Fix #5768, Corrected wrong formatted print of a trait member constraint.

[realvictorprm]

The bug was not that the constraint is ignored. It was that the compiler printed the suggested trait constraint wrongly.

As example, the signature a * b -> (c -> d) isn't the same as a * b -> c -> d because the return type differs (in case one it's (c -> d), in case two it's d). The compiler however didn't respect that behaviour and prints constraints of case 1 always like case 2 which is wrong. The fix was to enforce that trait constraint return type is always wrapped in brackets if the return type is a function.

@dsyme please have a look whether this assumption is fine? It's a short thing only about formatting.

Signed-off-by: realvictorprm mueller.vpr@gmail.com

[realvictorprm]

@cartermp I won't the neat command "@dotnetbot please test this" back 😭 Closing isn't nice.

[realvictorprm]

CI passes. This is ready.

[saul]

Can we add the example in #5768 as a test case to make sure we don't regress?

[realvictorprm]

I'll have a look how I can add it.

[dsyme]

Needs test, otherwise approved

[realvictorprm]

Regression test added.

[dsyme]

The fix is fine. The testing could be simpler I believe

[realvictorprm]

@dsyme I've changed it, please have a short look over it.

[realvictorprm]

CI passes, can this be merged?

[dsyme]

approved, please resolve conflict and we can merge

[realvictorprm]

done

[cartermp]

Ergh, looks like there were some build issues preventing this from getting merged and the codebase has diverged since. Sorry about that. @realvictorprm are you still interested in pursuing this? In theory it should be easy to adjust the tests and just copy the contents of this branch's neg111.fs and neg111.bsl into new files.

[abelbraaksma]

@cartermp, in an attempt not to lose this work, I've resurrected it here: #9728 (branched off this one to keep history intact). It does exactly what you suggest with the conflicting files.

[cartermp]

Thanks! Will close in favor of #9728