This PR was submitted for the 5.2 branch but it was squashed and merged into the 4.4 branch instead.
Discussion
----------
[Form] Use !isset for checks cause this doesn't falsely include 0
| Q | A
| ------------- | ---
| Branch? | 5.2
| Bug fix? | yes
| New feature? | no
| Deprecations? | no
| Tickets | Fix#40999
| License | MIT
By using isset 0 won't falsely included anymore
Commits
-------
cd541c5a63 [Form] Use !isset for checks cause this doesn't falsely include 0
This PR was squashed before being merged into the 4.4 branch.
Discussion
----------
[Mailer] Fix typo in README
| Q | A
| ------------- | ---
| Branch? | 4.4
| Bug fix? | yes
| New feature? | no
| Deprecations? | no
| Tickets |
| License | MIT
| Doc PR |
Commits
-------
14f2a2974c [Mailer] Fix typo in README
This PR was submitted for the 6.0 branch but it was merged into the 4.4 branch instead.
Discussion
----------
[VarDumper] fix tests
| Q | A
| ------------- | ---
| Branch? | 4.4
| Bug fix? | no
| New feature? | no
| Deprecations? | no
| Tickets | -
| License | MIT
| Doc PR | -
Fixing VarDumper on appveyor hopefully.
Commits
-------
bdf3222c60 [VarDumper] fix tests
This PR was merged into the 4.4 branch.
Discussion
----------
[DependencyInjection] keep container.service_subscriber tag on the decorated definition
| Q | A
| ------------- | ---
| Branch? | 4.4
| Bug fix? | yes
| New feature? | no
| Deprecations? | no
| Tickets |
| License | MIT
| Doc PR |
Commits
-------
b917fafc8d keep container.service_subscriber tag on the decorated definition
This PR was merged into the 4.4 branch.
Discussion
----------
[Filesystem] fix readlink() for Windows
| Q | A
| ------------- | ---
| Branch? | 4.4
| Bug fix? | yes
| New feature? | no
| Deprecations? | no
| Tickets |
| License | MIT
| Doc PR |
**How to reproduce**
Windows 10.0.19042.928, PHP 8.0.3, PHPUnit 9.5.4
run as Administrator
C:\php\php.exe ./phpunit --bootstrap ./vendor/autoload.php --configuration ./phpunit.xml.dist ./src/Symfony/Component/Filesystem/Tests
There were 2 failures:
1) Symfony\Component\Filesystem\Tests\FilesystemTest::testRemoveCleansInvalidLinks
Failed asserting that 'C:\Users\albat\AppData\Local\Temp\1618836823.005.2057903605\directory\dir\' is false.
D:\Z__PHP_PROJECT\symfony\src\Symfony\Component\Filesystem\Tests\FilesystemTest.php:379
2) Symfony\Component\Filesystem\Tests\FilesystemTest::testReadAbsoluteLink
Failed asserting that two strings are equal.
--- Expected
+++ Actual
@@ @@
-'C:\Users\albat\AppData\Local\Temp\1618836823.1681.131301953\dir\link'
+'C:\Users\albat\AppData\Local\Temp\1618836823.1681.131301953\file'
Commits
-------
f1b95d3c93 [Filesystem] fix readlink for Windows
This PR was merged into the 4.4 branch.
Discussion
----------
[HttpKernel] Fixes file_get_content in HttpCache's Store for PHP7.4+
| Q | A
| ------------- | ---
| Branch? | 4.4
| Bug fix? | yes
| New feature? | no
| Deprecations? | no
| Tickets | -
| License | MIT
| Doc PR | -
While trying to get 6.0 branch tests green, I encountered an error on multiple HttpKernel tests.
After investigating, it comes from the file_get_contents in the load method of the Store class from HttpCache. For obscure reasons, some calls are in errors and the content can't be read for "permission denied".
This is happening only on PHP 7.4.0+ (and that explains why the tests are green on the branch 5.4 as it runs php 7.2), adding the `@` error suppressor fixes this.
Commits
-------
c3ac11cc64 [HttpKernel] Fixes tests for PHP7.4+
This PR was merged into the 4.4 branch.
Discussion
----------
Fix markdown in README files
| Q | A
| ------------- | ---
| Branch? | 4.4
| Bug fix? | no
| New feature? | no <!-- please update src/**/CHANGELOG.md files -->
| Deprecations? | no <!-- please update UPGRADE-*.md and src/**/CHANGELOG.md files -->
| Tickets | Fix #... <!-- prefix each issue number with "Fix #", no need to create an issue if none exist, explain below instead -->
| License | MIT
| Doc PR | symfony/symfony-docs#... <!-- required for new features -->
I would up merge it and fix the upper branches while upmerging or via separate PR's, whatever is prefered
Commits
-------
ad0890c604 Fix markdown
This PR was merged into the 4.4 branch.
Discussion
----------
[Form] fix support for years outside of the 32b range on x86 arch
| Q | A
| ------------- | ---
| Branch? | 4.4
| Bug fix? | yes
| New feature? | no
| Deprecations? | no
| Tickets | -
| License | MIT
| Doc PR | -
Commits
-------
2353f19cc2 [Form] fix support for years outside of the 32b range on x86 arch
This PR was merged into the 5.2 branch.
Discussion
----------
[Security] Add an upgrade note about the removal of Serializable
| Q | A
| ------------- | ---
| Branch? | 5.2
| Bug fix? | no
| New feature? | no
| Deprecations? | no
| Tickets | N/A
| License | MIT
| Doc PR | N/A
Jumping from Symfony 5 on PHP 7.2/7.3 to Symfony 6 on PHP 8 directly will fry your sessions. I've added a note that PHP should be upgraded first. I believe that most apps will take that patch anyway, so this shouldn't be much of a footgun. 🤞🏻
Commits
-------
76572581aa Add an upgrade note about the removal of Serializable
This PR was merged into the 5.2 branch.
Discussion
----------
Make Mailgun Header compatible with other Bridges
| Q | A
| ------------- | ---
| Branch? | 5.2
| Bug fix? | no
| New feature? | no
| Deprecations? | no
| Tickets | -
| License | MIT
| Doc PR | -
This revert deprecating passing a header without the required Mailgun `h:` prefix.
And makes the bridge compatible with other bridges.
See:
- https://github.com/symfony/symfony/pull/36148#discussion_r619232776
- https://github.com/symfony/symfony/pull/41308#discussion_r637197208
Commits
-------
82f1f9c618 Make Mailgun Header compatible with other Bridges
This PR was merged into the 5.2 branch.
Discussion
----------
[Mailer] Remove dead code
| Q | A
| ------------- | ---
| Branch? | 5.2
| Bug fix? | no
| New feature? | no
| Deprecations? | no
| Tickets | -
| License | MIT
| Doc PR | -
Since #40481 the `async-aws/ses` package is required, thus the class always exists. We can remove this check
Commits
-------
52c1934717 [mailer] Remove useless code
This PR was squashed before being merged into the 4.4 branch.
Discussion
----------
[Messenger] Fix merging PrototypedArrayNode associative values
Now keys will not be considered as irrelevant.
| Q | A
| ------------- | ---
| Branch? | 4.4
| Bug fix? | yes
| New feature? | no
| Deprecations? | no
| Tickets | Fix#39846, fix#36874
| License | MIT
| Doc PR |
The problem was because in ConfigTree `options` are defined as `arrayNode` and has a `prototype('variable')`. So for combining config together `PrototypedArrayNode::mergeValues()` was called. And there was a trouble - for that `PrototypedArrayNode` `$rightSide` values were added without keys:
`\Symfony\Component\Config\Definition\PrototypedArrayNode::mergeValues`
```php
foreach ($rightSide as $k => $v) {
// prototype, and key is irrelevant, append the element
if (null === $this->keyAttribute) {
$leftSide[] = $v;
continue;
}
```
Commits
-------
11128c9c0b [Messenger] Fix merging PrototypedArrayNode associative values
This PR was merged into the 4.4 branch.
Discussion
----------
[WebProfilerBundle] Wrapping exception js in Sfjs check and also loading base_js Sfjs if needed
| Q | A
| ------------- | ---
| Branch? | 4.4
| Bug fix? | yes
| New feature? | no
| Deprecations? | no
| Tickets | Fix#41305
| License | MIT
| Doc PR | Not needed
This bug was introduced in #41168.
Best viewed with: https://github.com/symfony/symfony/pull/41346/files?w=1 - we could also remove the new indentation in `exception.js`.
In that PR, we innocently don't reload `Sfjs` if it doesn't exist. However, on an error page, that template ALSO adds an `Sfjs`, and that `Sfjs` is *smaller* - containing only a subset of the `Sfjs` functions. This is tricky, but this PR fixes it.
Here are the various situations:
1) I hit an exception page. The `Sfjs` from `exception.js` is loaded first. This adds the smaller `Sfjs`. Then the `base_js.html.twig` version is hit. Because `Sfjs` is missing the `loadToolbar()` method, that code DOES run again to reinitialize it.
2) I am on a normal page, then travel to an exception page with Turbo. In this case, the original `Sfjs` from `base_js.html.twig` was processed. Then, on the exception page, the `Sfjs` from `exception.js` is ignored, as is the 2nd execution of `base_js.html.twig`.
Overall, the JS could use some work for supporting things like Turbo (e.g. `DOMContentLoaded` doesn't work with Turbo, so the exception page JS is broken). But I wanted to focus on fixing the bug in this PR.
As a reminder, #41168 (avoiding overriding `Sfjs`) was done to help with the AJAX toolbar and Turbo - it's explained in point (1) on #41168.
Cheers!
Commits
-------
5fa42f11f3 Wrapping exception js in Sfjs check and also loading base_js Sfjs if needed
