Add ghc-9.0.1 to the build release script#1940
Conversation
|
looks good, yeah, could you test the build script creating a fake release in your repo? It is the default trigger of the script |
|
Oh, I should also enable the |
|
Here's the build for the test-release: https://github.com/anka-213/haskell-language-server/actions/runs/945876049 |
| optparse-applicative < 0.16 | ||
| -- These plugins doesn't work on GHC9 yet | ||
| , haskell-language-server -brittany -class -eval -fourmolu -modulename -ormolu -splice -stylishhaskell -tactic -refineImports | ||
| , haskell-language-server -brittany -class -eval -fourmolu -ormolu -splice -stylishhaskell -tactic -refineImports |
There was a problem hiding this comment.
should we also enable it in the test script? or tests are gonna fail?
There was a problem hiding this comment.
Oh, right! Thanks!
There was a problem hiding this comment.
Where is it disabled in the tests?
There was a problem hiding this comment.
Or did no-one add it to test.yaml at all?
There was a problem hiding this comment.
The later case seems to be the case. #1836 didn't modify test.yaml
There was a problem hiding this comment.
There are a lot of plugins that aren't tested:
$ for p in $(ls plugins); do grep -q $p .github/workflows/test.yml || echo "$p doesn't exist"; done
hls-explicit-imports-plugin doesn't exist
hls-floskell-plugin doesn't exist
hls-hlint-plugin doesn't exist
hls-module-name-plugin doesn't exist
hls-ormolu-plugin doesn't exist
hls-pragmas-plugin doesn't exist
hls-retrie-plugin doesn't exist
There was a problem hiding this comment.
Maybe we should add a test for this in a github action, to guarantee that people won't forget to add the test-suite to test.yml when they add a new plugin?
There was a problem hiding this comment.
not all plugins have the test suite in its own package, some of them are being tested in the the general hls test suite (under ./test). Plugins packages with its own test suite have their own test subdir (f.e. hlint does not have it but pragmas does)
Maybe we should add a test for this in a github action, to guarantee that people won't forget to add the test-suite to test.yml when they add a new plugin?
that is a nice idea
//cc @Ailrun
There was a problem hiding this comment.
Good point, here's the list with those without an internal testsuite filtered out:
$ for p in $(ls plugins); do [[ ! -d plugins/$p/test ]] || grep -q $p .github/workflows/test.yml || echo "$p doesn't exist"; done
hls-floskell-plugin doesn't exist
hls-module-name-plugin doesn't exist
hls-ormolu-plugin doesn't exist
hls-pragmas-plugin doesn't exist
And here is the list of plugins without an internal test-suite, for completeness:
grep -L test-suite plugins/*/*.cabal
plugins/hls-explicit-imports-plugin/hls-explicit-imports-plugin.cabal
plugins/hls-hlint-plugin/hls-hlint-plugin.cabal
plugins/hls-retrie-plugin/hls-retrie-plugin.cabal
There was a problem hiding this comment.
Good point. I will update the CI to check those missing tests.
just in case, i am afraid the build failed |
Yes, I noticed, thanks! I'm confused about why. Good thing we tested! |
|
Hmm, does anyone know why the tests on ghc-8.10.5, windows fails with ? |
|
Why does |
|
I think we need invalidate the cache, as the location of ghc have changed (due to some ghc vm update) |
|
@anka-213 many thanks again, this could be merged |
|
@mergify backport 1.2.0-hackage |
|
Command
Hey, I reacted but my real name is @Mergifyio |
|
Seems like this requires a manual merge, because workflows were modified. |
|
Command
|
* Add ghc-9.0.1 to the build release script * ghc9: Enable module-name plugin * Diagrams now supports ghc-9.0.1 * Upgrade cabal in build script * Cache reset updating hackage index Co-authored-by: jneira <atreyu.bbb@gmail.com> (cherry picked from commit 5f6ce4a)
* Add ghc-9.0.1 to the build release script * ghc9: Enable module-name plugin * Diagrams now supports ghc-9.0.1 * Upgrade cabal in build script * Cache reset updating hackage index Co-authored-by: jneira <atreyu.bbb@gmail.com> (cherry picked from commit 5f6ce4a)
See also discussion here: #1912 (comment)
Is this sensible, @jneira?