This PR was merged into the 5.1 branch.
Discussion
----------
[Messenger] fix ignored account & endpoint options amazon sqs connection
| Q | A
| ------------- | ---
| Branch? | 5.1<!-- see below -->
| Bug fix? | yes
| New feature? | no <!-- please update src/**/CHANGELOG.md files -->
| Deprecations? | no <!-- please update UPGRADE-*.md and src/**/CHANGELOG.md files -->
| Tickets | Fix#37598 <!-- prefix each issue number with "Fix #", no need to create an issue if none exist, explain below instead -->
| License | MIT
<!--
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/releases):
- Always add tests and ensure they pass.
- Never break backward compatibility (see https://symfony.com/bc).
- 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 master.
-->
Commits
-------
fc90a3b7c6 [Messenger] fix ignore account & endpoint options amazon sqs connection
This PR was squashed before being merged into the 3.4 branch.
Discussion
----------
[FrameworkBundle] set default session.handler alias if handler_id is not provided
| Q | A
| ------------- | ---
| Branch? | 3.4
| Bug fix? | yes
| New feature? | no
| Deprecations? | no
| Tickets | Fix#37245
| License | MIT
| Doc PR |
Without a configured [`handler_id`](https://symfony.com/doc/current/session.html#configuration), the `session.handler` alias is not created, and somes services depend on this alias specialy.
Related to the reported issue above (affected version 4.4.9), the `ServiceHandlerInterface` is an alias of `session.handler`, when execute :
- `bin/console debug:container session.handler` : The alias is not defined, so the `ContainerDebugCommand` look for a service containing the `session.handler` as part of his name (ContainerDebugCommand::findServiceIdsContaining() is called and returns by default `session.handler.native_file`)
- `bin/console debug:container SessionHandlerInterface` : The service is defined as an alias of `session.handler`, when calling `ContainerBuilder::getDefinition('session.handler')` the exception occured as the alias is not defined.
# Implemented solution
Create a default `session.handler` when the param `handler_id` is not provided (I choosed `session.handler.native_file` rather than `NullSessionHandler`).
# Affected versions
From [`3.4`](https://github.com/symfony/framework-bundle/blob/3.4/DependencyInjection/FrameworkExtension.php#L879) to [`5.1`](https://github.com/symfony/framework-bundle/blob/5.1/DependencyInjection/FrameworkExtension.php#L955)
Commits
-------
46c8c3791b [FrameworkBundle] set default session.handler alias if handler_id is not provided
This PR was merged into the 3.4 branch.
Discussion
----------
Fix checks for phpunit releases on Composer 2
| Q | A
| ------------- | ---
| Branch? | 3.4
| Bug fix? | yes
| New feature? | no
| Deprecations? | no
| Tickets | Fix#37601
| License | MIT
| Doc PR | n/a
`simple-phpunit` has a core assumption that any version of PHPUnit without a stable release will only have one dev version returned by Composer. Per https://github.com/symfony/symfony/issues/37601, it's possible for Composer 2 to list **more than one dev version**. This breaks that assumption and therefore prevents the installation of 9.3.* ([which is needed for testing on PHP 8](https://github.com/sebastianbergmann/phpunit/pull/4374#issuecomment-657029594)).
The fix implemented here is to remove any versions containing `dev-` or `-dev` from the list of possible versions to see if any stable versions remain.
Commits
-------
2bb3f08fba Fix checks for phpunit releases on Composer 2 (resolves#37601)
The password was already hidden in POST parameters, but still remained visible in raw content
[HttpKernel] added password hiding in request data collector raw content
The password was already hidden in POST parameters, but still remained visible in raw content
[HttpKernel] added password hiding in request data collector raw content
The password was already hidden in POST parameters, but still remained visible in raw content
This PR was squashed before being merged into the 5.2-dev branch.
