Opened 15 months ago

Last modified 9 months ago

#15270 new bug

TH doesn't verify name types during conversion

Reported by: bgamari Owned by:
Priority: normal Milestone: 8.10.1
Component: Template Haskell Version: 8.4.3
Keywords: Cc:
Operating System: Unknown/Multiple Architecture: Unknown/Multiple
Type of failure: None/Unknown Test Case:
Blocked By: Blocking:
Related Tickets: Differential Rev(s): Phab:D4859
Wiki Page:

Description (last modified by bgamari)

Angerman reported that a use of the `deriveEq` splice is causing GHC to abort with an assertion failure:

zonkExpr env (HsVar x (L l id))
  = ASSERT2( isNothing (isDataConId_maybe id), ppr id )
    return (HsVar x (L l (zonkIdOcc env id)))

I suspect the deriveEq1 is calling varE with a DataCon name. We should catch this case and throw a better error message.

Change History (10)

comment:1 Changed 15 months ago by RyanGlScott

If you could figure out which part of deriving-compat is doing something infelicitous with varE, let me know and I'll fix that separately. :)

comment:2 Changed 15 months ago by bgamari

I think my hypothesis may not be quite right, actually. For the record, the failing module is

comment:3 Changed 15 months ago by bgamari

Description: modified (diff)

comment:4 Changed 15 months ago by RyanGlScott

Actually, I think your hypothesis wasn't too far off, bgamari. Here's a minimal example which triggers the panic:

{-# LANGUAGE TemplateHaskell #-}
module Bug where

import Language.Haskell.TH.Lib
import Language.Haskell.TH.Syntax

main :: IO ()
main = print $(conE (mkNameG_v "ghc-prim" "GHC.Types" "True"))
$ ghc/inplace/bin/ghc-stage2 Bug2.hs
[1 of 1] Compiling Bug              ( Bug2.hs, Bug2.o )
ghc-stage2: panic! (the 'impossible' happened)
  (GHC version 8.5.20180616 for x86_64-unknown-linux):
        ASSERT failed!
  Call stack:
      CallStack (from HasCallStack):
        callStackDoc, called at compiler/utils/Outputable.hs:1164:37 in ghc:Outputable
        pprPanic, called at compiler/utils/Outputable.hs:1223:5 in ghc:Outputable
        assertPprPanic, called at compiler/typecheck/TcHsSyn.hs:655:67 in ghc:TcHsSyn

It turns out that I was using conE (mkNameG_v "ghc-prim" "GHC.Types" "True") in deriving-compat. Oops!

comment:5 Changed 15 months ago by bgamari

Interesting, I excluded my hypothesis after noticing that the following testcase indeed fails with a proper error:

{-# LANGUAGE TemplateHaskell #-}

module T10047A where

import Language.Haskell.TH

-- Passing var name to conE should fail.
x = $(conE 'id)

I didn't look much farther after this.

Quite interesting that mkNameG_v gets around the error. Are you going to continue looking into this or should I add it to my queue?

comment:6 Changed 15 months ago by RyanGlScott

Indeed, this is an odd one. Some other observations:

  • conE 'id fails with:
    • Illegal data constructor name: ‘id’
      When splicing a TH expression:
    • In the untyped splice: $(conE 'id)
  • varE 'True fails with:
    • Illegal variable name: ‘True’
      When splicing a TH expression: GHC.Types.True
    • In the untyped splice: $(varE 'True)
  • Finally, $(varE (mkNameG_d "base" "GHC.Base" "id")) fails with:
    • Can't find interface-file declaration for data constructor
        Probable cause: bug in .hi-boot file, or inconsistent .hi file
        Use -ddump-if-trace to get an idea of which file caused the error
    • In the expression: (
      In an equation for ‘it’: it = (

I suppose we want something like that last error if we try to use $(conE (mkNameG_v "ghc-prim" "GHC.Types" "True")) as in comment:4. That being said, I can't figure out why we don't currently get that error. After all, that error message is caused by looking up a Name in an EPS and failing, so if $(conE (mkNameG_v "ghc-prim" "GHC.Types" "True")) isn't erroring, does that mean that a value named True is located in the EPS?

I'm afraid I don't know how to proceed at this point, so do you think you could look at this, Ben?

comment:7 Changed 15 months ago by bgamari

Differential Rev(s): Phab:D4859

Have the beginning of some tests as Phab:D4859.

comment:8 Changed 15 months ago by bgamari


These won't be addressed in GHC 8.6.

comment:9 Changed 9 months ago by Ben Gamari <ben@…>

In 18b74cdb/ghc:

testsuite: Add tests for #15270

Reviewers: alpmestan

Reviewed By: alpmestan

Subscribers: rwbarton, carter

Differential Revision:

comment:10 Changed 9 months ago by osa1


Bumping milestones of low-priority tickets.

Note: See TracTickets for help on using tickets.