
Hello, I was just trying to cross compile ghc for my new amd64 box and was just wondering to what stage others have gotten with it? Personally I get a working ghc-inplace, and then I've tried to methods, neither of which work. First, I get a fresh source tree (all compilations were with 6.2.1) and do a configure passing --with-ghc=/path/to/ghc-inline, this works pretty well for a while, and eventually it cops out when it tries to find Data.Array and can't. Tried some other things with this method and they only got me marginally furhter (but I'm fairly sure it wasn't the proper way to go about it). Next, I tried, after using the supplied hc-boot script (which did complete), to do a make install... this also fails with: ghc-6.2.1: error: directory portion of "stage2/absCSyn/AbsCSyn.o" does not exist (used with "-o" option.) So stage2 isn't there, I verified that, but i'm not really familiar enough to know when/where it is created. Can anyone shed any light on this? Thanks!

Hello, scott west wrote:
Next, I tried, after using the supplied hc-boot script (which did complete), to do a make install... this also fails with: ghc-6.2.1: error: directory portion of "stage2/absCSyn/AbsCSyn.o" does not exist (used with "-o" option.)
As outlined in "Porting GHC", I did a second build --with-ghc=.../ghc/compiler/ghc-inplace, but this second build produces just corrupted binaries (they segfault immediatly). I'd be very happy if someone could provide a binary build for Linux/amd64 like there is for Linux/<almost every architecure>. Gabriel. -- Gabriel Ebner - reverse "ta.renbeleirbag@eg"

Well, I just poked and proded the build a bit more, and got it to complete and install, with the same result as you. Seg faults immediately and that's it. Hmm, any else had any experience with this?
participants (2)
-
Gabriel Ebner
-
scott west