Hi Konrad,

It turns out that getting GCC 4.6 to build on Darwin wasn’t so
difficult.  The patch below fixes that, by adding lipo(1) to stdenv.
However, since it modifies stdenv, it entails a full rebuild.

Would it be a problem for you?  What do others think?

If it’s a problem, we can still patch GCC directly in ‘trunk’ (similar
to what’s done for Valgrind), and apply the patch below only in
‘stdenv-updates’.

Besides, having a working GCC on MacOS X raises the question of whether
its stdenv should be use it in lieu of Apple’s GCC 4.2 (which is old,
buggy, and no longer maintained.)  I’m inclined to stick to Apple’s GCC
for now, because it’s good for portability testing.  But we could still
switch in the future.

Thoughts?

Thanks,
Ludo’.

Index: builder.sh
===================================================================
--- builder.sh	(revision 33620)
+++ builder.sh	(working copy)
@@ -1,6 +1,15 @@
 source $stdenv/setup
 
 mkdir -p $out/bin
-for i in ar as c++filt gprof ld nm nmedit ranlib size strings strip dsymutil libtool; do 
-    ln -s /usr/bin/$i $out/bin/
+for i in ar as c++filt gprof ld nm nmedit ranlib size strings strip dsymutil libtool lipo; do 
+    ln -s "/usr/bin/$i" "$out/bin/"
 done
+
+# MIG assumes the standard Darwin core utilities (e.g., `rm -d'), so
+# let it see the impure directories.
+cat > "$out/bin/mig" <<EOF
+#!/bin/sh
+export PATH="/usr/bin:/bin:\$PATH"
+exec /usr/bin/mig "\$@"
+EOF
+chmod +x "$out/bin/mig"
_______________________________________________
nix-dev mailing list
[email protected]
http://lists.science.uu.nl/mailman/listinfo/nix-dev

Reply via email to