Patrik Ragnarsson

Re: 1f7a775cd87325cfbaa3d7dc16a30902ff96f702 and 4f6b798e12c5737976dad0f6c022d5742a23f5ab, if done in one commit, git (on the command line, not github.com) can [detect moved lines](https://git-scm.com/docs/git-diff#Documentation/git-diff.txt---color-movedltmodegt). That's really useful IMHO.

We would also deviate from the Rack spec in its current form? @pdalberti what about using another header? This won't be solved in Puma before Rack 4 exists

In Rack 3.1, `HTTP_VERSION` was removed from the SPEC: https://github.com/rack/rack/commit/449ed4dd7abaf444a1e9b5bab34b79de0bfb0eda So I guess Puma could (should?) have a different behaviour for Rack 3.1 applications? https://github.com/rack/rack/issues/2141

@pdalberti are you using Rack 3.1?

I meant your app in production @pdalberti

Right, so even if Puma made a change for applications using Rack 3.1, it wouldn't have any effect until you start using Rack 3.1 in production. Hence you may want...

> Another example of workarounds is code in [Sinatra](https://github.com/sinatra/sinatra/blob/7b50a1bbb5324838908dfaa00ec53ad322673a29/lib/sinatra/base.rb#L309-L318): > > ```ruby > # SERVER_PROTOCOL is required in Rack 3, fall back to HTTP_VERSION > # for servers not updated...

> In Rack 3.1, `HTTP_VERSION` was removed from the SPEC: [rack/rack@449ed4d](https://github.com/rack/rack/commit/449ed4dd7abaf444a1e9b5bab34b79de0bfb0eda) > > So I guess Puma could (should?) have a different behaviour for Rack 3.1 applications? [rack/rack#2141](https://github.com/rack/rack/issues/2141) Also related...

I'm not sure what the ask is here... is there a problem? what is it? how can it be reproduced?

How to reproduce that situation?