Previous meeting notes
-
-
David asked if
import A.B.Ccould import theA.B.Cmodules from all packages that are made available in the Cabal file. Answer: not currently. Would at very least require a GHC proposal. -
Does the proposal address a question worth answering. Answer: yes, please.
-
Trevis: GHC API redesign. Good idea to have
GhcAPI.*to distinguish fromGHC.*. -
Proposal 1/2/3: succcessive refinements, not alternatives!
-
Starting with
Experimental.Wombat.Thinginghc-experimental, and migrating toWombat.Thingwhen it moves tobase, forces clients to update their imports. In some ways using the same module name inghc-experimentalas inbasecould make this switchover easier. (But there would be have to be an extended period in whichWombat.Thingis available in bothghc-experimentalandbase. And that might leave a period in whichimport Wombat.Thingis ambiguous. But maybe if it simply moves, without a deprecation period, that'd be OK -- providedbaseandghc-experimentalcome as a pair.... but that might conflict with reinstallablebasegoals.) -
Worked example of introduction of features and their life cycle.
- Tuple3
- HasField
- HasDict
-
Not everything in
ghc-experimentalwill end up inbase! -
Timescale: first release with this stuff in will be 9.10, in around 6 months time.
-
Limited goals!
-
Expectations around
ghc-experimental.- Major bump at every GHC release
- Migration cost when moves to
base. - Deprecation cycle is a reasonable expectation. Might prove impossible.
-
-
Extension lifecycle framework proposal
- Specifically, do we need to do anything else here?
- Nothing for us to do at this time
-
- Hackage trustees have historically been very reluctant to give control over part of the namespace to someone new, if the author can be contacted and doesn't want to - is this the right policy?
- This has frequently not ended well in other communities
- A vulnerability that goes unadressed for a long time in a security package that doesn't get updated is a big risk
- We will now have a place to advertise security issues - the need for a takeover is reduced by this
- Hackage trustees have historically been very reluctant to give control over part of the namespace to someone new, if the author can be contacted and doesn't want to - is this the right policy?
-
Add to
mtlmigration guide a way to work with prior and new versions- Previously holding the token: Chris
-
Template Haskell
- Adam started looking into #20828 (adding field names to TH AST) at Zurihac
- David mentions that adding
hlintchecks enforcing use of selectors instead of positional matching would be useful
-
Conversations about GHC nightlies
- Nightly releases are now available
- Need recipes for how to use this.
- Holding the token: Jappie
- Result: https://github.com/jappeace/haskell-nightly
- Nightly releases are now available
-
GHC API discussed previously
- Previously holding the token: David
-
Bulletin discussed previously
- Previously holding the token: Chris
-
GHC warning policy document as discussed previously
- Previously holding the token: Chris
-
Creating and maintaining a set of hlint rules to promote stability
- Previously holding the token: Trevis
- Opened issue asking if we can write rules for typeclasses.