-
-
Notifications
You must be signed in to change notification settings - Fork 369
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Prepare 0.7.0 release (includes bump ghcide to 0.6.0.1) #658
Conversation
jneira
commented
Dec 9, 2020
•
edited
Loading
edited
- This will not be the final commit (ghcide has to be released yet) but let's see what has to say ci.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Happy with this, once CI passes
Awesome! CI says there is an ambiguity for |
stack build in circleci using stack-8.8.3.yaml and stack-8.8.2.yaml (but no for stack-8.8.4.yaml) is failing with
As ghcide is not being built with ghc-8.8.3 or ghc-8.8.2 we dont get direct evidence if this error would be reproduced building ghcide standalone 🤔 stack-8.6.5.yaml and stack-8.6.4.yaml had built ghcide succesfully though. EDIT: github test builds with cabal and ghc-8.8.3 and ghc-8.8.2 has failed with the same error |
I've leveraged the new |
Some of nix builds (default, ghc8102) in github are failing with
I guess we have to change nix config files but i dont know how |
Does the comment at the top of haskell-language-server/shell.nix Lines 4 to 10 in 9120f4b
|
@pepeiborra yeah, I missed that config file, only took a quick look inside the nix dir |
Ok, two issues blocks the pr now:
@pepeiborra what do you think about add GADT's to the ghcide module to check if it fixes the issue and use that commit in hls? Ideally that version should be the last released one in hackage @peterwicksstringfield could you take a quick look to completion tetsts to see if they can be fixed without too much effort? |
I am wondering if the corners being cut in ghcide CI are worth it. |
Not related. We have cut some |
We will have to add 8.8.2 and 8.8.3 to the test matrix in ghcide CI. I'll see to that. |
Many thanks, ideally ghcide should have the same matrix than hls, at least with cabal. So maybe we should add ghc-8.6.4 and ghc-8.10.1 too as they are missing in ghcide afair. |
|
I have no objections to re-disabling these tests for now. In #668 I am attempting to fix the tests. |
You might have already seen it, but the failure on (ci/circleci: ghc-8.8.2) is something about a dependency on apply-refact. |
Not sure why
to
but it seems the actual behaviour is right iirc so i've updated the golden files. |
Nix is failing now with
I hope it is a transient one |
I think #672 should be the last pr to be merged before cut the release version so please, dont merge additional pr's to master until the tag is created (if you think someone else should be included comment it out here or in the pr itself) |
Waiting for ci... 🙄 |
completion/Context.hs used to generate a diagnostic for the unused "x", but no longer does so without -Wunused-binds. completion/Completion.hs used to generate a diagnostic for the unused import of Data.Maybe, but no longer does so without -Wunused-imports. We could add these flags like: {-# OPTIONS_GHC -Wunused-binds #-} But that would force us to update all the hardcoded line numbers in the unit tests. Instead we just add a redundant id, carefully positioned to avoid disturbing any hardcoded positions.
Snippets are back in ghcide 0.6.0.
Stop messing around the editing the file and reloading it. We are trying to test code completion, not editing files, and the extra complexity is making it harder to maintain the test.
|