This PR was merged into the 3.4 branch.
Discussion
----------
[Intl] Cleanup
| Q | A
| ------------- | ---
| Branch? | 3.4
| Bug fix? | no
| New feature? | no
| BC breaks? | no <!-- see https://symfony.com/bc -->
| 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 | symfony/symfony-docs#... <!-- required for new features -->
cleanup of #31365 for 3.4 + some other stuff to keep in sync across branches
Commits
-------
70a941ec0c [Intl] Cleanup
* 3.4:
[Console] Fix auto-complete for ChoiceQuestion (multi-select answers)
Translated form, security, validators resources into Belarusian (be)
[WebProfilerBundle] Don't filter submitted IP values
bumped Symfony version to 3.4.28
updated VERSION for 3.4.27
update CONTRIBUTORS for 3.4.27
updated CHANGELOG for 3.4.27
This PR was squashed before being merged into the 3.4 branch (closes#31377).
Discussion
----------
[Console] Fix auto-complete for ChoiceQuestion (multi-select answers)
| Q | A
| ------------- | ---
| Branch? | 3.4
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | https://github.com/symfony/symfony/issues/24072
| License | MIT
| Doc PR | -
Previously, a bug existed whereby for multi-select questions, the auto-complete would only work for the first answer supplied by the user. On all subsequent answers, the auto-complete would not appear.
Now it works as expected:
Commits
-------
59321fe031 [Console] Fix auto-complete for ChoiceQuestion (multi-select answers)
This PR was merged into the 3.4 branch.
Discussion
----------
Reword VarDumper description
| Q | A
| ------------- | ---
| Branch? | 3.4
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no <!-- don't forget to update UPGRADE-*.md and src/**/CHANGELOG.md files -->
| Tests pass? | yes <!-- please add some, will be required by reviewers -->
| Fixed tickets | n/a
| License | MIT
| Doc PR | n/a
It is hard to understand what "Built on top" refers to, and even when
knowing, the sentence looks weird.
The description of [the Github repository](https://github.com/symfony/var-dumper) has the same issue and should be changed.
Commits
-------
6024e16ea1 Reword VarDumper description
* 3.4:
Revert "bug #30620 [FrameworkBundle][HttpFoundation] make session service resettable (dmaicher)"
[Workflow] Fixed dumping when many transition with same name exist
fix ConsoleFormatter - call to a member function format() on string
This PR was merged into the 3.4 branch.
Discussion
----------
Revert "bug #30620 [FrameworkBundle][HttpFoundation] make session service resettable (dmaicher)"
| 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 | -
This reverts commit 029fb2e7e3, reversing
changes made to 9dad29d61c.
Reverts #30620
Replaces #31215
We don't need to solve this in 3.4
Making the session resettable should be done on master, by implementing `ResetInterface`.
On 3.4 apps, one should write a dedicated `SessionResetter` that would implement the reverted logic.
Commits
-------
4177331915 Revert "bug #30620 [FrameworkBundle][HttpFoundation] make session service resettable (dmaicher)"
* 3.4:
relax assertions in tests
[Validator] Translate messages into Japanese
Fix Thai translation in validators.th.xlf
[FramworkBundle] mark any env vars found in the ide setting as used
* 3.4:
[TwigBridge] Require twig ^1.40|^2.9
[Serializer] Fix tests
Use the apply tag instead of the filter tag
Updated some translation files
[Translator] Preserve default domain when extracting strings from php files
This PR was squashed before being merged into the 3.4 branch (closes#31275).
Discussion
----------
[Translator] Preserve default domain when extracting strings from php files
| 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 |
Use default domain when extracting translation strings with `null` as domain.
Commits
-------
ca77d9a33e [Translator] Preserve default domain when extracting strings from php files
* 3.4:
[Form] Fix author tag + exception messages
[TwigBridge] Fix deprecation on twig 2.9
[Validator] Fixed imprecise translations
[Validator] Add Dutch translations
Intercept redirections only for HTML format
Fix name and phpdoc of ContainerBuilder::removeBindings
[Intl] Update the ICU data to 64.2
This PR was squashed before being merged into the 4.2 branch (closes#31240).
Discussion
----------
Fix url matcher edge cases with trailing slash
| Q | A
| ------------- | ---
| Branch? | master / 4.2 (not sure whether this should've been against 4.2)
| Bug fix? | yes
| New feature? | no
| BC breaks? | no (I think... if so, in obscure route configurations like the ones that broke in 4.2.7 ;) )
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #30721
| License | MIT
As stated in #30721 the "redirecting" (compiled) UrlMatcher ignored host-requirements when doing a 'matched url except for trailing slash difference'-redirect.
With routes like this, you'd get 404's rather than 200's on the second routes.
```yaml
host1.withTrail:
path: /foo/ # host2/foo would become host2/foo/ due to this partial path-match
host: host1
host2.withoutTrail: # A request for host2/foo should match this route
path: /foo # host2/foo/ does not match this path
host: host2
```
This was caused by too eagerly testing whether a route could've worked with an additional trailing slash.
If it could be, that would result in an attempt to redirect _before_ testing the host.
The original url would get the additional slash and prior to actually redirecting, it'd be retested against the available routes. _That new url_ would actually match no routes, since now the host check for the first routes would actually be executed and fail the match for those routes. The adjusted path in the route does not match the second sets of routes...
This PR moves the trailing-slash check to after the host checks. I've added several tests with variants on these edge cases.
*Note:* This is a bug in 4.2 (and 4.3).
I fixed it against master. It appears 4.2's PhpMatcherTrait was renamed to CompiledUrlMatcherTrait in 4.3. But that made it loose its history.
So I'm not sure how to proceed from here. I can rebase it on 4.2 and do the change in PhpMatcherTrait, but that would probably fail to merge in master? I'm not nearly proficient enough in Symfony PR's to know how to solve all this ;)
@nicolas-grekas also asked for these or similar tests to be added to 3.4 as 'proof' those routes worked in that version as well. I have not (yet) done so.
The tests did work on the non-redirecting UrlMatcher without change (i.e. just running UrlMatcherTest), which implies - to me at least - the routes where properly defined and should not result in 404's simply because a RedirectableUrlMatcherInterface _can_ do redirects.
Actually, since I needed to change UrlMatcher as well, I'm not convinced these bugs where totally absent in 3.4. They didn't occur with the compiled version, since the host check was the very first if-statement in that humongous if-tree it generated.
PS, the branch name is a bit dramatic ;)
Commits
-------
4fcfa9d85c Fix url matcher edge cases with trailing slash
This PR was merged into the 3.4 branch.
Discussion
----------
Fix name and phpdoc of ContainerBuilder::removeBindings
| Q | A
| ------------- | ---
| Branch? | 3.4
| Bug fix? | yes
| New feature? | no
| BC breaks? | no <!-- see https://symfony.com/bc -->
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | https://github.com/symfony/symfony/pull/29944#discussion_r276718323
| License | MIT
| Doc PR | N/A
<!--
Write a short README entry for your feature/bugfix here (replace this comment block.)
This will help people understand your PR and can be used as a start of the Doc PR.
Additionally:
- Bug fixes must be submitted against the lowest 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 the master branch.
-->
Commits
-------
c93194d1db Fix name and phpdoc of ContainerBuilder::removeBindings
This PR was merged into the 4.2 branch.
Discussion
----------
[Form] resolve class name parameters
| Q | A
| ------------- | ---
| Branch? | 4.2
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #31052
| License | MIT
| Doc PR |
Commits
-------
5235be4aa3 resolve class name parameters
This PR was squashed before being merged into the 4.2 branch (closes#31023).
Discussion
----------
[Routing] Fix route URL generation in CLI context
| Q | A
| ------------- | ---
| Branch? | 4.2
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #30996
| License | MIT
| Doc PR | -
This fixes#30996 and makes URL generation in the CLI context behave the same as it does in the web context where the `LocaleListener` sets the default locale (to the router context).
The Travis CI failure is related to the fact that the constraint for `symfony/routing` should be bumped to `^4.2.6` in the composer.json of the FrameworkBundle (when it gets tagged).
Commits
-------
4a1ad4a5d6 [Routing] Fix route URL generation in CLI context
This PR was merged into the 3.4 branch.
Discussion
----------
[Intl] Update the ICU data to 64.2
| Q | A
| ------------- | ---
| Branch? | 3.4
| Bug fix? |no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes, including intl-data group
| Fixed tickets | -
| License | MIT
| Doc PR | -
It didn't take long for the ICU team to release a new version.
From their [changelog](http://site.icu-project.org/download/64#TOC-ICU4C-Download), 64.2 provides bug fixes and:
* Draft Unicode 12.1 data integration
* Updated locale data imported from CLDR 35.1
* Support for the new Japanese era "Reiwa" (令和).
Commits
-------
9aa2349ac6 [Intl] Update the ICU data to 64.2
This PR was merged into the 4.2 branch.
Discussion
----------
[Routing] fix trailing slash matching with empty-matching trailing vars
| Q | A
| ------------- | ---
| Branch? | 4.2
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | -
| License | MIT
| Doc PR | -
Reported by @bmack in https://github.com/symfony/symfony/pull/31107#issuecomment-484681404
This highlights a small inconsistency that exists for a long time (checked on 2.7 at least):
`new Route('/en-en/{b}', ['b' => 'bbb'], ['b' => '.*'])` matches `/en-en/`
`new Route('/en-en/{b}', ['b' => 'bbb'], ['b' => '.+'])` doesn't match it
(while both match `/en-en` and `/en-en/foo`)
This PR ensures the former behavior is preserved, while #31167 redirects the later to `/en-en`.
Commits
-------
d6da21ac19 [Routing] fix trailing slash matching with empty-matching trailing vars
This PR was merged into the 4.2 branch.
Discussion
----------
[Routing] fix trailing slash redirection with non-greedy trailing vars
| Q | A
| ------------- | ---
| Branch? | 4.2
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #30863, #31066
| License | MIT
| Doc PR | -
Fixes redirecting `/123/` to `/123` when the route is defined as `/{foo<\d+>}`
Commits
-------
d88833d27a [Routing] fix trailing slash redirection with non-greedy trailing vars
This PR was merged into the 4.2 branch.
Discussion
----------
[FrameworkBundle] decorate the ValidatorBuilder's translator with LegacyTranslatorProxy
| Q | A
| ------------- | ---
| Branch? | 4.2
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #31092, #31025
| License | MIT
| Doc PR | -
This allows defining a translator that implements only the new interface and use it with ValidatorBuilder.
ping @dvdknaap, @snebes since you were affected.
Commits
-------
a12656eaad [FrameworkBundle] decorate the ValidatorBuilder's translator with LegacyTranslatorProxy
This PR was merged into the 4.2 branch.
Discussion
----------
[HttpKernel] Fix get session when the request stack is empty
| Q | A
| ------------- | ---
| Branch? | 4.2
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | -
| License | MIT
This bug happen behind an exception on a kernel response event, when one collector (e.g. `RequestDataCollector`) is trying to get the request session and the request stack is currently empty.
**Reproducer**
https://github.com/yceruto/get-session-bug (`GET /`)
See logs on terminal:
```bash
Apr 15 20:29:03 |ERROR| PHP 2019-04-15T20:29:03-04:00 Call to a member function isSecure() on null
Apr 15 20:29:03 |ERROR| PHP PHP Fatal error: Uncaught Symfony\Component\Debug\Exception\FatalThrowableError: Call to a member function isSecure() on null in /home/yceruto/demos/getsession/vendor/symfony/http-kernel/EventListener/SessionListener.php:43
Apr 15 20:29:03 |DEBUG| PHP Stack trace:
Apr 15 20:29:03 |DEBUG| PHP #0 /home/yceruto/demos/getsession/vendor/symfony/http-kernel/EventListener/AbstractSessionListener.php(59): Symfony\Component\HttpKernel\EventListener\SessionListener->getSession()
Apr 15 20:29:03 |DEBUG| PHP #1 /home/yceruto/demos/getsession/vendor/symfony/http-foundation/Request.php(707): Symfony\Component\HttpKernel\EventListener\AbstractSessionListener->Symfony\Component\HttpKernel\EventListener\{closure}()
Apr 15 20:29:03 |DEBUG| PHP #2 /home/yceruto/demos/getsession/vendor/symfony/http-kernel/DataCollector/RequestDataCollector.php(65): Symfony\Component\HttpFoundation\Request->getSession()
Apr 15 20:29:03 |DEBUG| PHP #3 /home/yceruto/demos/getsession/vendor/symfony/http-kernel/Profiler/Profiler.php(167): Symfony\Component\HttpKernel\DataCollector\RequestDataCollector->collect(Object(Symfony\Component\HttpFoundation\Request), Object(Symfony\Component\HttpFoundation\Respo in /home/yceruto/demos/getsession/vendor/symfony/http-kernel/EventListener/SessionListener.php on line 43
```
Friendly ping @nicolas-grekas as author of the previous PR https://github.com/symfony/symfony/pull/28244
Commits
-------
d62ca37ab6 Fix get session when the request stack is empty
* 3.4:
Revert "bug #30423 [Security] Rework firewall's access denied rule (dimabory)"
[FrameworkBundle] minor: remove a typo from changelog
[VarDumper][Ldap] relax some locally failing tests
[Validator] #30192 Added the missing translations for the Tagalog ("tl") locale.
Make MimeTypeExtensionGuesser case insensitive
This PR was merged into the 3.4 branch.
Discussion
----------
[Validator] Added the missing translations for the Tagalog ("tl") locale
| Q | A
| ------------- | ---
| Branch? | 3.4
| Bug fix? | No
| New feature? | No
| BC breaks? | No
| Deprecations? | No
| Tests pass? | Yes
| Fixed tickets | #30192
| License | MIT
| Doc PR |
[Validator] This pull request will add the missing translations for the Tagalog ("tl") locale.
Commits
-------
6ab574b7c9 [Validator] #30192 Added the missing translations for the Tagalog ("tl") locale.
This PR was merged into the 3.4 branch.
Discussion
----------
[HttpFoundation] Make MimeTypeExtensionGuesser case insensitive
| Q | A
| ------------- | ---
| Branch? | 3.4
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #... <!-- #-prefixed issue number(s), if any -->
| License | MIT
| Doc PR | symfony/symfony-docs#... <!-- required for new features -->
Some mime types have a camelCase word in them.
The Apache HTTPD project list items are all lower case.
So I suggest making the $mimeType string lowercase while checking the array key.
That way, we can keep the list in sync.
Example: xlsm file mime type is `application/vnd.ms-excel.sheet.macroEnabled.12`
The key that matches the xlsm extension in the `$defaultExtensions` array is `application/vnd.ms-excel.sheet.macroenabled.12`
Example xlsm file:
https://github.com/vermeirentony/xlsm-example
Commits
-------
e294ee6b9a Make MimeTypeExtensionGuesser case insensitive
This PR was merged into the 3.4 branch.
Discussion
----------
[Security] Add a separator in the remember me cookie hash
Based on #89
Commits
-------
a29ce2817c [Security] Add a separator in the remember me cookie hash
* 3.4:
bumped Symfony version to 3.4.26
updated VERSION for 3.4.25
update CONTRIBUTORS for 3.4.25
updated CHANGELOG for 3.4.25
Workaround for \DateInterval::createFromDateString()
Missing Lithuanian translations added to validator component.
* 3.4:
Skip testing the phpunit-bridge on not-master branches when $deps is empty
more tests
[DI] Fixes: #28326 - Overriding services autowired by name under _defaults bind not working
This PR was merged into the 3.4 branch.
Discussion
----------
[DI] Overriding services autowired by name under _defaults bind not working
| Q | A
| ------------- | ---
| Branch? | 3.4
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #28326
| License | MIT
This is an implementation of ideas and suggestions of @nicolas-grekas and @GuilhemN.
