symfony/.git
11 months agoMerge pull request #35539 from fabpot/release-4.3.11 4.3 origin/4.3 v4.3.11
Fabien Potencier [Fri, 31 Jan 2020 12:38:52 +0000 (13:38 +0100)]
Merge pull request #35539 from fabpot/release-4.3.11

released v4.3.11

11 months agoupdated VERSION for 4.3.11
Fabien Potencier [Fri, 31 Jan 2020 12:38:38 +0000 (13:38 +0100)]
updated VERSION for 4.3.11

11 months agoupdated CHANGELOG for 4.3.11
Fabien Potencier [Fri, 31 Jan 2020 12:38:19 +0000 (13:38 +0100)]
updated CHANGELOG for 4.3.11

11 months agoMerge branch '3.4' into 4.3
Nicolas Grekas [Fri, 31 Jan 2020 09:56:21 +0000 (10:56 +0100)]
Merge branch '3.4' into 4.3

* 3.4:
  Bump phpunit-bridge cache

11 months agoBump phpunit-bridge cache
Nicolas Grekas [Fri, 31 Jan 2020 09:55:33 +0000 (10:55 +0100)]
Bump phpunit-bridge cache

11 months agoMerge branch '3.4' into 4.3
Nicolas Grekas [Fri, 31 Jan 2020 09:10:37 +0000 (10:10 +0100)]
Merge branch '3.4' into 4.3

* 3.4:
  [Validator] fix access to uninitialized property when getting value
  [HttpKernel] Fix stale-if-error behavior, add tests
  Improved error message when no supported user provider is found

11 months agobug #35530 [HttpClient] Fix regex bearer (noniagriconomie)
Nicolas Grekas [Fri, 31 Jan 2020 09:08:14 +0000 (10:08 +0100)]
bug #35530 [HttpClient] Fix regex bearer (noniagriconomie)

This PR was merged into the 4.3 branch.

Discussion
----------

[HttpClient] Fix regex bearer

| Q             | A
| ------------- | ---
| Branch?       | 4.3
| Bug fix?      | yes
| New feature?  | no
| Deprecations? | no
| Tickets       | Fix #34919
| License       | MIT
| Doc PR        | none

Small fix related to https://github.com/symfony/symfony/issues/34919#issuecomment-571145590

Commits
-------

cd0db78ab5 [HttpClient] Fix regex bearer

11 months agobug #35532 [Validator] fix access to uninitialized property when getting value (greed...
Fabien Potencier [Fri, 31 Jan 2020 08:56:30 +0000 (09:56 +0100)]
bug #35532 [Validator] fix access to uninitialized property when getting value (greedyivan)

This PR was squashed before being merged into the 3.4 branch (closes #35532).

Discussion
----------

[Validator] fix access to uninitialized property when getting value

| Q             | A
| ------------- | ---
| Branch?       | 3.4
| Bug fix?      | yes
| New feature?  | no
| Deprecations? | no
| Tickets       | Fix #35454
| License       | MIT
| Doc PR        |

In PHP 7.4, the type-hinted property is [uninitialized](https://wiki.php.net/rfc/typed_properties_v2#uninitialized_and_unset_properties) by default. So it needs to be checked before use.

Commits
-------

1edecf77c1 [Validator] fix access to uninitialized property when getting value

11 months ago[Validator] fix access to uninitialized property when getting value
Ivan Grigoriev [Thu, 30 Jan 2020 21:43:04 +0000 (00:43 +0300)]
[Validator] fix access to uninitialized property when getting value

11 months ago[HttpClient] Fix regex bearer
noniagriconomie [Thu, 30 Jan 2020 16:47:09 +0000 (17:47 +0100)]
[HttpClient] Fix regex bearer

11 months agobug #35305 [HttpKernel] Fix stale-if-error behavior, add tests (mpdude)
Fabien Potencier [Thu, 30 Jan 2020 16:05:58 +0000 (17:05 +0100)]
bug #35305 [HttpKernel] Fix stale-if-error behavior, add tests (mpdude)

This PR was squashed before being merged into the 3.4 branch (closes #35305).

Discussion
----------

[HttpKernel] Fix stale-if-error behavior, add tests

| Q             | A
| ------------- | ---
| Branch?       | 3.4
| Bug fix?      | yes
| New feature?  | no
| Deprecations? | no
| Tickets       | #24248
| License       | MIT
| Doc PR        |

This PR adds the first tests for `stale-if-error` logic in `HttpCache`.

It also fixes an observation from #24248: For responses that have been cached as `public` with an `ETag` but without a lifetime, in case of an error the stale response will be served forever (= as long as the error persists), even beyond the configured `stale-if-error` grace period.

Furthermore, it tries to improve compliance with RFC 7234: Stale responses must not be sent (under no condition) if one of
* `no-cache`
* `must-revalidate`
* `proxy-revalidate` or
* `s-maxage` (sic) is present.

This can be found in the corresponding chapters of Section 5.2.2 for these directives, but is also summarized in [Section 4.2.4](https://tools.ietf.org/html/rfc7234#section-4.2.4) as

 > A cache MUST NOT generate a stale response if it is prohibited by an explicit in-protocol directive (e.g., by a "no-store" or "no-cache" cache directive, a "must-revalidate" cache-response-directive, or an applicable "s-maxage" or "proxy-revalidate" cache-response-directive; see Section 5.2.2).

Because disabling of `stale-if-error` for `s-maxage` responses probably has a big impact on the usefulness of that feature in practice, it has to be enabled explicitly with a new config setting `strict_smaxage` (defaulting to `false`).

Commits
-------

ad5f427bed [HttpKernel] Fix stale-if-error behavior, add tests

11 months ago[HttpKernel] Fix stale-if-error behavior, add tests
Matthias Pigulla [Fri, 10 Jan 2020 23:18:38 +0000 (23:18 +0000)]
[HttpKernel] Fix stale-if-error behavior, add tests

11 months agobug #34808 [PhpUnitBridge] Properly handle phpunit arguments for configuration file...
Fabien Potencier [Thu, 30 Jan 2020 15:35:02 +0000 (16:35 +0100)]
bug #34808 [PhpUnitBridge] Properly handle phpunit arguments for configuration file (biozshock)

This PR was merged into the 4.3 branch.

Discussion
----------

[PhpUnitBridge] Properly handle phpunit arguments for configuration file

| Q             | A
| ------------- | ---
| Branch?       | 4.3
| Bug fix?      | yes
| New feature?  | no
| Deprecations? | no
| License       | MIT

PhpUnitBridge should properly parse cli arguments for configuration file.
After fixing #34300 the PhpUnitBridge stopped recognizing short `-c` option.
Also original PHPUnit allows to pass a directory as configuration parameter and read from either phpunit.xml or phpunit.xml.dist

Commits
-------

a7a5885661 Properly handle phpunit arguments for configuration file

11 months agobug #35480 [Messenger] Check for all serialization exceptions during message dec...
Tobias Schultze [Wed, 29 Jan 2020 14:30:24 +0000 (15:30 +0100)]
bug #35480 [Messenger] Check for all serialization exceptions during message dec… (Patrick Berenschot)

This PR was merged into the 4.3 branch.

