Tim Chevalier wrote:
On 7/12/07, Jonathan Cast <[EMAIL PROTECTED]> wrote:
No. Of course not. Before making wild guesses about how GHC is
implementing
your code, read (and understand[1]) the STG paper:
http://research.microsoft.com/copyright/accept.asp?path=/users/simonpj/papers/spineless-tagless-gmachine.ps.gz&pub=34
In this particular case, reading the simplifier paper would probably
be more relevant:
http://research.microsoft.com/copyright/accept.asp?path=/users/simonpj/papers/comp-by-trans.ps.gz&pub=18
Or even just understanding the syntax of Core, really.
[1] Understanding the STG paper is not a requirement to using Haskell,
just to
making wild (incorrect) guesses about how the compiler's going to
treat your
code. But, of course, making wild (incorrect) guesses about how the
compiler's going to treat your code is not a requirement to using
Haskell.
Amen!
Cheers,
Tim
I realize now that it was inappropriate to pose a question to this list
before I knew the answer! :(
Anyway, it turned out that the two ghc -ddump-simpl outputs were in fact
identical, but only after identifier renaming. It would be very useful
to have a tool that takes two core files and attempts to rename
identifiers from the second to match the first. Is there already such a
smart-diff tool out there?
Dan
_______________________________________________
Haskell-Cafe mailing list
Haskell-Cafe@haskell.org
http://www.haskell.org/mailman/listinfo/haskell-cafe
