Milan Miladinovic

Added a test that attempts to use: - 2 deflate blocks in a message - a deflate block without compression - a deflate block with `BFINAL` set to 1 Our...

Just changed the return type to `void` and assert that we get the correct return value: https://github.com/capnproto/capnproto/pull/1505/commits/ee92d4420c2931a029d409c9f4c4fe9e087c49d9

Rebased on master to get new commits, no new code otherwise.

> It occurs to me that we're going to want to let the application selectively opt into / out of compression when calling HttpService::Response::acceptWebSocket(). > > Specifically, when we're proxying...

Brief summary on build failures and what I think is happening: - linux-musl looks like it doesn't install `zlib` while building, so I figure we can fix that by including...

Rebase + force push, no code changes.

These fixups - https://github.com/capnproto/capnproto/pull/1505/commits/08a684372c2b81bc61ac072a1480909c198785c4 - https://github.com/capnproto/capnproto/pull/1505/commits/15096ece50753033b33d77045a13baff5d937915 will intentionally break the linux-bazel and MacOS builds respectively. I want to be sure they break _before_ pushing preprocessor changes so we know whether...

I've reordered the commits and moved the zlib build commit up before the commit that introduces zlib related code: https://github.com/capnproto/capnproto/pull/1505/commits/ac560be7e691a633916f6309f37c6b0acd48f3e8. Hopefully this fixes the build issue, if not, then it...

Rebased to get commits from https://github.com/capnproto/capnproto/pull/1504 as I wanted to finish a TODO related to that PR.

> So, I think we need to make sure that, at least as used in Workers, KJ respects exactly the response headers that the application gives it. If the application...