctrine Transport. It is now possible to specify a `SchemaConfig` that will be used when generating tables.
If Doctrine is configured as the following :
```yml
# config/packages/doctrine.yaml
doctrine:
dbal:
default_table_options:
charset: 'utf8mb4'
collate: 'utf8mb4_unicode_ci'
```
Then those options are used to create the table.
ping @weaverryan 😉
Commits
-------
93d0dc825e [Messenger] Retrieve table default options from the SchemaManager
This PR was merged into the 3.4 branch.
Discussion
----------
drop 4.2 branch from pull request template
| Q | A
| ------------- | ---
| Branch? | 3.4
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets |
| License | MIT
| Doc PR |
Commits
-------
48e7b14611 drop 4.2 branch from pull request template