#8310: Can we change the semantics of `Trustworthy`? -------------------------------------+------------------------------------ Reporter: ekmett | Owner: dterei Type: feature request | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.6.3 Resolution: | Keywords: Operating System: Unknown/Multiple | Architecture: Unknown/Multiple Type of failure: None/Unknown | Difficulty: Unknown Test Case: | Blocked By: Blocking: | Related Tickets: -------------------------------------+------------------------------------ Comment (by dterei): @ekmett yes importing a Trustworthy module still allows you to be inferred as `Safe-Inferred`. The question is, marking a module as `Trustworthy` has a different set of package-trust dependencies today than if you had left that module unmarked and let it be inferred as `Safe-Inferred`. If we change `Trustworthy` to be 'at-least' trustworthy and allow those modules to still be inferred `Safe-Inferred`, then what package-trust dependencies should apply? That isn't clear. The warning is a good idea and makes sense. I think the 'at-least' semantics may also be a good idea, just need to think through what package-trust dependencies should be applied and if that complication to the semantics is worth enough. -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/8310#comment:5 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler