On 16/03/2010 22:01, Matthias Kilian wrote:
On Mon, Mar 15, 2010 at 09:50:41AM +0000, Simon Marlow wrote:
Does anyone here care about portability? Some time ago (about three
years) someone told me that portability is an imporant goal for
ghc. Given that you now even need ghci to build ghc, this portability
claim is obviously a lie.
Woah, slow down there.
No, sorry. If I'm seeing unnecessary requirements and all those
little hurdles that get in your way trying to get GHC work on a
non-first-tier platform, I get really upset. For sure I appreciate
your work on the new build system, but what else happened/who else
cared?
The DPH package uses annotations, and annotations need GHCi, yes. GHCi
works on all the platforms we support, doesn't it? In particular, it
works on x86_64-unknown-openbsd, the platform you're on.
I've 7 machines at home. One amd64 (or x86_64), two i386. For the
other machines, I'd like to get GHC running some day far in the
future, and it's a royal PITA if you've to work around all those
little tiny itches like this dependency on GHCi.
I've actually been proceeding on the assumption that we can pretty
easily get GHCi working everywhere, so it's not a problem. That's why I
- made GHCi work unregisterised
- imported libffi
these measures were done precisely *for* portability, so forgive me if
I'm a little put off by your complaints!
Admittedly the linker is a slight sticking point, but for an ELF
platform it's not hard to make it work. Eventually you'd want it
anyway. If it really is a problem then we'll need to pursue the idea of
using shared libraries.
Cheers,
Simon
_______________________________________________
Cvs-ghc mailing list
Cvs-ghc@haskell.org
http://www.haskell.org/mailman/listinfo/cvs-ghc
