This PR was merged into the 2.5-dev branch.
Discussion
----------
[Console] hide output of ProgressHelper when isDecorated is false
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | yes
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #9511
| License | MIT
| Doc PR | n/a
Commits
-------
006cb81 [Console] show no output in ProgressHelper when isDecorated is false (fixes#9511)
This PR was submitted for the 2.2 branch but it was merged into the 2.3 branch instead (closes#9806).
Discussion
----------
[Security] Fix parent serialization of user object
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets |
| License | MIT
| Doc PR |
`parent::serialize()` and `parent::unserialize()`, which are used in the `AbstractToken` are [problematic](https://bugs.php.net/bug.php?id=62836) in PHP >= 5.4. [Cloning the object](https://gist.github.com/aurelijus/4713758) before serialization seems to fix this.
Commits
-------
e0bb891 [Security] Fix parent serialization of user object
This PR was merged into the 2.5-dev branch.
Discussion
----------
[Security][Acl] Add MutableAclProvider::updateUserSecurityIdentity
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | yes
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #5787
| License | MIT
| Doc PR | symfony/symfony-docs#3319
This provides a very simple function to enable the update of a User's username while keeping its associated ACEs (by updating its corresponding UserSecurityIdentity)
Developers can add a listener on the preUpdate of a user and remove all the related ACLs:
```php
if ($args->hasChangedField('username')) {
$aclProvider = $this->container->get('security.acl.provider');
$oldUsername = $args->getOldValue ('username');
$user = $args->getEntity();
$aclProvider->updateUserSecurityIdentity(UserSecurityIdentity::fromAccount($user) , $oldUsername);
}
```
Among the problems of not updating the UserSecurityIdentity:
- Inconsistent database, referring to a non-existent user.
- The user loses all its associated permissions
- If another user is created with the old username, it will inherit all the first user’s ACEs
This PR intends to fix Issue #5787 and is similar to and inspired from PR #8305.
This will also be heavily impacted by the outcome of #8848
Commits
-------
da53d92 [Security][Acl] Fix#5787 : Add MutableAclProvider::updateUserSecurityIdentity
This PR was merged into the 2.5-dev branch.
Discussion
----------
[PropertyAccess] Allowed non alphanumeric chars in object properties
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | https://github.com/symfony/symfony/issues/8930
| License | MIT
| Doc PR |
Commits
-------
20d4eb6 [PropertyAccess] Allowed non alphanumeric chars in object properties
This PR was merged into the 2.3 branch.
Discussion
----------
[DependencyInjection] fixed typo
Commits
-------
408d992 [DependencyInjection] fixed typo
This PR was submitted for the 2.4-dev branch but it was merged into the 2.4 branch instead (closes#9830).
Discussion
----------
[WCM][Tests] Improve tests running with specific requirements (except intl icu)
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | na
| License | MIT
| Doc PR | na
- [x] go over all skipped tests, take note and check they are reasonable
- [x] reenable memcache, mongodb, and memcached
We are keeping the icu intl related tests skipped because setting up icu 51.2 is totally time consuming in travis and it would require a custom distro box on travis because there are no ppa's available for the ubuntu version. I tried hard but it does not seem worth it. Same for plugging beta of memcached with pecl, it is just not reasonable to be running beta versions on travis. This then does not address #9797 but at least now we are aware.
This PR now can be merged as is as it improves tests that before were not ran. Not all but more than before. 👶
Commits
-------
6e0b2dc added condition to avoid skipping tests on JSON_PRETTY support
This PR was merged into the 2.3 branch.
Discussion
----------
[Tests|WCM] add memcache, memcached, and mongodb extensions to run skipped tests
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | na
| License | MIT
| Doc PR | na
- [x] go over all skipped tests, take note and check they are reasonable
- [x] reenable memcache, mongodb, and memcached
We are keeping the icu intl related tests skipped because setting up icu 51.2 is totally time consuming in travis and it would require a custom distro box on travis because there are no ppa's available for the ubuntu version. I tried hard but it does not seem worth it. Same for plugging beta of memcached with pecl, it is just not reasonable to be running beta versions on travis. This then does not address #9797 but at least now we are aware.
This PR now can be merged as is as it improves tests that before were not ran. Not all but more than before. 👶
Commits
-------
47a822d add memcache, memcached, and mongodb extensions to run skipped tests
This PR was merged into the 2.3 branch.
Discussion
----------
[DependencyInjection] Fixed support for backslashes in service ids.
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #9801
| License | MIT
| Doc PR |
This change is needed for consistency with `camelize()` which is used in [`ProxyDumper`](https://github.com/symfony/symfony/blob/2.3/src/Symfony/Bridge/ProxyManager/LazyProxy/PhpDumper/ProxyDumper.php#L69) and [`PhpDumper`](https://github.com/symfony/DependencyInjection/blob/2.3/Dumper/PhpDumper.php#L1275).
Either this PR needs to be merged for consistency or #9610 rolled back (if we don't want to support backslashes in service ids).
Anyone could tell me why we're not using the `camelize()` method internally in the `Container`?
Commits
-------
c6f210b [DependencyInjection] Fixed support for backslashes in service ids.
This PR was merged into the 2.5-dev branch.
Discussion
----------
[DependencyInjection] fixes#9815 Syntax error in PHP dumper
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #9815
| License | MIT
| Doc PR | -
I also updated the unit tests to test a namespaces class to prevent future issues like this one.
Commits
-------
e00b0f3 [DependencyInjection] fixes#9815 Syntax error in PHP dumper
This PR was merged into the 2.4 branch.
