Re: [GHC] #15783: Quoting an internal variable causes an error when splicing

19 Oct 2018

      #15783: Quoting an internal variable causes an error when splicing
-------------------------------------+-------------------------------------
        Reporter:  mpickering        |                Owner:  (none)
            Type:  bug               |               Status:  new
        Priority:  normal            |            Milestone:
       Component:  Template Haskell  |              Version:  8.6.1
      Resolution:                    |             Keywords:
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):

 Another curious thing. If you compile `B.hs` with `-ddump-simpl`, you'll
 get this:

 {{{
 $ /opt/ghc/8.6.1/bin/ghc B.hs -fforce-recomp -ddump-simpl
 [1 of 1] Compiling B                ( B.hs, B.o )

 ==================== Tidy Core ====================
 Result size of Tidy Core
   = {terms: 24, types: 9, coercions: 0, joins: 0/0}

 -- RHS size: {terms: 9, types: 1, coercions: 0, joins: 0/0}
 f :: Language.Haskell.TH.Syntax.Q
        (Language.Haskell.TH.Syntax.TExp Integer)
 [GblId]
 f = Language.Haskell.TH.Syntax.unsafeTExpCoerce
       @ Integer
       (Language.Haskell.TH.Lib.Internal.varE
          (Language.Haskell.TH.Syntax.mkNameG_v
             (GHC.CString.unpackCString# "main"#)
             (GHC.CString.unpackCString# "B"#)
             (GHC.CString.unpackCString# "d"#)))

 <Typeable stuff elided>
 }}}

 Now change the typed quote to an untyped one (i.e., change `f = [|| d ||]`
 to `f = [| d |]`) and recompile it:

 {{{
 $ /opt/ghc/8.6.1/bin/ghc B.hs -fforce-recomp -ddump-simpl
 [1 of 1] Compiling B                ( B.hs, B.o )

 ==================== Tidy Core ====================
 Result size of Tidy Core
   = {terms: 25, types: 7, coercions: 0, joins: 0/0}

 -- RHS size: {terms: 8, types: 0, coercions: 0, joins: 0/0}
 f :: Language.Haskell.TH.Lib.Internal.ExpQ
 [GblId]
 f = Language.Haskell.TH.Lib.Internal.varE
       (Language.Haskell.TH.Syntax.mkNameG_v
          (GHC.CString.unpackCString# "main"#)
          (GHC.CString.unpackCString# "B"#)
          (GHC.CString.unpackCString# "d"#))

 -- RHS size: {terms: 1, types: 0, coercions: 0, joins: 0/0}
 d :: Integer
 [GblId, Caf=NoCafRefs, Unf=OtherCon []]
 d = 0

 <Typeable stuff elided>
 }}}

 Notice that `d` actually makes an appearance this time! This makes me
 wonder if `d` is being mistakenly removed under the pretense of it being
 dead code.

-- 
Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/15783#comment:3
GHC http://www.haskell.org/ghc/
The Glasgow Haskell Compiler