Randy McMurchy wrote:
single test suite run successfully, I'm just trying to present a point of view that it is not reasonable to expect developers to to the amount of testing that Bruce suggests.
I agree with Randy here. It only makes sense after all. Throughout the life cycle of LFS we use a lot of different combinations of software. I think there are times projects like LFS and other from-source distributions end up testing a lot of combinations of software *by accident* that developers never thought to test *on purpose*.
And in reply to Jeremy's grep as-a-symlink comment in the previous email of this thread: I'm not sure which one is best to be honest.
If a symlink allows us to leave grep in its proper location in the alphabet, I don't see a problem with that. We just want to make sure that grep's "make install" replaces the symlink rather than overwrite the target in /tools.
Seeing /tools isn't deleted until at the very end of Chapter 6, the symlink method would be perfectly safe IMO.
I don't see a technical reason for symlink vs. no symlink (and subsequently moving the Grep package itself). I think it's a preference what is considered "nicer."
A symlink is less work than moving the package itself and the move has no added benefit as far as I can tell.
I'm changing my vote from moving grep to symlink grep unless a valid technical reason comes up that warrants not using a symlink. We should still move libtool.
-- Gerard Beekmans /* If Linux doesn't have the solution, you have the wrong problem */ -- http://linuxfromscratch.org/mailman/listinfo/lfs-dev FAQ: http://www.linuxfromscratch.org/faq/ Unsubscribe: See the above information page
