After retesting, it works with CurlHttpClient and not with NativeHttpClient... So it might not be the best solution.

[Form] Deprecate not honoring with_seconds => false for HTML5 datetime inputs in DateTimeType

[Console] fix clear of section with question

Added condition to always return the real Authenticator

[Serializer] Replace the MissingConstructorArgumentsException class with MissingConstructorArgumentException

add Sender to the list of bypassed headers

[HttpKernel] Fix setting the session on the main request when it's started by a subrequest

[Form] Check for RepeatedType child in PasswordHasherListener

Remove unused data provider

minor #49278 Remove unused data provider (OskarStark)

This PR was merged into the 5.4 branch.

Discussion

Remove unused data provider

| Q | A | ------------- | --- | Branch? | 5.4 | Bug fix? | no | New feature? | no | Deprecations? | no | Tickets | Refs https://github.com/symfony/symfony/pull/48668/files#r1098456404 | License | MIT | Doc PR | -

Commits

492fe4af0b Remove unused data provider

Fix: Split and clean up tests

[Semaphore] Fix test

[Tests] Migrate tests to static data providers

bug #49214 [Mailer] add Sender to the list of bypassed headers (xabbuh)

This PR was merged into the 5.4 branch.

Discussion

[Mailer] add Sender to the list of bypassed headers

| Q | A | ------------- | --- | Branch? | 5.4 | Bug fix? | yes | New feature? | no | Deprecations? | no | Tickets | Fix #48936 | License | MIT | Doc PR |

Commits

f50bec1d8d add Sender to the list of bypassed headers

[HttpClient] Fix data collector

Revert "[HttpClient] Add support for "friendsofphp/well-known-implementations""

This reverts commit b814ce8f2dd54571a5458682959f3dce730e50a4.

Use xxh128 instead of md5

Fix Psalm job

[Notifier] Fix notifier profiler when transport name is null

bug #49314 [HttpClient] Revert support for "friendsofphp/well-known-implementations" (nicolas-grekas)

This PR was merged into the 6.2 branch.

Discussion

[HttpClient] Revert support for "friendsofphp/well-known-implementations"

| Q | A | ------------- | --- | Branch? | 6.2 | Bug fix? | yes | New feature? | no | Deprecations? | no | Tickets | - | License | MIT | Doc PR | -

This reverts PR #47950

The major features of "friendsofphp/well-known-implementations" are now available in "php-http/discovery" v1.15.0, see:

• https://github.com/php-http/discovery/pull/209
• https://github.com/php-http/discovery/pull/208

Given the extremely low download stats of "friendsofphp/well-known-implementations", I propose to just remove support for it as a bugfix: https://packagist.org/packages/friendsofphp/well-known-implementations

This will save us from maintaining this integration, which is unused code in practice anyway.

Commits

081aa00f96 Revert "[HttpClient] Add support for "friendsofphp/well-known-implementations""

[HttpClient] Fix over-encoding of URL parts to match browser's behavior

bug #49299 [HttpClient] Fix over-encoding of URL parts to match browser's behavior (nicolas-grekas)

This PR was merged into the 5.4 branch.

Discussion

[HttpClient] Fix over-encoding of URL parts to match browser's behavior

| Q | A | ------------- | --- | Branch? | 5.4 | Bug fix? | yes | New feature? | no | Deprecations? | no | Tickets | Fix #48315 | License | MIT | Doc PR | -

Yes, the RFC says these chars should be url-encoded:

[ ] \ ^ ` { | }

But in practice, browser's don't encode them, and some servers don't expect them to be encoded either. See https://stackoverflow.com/questions/2366260/whats-valid-and-whats-not-in-a-uri-query for some pointers.

Commits

e79e331810 [HttpClient] Fix over-encoding of URL parts to match browser's behavior

[HttpClient] Add ServerSentEvent::getArrayData() to get the SSE's data decoded as an array directly

[HttpClient] Add ServerSentEvent::getArrayData() to get the SSE's data decoded as an array directly

Is the callback called when a real response is canceled?

Yes.

What about the "info" of the response? Don't they give the needed clue?

No, since the original mock response reference is lost, we can't access it and get the info from it.

| Q | A | ------------- | --- | Branch? | 6.3 | Bug fix? | no | New feature? | no | Deprecations? | no | Tickets | - | License | MIT | Doc PR | -

[HttpClient] Add ServerSentEvent::getArrayData() to get the SSE's data decoded as an array directly

| Q | A | ------------- | --- | Branch? | 6.3 | Bug fix? | no | New feature? | yes | Deprecations? | no | Tickets | - | License | MIT | Doc PR | -

Although the data field in SSEs can be "anything", it's really common that it's a JSON payload.

Just like we have a shortcut in responses (ResponseInterface::toArray()), I think it's worth having a shortcut in ServerSentEvent to avoid writing boilerplate code again and again :sweat_smile:

| Q | A | ------------- | --- | Branch? | 5.4 | Bug fix? | yes | New feature? | no | Deprecations? | no | Tickets | - | License | MIT | Doc PR | -

I want to assert that a response is canceled under certain circumstances.

My first approach was to get the info from the MockResponse I pass as the response factory to the MockHttpClient but this one is lost since a new instance is created when the request method is called.

My next approach was to wrap the MockHttpClient I pass to the class I'm testing with a TraceableHttpClient. But it doesn't work because the traced info is outdated. The problem seems to be that the on-progress callback is not called when a MockResponse is canceled.

[HttpClient] Fix not calling the on progress callback when canceling a MockResponse

To expose only needed data for debug/dump.

For debug yes, for var_dump() yes, but not for dumps in general IMO.