David Buchmann

@joelwurtz could you maybe help sort out what would be the correct behaviour?

yeah not much value in supporting it anymore. on the other hand, i usually only drop old versions when they hinder anything. if we would move to 8.0 or 8.1,...

thanks a lot! and as we just saw in #239 the legacy support does become cumbersome. please for now no BC breaks, so that we can have a 2.x version...

okay, we found a good reason to drop the php 7 versions. (that will also allow to reduce the range of allowed require-dev dependencies, we no longer will need 3...

i am not too familiar with HTTP/2 nor the decoder plugin. the code happens here: https://github.com/php-http/client-common/blob/676f98fb3785d2c9e1b16aa24d9fcc39066f3f06/src/Plugin/DecoderPlugin.php#L54 if the HTTP/2 forbids this header, is the plugin still useful and needed? if...

sure, happy to merge a proper solution if somebody can find out how that solution should look like.

seems reasonable to me. that was the main intention of the noop client. > What I'm not sure if we need todo `noop: false` in case the other proxy clients...

according to https://datatracker.ietf.org/doc/html/rfc7127, a "proposed standard" level RFC is already pretty solid. i prefer Toflars suggestion to extract the CacheControl into its own class. the HeaderBag mixes the 2 concerns...

according to the akamai blog, the reason for RFC9213 was that [edge architecture](https://www.w3.org/TR/edge-arch/) was not precise enough and incompatible implementations exist. the main difference seems to be that the targetting...

@nicolas-grekas you and I quickly discussed about this last week. if there is general interest in exploring this, alexander and me could have a look and propose changes, but it...