#11220: Stack overflow instead of type check failure in Servant route -------------------------------------+------------------------------------- Reporter: sopvop | Owner: Type: bug | Status: new Priority: high | Milestone: 8.0.1 Component: Compiler (Type | Version: 7.10.3 checker) | Resolution: | Keywords: Operating System: Unknown/Multiple | Architecture: Type of failure: Compile-time | Unknown/Multiple crash | Test Case: Blocked By: | Blocking: Related Tickets: | Differential Rev(s): Wiki Page: | -------------------------------------+------------------------------------- Changes (by bgamari): * priority: normal => high * failure: None/Unknown => Compile-time crash * milestone: => 8.0.1 -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/11220#comment:1 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler

#11220: Stack overflow instead of type check failure in Servant route -------------------------------------+------------------------------------- Reporter: sopvop | Owner: Type: bug | Status: new Priority: high | Milestone: 8.0.1 Component: Compiler (Type | Version: 7.10.3 checker) | Resolution: | Keywords: Operating System: Unknown/Multiple | Architecture: Type of failure: Compile-time | Unknown/Multiple crash | Test Case: Blocked By: | Blocking: Related Tickets: | Differential Rev(s): Wiki Page: | -------------------------------------+------------------------------------- Changes (by bgamari): * cc: goldfire (added) Old description:

GHC 7.10.3 eats a lot of memory and segfaults compiling this.

{{{#!hs {-# LANGUAGE DataKinds #-} {-# LANGUAGE EmptyDataDecls #-} {-# LANGUAGE TypeOperators #-} module Main where

import Data.Proxy

import Servant.API -- requires servant

data A

apiP :: Proxy Api apiP = Proxy

-- Users

type Api = "variants" :> Get '[JSON] ()

routeL :: URI routeL = safeLink apiP (Proxy :: Proxy ("variants" :> Get '[A] ())) -- Should result in type error -- '[JSON] in route but '[A] here

main = print routeL

}}}

There is '[ JSON ] in Api type, but '[A] in Proxy type passed to safeLink, so it should result in type error.

Correct program compiles without problems.

New description: GHC 7.10.3 eats a lot of memory and segfaults compiling this. {{{#!hs {-# LANGUAGE DataKinds #-} {-# LANGUAGE EmptyDataDecls #-} {-# LANGUAGE TypeOperators #-} module Main where import Data.Proxy import Servant.API -- requires servant data A apiP :: Proxy Api apiP = Proxy -- Users type Api = "variants" :> Get '[JSON] () routeL :: URI routeL = safeLink apiP (Proxy :: Proxy ("variants" :> Get '[A] ())) -- Should result in type error -- '[JSON] in route but '[A] here main = print routeL }}} There is `'[ JSON ]` in `Api` type, but `'[A]` in `Proxy` type passed to `safeLink`, so it should result in type error. Correct program compiles without problems. -- -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/11220#comment:3 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler

#11220: Stack overflow instead of type check failure in Servant route -------------------------------------+------------------------------------- Reporter: sopvop | Owner: Type: bug | Status: new Priority: high | Milestone: 8.0.1 Component: Compiler (Type | Version: 7.10.3 checker) | Resolution: | Keywords: Operating System: Unknown/Multiple | Architecture: Type of failure: Compile-time | Unknown/Multiple crash | Test Case: Blocked By: | Blocking: Related Tickets: | Differential Rev(s): Wiki Page: | -------------------------------------+------------------------------------- Comment (by goldfire): The new error is actually somewhat informative. I don't have the definitions of `IsSubList` and `And` to hand (though they're easy to imagine). Could they loop? If they are the simple definitions I imagine, this is indeed a GHC bug. If they reasonably can loop, on the other hand, then this error message might be acceptable. Need feedback from the original poster, and a test case that doesn't depend on all of `servant` would be great. (Just inline the definitions into this module?) -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/11220#comment:5 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler

#11220: Stack overflow instead of type check failure in Servant route -------------------------------------+------------------------------------- Reporter: sopvop | Owner: Type: bug | Status: new Priority: high | Milestone: 8.0.1 Component: Compiler (Type | Version: 7.10.3 checker) | Resolution: | Keywords: Operating System: Unknown/Multiple | Architecture: Type of failure: Compile-time | Unknown/Multiple crash | Test Case: Blocked By: | Blocking: Related Tickets: | Differential Rev(s): Wiki Page: | -------------------------------------+------------------------------------- Comment (by goldfire): You have this definition: {{{ type family IsSubList a b :: Constraint where IsSubList '[] b = () IsSubList '[x] (x ': xs) = () IsSubList '[x] (y ': ys) = IsSubList '[x] ys IsSubList (x ': xs) y = IsSubList '[x] y `And` IsSubList xs y }}} But `IsSubList '[Int] '[]` loops by triggering the fourth equation (and being apart from all previous ones). Note that the left conjunct of the `And` is the exact same thing that we started with. So I think GHC's behavior here is quite reasonable. -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/11220#comment:6 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler

#11220: Stack overflow instead of type check failure in Servant route -------------------------------------+------------------------------------- Reporter: sopvop | Owner: Type: bug | Status: closed Priority: high | Milestone: 8.0.1 Component: Compiler (Type | Version: 7.10.3 checker) | Resolution: invalid | Keywords: Operating System: Unknown/Multiple | Architecture: Type of failure: Compile-time | Unknown/Multiple crash | Test Case: Blocked By: | Blocking: Related Tickets: | Differential Rev(s): Wiki Page: | -------------------------------------+------------------------------------- Changes (by goldfire): * status: new => closed * resolution: => invalid Comment: OK. I'm closing this as invalid. Do reopen if you disagree. -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/11220#comment:8 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler