unknown symbol `hs_hpc_module'

[guest]

During a :load command I got: <interactive>: /home/foo/Bar/Baz.o: unknown symbol `hs_hpc_module'

No test case available

[AndyGill]

(unconfirmed) This happens when someone compilers a module with -fhpc, then uses ghci. ghci and hpc do not mix. A quick recompile fixes things.

Perhaps fixed for ghc-6.9? We want to think about how we offer hpc inside ghci anyway, so I'll leave the feature request as a place holder.

[igloo]

Can we easily print a more helpful error in 6.8.2?

A real fix can wait for 6.10, like Andy says.

[igloo]

See also #1907.

[igloo]

At this stage I think we'll just wait for a proper fix.

[AndyGill]

This no longer prints a message, and works just fine, in HEAD.

I compiled a module using -fhpc, then used ghci on another module that used the original module and it ran just fine.

[simonmar]

This is happening with 6.10 and HEAD, currently, on the qq005 test (see log below). Andy - any chance you could look into this?

=====> qq005(hpc)
cd ./quasiquotation/qq005 && '/64playpen/simonmar/nightly/HEAD-cam-04-unx/x86_64-unknown-linux/ghc/stage3-inplace/ghc' -fforce-recomp -dcore-lint -dcmm-lint -Dx86_64_unknown_linux  -dno-debug-output --make -o qq005 Main -O -fhpc   >qq005.comp.stderr 2>&1
Compile failed (status 256) errors were:
[1 of 2] Compiling Expr             ( Expr.hs, Expr.o )
[2 of 2] Compiling Main             ( Main.hs, Main.o )
ghc: ./Expr.o: unknown symbol `hs_hpc_module'
Loading package ghc-prim ... linking ... done.
Loading package integer ... linking ... done.
Loading package base ... linking ... done.
Loading package syb ... linking ... done.
Loading package array-0.2.0.0 ... linking ... done.
Loading package packedstring-0.1.0.1 ... linking ... done.
Loading package containers-0.2.0.0 ... linking ... done.
Loading package pretty-1.0.1.0 ... linking ... done.
Loading package template-haskell ... linking ... done.
Loading package parsec-2.1.0.1 ... linking ... done.

[igloo]

See also #2730.

[igloo]

I've made a minimal testcase, hpc_ghc_ghci.

In libraries/hpc/tests/ghc_ghci:

$ ../../../../ghc/stage2-inplace/ghc -v0 --interactive B.hs
*B> b
<interactive>: ./A.o: unknown symbol `hs_hpc_module'

$ ../../../../ghc/stage2-inplace/ghc -v0 --interactive B.hs ../../../../rts/Hpc.o 


GHCi runtime linker: fatal error: I found a duplicate definition for symbol
   hs_hpc_rootModule
whilst processing object file
   ../../../../rts/Hpc.o
This could be caused by:
   * Loading two different object files which export the same symbol
   * Specifying the same object file twice on the GHCi command line
   * An incorrect `package.conf' entry, causing some object to be
     loaded twice.
GHCi cannot safely continue in this situation.  Exiting now.  Sorry.

[spl]

I ran this minimal testcase manually on GHC 6.8.3 and 6.10.1 using MacPorts installations, and it passed both. I also just verified that it failed on GHCi 6.10.1 on Windows.

I think this problem is related to #1907, which we have been experiencing with  EMGM using 6.8.3 and 6.10.1. For both #1907 and this test case, Linux and Windows have failed while MacPorts has not. We have been using GHC, not GHCi, for EMGM.

Has the problem been determined, yet?

[simonmar]

I'm validating the fix for this.

[simonmar]

Fixed:

Mon Mar 30 01:44:14 PDT 2009  Simon Marlow <marlowsd@gmail.com>
  * FIX #1779 and qq005: export hs_hpc_module

[igloo]

We won't be merging non-essential fixes into the 6.10 branch any more.