Discussion
----------

[Messenger] Check for all serialization exceptions during message dec…

| Q             | A
| ------------- | ---
| Branch?       | 4.4
| Bug fix?      | yes
| New feature?  | no
| Deprecations? | no
| Tickets       | Fix #35446
| License       | MIT

Makes it so that the Messenger Serializer throws a `MessageDecodingFailedException` for any serializer exception.

Commits
-------

21fffcadd5 [Messenger] Check for all serialization exceptions during message dec…

11 months ago[Messenger] Check for all serialization exceptions during message dec…
Patrick Berenschot [Mon, 27 Jan 2020 11:46:05 +0000 (11:46 +0000)]
[Messenger] Check for all serialization exceptions during message dec…

11 months agominor #35487 [Messenger] Fix messenger argument (jderusse)
Robin Chalas [Mon, 27 Jan 2020 22:28:36 +0000 (23:28 +0100)]
minor #35487 [Messenger] Fix messenger argument (jderusse)

This PR was merged into the 4.3 branch.

Discussion
----------

[Messenger] Fix messenger argument

| Q             | A
| ------------- | ---
| Branch?       | 4.3
| Bug fix?      | yes
| New feature?  | no
| Deprecations? | no
| Tickets       | -
| License       | MIT
| Doc PR        | -

`Connection::DEFAULT_OPTIONS` is an key/value array. The exception `Unknown option found : [%s]. Allowed options are [%s]` should use allowed option's key instead of the value.

Commits
-------

ae0c6344b4 Fix exception message in Doctrine Messenger

11 months agoFix exception message in Doctrine Messenger
Jérémy Derussé [Mon, 27 Jan 2020 18:01:48 +0000 (19:01 +0100)]
Fix exception message in Doctrine Messenger

11 months agominor #35375 [Lock] Don't allow mysqli to be used as it doesn't work (duncan3dc)
Nicolas Grekas [Mon, 27 Jan 2020 10:29:37 +0000 (11:29 +0100)]
minor #35375 [Lock] Don't allow mysqli to be used as it doesn't work (duncan3dc)

This PR was merged into the 4.3 branch.

Discussion
----------

[Lock] Don't allow mysqli to be used as it doesn't work

| Q             | A
| ------------- | ---
| Branch?       | 4.3
| Bug fix?      | yes
| New feature?  | no
| Deprecations? | no
| Tickets       |
| License       | MIT
| Doc PR        |

Mysqli doesn't support named parameters, so if you pass a doctrine connection using `mysqli` then you get the following error:
`You have an error in your SQL syntax; check the manual that corresponds to your MySQL server version for the right syntax to use near ':id, :token, UNIX_TIMESTAMP() + 300)'`
This PR ensures a clear error is provided and suggests to use `pdo_mysql` instead

Commits
-------

ef3bcda5e3 Mysqli doesn't support the named parameters used by PdoStore

11 months agominor #35472 [Security] Improved error message when no supported user provider is...
Nicolas Grekas [Mon, 27 Jan 2020 09:47:05 +0000 (10:47 +0100)]
minor #35472 [Security] Improved error message when no supported user provider is found (wouterj)

This PR was submitted for the 4.4 branch but it was merged into the 3.4 branch instead.

Discussion
----------

[Security] Improved error message when no supported user provider is found

| Q             | A
| ------------- | ---
| Branch?       | 4.4
| Bug fix?      | no
| New feature?  | no
| Deprecations? | no
| Tickets       | Fix #35435
| License       | MIT
| Doc PR        | -

Commits
-------

6b2db6dc30 Improved error message when no supported user provider is found

11 months agoImproved error message when no supported user provider is found
Wouter de Jong [Sat, 25 Jan 2020 12:51:20 +0000 (13:51 +0100)]
Improved error message when no supported user provider is found

11 months agominor #35473 [Cache] Don't allow mysqli to be used as it doesn't work (duncan3dc)
Nicolas Grekas [Mon, 27 Jan 2020 09:22:54 +0000 (10:22 +0100)]
minor #35473 [Cache] Don't allow mysqli to be used as it doesn't work (duncan3dc)

This PR was merged into the 4.3 branch.

Discussion
----------

[Cache] Don't allow mysqli to be used as it doesn't work

| Q             | A
| ------------- | ---
| Branch?       | 4.3
| Bug fix?      | yes
| New feature?  | no
| Deprecations? | no
| Tickets       |
| License       | MIT
| Doc PR        |

As discussed in #35375 Mysqli doesn't support named parameters, so if you pass a doctrine connection using `mysqli` then you get errors for any queries using named parameters.
This PR ensures a clear error is provided and suggests to use `pdo_mysql` instead

Commits
-------

a7864489b0 Mysqli doesn't support the named parameters used by PdoAdapter

11 months agoMysqli doesn't support the named parameters used by PdoAdapter
Craig Duncan [Sat, 25 Jan 2020 13:06:31 +0000 (13:06 +0000)]
Mysqli doesn't support the named parameters used by PdoAdapter

12 months agominor #35471 Fix displaying anonymous classes on PHP >= 7.4.2 (nicolas-grekas)
Nicolas Grekas [Sat, 25 Jan 2020 12:38:50 +0000 (13:38 +0100)]
minor #35471 Fix displaying anonymous classes on PHP >= 7.4.2 (nicolas-grekas)

This PR was merged into the 4.3 branch.

Discussion
----------

Fix displaying anonymous classes on PHP >= 7.4.2

| Q             | A
| ------------- | ---
| Branch?       | 4.3
| Bug fix?      | yes
| New feature?  | no
| Deprecations? | no
| Tickets       | -
| License       | MIT
| Doc PR        | -

Commits
-------

dd94b386a9 Fix displaying anonymous classes on PHP >= 7.4.2

12 months agoFix displaying anonymous classes on PHP >= 7.4.2
Nicolas Grekas [Sat, 25 Jan 2020 12:32:28 +0000 (13:32 +0100)]
Fix displaying anonymous classes on PHP >= 7.4.2

12 months agominor #35396 [DoctrineBridge] [DX] Improve condition for exception text in ManagerReg...
Nicolas Grekas [Thu, 23 Jan 2020 10:55:20 +0000 (11:55 +0100)]
minor #35396 [DoctrineBridge] [DX] Improve condition for exception text in ManagerRegistry to avoid confusion (guilliamxavier)

This PR was submitted for the 4.4 branch but it was merged into the 4.3 branch instead.

Discussion
----------

[DoctrineBridge] [DX] Improve condition for exception text in ManagerRegistry to avoid confusion

| Q             | A
| ------------- | ---
| Branch?       | 4.4
| Bug fix?      | yesish
| New feature?  | no
| Deprecations? | no
| Tickets       | Fix https://github.com/symfony/symfony/pull/31047#issuecomment-575548372
| License       | MIT
| Doc PR        | -

(Targetting 4.4 because 4.3 is already almost EOM)

Commits
-------

0d47fdfb49 [DoctrineBridge] [DX] Improve condition for exception text in ManagerRegistry to avoid confusion

12 months ago[DoctrineBridge] [DX] Improve condition for exception text in ManagerRegistry to...
Guilliam Xavier [Mon, 20 Jan 2020 09:07:41 +0000 (10:07 +0100)]
[DoctrineBridge] [DX] Improve condition for exception text in ManagerRegistry to avoid confusion

12 months agoMerge branch '3.4' into 4.3
Nicolas Grekas [Thu, 23 Jan 2020 10:52:31 +0000 (11:52 +0100)]
Merge branch '3.4' into 4.3

* 3.4:
  Fix testing with mongodb
  bumped Symfony version to 3.4.38
  updated VERSION for 3.4.37
  update CONTRIBUTORS for 3.4.37
  updated CHANGELOG for 3.4.37

12 months agominor #35450 Fix testing with mongodb (nicolas-grekas)
Nicolas Grekas [Thu, 23 Jan 2020 10:33:58 +0000 (11:33 +0100)]
minor #35450 Fix testing with mongodb (nicolas-grekas)

This PR was merged into the 3.4 branch.

Discussion
----------

Fix testing with mongodb

| Q             | A
| ------------- | ---
| Branch?       | 3.4
| Bug fix?      | no
| New feature?  | no
| Deprecations? | no
| Tickets       | -
| License       | MIT
| Doc PR        | -

Commits
-------

28cd964ac9 Fix testing with mongodb

12 months agoFix testing with mongodb
Nicolas Grekas [Thu, 23 Jan 2020 10:22:20 +0000 (11:22 +0100)]
Fix testing with mongodb

12 months agobug #35428 [Cache] fix checking for igbinary availability (nicolas-grekas)
Fabien Potencier [Wed, 22 Jan 2020 09:38:05 +0000 (10:38 +0100)]
bug #35428 [Cache] fix checking for igbinary availability (nicolas-grekas)

This PR was merged into the 4.3 branch.

Discussion
----------

[Cache] fix checking for igbinary availability

| Q             | A
| ------------- | ---
| Branch?       | 4.3
| Bug fix?      | yes
| New feature?  | no
| Deprecations? | no
| Tickets       | -
| License       | MIT
| Doc PR        | -

As reported by @andrerom in https://github.com/symfony/symfony/commit/a2c924d77220011c5f5939579677d2a82b69a43d#r36427445

Commits
-------

09818e99ac [Cache] fix checking for igbinary availability

12 months agominor #35432 [4.3] Minor Travis configuration cosmetic patch (romainneutron)
Fabien Potencier [Wed, 22 Jan 2020 09:36:15 +0000 (10:36 +0100)]
minor #35432 [4.3] Minor Travis configuration cosmetic patch (romainneutron)

This PR was merged into the 4.3 branch.

Discussion
----------

[4.3] Minor Travis configuration cosmetic patch

| Q             | A
| ------------- | ---
| Branch?       | 4.3 (Does not appear in previous versions)
| Bug fix?      | no
| New feature?  | no
| Deprecations? | no

The Travis configuration is organized in steps: each step begins with a comment followed by the commands to run.
The ` # List all PHP extensions with versions` is empty, its content has been put in the next section, without comment title.

This patch fixes this minor issue

Commits
-------

3918f23307 Minor Travis cosmetic patch

12 months agoMinor Travis cosmetic patch
Romain Neutron [Wed, 22 Jan 2020 09:20:58 +0000 (10:20 +0100)]
Minor Travis cosmetic patch

12 months ago[Cache] fix checking for igbinary availability
Nicolas Grekas [Wed, 22 Jan 2020 07:15:02 +0000 (08:15 +0100)]
[Cache] fix checking for igbinary availability

12 months agobumped Symfony version to 4.3.11
Fabien Potencier [Tue, 21 Jan 2020 13:21:51 +0000 (14:21 +0100)]
bumped Symfony version to 4.3.11

12 months agoMerge pull request #35412 from fabpot/release-4.3.10 v4.3.10
Fabien Potencier [Tue, 21 Jan 2020 13:14:01 +0000 (14:14 +0100)]
Merge pull request #35412 from fabpot/release-4.3.10

released v4.3.10

12 months agoupdated VERSION for 4.3.10
Fabien Potencier [Tue, 21 Jan 2020 13:13:44 +0000 (14:13 +0100)]
updated VERSION for 4.3.10

12 months agoupdated CHANGELOG for 4.3.10
Fabien Potencier [Tue, 21 Jan 2020 13:13:32 +0000 (14:13 +0100)]
updated CHANGELOG for 4.3.10

12 months agobumped Symfony version to 3.4.38
Fabien Potencier [Tue, 21 Jan 2020 13:12:52 +0000 (14:12 +0100)]
bumped Symfony version to 3.4.38

12 months agoMerge pull request #35411 from fabpot/release-3.4.37 v3.4.37
Fabien Potencier [Tue, 21 Jan 2020 12:30:09 +0000 (13:30 +0100)]
Merge pull request #35411 from fabpot/release-3.4.37

released v3.4.37

12 months agoupdated VERSION for 3.4.37
Fabien Potencier [Tue, 21 Jan 2020 12:29:51 +0000 (13:29 +0100)]
updated VERSION for 3.4.37

12 months agoupdate CONTRIBUTORS for 3.4.37
Fabien Potencier [Tue, 21 Jan 2020 12:29:48 +0000 (13:29 +0100)]
update CONTRIBUTORS for 3.4.37

12 months agoupdated CHANGELOG for 3.4.37
Fabien Potencier [Tue, 21 Jan 2020 12:29:39 +0000 (13:29 +0100)]
updated CHANGELOG for 3.4.37

12 months agobug #35364 [Yaml] Throw on unquoted exclamation mark (fancyweb)
Nicolas Grekas [Tue, 21 Jan 2020 11:09:03 +0000 (12:09 +0100)]
bug #35364 [Yaml] Throw on unquoted exclamation mark (fancyweb)

This PR was merged into the 4.3 branch.

Discussion
----------

[Yaml] Throw on unquoted exclamation mark

| Q             | A
| ------------- | ---
| Branch?       | 4.3
| Bug fix?      | yes
| New feature?  | no
| Deprecations? |
| Tickets       | https://github.com/symfony/symfony/issues/35344
| License       | MIT
| Doc PR        | -

Commits
-------

6b4147c991 [Yaml] Throw on unquoted exclamation mark

12 months agoMerge branch '3.4' into 4.3
Nicolas Grekas [Tue, 21 Jan 2020 11:08:18 +0000 (12:08 +0100)]
Merge branch '3.4' into 4.3

* 3.4:
  Use supportsClass where possible

12 months agobug #35065 [Security] Use supportsClass in addition to UnsupportedUserException ...
Nicolas Grekas [Tue, 21 Jan 2020 11:02:57 +0000 (12:02 +0100)]
bug #35065 [Security] Use supportsClass in addition to UnsupportedUserException (linaori)

This PR was merged into the 3.4 branch.

Discussion
----------

[Security] Use supportsClass in addition to UnsupportedUserException

| Q             | A
| ------------- | ---
| Branch?       | 3.4+
| Bug fix?      | yes
| New feature?  | no
| Deprecations? | no
| Tickets       | Fix #35045
| License       | MIT
| Doc PR        | ~

This PR fixes the issue where user providers rely on just the UnsupportedUserException from `refreshUser()`, causing a flow where users are wrongfully re-authenticated.

There's one issue where `refreshUser()` can do far more sophisticated checks on the user class, which it will never reach if the class is not supported. As far as I know it was never intended to support instances that are rejected by `supportsClass()`, though people could've implemented this (by accident). So the question is more if we should add a BC layer for this; for example:

```php
try {
    $refreshedUser = $provider->refreshUser($user);
    $newToken = clone $token;
    $newToken->setUser($refreshedUser);

    if (!$provider->supportsClass($userClass)) {
        if ($this->shouldCheckSupportsClass) {
            continue;
        }
        // have to think of a proper deprecation here for 6.0
        @trigger_error('Provider %s does not support user class %s via supportsClass() while it does support it via refreshUser .. please set option X and fix %s::supportsUser() ', E_USER_DEPRECATED);
    }
```
This would prevent behavior from breaking but also means we can't fix this on anything less than 5.1.

Commits
-------

d3942cbe17 Use supportsClass where possible

12 months agominor #35409 [Messenger] Add missing documentation link in README (pyrech)
Nicolas Grekas [Tue, 21 Jan 2020 10:11:20 +0000 (11:11 +0100)]
minor #35409 [Messenger] Add missing documentation link in README (pyrech)

This PR was submitted for the 4.4 branch but it was merged into the 4.3 branch instead.

Discussion
----------

[Messenger] Add missing documentation link in README

| Q             | A
| ------------- | ---
| Branch?       | 4.4 (lower was experimental)
| Bug fix?      | no
| New feature?  | no
| Deprecations? | no
| Tickets       | -
| License       | MIT
| Doc PR        | -

Commits
-------

5714eaa0de Add link to messenger documentation in its README

12 months agoAdd link to messenger documentation in its README
Loïck Piera [Tue, 21 Jan 2020 08:22:22 +0000 (09:22 +0100)]
Add link to messenger documentation in its README

12 months ago[HttpKernel] restore compat with clock mocking
Nicolas Grekas [Tue, 21 Jan 2020 10:03:13 +0000 (11:03 +0100)]
[HttpKernel] restore compat with clock mocking

12 months agoMerge branch '3.4' into 4.3
Fabien Potencier [Tue, 21 Jan 2020 08:20:29 +0000 (09:20 +0100)]
Merge branch '3.4' into 4.3

* 3.4:
  chown and chgrp should also accept int as owner and group
  Fix RememberMe with null password
  [Validator] Fix plurals for sr_Latn (Serbian language written in latin script) validation messages
  [PhpUnitBridge][SymfonyTestsListenerTrait] Remove some unneeded code
  fix PHP const mapping keys using the inline notation
  Fix that no-cache requires positive validation with the origin, even for fresh responses

12 months agominor #35178 [PhpUnitBridge][SymfonyTestsListenerTrait] Remove some unneeded code...
Nicolas Grekas [Mon, 20 Jan 2020 12:28:41 +0000 (13:28 +0100)]
minor #35178 [PhpUnitBridge][SymfonyTestsListenerTrait] Remove some unneeded code (fancyweb)

This PR was merged into the 3.4 branch.

Discussion
----------

[PhpUnitBridge][SymfonyTestsListenerTrait] Remove some unneeded code

| Q             | A
| ------------- | ---
| Branch?       | 3.4
| Bug fix?      | no
| New feature?  | no
| Deprecations? | no
| Tickets       | -
| License       | MIT
| Doc PR        | -

Risky errors when there are no assertions are added before the test end listeners are called (ie, before the code in endTest is executed) so forcing beStrictAboutTestsThatDoNotTestAnything to false when there is a expectedDeprecation annotation is enough.

If the goal is to reset the value to the original value, then I think we should not do it since we basically "lie" to the next listeners. Let's assume that when a test expect a deprecation, it can have 0 assertions. Also this flag is not used anymore by PHPUnit after we reset it.

Ref https://github.com/symfony/symfony/pull/21786 btw

Commits
-------

fb48bbc05b [PhpUnitBridge][SymfonyTestsListenerTrait] Remove some unneeded code

12 months agoMysqli doesn't support the named parameters used by PdoStore
Craig Duncan [Fri, 17 Jan 2020 17:21:39 +0000 (17:21 +0000)]
Mysqli doesn't support the named parameters used by PdoStore

12 months ago[Yaml] Throw on unquoted exclamation mark
Thomas Calvet [Thu, 16 Jan 2020 14:27:16 +0000 (15:27 +0100)]
[Yaml] Throw on unquoted exclamation mark

12 months agominor #35369 [Filesystem] chown and chgrp should also accept int as owner and group...
Nicolas Grekas [Fri, 17 Jan 2020 11:07:25 +0000 (12:07 +0100)]
minor #35369 [Filesystem] chown and chgrp should also accept int as owner and group (3.4) (Slamdunk)

This PR was merged into the 3.4 branch.

Discussion
----------

[Filesystem] chown and chgrp should also accept int as owner and group (3.4)

| Q             | A
| ------------- | ---
| Branch?       | 3.4
| Bug fix?      | no
| New feature?  | no
| Deprecations? | no
| Tickets       |
| License       | MIT
| Doc PR        |

Reference: https://github.com/symfony/symfony/pull/35356#issuecomment-575526299

Commits
-------

6b811e6b4c chown and chgrp should also accept int as owner and group

12 months agochown and chgrp should also accept int as owner and group
Filippo Tessarotto [Thu, 16 Jan 2020 09:18:18 +0000 (10:18 +0100)]
chown and chgrp should also accept int as owner and group

12 months agominor #35340 [Validator] Fix plurals for sr_Latn validation messages (knezmilos13)
Fabien Potencier [Thu, 16 Jan 2020 09:23:40 +0000 (10:23 +0100)]
minor #35340 [Validator] Fix plurals for sr_Latn validation messages (knezmilos13)

This PR was merged into the 3.4 branch.

Discussion
----------

[Validator] Fix plurals for sr_Latn validation messages

| Q             | A
| ------------- | ---
| Branch?       | 3.4
| Bug fix?      | yes
| New feature?  | no
| Deprecations? | no
| Tickets       | Fix #35276
| License       | MIT

validators.sr_Latn.xlf (Serbian, written with latin script) has wrong plurals for all validation message translations that require them (only two where there should be three). This commit fixes that by adding the missing third plural-translation.

Commits
-------

207cdafd54 [Validator] Fix plurals for sr_Latn (Serbian language written in latin script) validation messages

12 months agobug #35343 [Security] Fix RememberMe with null password (jderusse)
Robin Chalas [Thu, 16 Jan 2020 00:14:41 +0000 (01:14 +0100)]
bug #35343 [Security] Fix RememberMe with null password (jderusse)

This PR was merged into the 3.4 branch.

Discussion
----------

[Security] Fix RememberMe with null password

| Q             | A
| ------------- | ---
| Branch?       | 3.4
| Bug fix?      | yes
| New feature?  | no
| Deprecations? | yes
| Tickets       | NA
| License       | MIT
| Doc PR        | NA

From `UserInterface` the method getPassword may return null, while generateCookieHash requires a string.
This PR changes the signature of the methods to allows null password

Similar to #35335 for branch 3.4

Commits
-------

820eb357c8 Fix RememberMe with null password

12 months agoProperly handle phpunit arguments for configuration file
Artem Lopata [Wed, 4 Dec 2019 10:31:52 +0000 (11:31 +0100)]
Properly handle phpunit arguments for configuration file

12 months agoFix RememberMe with null password
Jérémy Derussé [Tue, 14 Jan 2020 21:28:32 +0000 (22:28 +0100)]
Fix RememberMe with null password

12 months ago[Validator] Fix plurals for sr_Latn (Serbian language written in latin script) valida...
knezmilos13 [Tue, 14 Jan 2020 18:27:07 +0000 (19:27 +0100)]
[Validator] Fix plurals for sr_Latn (Serbian language written in latin script) validation messages

12 months agobug #34223 [DI] Suggest typed argument when binding fails with untyped argument ...
Nicolas Grekas [Tue, 14 Jan 2020 16:43:06 +0000 (17:43 +0100)]
bug #34223 [DI] Suggest typed argument when binding fails with untyped argument (gudfar)

This PR was merged into the 4.3 branch.

Discussion
----------

[DI] Suggest typed argument when binding fails with untyped argument

| Q             | A
| ------------- | ---
| Branch?       |  4.3
| Bug fix?      | yes
| New feature?  | no
| Deprecations? | no
| Tickets       |  #33470
| License       | MIT

I've added a condition that looks for arguments and if the typehint doesn’t match, throws an `InvalidArgumentException`

Commits
-------

0e92399daa [DI] Suggest typed argument when binding fails with untyped argument

12 months agominor #35334 [FrameworkBundle] remove messenger cache if not enabled (dmaicher)
Nicolas Grekas [Tue, 14 Jan 2020 16:22:33 +0000 (17:22 +0100)]
minor #35334 [FrameworkBundle] remove messenger cache if not enabled (dmaicher)

This PR was merged into the 4.3 branch.

Discussion
----------

[FrameworkBundle] remove messenger cache if not enabled

| Q             | A
| ------------- | ---
| Branch?       | 4.3
| Bug fix?      | yes/no
| New feature?  | no
| Deprecations? | no
| Tickets       | -
| License       | MIT
| Doc PR        | -
<!--
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/roadmap):
 - 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.
-->

I noticed on one of my apps that I have a cache pool related to the symfony messenger integration although I'm not using the messenger at all.

```
bin/console debug:container cache.messenger.restart_workers_signal

Information for Service "cache.messenger.restart_workers_signal"
================================================================

 An adapter that collects data about all cache calls.

 ---------------- --------------------------------------------------
  Option           Value
 ---------------- --------------------------------------------------
  Service ID       cache.messenger.restart_workers_signal
  Class            Symfony\Component\Cache\Adapter\TraceableAdapter
  Tags             cache.pool
                   kernel.reset (method: reset)
  Public           no
  Synthetic        no
  Lazy             no
  Shared           yes
  Abstract         no
  Autowired        no
  Autoconfigured   no
 ---------------- -----------------------------------------
```

So this PR removes the definition of the service in case the messenger integration is disabled.

Commits
-------

f81161df8d [FrameworkBundle] remove messenger cache if not enabled

12 months ago[FrameworkBundle] remove messenger cache if not enabled
David Maicher [Tue, 14 Jan 2020 11:21:57 +0000 (12:21 +0100)]
[FrameworkBundle] remove messenger cache if not enabled

12 months ago[PhpUnitBridge][SymfonyTestsListenerTrait] Remove some unneeded code
Thomas Calvet [Thu, 2 Jan 2020 18:17:31 +0000 (19:17 +0100)]
[PhpUnitBridge][SymfonyTestsListenerTrait] Remove some unneeded code

12 months agobug #35324 [HttpClient] Fix strict parsing of response status codes (Armando-Walmeric)
Nicolas Grekas [Mon, 13 Jan 2020 17:23:15 +0000 (18:23 +0100)]
bug #35324 [HttpClient] Fix strict parsing of response status codes (Armando-Walmeric)

This PR was submitted for the master branch but it was squashed and merged into the 4.3 branch instead.

Discussion
----------

[HttpClient] Fix strict parsing of response status codes

| Q             | A
| ------------- | ---
| Branch?       | 5.0
| Bug fix?      | yes
| New feature?  | no
| Deprecations? | no
| Tickets       | No ticket opened
| License       | MIT
| Doc PR        | No new feature

If you wait a strict Http code response by the server were you request, but the server response isn't strict and only return HTTP/1.1 200 we have a problem without sense to be strict on regular expression.

Commits
-------

4aa953600f [HttpClient] Fix strict parsing of response status codes

12 months ago[HttpClient] Fix strict parsing of response status codes
Armando [Mon, 13 Jan 2020 15:59:16 +0000 (16:59 +0100)]
[HttpClient] Fix strict parsing of response status codes

12 months agobug #35318 [Yaml] fix PHP const mapping keys using the inline notation (xabbuh)
Fabien Potencier [Mon, 13 Jan 2020 13:54:16 +0000 (14:54 +0100)]
bug #35318 [Yaml] fix PHP const mapping keys using the inline notation (xabbuh)

This PR was merged into the 3.4 branch.

Discussion
----------

[Yaml] fix PHP const mapping keys using the inline notation

| Q             | A
| ------------- | ---
| Branch?       | 3.4
| Bug fix?      | yes
| New feature?  | no
| Deprecations? | no
| Tickets       | Fix #35179
| License       | MIT
| Doc PR        |

Commits
-------

45461c73bf fix PHP const mapping keys using the inline notation

12 months agofix PHP const mapping keys using the inline notation
Christian Flothmann [Mon, 13 Jan 2020 08:00:59 +0000 (09:00 +0100)]
fix PHP const mapping keys using the inline notation

12 months agobug #35304 [HttpKernel] Fix that no-cache MUST revalidate with the origin (mpdude)
Fabien Potencier [Sat, 11 Jan 2020 07:35:43 +0000 (08:35 +0100)]
bug #35304 [HttpKernel] Fix that no-cache MUST revalidate with the origin (mpdude)

This PR was merged into the 3.4 branch.

Discussion
----------

[HttpKernel] Fix that no-cache MUST revalidate with the origin

| Q             | A
| ------------- | ---
| Branch?       | 3.4
| Bug fix?      | yes
| New feature?  | no
| Deprecations? | no
| Tickets       |
| License       | MIT
| Doc PR        |

From [RFC 7234 Section 5.2.2](https://tools.ietf.org/html/rfc7234#section-5.2.2)

> The "no-cache" response directive indicates that the response MUST NOT be used to satisfy a subsequent request without successful validation on the origin server.  This allows an origin server to prevent a cache from using it to satisfy a request without contacting it, even by caches that have been configured to send stale responses.

This is unconditional – the response must be revalidated right away.

(`must-revalidate`, to the contrary, requires revalidation only once the response has become stale.)

Commits
-------

c8bdcb3408 Fix that no-cache requires positive validation with the origin, even for fresh responses

12 months agoFix that no-cache requires positive validation with the origin, even for fresh responses
Matthias Pigulla [Fri, 10 Jan 2020 22:32:00 +0000 (22:32 +0000)]
Fix that no-cache requires positive validation with the origin, even for fresh responses

12 months agoMerge branch '3.4' into 4.3
Robin Chalas [Fri, 10 Jan 2020 21:48:14 +0000 (22:48 +0100)]
Merge branch '3.4' into 4.3

* 3.4:
  Avoid stale-if-error if kernel.debug = true, because it hides errors
  [Console] Fix SymfonyQuestionHelper tests sometimes failing on AppVeyor
  [DI] deferred exceptions in ResolveParameterPlaceHoldersPass

12 months agominor #35289 [Workflow] Fix configuration node reference for "initial_marking" (phansys)
Fabien Potencier [Fri, 10 Jan 2020 11:35:58 +0000 (12:35 +0100)]
minor #35289 [Workflow] Fix configuration node reference for "initial_marking" (phansys)

This PR was merged into the 4.3 branch.

Discussion
----------

[Workflow] Fix configuration node reference for "initial_marking"

| Q             | A
| ------------- | ---
| Branch?       | 4.3
| Bug fix?      | no
| New feature?  | no
| Deprecations? | no
| Related to       | #30890
| License       | MIT

Commits
-------

452f92540b [Workflow] Fix configuration node reference for "initial_marking"

12 months agobug #35299 Avoid `stale-if-error` in FrameworkBundle's HttpCache if kernel.debug...
Fabien Potencier [Fri, 10 Jan 2020 11:27:06 +0000 (12:27 +0100)]
bug #35299 Avoid `stale-if-error` in FrameworkBundle's HttpCache if kernel.debug = true (mpdude)

This PR was merged into the 3.4 branch.

Discussion
----------

Avoid `stale-if-error` in FrameworkBundle's HttpCache if kernel.debug = true

| Q             | A
| ------------- | ---
| Branch?       | 3.4
| Bug fix?      | yes
| New feature?  | no
| Deprecations? | no
| Tickets       | Fix #24248 (maybe?)
| License       | MIT
| Doc PR        |

When working with the `HttpCache` in development, error messages may not become visible if a `public` response has been successfully generated for the same URL before.

This is because the `HttpCache` from the `HttpKernel` component by default sets `stale_if_error` to 60 seconds.

At least when using the `HttpCache` subclass from the `FrameworkBundle`, we know about the `kernel.debug` setting and its intention to support local development. In that case, we could set the *default* `stale-if-error` value to 0.

Commits
-------

3a23ec89c3 Avoid stale-if-error if kernel.debug = true, because it hides errors

12 months agoAvoid stale-if-error if kernel.debug = true, because it hides errors
Matthias Pigulla [Fri, 10 Jan 2020 10:52:55 +0000 (10:52 +0000)]
Avoid stale-if-error if kernel.debug = true, because it hides errors

12 months agobug #35151 [DI] deferred exceptions in ResolveParameterPlaceHoldersPass (Islam93)
Fabien Potencier [Fri, 10 Jan 2020 08:08:19 +0000 (09:08 +0100)]
bug #35151 [DI] deferred exceptions in ResolveParameterPlaceHoldersPass (Islam93)

This PR was merged into the 3.4 branch.

Discussion
----------

[DI] deferred exceptions in ResolveParameterPlaceHoldersPass

| Q             | A
| ------------- | ---
| Branch?       | 3.4
| Bug fix?      | yes
| New feature?  | no
| Deprecations? | no
| Tickets       | Fix #30428
| License       | MIT
| Doc PR        | n/a

fixes case #30428
implemented as in AutowiringPass

Commits
-------

b3a2173c8e [DI] deferred exceptions in ResolveParameterPlaceHoldersPass

12 months agominor #35288 [Console] Fix SymfonyQuestionHelper tests sometimes failing on AppVeyor...
Fabien Potencier [Fri, 10 Jan 2020 08:03:47 +0000 (09:03 +0100)]
minor #35288 [Console] Fix SymfonyQuestionHelper tests sometimes failing on AppVeyor (fancyweb)

This PR was merged into the 3.4 branch.

Discussion
----------

[Console] Fix SymfonyQuestionHelper tests sometimes failing on AppVeyor

| Q             | A
| ------------- | ---
| Branch?       | 3.4
| Bug fix?      | no
| New feature?  | no
| Deprecations? | no
| Tickets       | https://github.com/symfony/symfony/issues/35035
| License       | MIT
| Doc PR        | -

The test uses heredoc for the expected part. Expected line returns are `"\n"` because that's how they are written in the source code file.
However, on Windows, the console outputs `"\r\n"` (`PHP_EOL`) for new lines.
`"qqq:\r\n"` does not contain `"qqq:\n"`.

I'm still wondering why this test is not *always* failing...

Commits
-------

474f3bef08 [Console] Fix SymfonyQuestionHelper tests sometimes failing on AppVeyor

12 months ago[Console] Fix SymfonyQuestionHelper tests sometimes failing on AppVeyor
Thomas Calvet [Thu, 9 Jan 2020 18:48:50 +0000 (19:48 +0100)]
[Console] Fix SymfonyQuestionHelper tests sometimes failing on AppVeyor

12 months ago[Workflow] Fix configuration node reference for "initial_marking"
Javier Spagnoletti [Thu, 9 Jan 2020 20:48:41 +0000 (17:48 -0300)]
[Workflow] Fix configuration node reference for "initial_marking"

12 months agobug #35278 [EventDispatcher] expand listener in place (xabbuh)
Nicolas Grekas [Thu, 9 Jan 2020 14:01:40 +0000 (15:01 +0100)]
bug #35278 [EventDispatcher] expand listener in place (xabbuh)

This PR was merged into the 4.3 branch.

Discussion
----------

[EventDispatcher] expand listener in place

| Q             | A
| ------------- | ---
| Branch?       | 4.3
| Bug fix?      | yes
| New feature?  | no
| Deprecations? | no
| Tickets       | Fix #35259
| License       | MIT
| Doc PR        |

Commits
-------

f5d407318d expand listener in place

12 months agoMerge branch '3.4' into 4.3
Nicolas Grekas [Thu, 9 Jan 2020 13:38:14 +0000 (14:38 +0100)]
Merge branch '3.4' into 4.3

* 3.4:
  [PhpUnitBridge] fix compat with older versions of composer

12 months ago[PhpUnitBridge] fix compat with older versions of composer
Nicolas Grekas [Thu, 9 Jan 2020 13:36:09 +0000 (14:36 +0100)]
[PhpUnitBridge] fix compat with older versions of composer

12 months agoexpand listener in place
Christian Flothmann [Thu, 9 Jan 2020 13:14:31 +0000 (14:14 +0100)]
expand listener in place

12 months agoMerge branch '3.4' into 4.3
Nicolas Grekas [Thu, 9 Jan 2020 12:45:17 +0000 (13:45 +0100)]
Merge branch '3.4' into 4.3

* 3.4:
  fix tests

12 months agofix tests
Nicolas Grekas [Thu, 9 Jan 2020 12:44:20 +0000 (13:44 +0100)]
fix tests

12 months agofix deps
Nicolas Grekas [Thu, 9 Jan 2020 12:29:25 +0000 (13:29 +0100)]
fix deps

12 months agoMerge branch '3.4' into 4.3
Nicolas Grekas [Thu, 9 Jan 2020 12:10:02 +0000 (13:10 +0100)]
Merge branch '3.4' into 4.3

* 3.4:
  Relax transient test

12 months agoRelax transient test
Nicolas Grekas [Thu, 9 Jan 2020 12:09:28 +0000 (13:09 +0100)]
Relax transient test

12 months agoMerge branch '3.4' into 4.3
Nicolas Grekas [Thu, 9 Jan 2020 11:56:22 +0000 (12:56 +0100)]
Merge branch '3.4' into 4.3

* 3.4:
  Added more tests for WebProfilerBundle

12 months agominor #31522 Added more tests for WebProfilerBundle (javiereguiluz)
Nicolas Grekas [Thu, 9 Jan 2020 11:44:37 +0000 (12:44 +0100)]
minor #31522 Added more tests for WebProfilerBundle (javiereguiluz)

This PR was merged into the 3.4 branch.

Discussion
----------

Added more tests for WebProfilerBundle

| Q             | A
| ------------- | ---
| Branch?       | 3.4
| Bug fix?      | no
| New feature?  | no <!-- please update src/**/CHANGELOG.md files -->
| BC breaks?    | no     <!-- see https://symfony.com/bc -->
| Deprecations? | no <!-- please 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        | -

Thanks to @jpauli Code Coverage info about Symfony (http://cov.jpauli.tech/) I found that WebProfiler's controllers are pretty badly covered:

![image](https://user-images.githubusercontent.com/73419/57919817-ec390500-7899-11e9-81b7-763a0b35d0ec.png)

This PR focuses on testing the main controller class:

![image](https://user-images.githubusercontent.com/73419/57919877-04108900-789a-11e9-8a93-3466b672d873.png)

Commits
-------

2f7a820edd Added more tests for WebProfilerBundle

12 months agoAdded more tests for WebProfilerBundle
Javier Eguiluz [Fri, 17 May 2019 09:48:13 +0000 (11:48 +0200)]
Added more tests for WebProfilerBundle

12 months agoMerge branch '3.4' into 4.3
Nicolas Grekas [Thu, 9 Jan 2020 09:48:33 +0000 (10:48 +0100)]
Merge branch '3.4' into 4.3

* 3.4:
  [PHPUnitBridge] file_get_contents() expects parameter 3 to be resource
  [PHPUnit-Bridge] Fail-fast in simple-phpunit if one of the passthru() commands fails

12 months agominor #35265 [Process] - update @throws phpdoc (Benhssaein Youssef)
Nicolas Grekas [Thu, 9 Jan 2020 09:39:14 +0000 (10:39 +0100)]
minor #35265 [Process] - update @throws phpdoc (Benhssaein Youssef)

This PR was merged into the 4.3 branch.

Discussion
----------

[Process] - update @throws phpdoc

| Q             | A
| ------------- | ---
| Branch?       | 4.4
| Bug fix?      | no
| New feature?  | no
| Deprecations? | no
| Tickets       | no
| License       | MIT
| Doc PR        |

Update PHPDoc throws Exceptions on https://github.com/symfony/symfony/blob/4.4/src/Symfony/Component/Process/Process.php

Commits
-------

2be7029b70 [Process] - update @throws phpdoc

12 months ago[Process] - update @throws phpdoc
Benhssaein Youssef [Wed, 8 Jan 2020 15:02:14 +0000 (16:02 +0100)]
[Process] - update @throws phpdoc

12 months agominor #35274 [PHPUnitBridge] file_get_contents() expects parameter 3 to be resource...
Nicolas Grekas [Thu, 9 Jan 2020 09:21:19 +0000 (10:21 +0100)]
minor #35274 [PHPUnitBridge] file_get_contents() expects parameter 3 to be resource (plozmun)

This PR was squashed before being merged into the 3.4 branch.

Discussion
----------

[PHPUnitBridge] file_get_contents() expects parameter 3 to be resource

| Q             | A
| ------------- | ---
| Branch?       | 3.4
| Bug fix?      | yes
| New feature?  | no
| Deprecations? | no
| Tickets       | -
| License       | MIT
| Doc PR        | -

file_get_contents 3rd parameter (context) expects resource or NULL to ignore them

Commits
-------

a28a42187c [PHPUnitBridge] file_get_contents() expects parameter 3 to be resource

12 months ago[PHPUnitBridge] file_get_contents() expects parameter 3 to be resource
Pablo Lozano [Thu, 9 Jan 2020 08:55:05 +0000 (09:55 +0100)]
[PHPUnitBridge] file_get_contents() expects parameter 3 to be resource

12 months agobug #35254 [PHPUnit-Bridge] Fail-fast in simple-phpunit if one of the passthru()...
Nicolas Grekas [Thu, 9 Jan 2020 09:20:23 +0000 (10:20 +0100)]
bug #35254 [PHPUnit-Bridge] Fail-fast in simple-phpunit if one of the passthru() commands fails (mpdude)

This PR was squashed before being merged into the 3.4 branch.

Discussion
----------

[PHPUnit-Bridge] Fail-fast in simple-phpunit if one of the passthru() commands fails

| Q             | A
| ------------- | ---
| Branch?       | 3.4
| Bug fix?      | yes
| New feature?  | no
| Deprecations? | no
| Tickets       |
| License       | MIT
| Doc PR        |

Some commands executed by the `simple-phpunit` script are not checked for success. For example [here](https://travis-ci.org/twigphp/Twig/jobs/634110681), Composer fails with the message

```
  [InvalidArgumentException]
  Could not find package phpunit/phpunit with version 7.5.* in a version inst
  allable using your PHP version 7.0.25.
```

Yet, the `simple-phpunit` script happily continues, going over failing `chdir()`, `file_get_contents()` and `include()` calls and eventually returns a successful `0` exit code. So CI tests look OK when in fact PHPUnit was not even downloaded.

Commits
-------

576e18561f [PHPUnit-Bridge] Fail-fast in simple-phpunit if one of the passthru() commands fails

12 months ago[PHPUnit-Bridge] Fail-fast in simple-phpunit if one of the passthru() commands fails
Matthias Pigulla [Wed, 8 Jan 2020 13:37:17 +0000 (14:37 +0100)]
[PHPUnit-Bridge] Fail-fast in simple-phpunit if one of the passthru() commands fails

12 months agobug #35261 [Routing] Fix using a custom matcher & generator dumper class (fancyweb)
Nicolas Grekas [Wed, 8 Jan 2020 17:23:22 +0000 (18:23 +0100)]
bug #35261 [Routing] Fix using a custom matcher & generator dumper class (fancyweb)

This PR was merged into the 4.3 branch.

Discussion
----------

[Routing] Fix using a custom matcher & generator dumper class

| Q             | A
| ------------- | ---
| Branch?       | 4.3
| Bug fix?      | yes
| New feature?  | no
| Deprecations? | no
| Tickets       | -
| License       | MIT
| Doc PR        | -

This PR fixes a BC break I encountered while upgrading an existing project from 4.2 to 4.4. In this project I use a custom `generator_dumper_class` that is not a `CompiledUrlGeneratorDumper` (it didn't exist yet). I faced 2 problems:
- The generator is considered "compiled" while it is not. This is because we don't check if the `generator_dumper_class` is effectively a `CompiledUrlGeneratorDumper` to compute the `$compiled` variable. That result in a `\TypeError: Return value of Symfony\Component\Routing\Router::getCompiledRoutes() must be of the type array, int returned`
- My custom dumper is not used at all. This is because of https://github.com/symfony/symfony/pull/31964. I altered the condition to fall back only in one way and not the other. The original issue is still fixed (if one uses a classic `UrlGenerator` + a `CompiledUrlGeneratorDumper`, it fall backs on `PhpGeneratorDumper`). However, if one uses a `CompiledUrlGenerator` + a classic `PhpGeneratorDumper` (my case), the classic dumper is still returned. Since `$compiled` is now correctly computed, this case works fine. The Router won't try to get the compiled routes and will use the "old" way.

Commits
-------

3a840a9796 [Routing] Fix using a custom matcher & generator dumper class

12 months agoMerge branch '3.4' into 4.3
Nicolas Grekas [Wed, 8 Jan 2020 17:18:34 +0000 (18:18 +0100)]
Merge branch '3.4' into 4.3

* 3.4:
  [Debug] fix ClassNotFoundFatalErrorHandler
  [Dotenv] Fixed infinite loop with missing quote followed by quoted value
  [TwigBridge] button_widget now has its title attr translated even if its label = null or false
  [PhpUnitBridge] When using phpenv + phpenv-composer plugin, composer executable is wrapped into a bash script
  [Security] Prevent canceled remember-me cookie from being accepted
  [FrameworkBundle][TranslationUpdateCommand] Do not output positive feedback on stderr

12 months agobug #34643 [Dotenv] Fixed infinite loop with missing quote followed by quoted value...
Nicolas Grekas [Wed, 8 Jan 2020 17:02:35 +0000 (18:02 +0100)]
bug #34643 [Dotenv] Fixed infinite loop with missing quote followed by quoted value (naitsirch)

This PR was merged into the 3.4 branch.

Discussion
----------

[Dotenv] Fixed infinite loop with missing quote followed by quoted value

| Q             | A
| ------------- | ---
| Branch?       | 3.4
| Bug fix?      | yes
| New feature?  | no
| Deprecations? | no
| Tickets       | Fix #34642
| License       | MIT
| Doc PR        |

If there's a quote missing to end a value and in the next line there's again a quoted value Dotenv will run into an infinite loop. An .env file with the following content will result in this error:
```
FOO="foo
BAR="bar"
```
See #34642 for more details.

Commits
-------

eb69e135b2 [Dotenv] Fixed infinite loop with missing quote followed by quoted value

12 months agobug #35239 [Security\Http] Prevent canceled remember-me cookie from being accepted...
Nicolas Grekas [Wed, 8 Jan 2020 17:02:00 +0000 (18:02 +0100)]
bug #35239 [Security\Http] Prevent canceled remember-me cookie from being accepted (chalasr)

This PR was merged into the 3.4 branch.

Discussion
----------

[Security\Http] Prevent canceled remember-me cookie from being accepted

| Q             | A
| ------------- | ---
| Branch?       | 3.4
| Bug fix?      | yes
| New feature?  | no
| Deprecations? | no
| Tickets       | Fix #35198
| License       | MIT
| Doc PR        | -

`RememberMeServices::autoLogin()` only checks that the cookie exists in `$request->cookies` while `loginFail()` only alter `$request->attributes` (which allows child implementations to read the canceled cookie for e.g. removing a persistent one).
This makes `autoLogin()` checks for `request->attributes` first, which fixes the linked issue.

Failure expected on deps=high build.

Commits
-------

9b711b87fe [Security] Prevent canceled remember-me cookie from being accepted