Commits
-------
7e805eae2b more tests
35a40ace6f [DI] Fixes: #28326 - Overriding services autowired by name under _defaults bind not working
* 3.4:
minor: the meaning of the data breach was not correct
Optimize SVGs
property normalizer should also pass format and context to isAllowedAttribute
This PR was merged into the 3.4 branch.
Discussion
----------
Optimize SVGs
| Q | A
| ------------- | ---
| Branch? | 3.4 <!-- see below -->
| Bug fix? | yes
| New feature? | no <!-- don't forget to update src/**/CHANGELOG.md files -->
| BC breaks? | no <!-- see https://symfony.com/bc -->
| Deprecations? | no <!-- don't forget to update UPGRADE-*.md and src/**/CHANGELOG.md files -->
| Tests pass? | yes <!-- please add some, will be required by reviewers -->
| Fixed tickets | / <!-- #-prefixed issue number(s), if any -->
| License | MIT
| Doc PR | / <!-- required for new features -->
Used [svgo](https://github.com/svg/svgo) to optimize the svgs. I kept the `viewBox` attribute to keep the aspects when SVGs are rescaled.
I also added `insert_final_newline = false` to the `.editorconfig` file because the newlines are removed from the SVGs and there's only one line left.
Commits
-------
4614cea9d2 Optimize SVGs
This PR was merged into the 3.4 branch.
Discussion
----------
property normalizer should also pass format and context to isAllowedAttribute
| Q | A
| ------------- | ---
| Branch? | 3.4
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | found while working on https://github.com/symfony/symfony/pull/30888
| License | MIT
| Doc PR | -
the context and format are optional parameters to `isAllowedAttribute`, but should be forwarded. due to this omission, the PropertyNormalizer was ignoring the 'attributes' context option (and does in version 4 also ignore the 'ignore_attributes' context option - that one is a property on the normalizer class in version 3 and therefore not ignored here)
Commits
-------
13e2fb735d property normalizer should also pass format and context to isAllowedAttribute
* 3.4:
Show more accurate message in profiler when missing stopwatch
CS Fixes: Not double split with one array argument
Remove redundant animation prefixes
Remove redundant `box-sizing` prefixes
Rework firewall access denied rule
fixed CS
Fix missing $extraDirs when open_basedir returns
This PR was merged into the 3.4 branch.
Discussion
----------
CS Fixes: Not double split with one array argument
| Q | A
| ------------- | ---
| Branch? | 3.4 (master from #31063)
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | None
| License | MIT
| Doc PR | None
Keep to use the same CS in all the Symfony code base.
Use:
```php
$resolver->setDefaults([
'compound' => false
]);
```
Instead of:
```php
$resolver->setDefaults(
[
'compound' => false,
]
);
```
Keep the double split when the method has two or more arguments.
I miss a PSR with this rule.
Commits
-------
a56bf552ad CS Fixes: Not double split with one array argument
Keep to use the same CS in all the Symfony code base.
Use:
```php
$resolver->setDefaults([
'compound' => false
]);
```
Instead of:
```php
$resolver->setDefaults(
[
'compound' => false,
]
);
```
Keep the double split when the method has two or more arguments.
I miss a PSR with this rule.
This PR was squashed before being merged into the 4.2 branch (closes#31026).
Discussion
----------
[Serializer] Add default object class resolver
| Q | A
| ------------- | ---
| Branch? | 4.2
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | -
| License | MIT
| Doc PR | -
The commit 1d8b5af3f0 introduce a BC break because before that commit the `extractAttributes` the `$object` can be a string which contain the fully qualified name of an object.
To fix the BC break and preserve the new feature, I suggest to create a default object class resolver if it is not set by the developer.
Commits
-------
dd5b8f16f5 [Serializer] Add default object class resolver
This PR was merged into the 4.2 branch.
Discussion
----------
[Serializer] MetadataAwareNameConverter: Do not assume that property names are strings
| Q | A
| ------------- | ---
| Branch? | 4.2 (class introduced in v4.2.3)
| Bug fix? | yes
| New feature? | no
| BC breaks? | no <!-- see https://symfony.com/bc -->
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | https://github.com/api-platform/core/pull/2709
| License | MIT
| Doc PR | n/a
When this class was introduced, there was an assumption made about the type of `propertyNames` and therefore a `: ?string` return type was introduced in the fallbacks/normalization private methods. Because symfony doesn't use strict mode yet (compatibility issues with php IIRC), when using a non-string property name (for example the integer `0` which is a valid property name in an array), it will convert the integer to a string.
This is not good, especially if you have a name converter that returns the given property name (ie no transformation) you'll have it's type changed which isn't correct.
I've discovered this bug while working on adding this name converter in api platform (https://github.com/api-platform/core/pull/2709).
Commits
-------
af1e136ca0 MetadataAwareNameConverter: Do not assume that property names are strings
This PR was merged into the 3.4 branch.
Discussion
----------
[Security] Rework firewall's access denied rule
| Q | A
| ------------- | ---
| Branch? | 3.4
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | ~~#30099~~, #28229
| License | MIT
| Doc PR |
Follow tickets provided above to reproduce bugs. (there are also some project examples)
~~In addition, I'm looking for someone who knows an answer to [this](https://github.com/symfony/symfony/issues/30099#issuecomment-468693492) regarding rework in this PR.~~
Commits
-------
5790859275 Rework firewall access denied rule
This PR was merged into the 4.2 branch.
Discussion
----------
[Console] fix buildTableRows when Colspan is use with content too long
| Q | A
| ------------- | ---
| Branch? | 4.2 for bug fixes
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes ( new test added TableTest::testWithColspanAndMaxWith)
| Fixed tickets | https://github.com/symfony/symfony/issues/30701
| License | MIT
| Doc PR | no
<!-- fix for keeping ColumnMaxwith when Content is too long
Commits
-------
1cf9659b5f fix buildTableRows when Colspan is use with content too long
This PR was merged into the 4.2 branch.
Discussion
----------
[Serializer] take setIgnoredAttributes() deprecation into account
| Q | A
| ------------- | ---
| Branch? | 4.2
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets |
| License | MIT
| Doc PR |
Commits
-------
4134be127a take setIgnoredAttributes() deprecation into account
* 3.4:
[Serializer] Respect ignored attributes in cache key of normalizer
fix resetting the COLUMN environment variable
Fix TestRunner compatibility to PhpUnit 8
prevent mixup of the object to populate
This PR was squashed before being merged into the 3.4 branch (closes#30907).
Discussion
----------
[Serializer] Respect ignored attributes in cache key of normalizer
EUFOSSA
| 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 | -
Do not share the attributes cache in object normalizer when using a different setting for the ignoredAttributes setting.
In Symfony 4.2, the setter is deprecated in favor of the ignored_attibutes option in the $context. When merging this up, we will however still need to respect the field as well for BC, the cache key does not look at the default context (apart from the deprecated modifiers, the default context is immutable)
There might be performance regression for some use cases, but also could be a performance improvement when using 'attributes' in the context with lists of objects of the same class.
Commits
-------
926d228877 [Serializer] Respect ignored attributes in cache key of normalizer
This PR was merged into the 3.4 branch.
Discussion
----------
[serializer] prevent mixup in normalizer of the object to populate
EUFOSSA
| 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 | -
OBJECT_TO_POPULATE is meant to specify the top level object. The implementation left the option in the context and it would be used whenever we have the first element that matches the class. #30607 (to master) introduces the feature to also keep the instances of attributes to deeply populate an existing object tree. In both cases, we do not want the mix up to happen with what the current OBJECT_TO_POPULATE is.
Commits
-------
fdb668e051 prevent mixup of the object to populate
This PR was merged into the 3.4 branch.
Discussion
----------
Make tests independent from each other
| Q | A
| ------------- | ---
| Branch? | 3.4
| Bug fix? | yes
| New feature? | yes
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | none
| License | MIT
| Doc PR | n/a
Environment variables set in a test need to be restored to their
previous values or unset if we want to be able to run tests
independently.
Credits to @ostrolucky for spotting this issue, I'm available for help when merging this in more recent branch (issues may appear then).
Created during the EU-FOSSA hackathon
Commits
-------
00883fc409 Make tests independent from each other
