Conversation
erikd
force-pushed
the
erikd/ghc-9.14
branch
2 times, most recently
from
0fe8598 to
acac990
Compare
Manually tested that io-classes ^>=1.9 was ok.
erikd
force-pushed
the
erikd/ghc-9.14
branch
3 times, most recently
from
4b16e72 to
efdb4e4
Compare
* Remove usage of -Wno-x-partial from cabal file where it is not needed. * Disable -Wredundant-constraint warning for ghc >= 9.14. * Disable -Wpattern-namespace-specifier for ghc >= 9.14.
|
I have logged a GHC ticket https://gitlab.haskell.org/ghc/ghc/-/issues/27005 asking if this is a If it is a bug this PR still makes sense, but we might want to wait for a fix before adding |
jorisdral
left a comment
jorisdral
left a comment
There was a problem hiding this comment.
Thank you!
Could you also add changelog entries for blockio and lsm-tree? Something along the lines of:
* Support `ghc-9.14`. See [PR #827](https://github.com/IntersectMBO/lsm-tree/pull/827)| -- ghc-9.14 gives redundant constraint warnings on some constraints | ||
| -- that are needed for earlier compilers. | ||
| if impl(ghc >=9.14) | ||
| ghc-options: -Wno-redundant-constraints |
There was a problem hiding this comment.
It might be better to only selectively disable this warning per module, so that we still get some warnings in modules where they do matter
| , fs-sim | ||
| , io-classes | ||
| , io-sim | ||
| , io-sim <1.10 |
There was a problem hiding this comment.
What is this upper bound for? Is there something incompatible in io-sim-1.10?
There was a problem hiding this comment.
Yes, there is. I had a comment in the file but the auto format did not like the comment.
io-sim-1.10 drops a method from a typeclass.
| , deepseq ^>=1.4 || ^>=1.5 | ||
| , fs-api ^>=0.4 | ||
| , io-classes ^>=1.6 || ^>=1.7 || ^>=1.8.0.1 | ||
| , io-classes ^>=1.6 || ^>=1.7 || ^>=1.8.0.1 || ^>=1.9 |
There was a problem hiding this comment.
Note that #819 also adds support for io-classes-1.9 and io-classes-1.10, so there might be some conflicts
|
BTW, I might hold off on merging this PR just until we know the fix for |
|
Turns out that the tests in this repo uncovered a bug in GHC. |
|
This PR is on hold until I will switch it to "draft" mode. |
2 participants
Description
Support ghc-9.14 exncluding adding it to CI.
The tests found a bug in the code which is only triggered when compiling with
ghc-9.14due to changes in strictness analysis and optimisation behavior. Using Claude I have come up with a fix for that, but I am still validating that fix which will be in a later PR and I will then includeghc-9.14to CI.Checklist