#14899: Significant compilation time regression between 8.4 and HEAD due to coverage checking -------------------------------------+------------------------------------- Reporter: RyanGlScott | Owner: (none) Type: bug | Status: new Priority: highest | Milestone: 8.6.1 Component: Compiler | Version: 8.5 Resolution: | Keywords: | PatternMatchWarnings Operating System: Unknown/Multiple | Architecture: | Unknown/Multiple Type of failure: None/Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | Differential Rev(s): Wiki Page: | -------------------------------------+------------------------------------- Comment (by RyanGlScott): This is bad. Really bad. I looked briefly into the GHC source, and found [http://git.haskell.org/ghc.git/blob/d9d463289fe20316cff12a8f0dbf414db678fa72... this ominous Note]: {{{ Note [Translate CoPats] ~~~~~~~~~~~~~~~~~~~~~~~ The pattern match checker did not know how to handle coerced patterns `CoPat` efficiently, which gave rise to #11276. The original approach translated `CoPat`s: pat |> co ===> x (pat <- (e |> co)) Instead, we now check whether the coercion is a hole or if it is just refl, in which case we can drop it. Unfortunately, data families generate useful coercions so guards are still generated in these cases and checking data families is not really efficient. }}} If that is to be believed, then is coverage-checking data family instances really doomed to be slow? -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/14899#comment:2 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler