Too many negations 🙃

The strange case of the misplaced parenthesis. Thank you, that did the trick.

@fendor Now I can ask for your opinion on the failing test.

Ill have a look, but it is not my area of expertise. Maybe @jhrcek has some pointers, too?

Looking at the test failure, it looks like the newer apply-refact produces very slightly different (but okay) output. Sadly that means we probably need to have the test output be conditional on the GHC version, as we do in various places.

Looking at the test failure, it looks like the newer apply-refact produces very slightly different (but okay) output. Sadly that means we probably need to have the test output be conditional on the GHC version, as we do in various places.

That was my intuition. Thanks for confirming.

Why does apply-refact not work with 9.10? I'm a bit confused by that if it works for 9.12?

Why does apply-refact not work with 9.10? I'm a bit confused by that if it works for 9.12?

First, thanks for the invite. It's an honour.

For apply-refact, the owner (and @jhrcek ) have had difficulties with ghc-exactprint.

Looking into it is on my todo list, but if jhrcek is stumped, I guess I'm not going to do any better ;)

Anyhow, we should be good now.

Could you update the plugin support table? In particular https://github.com/haskell/haskell-language-server/blob/master/docs/support/plugin-support.md?plain=1#L58

I know hls-refactor-plugin already supports GHC 9.12.2, I just forgot to update the table myself 😬

EDIT: if you fixed the hls-refactor-plugin entry in a separate commit in this, I would be grateful.

@fendor: can I commit directly that kind of small changes?

To master? No, I don't think you have the permissions to do that (no one should have these permissions, I believe), I meant in this PR

To master? No, I don't think you have the permissions to do that (no one should have these permissions, I believe), I meant in this PR

Oh right. I thought that @michaelpj 's invitation gave me write access (which seemed a lot, to be honest). Anyhow, I made a separate PR and you merged it, so we're good.

Sadly, I think some tests need to be rerun, there was a pipe problem.

You've got write access (you've contributed a lot already! :) ), but even people with write access can't commit directly to master, we require everything to go via PRs. You can merge things though!

Makes sense !