Bernhard Fischer wrote: > On Wed, Jun 11, 2008 at 04:10:23PM +0200, Bernd Schmidt wrote: > >> Daniel Jacobowitz wrote: >> >> >>> I have no comment on the patch itself, but I like the approach - I've >>> concluded before that this is the only sane way to test toolchain >>> pieces, especially compiler or C library. We do all of our testing >>> after installation here. >>> >> My patch doesn't really create a full installation tree that looks >> identical to the one created by 'make install'; that might require >> somewhat more effort (unless we want to just $(MAKE) install >> DESTDIR=somewhere/in/tree for every make test). >> > > A better approach would be to just > > test-includes: test/include/bits/uClibc_config.h > $(make) -C $(top_builddir) PREFIX=test/ RUNTIME_PREFIX=test/ \ > DEVEL_PREFIX=/ \ > HOSTCC="$(HOSTCC)" \ > eventual_flags_passing_here \ > install_dev > > i.e. do a real, full install an no fakery. > Hi, I've hacked Makefile based on your both inputs and tested a working solution. Moreover I think this approach should be used for building the utils too, this why I think dev files (headers/lib*.a and linker script lib*.so) should be installed in a dir (let's say 'install_dir') undert the builddir instead of the test folder.
I'll came soon with a complete patch for you wider review. Cheers, Carmelo > _______________________________________________ > uClibc mailing list > [email protected] > http://busybox.net/cgi-bin/mailman/listinfo/uclibc > > _______________________________________________ uClibc mailing list [email protected] http://busybox.net/cgi-bin/mailman/listinfo/uclibc
