http://www.cse.unsw.edu.au/~sseefried/files/ghc-nightly/logs/STABLE- cmis
**** Build finished successfully at Tue Sep 28 06:44:33 EST 2004
Looks like you have some modifications to the nightly build scripts. Is
there anything we should merge into the CVS sources?
No. Actually I've written my own versions of the the build
scripts. You may wonder why I bothered to do this. The nightly build system uses a whole lot of options that correspond directly to something you can just write manually into a build.mk, but with slightly different names.
Back when I was compiling GHC manually I used to do it by writing two things: a custom configure script, which is simply a wrapper around a call to configure with whatever flags I wish to pass to it, and a build.mk.
I'd then try to build it. If there were any problems I'd usually just twiddle a couple of flags in the custom configure script or build.mk and try again. Naturally I didn't have to start from scratch though - many of the file had already compiled.
I got annoyed that every time something went wrong when I was using 'nightly' build scripts I had to twiddle a couple of variables and start again. And I was having a lot of trouble getting GHC to build on my Mac.
My nightly build system, which I have called buildghc, is a very simple set of scripts that take a custom configure and a build.mk, put them in the right place, and then orchestrate the build process (much like 'nightly'). But the thing is I can work out whether the build will work first manually and once that works simply copy the custom configure and build.mk to the 'site' directory of my buildghc system. And this time I'll know it's going to work.
I've tried to make the output look much like that of 'nightly' so as not to confuse readers of the cvs-ghc list.
Sean
_______________________________________________ Cvs-ghc mailing list [EMAIL PROTECTED] http://www.haskell.org/mailman/listinfo/cvs-ghc
