Compatibility with GHC 9.0 #96
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
The main difference in GHC 9.0 is the simplified subsumption rule, which affects some of our dependencies as well as our own code. Summary: - Adds a `stack-ghc9.yaml` to check 9.0+ compatibility. Notice how we need to use a development version of `generic-lens`. - Does a few manual 𝜂-expansions required by the simplified subsumption rule. - Remove the `Writer.Discouraged` module. It contained a single instance for `Lift`. This instance was not particularly useful as it depended on `monad-unlift`, which basically meant that it can only lift `HasWriter` through a `ReaderT` (the `mtl` has other instances of `MonadWriter`, but the `listen` function is written in an _ad hoc_ manner for each). The `monad-unlift` package is deprecated by its author, and doesn't compile in GHC 9.0. Moreover the documentation of this instance said to prefer using `Lift` directly on the underlying state capability (which works for any monad transformer and is more efficient), and, since the only way that we provide to make a `HasWriter` capability is via a state capability, there were not many opportunities to use this instance. All in all, it seemed best to remove the instance. The only alternative I could think of being to internalise the `monad-unlift` constraint which doesn't make a ton of sense in my opinion.
aherrmann
reviewed
Jul 19, 2021
| @@ -58,7 +58,6 @@ library | |||
| Capability.Stream | |||
| Capability.TypeOf | |||
| Capability.Writer | |||
| Capability.Writer.Discouraged | |||
There was a problem hiding this comment.
This module is also referenced here. I think that paragraph can be removed now, given that Capability.Writer.Discouraged was the only such discouraged module.
| @@ -0,0 +1,12 @@ | |||
| resolver: nightly-2021-07-04 | |||
There was a problem hiding this comment.
Can we test this configuration on CI?
Now that the one remaining Discouraged module has been removed.
The GHC 9.0 build doesn't do any test, as they are performed by the main branch. This commit also fixes a caching bug in the Github action. It's too small for a separate commit.
aspiwack
force-pushed
the
ghc-9.0-compatibility
branch
from
67a6128 to
ba61041
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
The main difference in GHC 9.0 is the simplified subsumption rule,
which affects some of our dependencies as well as our own code.
Summary:
Adds a
stack-ghc9.yamlto check 9.0+ compatibility. Notice how weneed to use a development version of
generic-lens.Does a few manual 𝜂-expansions required by the simplified
subsumption rule.
Remove the
Writer.Discouragedmodule. It contained a singleinstance for
Lift. This instance was not particularly useful as itdepended on
monad-unlift, which basically meant that it can onlylift
HasWriterthrough aReaderT(themtlhas other instancesof
MonadWriter, but thelistenfunction is written in an adhoc manner for each). The
monad-unliftpackage is deprecated byits author, and doesn't compile in GHC 9.0. Moreover the
documentation of this instance said to prefer using
Liftdirectlyon the underlying state capability (which works for any monad transformer
and is more efficient), and, since the only way that we provide to
make a
HasWritercapability is via a state capability, there werenot many opportunities to use this instance.
All in all, it seemed best to remove the instance. The only
alternative I could think of being to internalise the
monad-unliftconstraint which doesn't make a ton of sense in my opinion.