-
Notifications
You must be signed in to change notification settings - Fork 1.1k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[new release] conduit, conduit-mirage, conduit-lwt, conduit-lwt-unix and conduit-async (6.2.0) #23344
Conversation
…and conduit-async (6.2.0) CHANGES: * conduit-lwt-unix: adapt tls 0.16 split: tls.lwt is now tls-lwt (mirage/ocaml-conduit#419 @hannesm) * conduit-mirage: adapt dns-client 7.0.0 split: dns-client-mirage (mirage/ocaml-conduit#419 @hannesm)
I struggle a bit with reverse dependency failures of sihl (e.g. 0.1.10) fails https://opam-ci.ci3.ocamllabs.io/github/ocaml/opam-repository/commit/45bdebd1ff0056182d5d140c21e7422fb978bb4d/variant/compilers,5.0,conduit-lwt-unix.6.2.0,revdeps,sihl.0.1.10 with:
Now, I've no clue which package provided the mirage-crypto dependency for sihl, and/or how to restrict it properly. Related is @kit-ty-kate comment at #23315 (comment) Looking upstream at sihl, I'm as well not sure whether it is still active (there's a huge PR with not much activity recently), maybe @joseferben can shed some light into this? |
Another revdep failure is current_gitlab:
And I've no idea which package provided the ptime dependency (this unclear to me what to restrict) /cc @tmcgilchrist |
That's coming from ocurrent projects (current_github and current_gitlab), which needs a lower bound on ptime. |
The deprecations that stihl is warning about were introduced in caqti.1.8.0. See https://github.com/paurkedal/ocaml-caqti/blob/v1.8.0/caqti/lib/caqti_request.mli#L285 |
Thanks @tmcgilchrist for your ocurrent investigations. I don't quite understand what you mean with sihl and caqti. From my impression, the error in sihl is: |
Any guidance what to do? From my perspective, it is sad that sihl 0.1.x does no longer work, but on the other hand, they specify rather weakly dependencies. Also, this is quite old, with version 3.0.5 being the last released one. So my investment in trying to find out which constraint to add to sihl 0.1.x is near its end. Would it be fine for the repo maintainers to add a "conflict: conduit-lwt-unix >= 6.2.0" (and hope that fixes it)? |
yeah i agree. I've tried to figure it out but also failed to do so in a reasonable amount of time. I'll open a PR to do just that. |
Thanks @kit-ty-kate. IMHO this is fine to be merged! :) |
Sadly it's still not fine. I've spent some time tracking down the infinit loop in ocluster's tests and it seems to be coming from that error:
cc @talex5 do you know where the missing |
Ok i understand what's happening (and i think it might be holding true for the rest of the revdeps failures we had previously):
I'll open a PR |
Thanks |
Thanks a lot @kit-ty-kate for tracking this down! :) It shows that splitting packages is a huge pain (and still not clear whether it was worth the effort). |
A network connection establishment library
CHANGES: