search

haskell-grpc-native / http2-client

A native HTTP2 client in Haskell
BSD 3-Clause "New" or "Revised" License
43 stars 27 forks source link

Please publish a new version to Hackage #90

Closed domMayhew closed 6 months ago

domMayhew commented 8 months ago

I am writing a library that depends on this package.

Please publish a new version of http2-client to Hackage. If that's not possible, please update the dependencies listed on Hackage to prevent the build breaking on mtl versions >= 2.3. It seems that many of the dependencies in the Hackage metadata for this package have inadequate bounds, allowing the use of a major version greater than specified. (E.g., time >=1.8 && <2 should be time >= 1.8 && < 1.9).

Since I am writing a library and not and end-use project, depending on commits in this GitHub repo is not a viable solution.

Thanks in advance!

lucasdicioccio commented 8 months ago

Hi!, I'll try to come up with some updates in the upcoming week/week-end. Sorry for being a bit slow, life makes it a bit hard to get some spare-time budget these days.

domMayhew commented 8 months ago

Thank you! Let me know if I can help with PRs or anything!

lucasdicioccio commented 8 months ago

@domMayhew I've edited the 0.10.0.0 revision package bounds with your two suggested changes

I see that #89 has some fixes for warnings and some package bound checks. I'm considering merging that after some test and doing another version. Stay tuned.

domMayhew commented 8 months ago

Awesome, thanks! Looking forward to your PR merging and being published. Thanks for working on this!

lucasdicioccio commented 6 months ago

hi @domMayhew I've published 0.10.0.1 with increased compatibility over dependent packages (and ghc-version) https://hackage.haskell.org/package/http2-client

domMayhew commented 6 months ago

Thanks @lucasdicioccio !

I've encountered one problem: the lower bounds on tls is not compatible with the upper bounds on tls in http2-client-grpc. I've tested using tls-1.8.0 (with stack build and stack test) and it was successful. Not sure if there are other tests that need to be run though.

Is it possible to publish a meta-revision with relaxed upper bounds? I've made a PR (#93) with the one-line change.

Thank you!