atomb at soe.ucsc.edu
Mon Nov 19 22:16:07 EST 2007
Actually, that file should be fine. I updated the External Core output
path to include all of the constructs in modern GHC core. It's the
input path I've been slow to finish (and, when I do finish it, the
output format will likely be totally different, too :) )
So, there's little formal documentation of the format, but the files
can be generated, and you can use them if you like. The document Tim
pointed to describes something pretty close, but not identical, to
what it GHC currently generates.
However, I would be wary of investing too much time in writing code to
process the current .hcr files, because they're likely to change
significantly in 6.10.
On Nov 19, 2007, at 6:53 PM, Guilherme Avelino wrote:
> I didn´t understand. I´m using the GHC 6.8.1 and when I use -fext-
> core is generated a file (.hcr). Is there any problem with the
> generated code? Which version should I use?
> 2007/11/19, Tim Chevalier <catamorphism at gmail.com>: On 11/19/07,
> Guilherme Avelino <gavelino at gmail.com> wrote:
> > I have two questions about the code generated by -fext-core:
> > Is there one place where I can obtain the grammar for it?
> > How is represented the free-vars?
> "An External Representation for the GHC Core Language" (2001)
> Andrew Tolmach
> However, as Simon noted, -fext-core is not working in recent versions
> of GHC. (If you don't mind using an older version, then you would be
> able to use it.) We hope to have it working for 6.10.
> Tim Chevalier * catamorphism.org * Often in error, never in doubt
> "It is easy to keep secrets by being honest in an ironic tone of
> voice." -- Andrew Solomon
> Cvs-ghc mailing list
> Cvs-ghc at haskell.org
More information about the Cvs-ghc