How bad is that? Does tooling really parse error messages? (If so
#11208: GHCi doesn't qualify types anymore -------------------------------------+------------------------------------- Reporter: hvr | Owner: Type: bug | Status: new Priority: highest | Milestone: 8.0.1 Component: GHCi | Version: 7.11 Resolution: | Keywords: regression Operating System: Unknown/Multiple | Architecture: | Unknown/Multiple Type of failure: Other | Test Case: Blocked By: | Blocking: Related Tickets: | Differential Rev(s): Wiki Page: | -------------------------------------+------------------------------------- Comment (by hvr): Replying to [comment:4 simonpj]: there should be a Better Way!) While there *should* be a better way, there isn't yet... GHC doesn't ship with anything better than GHC(i) right now Here's two ways how this regression affects Emacs' haskell-mode: - For warnings about missing type-signatures, Emacs parses the warning message and then inserts the extracted missing type-signature - Another feature is based on GHCi, which queries the symbol under the cursors for its type, and then inserts that above as a type-signature There may be other features I may have forgotten about which also require to know renamer-information. In both cases it's important to know how to qualify (if needed) symbols. And in passive mode (e.g. when parsing GHC's warning messages w/o an active GHCi session attached) Emacs has no way to query the module namespace. In active mode (e.g. when using GHCi to resolve/introspect information) Emacs uses `:type`/`:browse`/`:info` as well as other commands newly added (soon) to GHC 8.0 (see #10874). Those now are missing the module qualifications -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/11208#comment:5 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler