Closed cdepillabout closed 6 years ago
I haven't actually tried compiling this PR with older versions of GHC/http-conduit/conduit yet. I'll do that tomorrow.
Also, if the CI tests fail I'll try to figure out why.
Okay, it looks like with commit 798e90d, everything should be building successfully.
This PR should be ready for review.
It looks great! Thanks a lot. I'll merge and release it to hackage sooner.
This PR adds support for both http-conduit-2.3 and conduit-1.3.
At work we are trying to use twitter-conduit with stackage's lts-11, which includes http-conduit-2.3 and conduit-1.3. Once this gets merged it, it would be great if you could cut a release!
It seems like there are quite a few changes in conduit-1.3.
ResumableSource
has been removed, so now some of the http-conduit functions just return a normalSource
instead. (Also,Source
has been deprecated, and the conduit docs say to use the fullConduitT
type.) Now some of the functions in this package are returning aConduitT
instead of aResumableSource
.Also, there was a problem with using
MonadBase IO
. To be honest, I'm not sure what the problem was. I think some type has lost aMonadBase
instance, but I didn't really look into what has happened. Instead, I just switched everything to usingMonadIO
. I don't think there should be a problem with this, but if there is, let me know.I think the Snoyman-family of libraries have been moving away from
MonadBase
andMonadBaseControl
and instead settling onMonadIO
andUnliftIO
. I thought this might be the reason for theMonadBase IO
problem, but like I said, I didn't look into it.I also modified the travis.yml file to test on ghc-8.4.3.
This PR would require a major version bump before being released to Hackage.