Meeting URL: https://meet.jit.si/StableHaskellMeetBiWeekly
Previous meeting notes
-
- What about the GHC API. Invite Alexis?
- Ben: will add notes in the user manual to explain how plugins can avoid relying on hard-coding name locations.
- Ben is optimistic about low breakage with GHC 9.10.1
-
What do we think about #24453? (Merge
ghc-prim,ghc-bignum,ghc-internal)- Currently we get no gain, but we do get pain.
- In future, it would be great to split out stuff, such as the IO library, thereby making them re-installable. Yes yes! But that's quite different to what we have now!
- Conclusion:
- thumbs up to general direction of #24453
- explicitly flag up the goal of splitting out more re-installable packages, e.g. I/O.
-
Lots happening on the TH front -- invite Sebastian and Teo next time?
- #24783: tracking ticket
- #24021: Making TH more stable
- #24713: hygiene
- !12479: moving stuff into
ghc-internalto maketemplate-haskell(more) reinstallable. - #24766 exposing stuff from L.H.TH.Lib.Internal
- etc
- Twitter discussion
- Conclusion Invite Teo and Sebastian to the next meeting.
-
Making Prelude reinstallable? (Discussed last time.) Design work needed.
- Investigate head.hackage to answer this question: for each patch in head.hackage, which of (1)-(5) in the state-of-play doc caused the need for the patch?
- Are there others not listed?
- Holding the token: Ben
- Patches: https://gitlab.haskell.org/ghc/head.hackage/-/tree/master/patches?ref_type=heads
- Investigate head.hackage to answer this question: for each patch in head.hackage, which of (1)-(5) in the state-of-play doc caused the need for the patch?
-
Making base reinstallable?
-
JMCT: What would a GitHub mirror that posts to GHC GitLab look like?
-
Extension lifecycle framework proposal #601
- Has been renovated!
- Bulletin discussed previously
- Previously holding the token: Chris
- (jappie) go through head.hackage categorise according to the 5 categories + as of today + as of one year ago + as of two years ago