Skip to content

Update Stack LTS to 19.33 (GHC 9.0.2)#53

Closed
sodic wants to merge 11 commits intomainfrom
update-ghc-to-902
Closed

Update Stack LTS to 19.33 (GHC 9.0.2)#53
sodic wants to merge 11 commits intomainfrom
update-ghc-to-902

Conversation

@sodic
Copy link
Contributor

@sodic sodic commented Oct 10, 2025

No description provided.

@sodic sodic changed the title Update GHC to 9.0.2 Update Stack LTS to 19.33 (GHC 9.0.2) Oct 10, 2025
@sodic sodic marked this pull request as draft October 10, 2025 08:41
Copy link
Contributor Author

@sodic sodic left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Some questions for @Martinsos

@@ -35,7 +35,7 @@ library:
- path >=0.9.2 && <0.10
- exceptions >=0.10 && <0.11
Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@Martinsos

If I got it right, the lower bounds should match the version set in workflows/ci.yaml which was 18.21. I am basing this off the comment above that line and your article.

The lower bounds did match that LTS for the most part, but they ignored minor (third digit) changes. For example:

  • The LTS 18.21 version for exceptions is 0.10.4 (source).
  • package.yaml lists >=0.10 as the lower end of the range.

Was this done on purpose as a way to "broaden" the range?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

We've concluded "yes"

- tasty-hspec >=1.2 && <1.3
- tasty-quickcheck >=0.10 && <0.11
- tasty-discover >=4.2 && <4.3
- hspec >=2.7 && <2.10
Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Why did hspec allow for 3 breaking change versions in its range? Is it a case of an "adventurous upper bound?"

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Check changelog.

# NOTE: If you modify this field, make sure to update the tested-with
# field in package.yaml to contain the corresponding GHC versions.
resolver: nightly-2022-01-04
resolver: lts-19.33
Copy link
Contributor Author

@sodic sodic Oct 10, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I've set both this resolver and the one in workflows/ci.yaml to the same version. If I got it right, this means we're only testing Strong Path against a single version of GHC.

This PR only updates GHC to 9.0.2 and I've mostly done it to get a hang of the procedure, so picking the right LTS versions is not yet important.

Once I actually fully update StrongPath to 9.6.7, which ranges would you like to support and test for? In other words, which LTSs should I list in ci.yaml and stack.yaml?
Perhaps from 9.4.x (2 years ago) to 9.6.7 (current recommended). Or should we go until current latest?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

We've decided to support 9.4.5 to 9.12.2. Hanled in #54

Comment on lines +8 to +10
concurrency:
group: ${{ github.workflow }}-${{ github.ref }}
cancel-in-progress: true
Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This automatically cancels in progress jobs, making our cancel step below redundant.

runs-on: ${{ matrix.os }}
needs: code-formatter
strategy:
fail-fast: false
Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

With this, failure on one job in the matrix does not cause other jobs to fail.

@sodic sodic force-pushed the update-ghc-to-902 branch from 66aa9b2 to 3411d89 Compare October 10, 2025 15:25
@sodic
Copy link
Contributor Author

sodic commented Oct 29, 2025

Closed in favor of #54. This one was just a testing ground.

@sodic sodic closed this Oct 29, 2025
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant