GHC-5.04.2 generates faulty Core code?

[Simon Peyton-Jones]

Yes, this is a bug (GHC is failing to output some important implicit
definitions). It's fixed in the HEAD, so you can use that (if you build
it from source).  I'm not quite sure when it got fixed.

Simon

| -----Original Message-----
| From: Tobias Gedell [mailto:d99getob at dtek.chalmers.se]
| Sent: 20 January 2003 01:16
| To: glasgow-haskell-bugs at haskell.org
| Subject: GHC-5.04.2 generates faulty Core code?
| 
| I am currently working on my master thesis which is designing and
| implementing an usage analysis for Core. I have been working with
| ghc-5.02 but want to start use ghc-5.04.2 instead. The problem is that
| it seems like ghc-5.04.2 generates faulty Core code.
| 
| I need to generate Core for the entire prelude and parts of the hslibs
| modules.
| 
| What I do is:
| 
| ghc -fno-implicit-prelude -fno-code -cpp -fext-core -fglasgow-exts
Base.lhs
| 
| 
| Then I try to analyse Base.hcr using my own program. I then get an
error:
| 
| Base.hcr:1261:
| Invalid variable name in getCoreType, (Just "GHCziBase","zeze")
| 
| 
| This occurs since my program cannot find the type for GHCziBase.zeze
| which is refered to at line 1261 in Base.hcr.
| 
| If I look at that line in Base.hcr I find:
| 
| GHCziBase.zaza (GHCziBase.zeze @ a zddEq x y) (zeze5 xs ys)};
| 
| 
| But there is no definition for GHCziBase.zeze in the file Base.hcr. It
| seems to me that GHC generates faulty Core code.
| 
| Am I doing something wrong? Should I try to use another version of
GHC?
| 
| 
| Thankful for help
| 
| Sincerely,
|   Tobias
| 
| _______________________________________________
| Glasgow-haskell-bugs mailing list
| Glasgow-haskell-bugs at haskell.org
| http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs