Bump upper bounds #104
Bump upper bounds #104
Conversation
| @@ -75,17 +75,17 @@ test-suite tests | |||
|
|
|||
| build-depends: | |||
| base >=4.9 && <=5.0 | |||
| , bytestring >=0.10.6.0 && <0.12.0 | |||
| , bytestring >=0.10.6.0 && <0.13.0 | |||
There was a problem hiding this comment.
Should the bounds really be duplicated for packages that are dependencies of the library and the tests? Easy to accidentally introduce incoherencies
There was a problem hiding this comment.
Well, I accidentally forgot to update the bound here and then had trouble building the test elsewhere.
The test depends on the library and IIUC they're configured at the same time, so we actually get the intersection of bounds when running the test suite.
If the test suite a looser bound it's ignored because the library one is applied, but if the test has a more narrow bound, then the library goes un-tested with some versions it builds for.
There was a problem hiding this comment.
packages that are dependencies of the library and the tests?
Huh this sentence makes no sense. I meant components that depend on the library
Yes, CI needs to test newer versions of GHC. I'll work on that in a separate PR. |
|
@alexfmpe Thank you for the contribution. |
We now test up through 9.8 as of commit cef201d . |
9.8 comes with
bytestring 0.12.1.0so I was getting build failures when building out of nixpkgs.Not sure how #102 avoids that issue
I see CI only building up to 9.4, should that be updated?