amesgen

@hasufell That is the same issue as in https://github.com/cdepillabout/servant-rawm/issues/7, see the penultimate comment for a solution. Actually, it is one of the motivations I mentioned [above](https://github.com/haskell-servant/servant/pull/1551#issuecomment-1064110949) to merge this PR:...

We now disabled the test in #3942 for the time being, but the core issue remains.

With #3743, this will no longer annoy us in CI for the time being, but this should be revisited in the future.

The team discussed this today, and came to the following conclusions: 1. Reduce the ChainDB API by making `getIsValid` redundant: #3974 2. We agreed that not reverting #3743 in some...

1. People could write very weird `TH.Lift` instances. :grin: 2. A bit less contrived: Consider ```haskell test = (1 ,2) ``` which one might try to format via `ormolu --start-line...

@buggymcbugfix The cases you mention are already correctly handled by Ormolu, or am I missing something in your comment?

Yeah, but the cases you mention are not top-level tuples/lists, but different constructs. But the asymmetry between tuple expressions and tuple patterns is indeed surprising: ```haskell a = (a ,a)...

Hmm, I think that Ormolu's output is actually "Haddock-preserving" here. Consider these rendered Haddock outputs Source Rendered Haddock ```haskell module MyLib ( -- $intro -- * Connections Foo (..) )...

> That said, I do find Ormolu makes it easy to now just lose the comment entirely if you're not careful reading the diff Agreed; so we basically have two...

Yes, blank lines are sometimes useful for separation. To clarify, I would like both of my code examples above to be unchanged when formatted, as well as ```haskell data Foo...