This PR was merged into the 4.3 branch.
Discussion
----------
[HttpClient] fallbackto CURLMOPT_MAXCONNECTS when CURLMOPT_MAX_HOST_CONNECTIONS is not available
| Q | A
| ------------- | ---
| Branch? | 4.3
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #33462
| License | MIT
| Doc PR | -
Commits
-------
0aae1d7c0a [HttpClient] fallbackto CURLMOPT_MAXCONNECTS when CURLMOPT_MAX_HOST_CONNECTIONS is not available
This PR was squashed before being merged into the 3.4 branch (closes#33504).
Discussion
----------
SCA: minor code tweaks
| Q | A
| ------------- | ---
| Branch? | 3.4
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | n/a
| License | MIT
| Doc PR | n/a
Minor code tweaks: disabled sorting in glob (the order is not important there), used implode() negative index support in order to drop an array_pop() call.
Commits
-------
32ea449679 SCA: minor code tweaks
This PR was merged into the 3.4 branch.
Discussion
----------
[HttpKernel] Fix getFileLinkFormat() to avoid returning the wrong URL in Profiler
| Q | A
| ------------- | ---
| Branch? | >= 3.4 <!-- see below -->
| Bug fix? | yes
| 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 | #32444 <!-- #-prefixed issue number(s), if any -->
| License | MIT
| Doc PR | N/A <!-- required for new features -->
I had this problem and I noticed that an issue had already been registered.
So I registered the pull request.
Commits
-------
9ed5f03b98 Fix getFileLinkFormat() to avoid returning the wrong URL in Profiler
This PR was squashed before being merged into the 3.4 branch (closes#33487).
Discussion
----------
[HttpKernel] Fix Apache mod_expires Session Cache-Control issue
| Q | A
| ------------- | ---
| Branch? | 3.4 for bug fixes <!-- see below -->
| Bug fix? | yes
| 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 -->
| License | MIT
Apaches's [mod_expires](https://httpd.apache.org/docs/current/mod/mod_expires.html) is a widely used module to set HTTP caching headers. It allows you to set a default cache lifetime as well as lifetimes by mime_type.
When an application server has set a `Cache-Control` header, mod_expires ignores this and sets its own, resulting in duplicate `Cache-Control` headers and conflicting information. It does this _unless_ the application server sets an `Expires` header, in which case mod_expires does nothing. This is documented on the link above:
> When the `Expires` header is already part of the response generated by the server, for example when generated by a CGI script or proxied from an origin server, this module does not change or add an `Expires` or `Cache-Control` header.
Symfony automatically sets a `Cache-Control` header if a session exists. This patch adds an `Expires` header to ensure it's respected by mod_expires.
## Example 1
With the following Apache config:
```apache
<IfModule mod_expires.c>
ExpiresActive on
ExpiresDefault "access plus 1 month"
</IfModule>
```
The HTTP response headers are:
### Without the patch
```
HTTP/1.1 200 OK
Date: Fri, 06 Sep 2019 08:02:02 GMT
Server: Apache/2.4.37 (Ubuntu)
Cache-Control: max-age=0, must-revalidate, private
Cache-Control: max-age=2592000
Expires: Sun, 06 Oct 2019 08:02:00 GMT
Vary: Accept-Encoding
Content-Encoding: gzip
Content-Length: 13099
Connection: close
Content-Type: text/html; charset=UTF-8
```
### With the patch
```
HTTP/1.1 200 OK
Date: Fri, 06 Sep 2019 08:21:34 GMT
Server: Apache/2.4.37 (Ubuntu)
Cache-Control: max-age=0, must-revalidate, private
Expires: Fri, 06 Sep 2019 08:21:34 GMT
Vary: Accept-Encoding
Content-Encoding: gzip
Content-Length: 13098
Connection: close
Content-Type: text/html; charset=UTF-8
```
## Example 2
With the following Apache config:
```apache
<IfModule mod_expires.c>
ExpiresActive on
ExpiresDefault "access plus 1 month"
ExpiresByType text/html "access plus 0 seconds"
</IfModule>
```
### Without the patch
```
HTTP/1.1 200 OK
Date: Fri, 06 Sep 2019 08:18:40 GMT
Server: Apache/2.4.37 (Ubuntu)
Cache-Control: max-age=0, must-revalidate, private
Cache-Control: max-age=0
Expires: Fri, 06 Sep 2019 08:18:39 GMT
Vary: Accept-Encoding
Content-Encoding: gzip
Content-Length: 13099
Connection: close
Content-Type: text/html; charset=UTF-8
```
### With the patch
```
HTTP/1.1 200 OK
Date: Fri, 06 Sep 2019 08:20:40 GMT
Server: Apache/2.4.37 (Ubuntu)
Cache-Control: max-age=0, must-revalidate, private
Expires: Fri, 06 Sep 2019 08:20:40 GMT
Vary: Accept-Encoding
Content-Encoding: gzip
Content-Length: 13100
Connection: close
Content-Type: text/html; charset=UTF-8
```
Commits
-------
9e942768c9 [HttpKernel] Fix Apache mod_expires Session Cache-Control issue
This PR was merged into the 4.3 branch.
Discussion
----------
Add BC break note to UPGRADE-4.2.md
| Q | A
| ------------- | ---
| Branch? | 4.4
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| License | MIT
Added a note about upgrade consequences. See #33473
Commits
-------
89f7d77ea0 Add BC break note to UPGRADE-4.2.md
This PR was merged into the 4.3 branch.
Discussion
----------
[Router] routing cache crash when using generator_class
| Q | A
| ------------- | ---
| Branch? | 4.3
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #31807
| License | MIT
Since #28865 the Router use, by default, new generator, matcher, and dumpers.
This leads to crash when the Router use a custom generator, or matcher based on the old ones.
Commits
-------
a5b46e5390 Fix routing cache broken when using generator_class
This PR was merged into the 4.3 branch.
Discussion
----------
Fix test fixtures with deprecated method signatures
| Q | A
| ------------- | ---
| Branch? | 4.3
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #33483 (partly)
| License | MIT
| Doc PR | N/A
This PR upgrades two fixtures that implemented deprecated method signatures. As far as I can tell, they are used in tests that do not specifically test legacy behavior, so the fixtures should be up to date. Currently, these fixtures cause failing tests on the 4.4 branch.
Commits
-------
cc3e3d54ea Fix more bad tests
592aacff6f Fix test fixtures with deprecated method signatures.
This PR was merged into the 4.3 branch.
Discussion
----------
[Messenger] fix empty amqp body returned as false
| Q | A
| ------------- | ---
| Branch? | 4.3
| Bug fix? | yes
| 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 |
| License | MIT
| Doc PR |
Having `false` in the body breaks typehints in the serializer and is not consistent with other transports like doctrine. See https://github.com/pdezwart/php-amqp/issues/351
Commits
-------
ee5cbe3658 [Messenger] fix empty amqp body returned as false
This PR was merged into the 4.3 branch.
Discussion
----------
[Mailer] maintain sender/recipient name in SMTP envelopes
| Q | A
| ------------- | ---
| Branch? | 4.4
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #33383
| License | MIT
| Doc PR |
Commits
-------
46ed0e84cd maintain sender/recipient name in SMTP envelopes
This PR was merged into the 4.3 branch.
Discussion
----------
Fix gmail relay
| Q | A
| ------------- | ---
| Branch? | 4.3
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #32148
| License | MIT
This tiny PR fixes#32148 by removing the emission of a notice which prevented a `TransportException` to be thrown when something wrong occured.
Commits
-------
6c90e08368Fix#32148 TransportException was not thrown
This PR was merged into the 4.3 branch.
Discussion
----------
[HttpClient] fix support for 103 Early Hints and other informational status codes
| Q | A
| ------------- | ---
| Branch? | 4.3
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | -
| License | MIT
| Doc PR | -
I learned quite recently how 1xx status codes work in HTTP 1.1 when I discovered the [103 Early Hint](https://evertpot.com/http/103-early-hints) status code from [RFC8297](https://tools.ietf.org/html/rfc8297)
This PR fixes support for them by adding a new `getInformationalStatus()` method on `ChunkInterface`. This means that you can now know about 1xx status code by using the `$client->stream()` method:
```php
$response = $client->request('GET', '...');
foreach ($client->stream($response) as $chunk) {
[$code, $headers] = $chunk->getInformationalStatus();
if (103 === $code) {
// $headers['link'] contains the early hints defined in RFC8297
}
// ...
}
```
Commits
-------
34275bba1c [HttpClient] fix support for 103 Early Hints and other informational status codes
* 3.4:
[Validator] Add ConstraintValidator::formatValue() tests
[Validator] Sync string to date behavior and throw a better exception
Check phpunit configuration for listeners
This PR was merged into the 4.3 branch.
Discussion
----------
[HttpClient] Fix a bug preventing Server Pushes to be handled properly
| Q | A
| ------------- | ---
| Branch? | 4.3
| Bug fix? | yes
| 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 | n/a
| License | MIT
| Doc PR | n/a
Follows https://github.com/symfony/symfony/pull/33444.
Commits
-------
e1fbaeb65c [HttpClient] Fix a bug preventing Server Pushes to be handled properly
This PR was merged into the 4.3 branch.
Discussion
----------
[HttpClient] improve handling of HTTP/2 PUSH, disable it by default
| Q | A
| ------------- | ---
| Branch? | 4.3
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | -
| License | MIT
| Doc PR | -
This follows discussions with @dunglas
For the test cases, https://http2-push.io is down, let's use Akamai instead
This PR now considers the proxy settings before accepting a pushed response.
It also splits the responsibility of dealing with accepting pushed responses in method `acceptPushForRequest`.
The logic in this method could also be delegated to a userland callback passed as an option. Let's wait for someone with an actual use case before adding the option.
This PR also disables HTTP/2 PUSH by default because it is not stable: locally, with the latest curl version, enabling this on a server that pushes things fails with `Failure when receiving data from the peer`. This is not ready for prime time in either ext-curl or the underlying libcurl. You can still enable it explicitly by passing some positive number to the constructor.
Commits
-------
019bce7230 [HttpClient] improve handling of HTTP/2 PUSH
This PR was squashed before being merged into the 3.4 branch (closes#33434).
Discussion
----------
[Validator] Add ConstraintValidator::formatValue() tests
| Q | A
| ------------- | ---
| Branch? | 3.4
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | -
| License | MIT
| Doc PR | -
So https://github.com/symfony/symfony/pull/33401 tests can be built on top of this.
Commits
-------
b688aa31ec [Validator] Add ConstraintValidator::formatValue() tests
This PR was merged into the 4.3 branch.
Discussion
----------
[Validator] Only handle numeric values in DivisibleBy
| Q | A
| ------------- | ---
| Branch? | 4.3
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | -
| License | MIT
| Doc PR | -
Currently it probably breaks because `abs` throws a notice on objects.
Commits
-------
f974add66a [Validator] Only handle numeric values in DivisibleBy
This PR was merged into the 4.3 branch.
Discussion
----------
Fix#33427
| Q | A
| ------------- | ---
| Branch? | 4.3
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #33427
| License | MIT
| Doc PR | none
Fix#33427 by checking if the message returned by the intl-icu catalog is empty. If yes then the translator returns an empty string instead of running `formatIntl()` which uses the constructor of `MessageFormatter` which throws an exception with empty strings.
Commits
-------
414dcebfc4Fix#33427
This PR was merged into the 3.4 branch.
Discussion
----------
[Validator] Sync string to date behavior and throw a better exception
| Q | A
| ------------- | ---
| Branch? | 3.4
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | -
| License | MIT
| Doc PR | -
`\DateTimeImmutable` are not compared with `\DateTimeImmutable` in the `RangeValidator` class (contrary to the behavior in the `AbstractComparisonValidator` class).
Also, let's throw a dedicated exception when the provided string value cannot be parsed to a `\DateTime` or `\DateTimeImmutable`. It's better than the default exception IMO.
Commits
-------
28d7d9444f [Validator] Sync string to date behavior and throw a better exception
This PR was merged into the 4.3 branch.
Discussion
----------
[DI] fix support for "!tagged_locator foo"
| Q | A
| ------------- | ---
| Branch? | 4.3
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | -
| License | MIT
| Doc PR | -
Spotted during the workshop at WebSummerCamp
Commits
-------
a9f75692c9 [DI] fix support for "!tagged_locator foo"
This PR was merged into the 3.4 branch.
Discussion
----------
[PHPUnit Bridge] Avoid registering listener twice
The listener can be registered via configuration by the user. In that
case, we do not want to add it again to the list of listeners.
Closes#31649
| Q | A
| ------------- | ---
| Branch? | 3.4
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | n/a
| License | MIT
| Doc PR | n/a
Commits
-------
b190536205 Check phpunit configuration for listeners
The bridge listener can be registered via configuration by the user. In that
case, we do not want to add it again to the list of listeners.
Closes#31649