Re: impossible happened: splitTyConApp a{tv i6hr}
Here's another example of the same bug: $ ghc -O3 -threaded -package Futility --make scmp.hs -o scmp install scmp ~/bin/ Chasing modules from: scmp.hs Skipping Engines ( ./Engines.hs, ./Engines.o ) Skipping SCmp ( ./SCmp.hs, ./SCmp.o ) Compiling Fitting ( ./Fitting.hs, ./Fitting.o ) ghc-6.4.2: panic! (the `impossible' happened, GHC version 6.4.2): splitTyConApp a{tv a8OQ} Please report this as a compiler bug. See: http://www.haskell.org/ghc/reportabug The code: http://ofb.net/~frederik/scmp-ghc-bug.tar.gz http://ofb.net/~frederik/futility-ghc-bug.tar.gz It's also linked to my GSLHaskell2 library which is still exhibiting the bug. Frederik -- http://ofb.net/~frederik/ ___ Glasgow-haskell-bugs mailing list Glasgow-haskell-bugs@haskell.org http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs
RE: impossible happened: splitTyConApp a{tv i6hr}
OK, I can reproduce it now. (Previously I was not compiling with -O2.) It's a bug in 6.4.2 related to GADTs, specifically in the typing of interface files. You get a much more informative message if you use the flag -dcore-lint. I think it's fixed in the upcoming 6.4.3 release; at least it works for me. Can you try to download a snapshot distribution of 6.4.3 (the STABLE branch) from http://www.haskell.org/ghc/download.html and try that? Simon | -Original Message- | From: Frederik Eaton [mailto:[EMAIL PROTECTED] | Sent: 16 August 2006 15:48 | To: Simon Peyton-Jones | Subject: Re: impossible happened: splitTyConApp a{tv i6hr} | | Here: | | http://ofb.net/~frederik/gslhaskell2-ghc-bug1.tar.gz | | The file General.hs is the subdirectory Vector. | | The problem only happens when I edit General.hs and recompile... | | Frederik | | On Wed, Aug 16, 2006 at 03:30:02PM +0100, Simon Peyton-Jones wrote: | I'd like to reproduce this, but to do so I'll need either a cut-down | version, or all the source code need to get to the point where it fell | over. Could you manage the former, or if not, the latter? Thanks | | S | | | -Original Message- | | From: [EMAIL PROTECTED] | [mailto:[EMAIL PROTECTED] | | On Behalf Of Frederik Eaton | | Sent: 16 August 2006 14:52 | | To: glasgow-haskell-bugs@haskell.org | | Subject: impossible happened: splitTyConApp a{tv i6hr} | | | | I've attached the code. | | | | Skipping Vector ( Vector.hs, dist/build/Vector.o ) | | Compiling Vector.General ( ./Vector/General.hs, | dist/build/Vector/General.o ) | | | | ./Vector/General.hs:29:0: | | Warning: No explicit method nor default method for | `vectorUpdate' | | In the instance declaration for `Vector (GVector e) e' | | | | ./Vector/General.hs:29:0: | | Warning: No explicit method nor default method for `byRow' | | In the instance declaration for `Vector (GVector e) e' | | ghc-6.4.2: panic! (the `impossible' happened, GHC version 6.4.2): | | splitTyConApp a{tv i6hr} | | | | Please report this as a compiler bug. See: | |http://www.haskell.org/ghc/reportabug | | | | -- | | http://ofb.net/~frederik/ | | | -- | http://ofb.net/~frederik/ ___ Glasgow-haskell-bugs mailing list Glasgow-haskell-bugs@haskell.org http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs
Re: impossible happened: splitTyConApp a{tv i6hr}
Compiling with 6.4.3, I have no problems, even with optimisation on. I'll let you know if it happens again. Thanks! Frederik On Thu, Aug 17, 2006 at 05:47:21PM +0100, Simon Peyton-Jones wrote: OK, I can reproduce it now. (Previously I was not compiling with -O2.) It's a bug in 6.4.2 related to GADTs, specifically in the typing of interface files. You get a much more informative message if you use the flag -dcore-lint. I think it's fixed in the upcoming 6.4.3 release; at least it works for me. Can you try to download a snapshot distribution of 6.4.3 (the STABLE branch) from http://www.haskell.org/ghc/download.html and try that? Simon | -Original Message- | From: Frederik Eaton [mailto:[EMAIL PROTECTED] | Sent: 16 August 2006 15:48 | To: Simon Peyton-Jones | Subject: Re: impossible happened: splitTyConApp a{tv i6hr} | | Here: | | http://ofb.net/~frederik/gslhaskell2-ghc-bug1.tar.gz | | The file General.hs is the subdirectory Vector. | | The problem only happens when I edit General.hs and recompile... | | Frederik | | On Wed, Aug 16, 2006 at 03:30:02PM +0100, Simon Peyton-Jones wrote: | I'd like to reproduce this, but to do so I'll need either a cut-down | version, or all the source code need to get to the point where it fell | over. Could you manage the former, or if not, the latter? Thanks | | S | | | -Original Message- | | From: [EMAIL PROTECTED] | [mailto:[EMAIL PROTECTED] | | On Behalf Of Frederik Eaton | | Sent: 16 August 2006 14:52 | | To: glasgow-haskell-bugs@haskell.org | | Subject: impossible happened: splitTyConApp a{tv i6hr} | | | | I've attached the code. | | | | Skipping Vector ( Vector.hs, dist/build/Vector.o ) | | Compiling Vector.General ( ./Vector/General.hs, | dist/build/Vector/General.o ) | | | | ./Vector/General.hs:29:0: | | Warning: No explicit method nor default method for | `vectorUpdate' | | In the instance declaration for `Vector (GVector e) e' | | | | ./Vector/General.hs:29:0: | | Warning: No explicit method nor default method for `byRow' | | In the instance declaration for `Vector (GVector e) e' | | ghc-6.4.2: panic! (the `impossible' happened, GHC version 6.4.2): | | splitTyConApp a{tv i6hr} | | | | Please report this as a compiler bug. See: | |http://www.haskell.org/ghc/reportabug | | | | -- | | http://ofb.net/~frederik/ | | | -- | http://ofb.net/~frederik/ -- http://ofb.net/~frederik/ ___ Glasgow-haskell-bugs mailing list Glasgow-haskell-bugs@haskell.org http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs