When I remove the strictness annotations and compile and run the code, I get the following runtime error: demo_code.c:2863: case fell off I know how to leave the C code, so I can see where this error comes from, but it's opaque enough that I've no idea what's going wrong. :( David On Fri, Sep 4, 2009 at 7:08 AM, David Roundy wrote:

Agreed.  Removing all strictness annotations from data structures eliminates the error.

David

On Thu, Sep 3, 2009 at 11:23 PM, John Meacham wrote:

...

Hmm.. it looks like that for some reason not all of the TID newtype constructors are being removed properly. I'll take a look. I am thinking it might have something to do with the strictness annotation and UNPACKing of strict fields... hmm...

       John

-- John Meacham - ⑆repetae.net⑆john⑈ - http://notanumber.net/ _______________________________________________ jhc mailing list jhc@haskell.org http://www.haskell.org/mailman/listinfo/jhc

-- David Roundy

-- David Roundy

It turned out that after a few recompiles of jhc, this "case fell off" problem disappears. Presumably it was something that I messed up... although I've no idea what that might have been. But as you've gathered, I've been messing with the libraries and recompiling pretty often... and not always remembering to do a make install. David On Fri, Sep 4, 2009 at 9:13 PM, John Meacham wrote:

On Fri, Sep 04, 2009 at 07:18:53AM -0400, David Roundy wrote:

...

When I remove the strictness annotations and compile and run the code, I get the following runtime error:

demo_code.c:2863: case fell off

I know how to leave the C code, so I can see where this error comes from, but it's opaque enough that I've no idea what's going wrong.  :(

Hmm.. this shouldn't happen. When debugging it is also a good idea to do a -dgrin-final and -dc, reading the final grin code is often easier than the c, variable numbers stay the same between the versions, so finding cooresponding places is pretty easy just by searching for the variable names.

I tried to reproduce your test case but wasn't able to. do you have a version that reliably triggers this? are you compiling with -flint? that is usually a good thing to do if you suspect a bug in the compiler.

       John

-- John Meacham - ⑆repetae.net⑆john⑈ - http://notanumber.net/ _______________________________________________ jhc mailing list jhc@haskell.org http://www.haskell.org/mailman/listinfo/jhc

-- David Roundy