Discussion
----------
[PhpUnitBridge] Polyfill new phpunit 9.1 assertions
| Q | A
| ------------- | ---
| Branch? | master
| Bug fix? | no
| New feature? | yes
| Deprecations? | no
| Tickets | Fix#37320
| License | MIT
| Doc PR | n/a
This PR adds the following new assertions that was added to PhpUnit 9.1:
- Implement **`assertIsNotReadable()`** as alternative for `assertNotIsReadable()`
- Implement **`assertIsNotWritable()`** as alternative for `assertNotIsWritable()`
- Implement **`assertDirectoryDoesNotExist()`** as alternative for `assertDirectoryNotExists()`
- Implement **`assertDirectoryIsNotReadable()`** as alternative for `assertDirectoryNotIsReadable()`
- Implement **`assertDirectoryIsNotWritable()`** as alternative for `assertDirectoryNotIsWritable()`
- Implement **`assertFileDoesNotExist()`** as alternative for `assertFileNotExists()`
- Implement **`assertFileIsNotReadable()`** as alternative for `assertFileNotIsReadable()`
- Implement **`assertFileIsNotWritable()`** as alternative for `assertFileNotIsWritable()`
- Implement **`assertMatchesRegularExpression()`** as alternative for `assertRegExp()`
- Implement **`assertDoesNotMatchRegularExpression()`** as alternative for `assertNotRegExp()`
Commits
-------
51c0bf0d0d [PhpUnitBridge] Polyfill new phpunit 9.1 assertions
This PR was merged into the 4.4 branch.
Discussion
----------
[Mailer] Fix mandrill raw http request setting from email/name
| Q | A
| ------------- | ---
| Branch? | 4.4, 5.0, 5.1
| Bug fix? | yes
| New feature? | no
| Deprecations? | no
| Tickets | Fix#36879
| License | MIT
| Doc PR | None
As describe in https://github.com/symfony/symfony/issues/36879 there is a bug in sending raw http request to mandrill it will not set from email/name correct.
As you can see i make sure to set `from_email` and `from_name` correct now and changed the unit test to check correct you can see the doc that the format is correct https://mandrillapp.com/api/docs/messages.curl.html#method-send-raw
Commits
-------
6128dd0b75 ticket_36879 - Fix mandrill raw http request setting from email/name
This PR was merged into the 5.2-dev branch.
Discussion
----------
[HttpFoundation] Added File::getContent()
| Q | A
| ------------- | ---
| Branch? | master
| Bug fix? | no
| New feature? | yes
| Deprecations? | no
| Tickets |
| License | MIT
| Doc PR |
---
This is a very simple PR but I think it increases the DX when working
with uploaded files: I looked at for this method and I did not found it.
So I had to check witch method returns the "pathname" or to remember I
can cast the object to string to get its pathname.
It's a small detail, but IMHO it's better with it.
About file_get_contents vs stream: let's be simple here. If one want a
stream, they can use the code they always used for it :)
Commits
-------
50d5167a66 [HttpFoundation] Added File::getContent()
This PR was squashed before being merged into the 4.4 branch.
Discussion
----------
[Mailer] Fix reply-to functionality in the SendgridApiTransport
| Q | A
| ------------- | ---
| Branch? | 4.4
| Bug fix? | yes
| New feature? | no
| Deprecations? | no
| Tickets |
| License | MIT
| Doc PR |
When sending a message using the `SendgridApiTransport`, the reply-to address was being ignored. In other transports, the reply to can be set using headers, but SendGrid requires that certain fields be added explicitly to the API payload. This is already handled for From, To, Cc, Bcc, and Subject, but was not handled for Reply-To. This change extracts the reply to address from the `Email` object and adds it to the payload.
Note that the `Email` object allows for multiple Reply-To addresses, but SendGrid only supports a single one, so I am just using the first element of the array.
I also fixed a link in a comment to SendGrid's documentation explaining the reserved headers that are not allowed.
Commits
-------
2cf25d1055 [Mailer] Fix reply-to functionality in the SendgridApiTransport
* 5.1:
[DependencyInjection][Config] Use several placeholder unique prefixes for dynamic placeholder values
[Mime] Keep Sender full address when used by non-SMTP transports
Update MimeTypes.php
* 5.0:
[DependencyInjection][Config] Use several placeholder unique prefixes for dynamic placeholder values
[Mime] Keep Sender full address when used by non-SMTP transports
Update MimeTypes.php
* 4.4:
[DependencyInjection][Config] Use several placeholder unique prefixes for dynamic placeholder values
[Mime] Keep Sender full address when used by non-SMTP transports
Update MimeTypes.php
This PR was merged into the 4.4 branch.
Discussion
----------
[Mime] Keep Sender full address when used by non-SMTP transports
| Q | A
| ------------- | ---
| Branch? | 4.4 <!-- see below -->
| Bug fix? | yes
| New feature? | no <!-- please update src/**/CHANGELOG.md files -->
| Deprecations? | no <!-- please update UPGRADE-*.md and src/**/CHANGELOG.md files -->
| Tickets | n/a <!-- prefix each issue number with "Fix #", no need to create an issue if none exist, explain below instead -->
| License | MIT
| Doc PR | n/a
refs #36178
The `Envelope` is an SMTP concept. The Sender is used in the `MAIL FROM` SMTP command, where only an address is supported. But we are also supporting non-SMTP transports, where the Sender might also be used as the `From` header, where a full mailbox is supported.
To take into account the 2 usages, this PR keeps the full mailbox in the Envelope and let the SMTP class only use the address (which was already the case anyway).
Commits
-------
1ca9be77aa [Mime] Keep Sender full address when used by non-SMTP transports
This PR was squashed before being merged into the 5.2-dev branch.
Discussion
----------
[Mime] switching source of mime types from Apache to "MIME-db"
| Q | A
| ------------- | ---
| Branch? | master
| Bug fix? | no
| New feature? | no
| Deprecations? | no
| Tickets | Fix#37412
| License | MIT
| Doc PR | no
Commits
-------
ebea452f57 [Mime] switching source of mime types from Apache to "MIME-db"
This PR was squashed before being merged into the 5.2-dev branch.
Discussion
----------
Remove some magic from TypeValidator logic and OptionsResolver type verify logic
| Q | A
| ------------- | ---
| Branch? | 5.0
| Bug fix? | yes
| New feature? | no
| Deprecations? | no
| License | MIT
Following discussion on https://twitter.com/Ocramius/status/1260236117129015297
The changes in this PR could ensure that types that can be verified by the TypeValidator is limited and there is no way to magically use it.
For example you can break it by defining you own function like:
```php
function is_usercollection {
return true;
}
```
that will validate any value that have the validator type as 'UserCollection'.
Another example would be defining `is_lower` or `ctype_int` that would also break the implementation.
Commits
-------
e8c9049a5a Remove some magic from TypeValidator logic and OptionsResolver type verify logic
* 5.1:
[Cache] Use the default expiry when saving (not when creating) items
Fix typo
Fix DBAL deprecation
[Form] Fix ChoiceType translation domain
Add Tagalog translations for new form messages
[Form] Add missing vietnamese translations
sync translations from master
[OptionsResolver] Fix force prepend normalizer
add vietnamese translation for html5 color validation
* 5.0:
[Cache] Use the default expiry when saving (not when creating) items
Fix typo
Fix DBAL deprecation
[Form] Fix ChoiceType translation domain
Add Tagalog translations for new form messages
[Form] Add missing vietnamese translations
sync translations from master
[OptionsResolver] Fix force prepend normalizer
add vietnamese translation for html5 color validation
* 4.4:
[Cache] Use the default expiry when saving (not when creating) items
Fix typo
Fix DBAL deprecation
[Form] Fix ChoiceType translation domain
Add Tagalog translations for new form messages
[Form] Add missing vietnamese translations
sync translations from master
[OptionsResolver] Fix force prepend normalizer
add vietnamese translation for html5 color validation
This PR was merged into the 3.4 branch.
Discussion
----------
[Cache] Use the default expiry when saving (not when creating) items
| Q | A
| ------------- | ---
| Branch? | 3.4
| Bug fix? | yes
| New feature? | no
| Deprecations? | no
| Tickets | Fix#37263
| License | MIT
| Doc PR |
This PR is for align expiration handling in ChainAdapter with ChainCache as proposed in #37263.
Commits
-------
49e9f3229c [Cache] Use the default expiry when saving (not when creating) items
This PR was squashed before being merged into the 3.4 branch.
Discussion
----------
[Form] Fix ChoiceType translation domain
| Q | A
| ------------- | ---
| Branch? | 3.4
| Bug fix? | yes
| New feature? | no
| Deprecations? |no
| Tickets | Fix #...
| License | MIT
| Doc PR | symfony/symfony-docs#...
When using
```
->add('foo', ChoiceType::class, [
'label' => 'label',
'translation_domain' => false,
'choices' => [
'choice.no' => false,
'choice.yes' => true,
],
'choice_translation_domain' => 'messages',
'expanded' => true,
'required' => false,
'placeholder' => false,
]);
```
I discovered that the choices was not translated.
Seems like it's because the subForm is using the `translation_domain` instead of the `choice_translation_domain`.
Commits
-------
2effda79ea [Form] Fix ChoiceType translation domain
This PR was merged into the 3.4 branch.
Discussion
----------
[Form] sync translations from master
| Q | A
| ------------- | ---
| Branch? | 3.4
| Bug fix? | no
| New feature? | no
| Deprecations? | no
| Tickets |
| License | MIT
| Doc PR |
Commits
-------
acea9be6d3 sync translations from master
This PR was squashed before being merged into the 3.4 branch.
Discussion
----------
Add Tagalog translations for new form messages
| Q | A
| ------------- | ---
| Branch? | 3.4
| Bug fix? | no
| New feature? | no <!-- please update src/**/CHANGELOG.md files -->
| Deprecations? | no <!-- please update UPGRADE-*.md and src/**/CHANGELOG.md files -->
| Tickets |
| License | MIT
| Doc PR |
This PR adds Tagalog translations for the new form messages @xabbuh added in https://github.com/symfony/symfony/pull/37552.
~~As I'm not aware of any Filipinos that regularly review Symfony pull requests, I'm going to ask my mother to proof-read this first before I remove the "WIP" label from the title. :)~~
Commits
-------
1809b7c5eb Add Tagalog translations for new form messages
This PR was merged into the 5.2-dev branch.
Discussion
----------
[Mime] allow non-ASCII characters in local part of email
| Q | A
| ------------- | ---
| Branch? | 4.4
| Bug fix? | yes
| New feature? | no
| Deprecations? | no
| Tickets | https://github.com/symfony/symfony/issues/34932
| License | MIT
| Doc PR | -
This fixes https://github.com/symfony/symfony/issues/34932 by allowing non-ASCII characters in the local part of emails.
I tried this using 3 different smtp servers (gmail, mailgun and local postfix) and for me this just works in case there are non-ASCII characters in the local part of emails. Emails are correctly delivered.
PHPMailer does this in the same way: https://github.com/PHPMailer/PHPMailer/blob/master/src/PHPMailer.php#L1411
This is also in line with the behavior of Swiftmailer (< 6.1) **before** this commit that introduced the `IdnAddressEncoder`: 6a87efd39b (diff-e5f85d26733017e183b2633ae3c433f0R31)
I'm not an expert when it comes to SMTP and all the different RFCs out there 😕 But for me this exception seems not needed.
Maybe @c960657 can help here?
Commits
-------
d057dffcd6 [Mime] allow non-ASCII characters in local part of email
* 5.1:
cs fix
add german translation of the html5 color validation
Fix PHP 8 deprecations
ensure compatibility with PHP 8 stack traces
clean up HHVM instructions
sync translations from master branch
silently ignore uninitialized properties when mapping data to forms
* 5.0:
cs fix
add german translation of the html5 color validation
Fix PHP 8 deprecations
ensure compatibility with PHP 8 stack traces
clean up HHVM instructions
sync translations from master branch
silently ignore uninitialized properties when mapping data to forms
* 4.4:
cs fix
add german translation of the html5 color validation
Fix PHP 8 deprecations
ensure compatibility with PHP 8 stack traces
clean up HHVM instructions
sync translations from master branch
silently ignore uninitialized properties when mapping data to forms
* 3.4:
cs fix
add german translation of the html5 color validation
sync translations from master branch
silently ignore uninitialized properties when mapping data to forms
This PR was merged into the 3.4 branch.
Discussion
----------
[Form] silently ignore uninitialized properties when mapping data to forms
| Q | A
| ------------- | ---
| Branch? | 3.4
| Bug fix? | yes
| New feature? | no
| Deprecations? | no
| Tickets | #36492
| License | MIT
| Doc PR |
Commits
-------
b4616c810c silently ignore uninitialized properties when mapping data to forms
A public method declared in a trait can be used by a class without
importing the trait itself when the trait is already imported by one
of the parent classes from the inheritance chain.
This PR was squashed before being merged into the 5.1 branch.
Discussion
----------
[PhpUnitBridge] Fix expectDeprecation() in isolation
| Q | A
| ------------- | ---
| Branch? | 5.1
| Bug fix? | yes
| 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 -->
Tests like
```
/**
* Do not remove this test in the next major version.
*
* @group legacy
* @runInSeparateProcess
*/
public function testOneInIsolation()
{
$this->expectDeprecation('foo');
@trigger_error('foo', E_USER_DEPRECATED);
}
```
will fail due to:
```
Testing Symfony\Bridge\PhpUnit\Tests\ExpectDeprecationTraitTest
R 1 / 1 (100%)R
Time: 111 ms, Memory: 6.00 MB
There were 2 risky tests:
1) Symfony\Bridge\PhpUnit\Tests\ExpectDeprecationTraitTest::testOneInIsolation
This test did not perform any assertions
/Users/alex/dev/symfony/src/Symfony/Bridge/PhpUnit/Tests/ExpectDeprecationTraitTest.php:38
2) Symfony\Bridge\PhpUnit\Tests\ExpectDeprecationTraitTest::testOneInIsolation
This test did not perform any assertions
OK, but incomplete, skipped, or risky tests!
Tests: 1, Assertions: 0, Risky: 2.
```
Commits
-------
e7e2ee75d3 Expect deprecations in isolation
* 5.1:
[String] Added a help message when translation-contracts is not installed
Add: ExcludeList usage for PHPUnit 9.4
[Cache] Fix compat wth DBAL v3
[String] throw when Alpine is used and translit fails
* 5.0:
[String] Added a help message when translation-contracts is not installed
Add: ExcludeList usage for PHPUnit 9.4
[Cache] Fix compat wth DBAL v3
[String] throw when Alpine is used and translit fails
This PR was squashed before being merged into the 5.2-dev branch.
Discussion
----------
[ErrorHandler] Allow override of the default non-debug template
| Q | A
| ------------- | ---
| Branch? | master
| Bug fix? | no
| New feature? | yes - a very tiny one
| Deprecations? | no
| License | MIT
| Doc PR | symfony/symfony-docs#... <!-- required for new features -->
Add new feature to allow the default hardcoded non-debug template to be overridden when `symfony/error-handler`used as a component in other projects.
Currently, when used as a component, its not possible (that I could see) to override the standard default template of `views/error.html.php` therefore the look and feel, and text is hard coded. There is much written on how to customise the error pages when using the full symfony stack, but not when using individual `symfony/error-handler` component on its own.
This PR is related to the use of `symfony/error-handler` as a component in other projects - specifically the Joomla Content Management System, where, in non debug mode, a generic template **needs to be translatable, and customisable** even though most of the time when the error occurs will bt at boot time and not much available to it. https://github.com/joomla/joomla-cms/issues/29968
This PR allows the use of code such as
```php
<?php
require 'vendor/autoload.php';
use Symfony\Component\ErrorHandler\ErrorHandler;
use Symfony\Component\ErrorHandler\ErrorRenderer\HtmlErrorRenderer;
ErrorHandler::register();
HtmlErrorRenderer::setTemplate('/path/to/custom/mine.php');
throw new Exception('myException is this', 500);
```
and then for further customisation of the "look and feel" to be done in `mine.php`
Due to the unique way the error handler component is init'ed, the use of static calls has been employed rather than any DI approach, I hope this is ok.
Commits
-------
6e1d16b44d [ErrorHandler] Allow override of the default non-debug template
* 5.1:
minor #37121 [Contracts] Add missing "extra.thanks" entries in composer.json (nicolas-grekas)
[Process] Fix Permission Denied error when writing sf_proc_00 lock files on Windows
fix handling null as empty data
[Security\Http] Skip remember-me logout on empty token
Missing return in loadValuesForChoices method
No need to create an issue when creating a PR
Use ">=" for the "php" requirement
[HttpClient] Fix promise behavior in HttplugClient
[Console] Fixes question input encoding on Windows
* 5.0:
minor #37121 [Contracts] Add missing "extra.thanks" entries in composer.json (nicolas-grekas)
[Process] Fix Permission Denied error when writing sf_proc_00 lock files on Windows
fix handling null as empty data
No need to create an issue when creating a PR
Use ">=" for the "php" requirement
[HttpClient] Fix promise behavior in HttplugClient
[Console] Fixes question input encoding on Windows
* 4.4:
minor #37121 [Contracts] Add missing "extra.thanks" entries in composer.json (nicolas-grekas)
[Process] Fix Permission Denied error when writing sf_proc_00 lock files on Windows
fix handling null as empty data
No need to create an issue when creating a PR
[Console] Fixes question input encoding on Windows
This PR was merged into the 5.2-dev branch.
Discussion
----------
[Contracts] Add missing "extra.thanks" entries in composer.json
| Q | A
| ------------- | ---
| Branch? | master
| Bug fix? | no
| New feature? | no
| Deprecations? | no
| Tickets | -
| License | MIT
| Doc PR | -
Something we forgot and that might help give some more visibility to symfony/contracts.
Commits
-------
bd04f0cce6 [Contracts] Add missing "extra.thanks" entries in composer.json
* 3.4:
[Process] Fix Permission Denied error when writing sf_proc_00 lock files on Windows
fix handling null as empty data
No need to create an issue when creating a PR
[Console] Fixes question input encoding on Windows
This PR was merged into the 3.4 branch.
Discussion
----------
[Process] Fix Permission Denied error when writing sf_proc_00 lock files on Windows
| Q | A
| ------------- | ---
| Branch? | 3.4
| Bug fix? | yes
| New feature? |no
| Deprecations? | no
| Tickets | Fix#37294
| License | MIT
| Doc PR |
Passes current Process unit tests.
On Windows systems, a new set of sf_proc_## files are generated in the system temp directory for each WindowsPipes object. These files are removed when the WindowsPipes object is destroyed.
This avoids receiving a permission denied error when attempting to write to sf_proc_## files between multiple sites running as different users, when the users do not have permissions to modify each others files.
Changes
- [Process] WindowsPipes always creates new sf_proc_## files in constructor
- [Process] WindowsPipes removes its sf_proc_## files in destructor
Commits
-------
220be89f5c [Process] Fix Permission Denied error when writing sf_proc_00 lock files on Windows
This PR was merged into the 3.4 branch.
Discussion
----------
[Form] fix handling null as empty data
| Q | A
| ------------- | ---
| Branch? | 3.4
| Bug fix? | yes
| New feature? | no
| Deprecations? | no
| Tickets | Fix#37493
| License | MIT
| Doc PR |
Commits
-------
b5aa55dab9 fix handling null as empty data
This PR was merged into the 3.4 branch.
Discussion
----------
[Console] Fixes question input encoding on Windows
| Q | A
| ------------- | ---
| Branch? | 3.4 <!-- see below -->
| Bug fix? | yes
| New feature? | no <!-- please update src/**/CHANGELOG.md files -->
| Deprecations? | no <!-- please update UPGRADE-*.md and src/**/CHANGELOG.md files -->
| Tickets | Fix#35842, Fix#36324, Fix#37495 and Fix#37278 <!-- prefix each issue number with "Fix #", if any -->
| License | MIT
| Doc PR | no <!-- 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/releases):
- Always add tests and ensure they pass.
- Never break backward compatibility (see https://symfony.com/bc).
- 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 master.
-->
To ask a question to a user, the [QuestionHelper](https://github.com/symfony/symfony/blob/master/src/Symfony/Component/Console/Helper/QuestionHelper.php) use [`fgets`](https://www.php.net/manual/en/function.fgets.php). However, special characters are not supported on Windows with this function (like accents: `é`, `à`, `ö`). The solution is to set a special encoding with [`sapi_windows_cp_get`](https://www.php.net/manual/en/function.sapi-windows-cp-get).
> Before
```php
$stream = fopen('php://stdin', 'r');
$input = fgets($stream);
echo $input;
// input: "Bonjour à tous"
// output: 'Bonjour \ tous" or "Bonjour tous"
```
> After
```php
// Check if the function exists because it only exists on Windows
if(function_exists('sapi_windows_cp_set')) {
sapi_windows_cp_get(437);
}
$stream = fopen('php://stdin', 'r');
$input = fgets($stream);
echo $input;
// input: "Bonjour à tous"
// output: 'Bonjour à tous"
```
*Thanks to @bnjmnfnk for the solution 😉*
Commits
-------
4288df4f74 [Console] Fixes question input encoding on Windows
This PR was squashed before being merged into the 4.4 branch.
Discussion
----------
[HttpClient] Fix promise behavior in HttplugClient
| Q | A
| ------------- | ---
| Branch? | 4.4 & up <!-- see below -->
| Bug fix? | yes
| New feature? | no <!-- please update src/**/CHANGELOG.md files -->
| Deprecations? | no <!-- please update UPGRADE-*.md and src/**/CHANGELOG.md files -->
| Tickets | Fix#37488 <!-- prefix each issue number with "Fix #", 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/releases):
- Always add tests and ensure they pass.
- Never break backward compatibility (see https://symfony.com/bc).
- 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 master.
-->
## The Problem
Promises have 2 important methods: `then` and `wait`.
To implement Httplug's promise interface, we built `HttplugPromise` on top of Guzzle promise.
However, when an error occurred (Httplug `NetworkException` thrown) while init the request/before actually sending the request,
`HttplugClient::sendAsyncRequest` will return a `Http\Promise\RejectedPromise`, which is a dummy implementation.
If the `then` callable returns a promise-like object, `Http\Promise\RejectedPromise` will treat it as plain value.
Guzzle promise will try to resolve the promise-like value, which is an object that has `then` method on it.
bbf3b200bc/src/Promise.php (L116)
To fix this, I edited `src/Symfony/Component/HttpClient/HttplugClient.php`.
Next, let me explain why to edit `src/Symfony/Component/HttpClient/Response/HttplugPromise.php`.
After the previous fix, when a Guzzle promise returned by the `then` callable, things will work.
However, If I return a `HttplugPromiseInterface`, it doesn't work, because Guzzle promise `wait` the return value (result)
only if it's a Guzzle promise.
bbf3b200bc/src/Promise.php (L63)
To fix this, I referenced the `wait` code of Guzzle promise and edited our `HttplugPromise`.
## How this fix make sense
So, why to return a promise from the `then` callable?
This let us change the promise chain according to current promise's result (fulfilled/rejected).
For example, we can retry an HTTP request if it failed.
Please take a look at my test code.
Commits
-------
147b6adc39 [HttpClient] Fix promise behavior in HttplugClient
* 5.1:
[HttpClient][CurlHttpClient] Fix http_version option usage
[HttpClient] fix parsing response headers in CurlResponse
[PropertyAccess] Remove inflector component
[Console] Do not check for "stty" using "exec" if that function is disabled
[Console] always use stty when possible to ask hidden questions
* 5.0:
[HttpClient][CurlHttpClient] Fix http_version option usage
[HttpClient] fix parsing response headers in CurlResponse
[Console] Do not check for "stty" using "exec" if that function is disabled
[Console] always use stty when possible to ask hidden questions
* 4.4:
[HttpClient][CurlHttpClient] Fix http_version option usage
[HttpClient] fix parsing response headers in CurlResponse
[Console] Do not check for "stty" using "exec" if that function is disabled
[Console] always use stty when possible to ask hidden questions
This PR was merged into the 4.4 branch.
Discussion
----------
[Console] always use stty when possible to ask hidden questions
| Q | A
| ------------- | ---
| Branch? | 4.4
| Bug fix? | yes
| New feature? | no
| Deprecations? | no
| Tickets | Fix#36565, replaces #36590
| License | MIT
| Doc PR | -
The current code doesn't make much sense: we check `hasSttyAvailable()`, and if the answer is `false`, we still use `stty` directly.
This PR relies on `stream_isatty` and equivalent fallback checks to decide if the password can be hidden or not.
Best reviewed [ignoring whitespaces](https://github.com/symfony/symfony/pull/37469/files?w=1).
Commits
-------
055b605e30 [Console] always use stty when possible to ask hidden questions
This PR was merged into the 4.4 branch.
Discussion
----------
[HttpClient] fix parsing response headers in CurlResponse
| Q | A
| ------------- | ---
| Branch? | 4.4
| Bug fix? | yes
| New feature? | no
| Deprecations? | no
| Tickets | -
| License | MIT
| Doc PR | -
This fixes rare notices that look like:
`Argument 2 passed to Symfony\Component\HttpClient\Chunk\DataChunk::__construct() must be of the type string, null given`
I'm unable to provide a test case since I'm unable to provide a simple reproducer.
It happens that curl can call the header-function with multiple lines at once apparently, while most of the time it does so with one at a time.
Commits
-------
eb70fb2f26 [HttpClient] fix parsing response headers in CurlResponse
This PR was squashed before being merged into the 5.2-dev branch.
Discussion
----------
[Workflow] Added Function (and Twig extension) to retrieve a specific transition
You can now easily retrieve a specific transition. Useful when you want the metadata of a specific transition.
| Q | A
| ------------- | ---
| Branch? | master
| Bug fix? | no
| New feature? | yes
| Deprecations? | no
| Tickets | hmmmm, should I create a ticket first ?
| License | MIT
| Doc PR | https://github.com/symfony/symfony-docs/pull/13907
I needed to get the metadata of a transition, but in order to do that, you need the transition object. So here is a PR to easily get the transition object
Commits
-------
0eebe74259 [Workflow] Added Function (and Twig extension) to retrieve a specific transition
This PR was merged into the 5.2-dev branch.
Discussion
----------
[HttpClient] Add MockResponse::getRequestMethod() and getRequestUrl() to allow inspecting which request has been sent
| Q | A
| ------------- | ---
| Branch? | 4.4
| Bug fix? | no
| New feature? | yes <!-- please update src/**/CHANGELOG.md files -->
| Deprecations? | no <!-- please update UPGRADE-*.md and src/**/CHANGELOG.md files -->
| Tickets | - <!-- prefix each issue number with "Fix #", if any -->
| License | MIT
| Doc PR | - <!-- 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/releases):
- Always add tests and ensure they pass.
- Never break backward compatibility (see https://symfony.com/bc).
- 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 master.
-->
---
As same as the `MockResponse` class has `getRequestOptions()` when doing a request; I've added:
* `getRequestUrl()` - returns the URL used when doing the request
* `getRequestMethod()` - returns the HTTP method used when doing the request
With these two getters, we would be able to assert that the method and URL passed to the `HttpClient` were well generated. I've tried to assert the URL generated in a unit test of a class with a `SymfonyHttpClient` injected and it wasn't possible. Calling `$mock->getInfo('url')` returns `null`.
Example, if we have a class with `HttpClientInterface` injected like this:
```php
final class SymfonyHttpUserClient
{
private HttpClientInterface $httpClient;
private string $baseUri;
public function __construct(
HttpClientInterface $httpClient,
string $baseUri
) {
$this->httpClient = $httpClient;
$this->baseUri = $baseUri;
}
public function getById(string $userId): void
{
$this->httpClient->request(
'GET',
$this->baseUri . $customerId
);
}
}
```
And if we want to do a unit test of `SymfonyHttpUserClient` right now we are not able to check if the URL of the request was well-formed passing a `MockResponse`. Also, we weren't able to assert the HTTP method (maybe this piece is not so critical to assert in the unit test).
```php
class SymfonyHttpUserClientTests extends TestCase
{
public function testGetById(): void
{
$baseUri = 'https://user-api.test/api/v1/users/';
$mockResponse = new MockResponse();
$symfonyHttpUserClient = new SymfonyHttpUserClient(
new MockHttpClient(
[$mockResponse],
$baseUri
),
$baseUri
);
// test get by id:
$symfonyHttpUserClient->getById('some-user-id');
// cannot be asserted right now:
$this->assertSame($mockResponse->getInfo('url'), $baseUri . 'some-user-id'); // fail
$this->assertSame($mockResponse->getInfo('http_method'), 'GET'); // fail
// it could be with the new getters:
$this->assertSame($mockResponse->getRequestUrl(), $baseUri . 'some-user-id');
$this->assertSame($mockResponse->getRequestMethod(), 'GET');
}
}
```
This only happens if the class has injected the HttpClient and if it is used inside void methods with no being able to return the response. If the class returns the response, `url` and `http_method` are available with `$response->getInfo()` call. But this response object is a new one, is not the mock passed by argument when you instance the MockHttpClient.
Var dumps of `getInfo` array:
```
$response->getInfo() from MockClient:
..array(10) {
["start_time"]=>
float(1587109014.7985)
["user_data"]=>
NULL
["http_code"]=>
int(200)
["response_headers"]=>
array(0) {
}
["error"]=>
NULL
["canceled"]=>
bool(false)
["redirect_count"]=>
int(0)
["redirect_url"]=>
NULL
["http_method"]=>
string(3) "GET"
["url"]=>
string(47) "https://user-api.test/api/v1/users/some-user-id"
}
$mock->getInfo()
array(4) {
["http_code"]=>
int(200)
["response_headers"]=>
array(0) {
}
["error"]=>
NULL
["canceled"]=>
bool(false)
}
```
This is a minor change, I opened the PR with `4.4` as base branch; but not sure if it should be opened with `5.0` as base branch or even `master` taking account is a feature (add two new getters in MockResponse class). Let me know if I didn't follow right the instructions (first PR on Symfony project 😃)
Thanks!
Commits
-------
7a250d80c1 [HttpClient] Add MockResponse::getRequestMethod() and getRequestUrl() to allow inspecting which request has been sent
This PR was merged into the 5.2-dev branch.
Discussion
----------
Move event alias mappings to their components
| Q | A
| ------------- | ---
| Branch? | master
| Bug fix? | no
| New feature? | no
| Deprecations? | no
| Tickets | N/A
| License | MIT
| Doc PR | N/A
For a small console tool, I wanted to use the Console, EventDispatcher and DI components standalone. In order to make use of aliased events, I had to copy over the mapping information from FrameworkBundle, which was a bit unfortunate.
With this PR, I'd like to suggest to store all alias mappings on the `*Events` classes of each component instead. This change will make the mapping reusable outside of the full-stack framework.
Note: I've bumped the dependencies of FrameworkBundle/SercurityBundle in order to gain access to the moved mapping information. If any of the bumps is too heavy, please tell me and I'll implement a fallback instead.
Commits
-------
28e6f6f72c Move event alias mappings to their components.