Discussion
----------
[Form] Improved debugger icon
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| License | MIT
Changes the form icon in the toolbar to be pixel perfect.
| before | after |
| ---------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------- |
| ![before](https://f.cloud.github.com/assets/890645/1558707/8c4e2d42-4fa4-11e3-9397-376f761869a2.png) | ![after](https://f.cloud.github.com/assets/890645/1558708/90257844-4fa4-11e3-8a83-73d163d47261.png) |
Commits
-------
74e59fe Changes the form icon in the toolbar to be pixel perfect.
This PR was submitted for the 2.2 branch but it was merged into the 2.3 branch instead (closes#9826).
Discussion
----------
fix#9356 [Security] Logger should manipulate the user reloaded from provider
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #9356
| License | MIT
PR for well explained ticket #9356 by nikolajosipovic
Any feedback welcome, this is my first PR to symfony
Commits
-------
17b0b34 fix#9356 [Security] Logger should manipulate the user reloaded from provider
This PR was merged into the 2.5-dev branch.
Discussion
----------
[EventDispatcher][HttpKernel] Move RegisterListenersPass from HttpKernel to EventDispatcher.
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | -
| License | MIT
| Doc PR | -
I've been recently in the position to use the DependencyInjection Component and the EventDispatcher together but not HttpKernel. To make it easier to use these two together I propose moving RegisterListenersPass to the EventDispatcher component where the ContainerAwareEventDispatcher already resides.
Commits
-------
89b8e0a [EventDispatcher][HttpKernel] Move RegisterListenersPass from HttpKernel to EventDispatcher.
This PR was squashed before being merged into the 2.5-dev branch (closes#9668).
Discussion
----------
[DepdencyInjection] forgot to add definition of dumped container member variable parameters
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #5408
| License | MIT
| Doc PR | n/a
this fixes the allowing of access to parameters property of the dumped container
Commits
-------
5b02d3f [DepdencyInjection] forgot to add definition of dumped container member variable parameters
This PR was merged into the 2.5-dev branch.
Discussion
----------
[DependencyInjection] fix a regression introduced in #9807
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | -
| License | MIT
| Doc PR | -
As @stof pointed out in #9807 we always need to dump the fully qualified class name, to support dumping containers in a namespace.
Commits
-------
0d78776 [DependencyInjection] fix a regression introduced in #9807
This PR was merged into the 2.5-dev branch.
Discussion
----------
[HttpKernel] Remove FrameworkBundle dependency in BundleTest
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| License | MIT
`HttpKernel` tests fatal due to missing dependency `AddConsoleCommandPass`.
Also removed unused import.
Commits
-------
20a064f [HttpKernel] Remove FrameworkBundle dependency in BundleTest
This PR was merged into the 2.5-dev branch.
Discussion
----------
[Console] Added a way to set the process title
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | yes
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | -
| License | MIT
* I did not write test, because I want to RFC the idea before
* Do we want the process update its name when a sub command is executed by a master command ?
Commits
-------
204a25e [Console] Added a way to set the process title
This PR was squashed before being merged into the 2.4 branch (closes#9775).
Discussion
----------
[FrameworkBundle] Added extra details in XMLDescriptor to improve container description
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets |
| License | MIT
| Doc PR |
Added some new attributes to XMLDescriptor used by container:debug command.
~~This PR is based on #9768 PR.~~
Commits
-------
8edc855 [FrameworkBundle] Added extra details in XMLDescriptor to improve container description
This PR was merged into the 2.5-dev branch.
Discussion
----------
[DependencyInjection] Avoid call_user_func in dumped containers.
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | -
| License | MIT
| Doc PR | -
This is the second commit from #9432.
When using static methods for either the factory or as the configurator we can avoid using call_user_func and directly use the class:method notation. This is faster (about 5 times, but we're talking milliseconds here) but I think the resulting code is also much easier to read.
The code to use call_user_func has to remain in PhpDumper because in the uncompiled container they still get used.
Commits
-------
be1eaaa [DependencyInjection] Avoid call_user_func in dumped containers.
This PR was squashed before being merged into the 2.4 branch (closes#9771).
Discussion
----------
Crawler default namespace fix
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | no
| BC breaks? | yes
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #9732, #6650
| License | MIT
| Doc PR | symfony/symfony-docs/pull/2979
Fix backwards compatibility of xml namespaces for having only one default namespace.
Commits
-------
cfff054 Crawler default namespace fix
This PR was squashed before being merged into the 2.5-dev branch (closes#9773).
Discussion
----------
[Form] Added delete_empty option to allow proper emptyData handling of collections
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | yes/no?
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #9375
| License | MIT
| Doc PR | symfony/symfony-docs#3338
This PR adresses the issue that if you have a form collection with the option required => false and you submit an empty collection so you will get a ArrayCollection that contains en element with the value null.
This behaviour was introduced with the following changes from symfony/symfony#3257
In addition to the above mentioned ticket there is also a description about the same issue here: http://stackoverflow.com/questions/19474872/symfony2-form-collection-allow-add-and-allow-delete-null-error-silex
With the changes of this PR the new option empty_data is introduced. With this option you will be able to ignore/delete such empty collections, so they will neither be validated nor appear as empty field in the result.
The option will remove/ignore such empty collections if you add them newly and if allow_add is enabled and
removes such empty collections only if allow_delete is enabled.
With setting required and empty_data accordingly it will be now flexible to customize to the outcome you want to achieve.
Thanks to @bschussek for the great work together - We have to discuss how to name this option so if delete
or ignore is the appropriate name.
Commits
-------
8bdb7a0 [Form] Added delete_empty option to allow proper emptyData handling of collections