Ticket #1845 (new bug)

Opened 1 year ago

Last modified 4 months ago

unconditional relative branch out of range (GHC version 6.8.1/6.8.2 for powerpc_apple_darwin)

Reported by: guest Assigned to: thorkilnaur
Priority: normal Milestone: 6.10.1
Component: Compiler Version: 6.8.2
Severity: major Keywords:
Cc: Christian.Maeder@dfki.de, alfonso.acosta@gmail.com Difficulty: Unknown
Test Case: Architecture: powerpc
Operating System: MacOS X

Description

Having build ghc-6.8.1 from sources revealed the following problem after installation: 

> ghci -package ghc
GHCi, version 6.8.1: http://www.haskell.org/ghc/  :? for help
Loading package base ... linking ... done.
Loading package old-locale-1.0.0.0 ... linking ... done.
Loading package old-time-1.0.0.0 ... linking ... done.
Loading package filepath-1.1.0.0 ... linking ... done.
Loading package directory-1.0.0.0 ... linking ... done.
Loading package array-0.1.0.0 ... linking ... done.
Loading package containers-0.1.0.0 ... linking ... done.
Loading package hpc-0.5.0.0 ... linking ... done.
Loading package bytestring-0.9.0.1 ... linking ... done.
Loading package pretty-1.0.0.0 ... linking ... done.
Loading package packedstring-0.1.0.0 ... linking ... done.
Loading package template-haskell ... linking ... done.
Loading package unix-2.2.0.0 ... linking ... done.
Loading package process-1.0.0.0 ... linking ... done.
Loading package readline-1.0.1.0 ... linking ... done.
Loading package Cabal-1.2.2.0 ... linking ... done.
Loading package random-1.0.0.0 ... linking ... done.
Loading package haskell98 ... linking ... done.
ghc-6.8.1: internal error: unconditional relative branch out of range: jump island out of range
    (GHC version 6.8.1 for powerpc_apple_darwin)
    Please report this as a GHC bug:  http://www.haskell.org/ghc/reportabug
Abort trap

Change History

11/07/07 02:36:26 changed by guest

  • cc set to Christian.Maeder@dfki.de.

11/08/07 07:15:12 changed by igloo

  • difficulty set to Unknown.
  • milestone set to 6.8.2.

Thanks for the report!

11/09/07 05:37:04 changed by ChrisKuklewicz

This *might* be related to bug #1843 on OS X 10.5 (Leopard) on ppc.

11/24/07 14:50:41 changed by thorkilnaur

I can reproduce this using the binary package http://haskell.org/ghc/dist/6.8.1/maeder/ghc-6.8.1-powerpc-apple-darwin.tar.bz2 by Christian Maeder, both on PPC Mac OS X 10.4 Tiger and PPC Mac OS X 10.5 Leopard.

Best regards Thorkil

12/18/07 08:36:46 changed by igloo

  • milestone changed from 6.8.2 to 6.8.3.

12/19/07 04:52:16 changed by thorkilnaur

  • owner set to thorkilnaur.

I will see if I can dig out some additional details here.

Best regards Thorkil

12/30/07 13:04:08 changed by maeder

I'm currently trying if passing -mlongcall to gcc helps. http://developer.apple.com/documentation/DeveloperTools/gcc-4.0.1/gcc/RS_002f6000-and-PowerPC-Options.html

01/02/08 04:10:19 changed by maeder

Just adding -mlongcall to extra-gcc-opts did not help. When trying to compile with 

SRC_HC_OPTS += -fvia-C

in mk/build.mk I got the following error: 

 ../compiler/ghc-inplace -H16m -O -fvia-C -optc-O2 -package-name rts -static -I../gmp/gmpbuild -I. -#include HCIncludes.h -dcmm-lint     -c Apply.cmm -o Apply.o
/tmp/ghc24169_0/ghc24169_0.s:unknown:Undefined local symbol L___DISCARD__$stub
make[1]: *** [Apply.o] Error 1
make: *** [stage1] Error 1

(follow-up: ↓ 10 ) 04/18/08 16:48:37 changed by fons

  • cc changed from Christian.Maeder@dfki.de to Christian.Maeder@dfki.de, alfonso.acosta@gmail.com.
  • version changed from 6.8.1 to 6.8.2.
  • component changed from GHCi to Compiler.
  • severity changed from normal to major.
  • summary changed from unconditional relative branch out of range (GHC version 6.8.1 for powerpc_apple_darwin) to unconditional relative branch out of range (GHC version 6.8.1/6.8.2 for powerpc_apple_darwin).

I'm suffering exactly the same problem with ghc 6.8.2 in Leopard/ppcG4.

I can reproduce the bug both with macports ghc and http://www.informatik.uni-bremen.de/agbkb/forschung/formal_methods/CoFI/hets/mac/ghcs/ghc-6.8.2-powerpc-apple-darwin-static-libs.tar.bz2

Furthermore, the problem is not limited to interactive mode. Compilation fails for this cabal package http://hackage.haskell.org/cgi-bin/hackage-scripts/package/type-level-0.1

Removing the genAliases splice from Data.TypeLevel?.Num.Aliases fixes the problem.

(in reply to: ↑ 9 ) 04/18/08 17:19:57 changed by fons

Replying to fons:

Removing the genAliases splice from Data.TypeLevel?.Num.Aliases fixes the problem.

The problem can also be fixed by removing the unnnecesary ghc package dependency from the cabal description (its intention is to make clear to other compilers that the code was ghc-dependent)

This makes more sense.

However, other cabal packages with the ghc dependency can be compiled without problems (e.g. http://hackage.haskell.org/cgi-bin/hackage-scripts/package/parameterized-data-0.1 )

04/22/08 11:01:25 changed by thorkilnaur

Thanks a lot for this information. In addition, the problem occurs for test case ghci024(ghci) as well, both HEAD and STABLE, for the tnaur PPC OSX builders which are Mac OS X 10.4 (Tiger).

Best regards Thorkil

06/20/08 10:28:53 changed by igloo

  • milestone changed from 6.8.3 to 6.10.1.