This PR was merged into the 2.2 branch.
Discussion
----------
[HttpFoundation] Do not return an empty session id if the session was closed
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #7936
| License | MIT
| Doc PR | ~
A regression from the bug fix#9246 was introduced due to an incomplete fix ;
As the `started` flag on the NativeSessionStorage was not `true` anymore when saving the session, the session id was always empty when saving it, and thus when sending the `PHPSESSID` cookie.
The previous PR I made (#9530) was another approach to solve this regression, but this one should keep the fix made by @tecbot on #9246, and finish it with another verification. I may miss another place where `started` is used, but I don't really see which other conditions depending on this property should be altered...
**Note : this is #9611 rebased on 2.2**
This should be mergeable.
Commits
-------
5b9a727 When getting the session's id, check if the session is not closed
This PR was submitted for the master branch but it was merged into the 2.3 branch instead (closes#9636).
Discussion
----------
[Translation] Fixes#9633, Removed dependency to Symfony\Bundle\FrameworkBundle\Tests\TestCase
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #9633
| License | MIT
| Doc PR |
Commits
-------
2669486 [Translation] Fixes#9633, Removed dependency to Symfony\Bundle\FrameworkBundle\Tests\TestCase
This PR was merged into the 2.3 branch.
Discussion
----------
[Validator] Replaced inexistent interface
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | no
| Fixed tickets | #9631
| License | MIT
| Doc PR | -
`ClassMetadataFactoryInterface` was removed in 2.3 (#6096) and replaced with `MetadataFactoryInterface`.
Commits
-------
8fd3256 [Validator] Replaced inexistent interface.
This introduced a regression from #9246, with an incomplete fix ;
As the `started` flag on the NativeSessionStorage was not `true`
anymore when saving the session, the session id was always empty
when saving it, and thus when sending the `PHPSESSID` cookie
This PR was merged into the 2.3 branch.
Discussion
----------
Adjusting CacheClear Warmup method to namespaced kernels
Backported the patch in #9525 to the 2.3 branch.
| Q | A
| ------------- | ---
| Bug fix? | [yes]
| New feature? | [no]
| BC breaks? | [no]
| Deprecations? | [no]
| Tests pass? | have not tried yet
| License | MIT
| Fixed tickets | further fixes#1431
My kernel has been moved and namespaced to `Cfs\Bundle\MultiSiteBundle\Kernel\CfsKernel`. This worked fine until a change was made to how the kernel temp stuff is handled in the warmup phase.
When the app generates its own cache (i.e you run cache without warmup and access the site) everything is generated ok and the .meta files generate the proper reference to the FQN of the Kernel.
However if the warmup is used, it uses `Cfs\Bundle\MultiSiteBundle\Kernel\CfsKerne_` as the temporary Kernel, and when it does "fix references to the Kernel in .meta files" it generates 2 errors.
1. It does not use a string safe tempKernel name, so it never finds the reference to the kernel
2. If you fix that, then it replaces the FQN of the tempKernel with `CfsKernel`, the non-namespaced name of the proper Kernel (it also leaves the character count wrong in the serialization `C:43:<class>` where 43 is the char count for the FQN above)
The two changes above fix this, by escaping the string and replacing it with a FQN Kernel Class name.
What are your thoughts on this? If this sounds reasonable i'll do further enhancements and check tests.
Commits
-------
00d79d5 Adjusting CacheClear Warmup method to namespaced kernels
This PR was submitted for the master branch but it was merged into the 2.3 branch instead (closes#9616).
Discussion
----------
Fixed mistake in upgrade docu
Commits
-------
821006b Fixed mistake in upgrade docu
This PR was submitted for the 2.4 branch but it was merged into the 2.3 branch instead (closes#9610).
Discussion
----------
Container::camelize also takes backslashes into consideration
I [complained](https://github.com/symfony/symfony/pull/8494) about unreported and unnoticed BC break.
Service names in our DI container contain backslashes and that pull request rendered newer versions of Symfony unusable for us.
This commit fixes this - it normalizes backslashes in service names into underscores.
Commits
-------
0f9bb0a Container::camelize also takes backslashes into consideration
This PR was merged into the 2.2 branch.
Discussion
----------
Fixed some typos
Commits
-------
5c6edb3 fixed some typos
8d69bb5 fixed @expectedException class names
This PR was submitted for the master branch but it was merged into the 2.3 branch instead (closes#9599).
Discussion
----------
[Validator] Typo and better wording for german validator translation
Follow up of #9588 as supposed by @craue
Commits
-------
263b618 Typo and better wording for german validator translation
* 2.2:
No Entity Manager defined exception
fixed CS
[Acl] Fix for issue #9433
[Validator] fix docblock typos
[DependencyInjection] removed the unused Reference and Parameter classes use statements from the compiled container class
Fix mistake in translation's service definition.
if handler_id is identical to null fix
CS fix
Fixed ModelChoiceList tests in Propel1 bridge.
[AclProvider] Fix incorrect behaviour when partial results returned from cache
Check if the pipe array is empty before calling stream_select()
re-factor Propel1 ModelChoiceList
[Locale] fixed the failing test described in #9455
[Process] fix phpdoc and timeout of 0
bug #9445 [BrowserKit] fixed protocol-relative url redirection
Conflicts:
src/Symfony/Component/BrowserKit/Tests/ClientTest.php
src/Symfony/Component/Locale/Tests/Stub/StubIntlDateFormatterTest.php
This PR was merged into the 2.3 branch.
Discussion
----------
Add missing german translation for validators
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #9575
| License | MIT
| Doc PR |
Some validators do not have german translation yet.
Commits
-------
11fd126 Add german translation for several validators (Greater/Equal/Less)
This PR was merged into the 2.2 branch.
Discussion
----------
[BrowserKit] fixed protocol-relative url redirection
Fixed redirects to protocol relative URLs, e.g. //www.example.org. Previously the code would treat this as a relative URL.
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #9445
| License | MIT
| Doc PR |
Commits
-------
e8b5c84 bug #9445 [BrowserKit] fixed protocol-relative url redirection
This PR was submitted for the master branch but it was merged into the 2.2 branch instead (closes#9535).
Discussion
----------
No Entity Manager defined exception
This PR is linked to this issue : https://github.com/symfony/symfony/issues/9534
Commits
-------
719ccb7 No Entity Manager defined exception
This PR was squashed before being merged into the 2.2 branch (closes#9485).
Discussion
----------
[Acl] Fix for issue #9433
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #9433
| License | MIT
| Doc PR |
Two new test for issue #9433 :
`testUpdateAclInsertingMultipleObjectFieldAcesThrowsDBConstraintViolations()`
`testUpdateAclDeletingObjectFieldAcesThrowsDBConstraintViolations()`
The change to `updateAces()` line 857 is enough to make the first test succeed. When changing the `order` field value to a higher value, we must first change the value of the next entry (and all the next entries recursively) to preserve uniqueness of the `order` field in the database.
All the other changes are for the second test. In the former `updateAcl()` method, we commit the changes of the existing ACEs to the database before deleting or adding the new ones. We must delete the old ACEs before changing the existing ACEs in order to preserve uniqueness of the `order` field in the database.
Commits
-------
a38fab9 [Acl] Fix for issue #9433
This PR was merged into the 2.2 branch.
Discussion
----------
[Process] fix phpdoc and timeout of 0
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | -
| License | MIT
| Doc PR | n/a
getTimeout says it returns null when it's disabled. But passing 0 or 0.0 also disabled timeouts. So it should be treated as null as well. also it says it returns integer whereas it returns float.
Commits
-------
1758010 [Process] fix phpdoc and timeout of 0
This PR was merged into the 2.2 branch.
Discussion
----------
[AclProvider] Fix incorrect behavior when partial results returned from cache
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #9515
| License | MIT
| Doc PR | none
[AclProvider] Fix incorrect behavior when partial results returned from cache
findAcls wasn't throwing the NotAllAclsFoundException with the partial result when some acls were retrieved from the cache. This will fix that behavior. Ideally we would update it to cache items that were not found so it wouldn't have to query the db every time for non existent items, but I'm not sure if that is possible.
I'm not sure if the ACL cache is being used very much, I didn't find any docs about it but it seems to work except for this issue. It might be more appropriate to merge the fix in master if no one is relying on it to work in 2.2.
Commits
-------
edae59c [AclProvider] Fix incorrect behaviour when partial results returned from cache
This PR was merged into the 2.2 branch.
Discussion
----------
[DependencyInjection] removed the unused Reference and Parameter classes use statements from the compiled container class
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | ~
| License | MIT
| Doc PR | ~
Commits
-------
7730d8d [DependencyInjection] removed the unused Reference and Parameter classes use statements from the compiled container class
This PR was merged into the 2.3 branch.
Discussion
----------
[Intl] make currency bundle merge fallback locales when accessing data, ...
...allowing use of country-specific locales
Fixes#9262.
Commits
-------
edc287b [Intl] make currency bundle merge fallback locales when accessing data, allowing use of country-specific locales
This PR was merged into the 2.2 branch.
Discussion
----------
CS fix
I hope this is not BC break though handler_id checking should be done with identical comparison operator I believe.
Commits
-------
1b6b1e9 if handler_id is identical to null fix
This PR was submitted for the master branch but it was merged into the 2.3 branch instead (closes#9526).
Discussion
----------
[HttpFoundation] Removed useless check if self::$trustProxies is set
In Request::getClientIps() on line 772 there is a check if self::$trustedProxies is not set. If this condition evaluates to true the method will return.
Because of this, the second identical check on line 783 will never evaluate to true.
Commits
-------
21f2aee [HttpFoundation] Removed useless check if self::$trustProxies is set
In Request::getClientIps() on line 772 there is a check if self::$trustedProxies is not set. If this condition evaluates to true the method will return.
Because of this the second identical check on line 783 will never evaluate to true, as when reaching this position self::$trustedProxies must be set.
This PR was submitted for the 2.3 branch but it was merged into the 2.2 branch instead (closes#9537).
Discussion
----------
[FrameworkBundle] Fix mistake in translation's service definition.
Wrong class parameter fixed.
Commits
-------
b542e84 [FrameworkBundle] Fix mistake in translation's service definition.
This PR was merged into the 2.2 branch.
Discussion
----------
[2.2] Fixed ModelChoiceList tests in Propel1 bridge
Tests on 2.2 are broken since #9469 was merged.
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | -
| License | MIT
| Doc PR | -
Commits
-------
1e2d14e Fixed ModelChoiceList tests in Propel1 bridge.
This PR was submitted for the master branch but it was merged into the 2.2 branch instead (closes#9367).
Discussion
----------
[Process] Check if the pipe array is empty before calling stream_select()
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #9280
| License | MIT
| Doc PR |
ProcessPipes generates a warning frequently which can cause issues for custom shutdown functions. Adding a check to see if the pipe array is empty should be functionally equivalent without having to generate the error.
Fixes: #9280
Commits
-------
12f95e2 [Process] Check if the pipe array is empty before calling stream_select()
