Closed andreasabel closed 10 months ago
Thanks. Better to remove this code completely: the comment in the commit message is sufficient. Even better would be to extend the existing message so that it matches its original targets but not the new "refactor" message. The tricky thing is finding examples of the former!
(Just passing by) Is it difficult for haskell-mode
to filter anything remotely suspicious to be a language extension through ghc --supported-extensions
? This would be a more robust solution than adjusting regexps here and there.
@purcell : Thanks for the quick review. I will leave the decision to you how to fix the problem. @Bodigrim 's advice is good; the current heuristics is just too brittle.
Sounds reasonable. I think there may even be a list of supported extensions in the code which gets updated periodically — using that would be better than a fresh process invocation if so. I'm unlikely to work on this myself though, sorry.
@purcell : You may at least merge my hotfix...
You may at least merge my hotfix...
Thanks! 😉
Alright, have applied my suggestion above and will merge once CI passes. Might have to close and re-open this PR briefly to trigger a rebase on top of changes I just made in master
to fix CI.
Thanks Steve @purcell!
Remove bad regex that was supposed to recognize a LANGUAGE suggestion.
This regex fires unfortunately on GHC 9.8's new
x-partial
warning, destroying the UX with obnoxious questions whether to "add LANGUAGE Data".Closes #1834.