David Buchmann

thanks for the constructive suggestion. i intend to soon organize a hackday at liip to tackle this. we have a few people interested here who have not so far contributed....

i am not sure i understand? PSR-17 is the newer standard, but php-http/message-factory is not the same as PSR-17. if a consumer requires php-http/message-factory, they need it to be installed...

hm, the symfony http client in version 5.4 does require php-http/message-factory, but only as a [develop dependency](https://github.com/symfony/http-client/blob/5.4/composer.json#L41) maybe @nicolas-grekas has an idea why this is being installed and if its...

i am not sure if i fully understand. the TagListener gathers tags from the annotations and sets the tags on the response object. do we have logic in this bundle...

ah, now i get it, thank you! if we would calculate the tags in the ControllerArgumentsEvent, the arguments would be of the correct type, right? then we should indeed change...

thanks for the analysis. i agree about the BC break risk, as newer users probably rely on that behaviour. we could do some prefix for the controller argument names but...

the named collection could indeed hide things, and the shorter its name the more likely. that said, we can totally do a new major version to avoid BC breaks. thinking...

okay, lets go with the full words. symfony `Request` calls it attributes, so i would go with plural for attributes and arguments. the attributes would also be available as request.attributes.XY...

in #611 we removed the annotation support and solely use attributes. i hope to release 3.x soon, so if you find time to look into it, now would be a...

it is not a blocker. would be nice to find a good solution, but better live with limited functionality here than rush a bad solution now that needs more BC...