Hi, I'm tackling #3002 (upgrading Util-Linux to 2.21). Bruce suggested the use of '--enable-new-mount' so that we will use the new libmount based version of 'mount'. It's marked as EXPERIMENTAL in configure's help, but I'm happy enough to put it in as Util-Linux is now well maintained and I fully expect that it will be moved from EXPERIMENTAL before the next LFS release.
That led me to wonder why we bother passing the other '--enable' options (partx, arch and write). For reference, Bruce brought up 'arch' in http://www.linuxfromscratch.org/pipermail/lfs-dev/2012-February/065811.html and partx/write are discussed in http://www.linuxfromscratch.org/pipermail/lfs-dev/2008-October/061763.html. My feelings on these are that we should use upstream's defaults wherever possible, and only diverge from them when absolutely necessary. So, based on that, I'd suggest the following change to Util-Linux's configure invocation: Current: ./configure --enable-arch --enable-partx --enable-write New: ./configure --enable-new-mount This would also require a change to coreutils' instructions to build its version of 'arch'. Aside from Alexander's somewhat contrived example of `dd`ing a disk image and having to invoke `partx` to inform the kernel about the partitions contained within it, I can't figure out when else I'd need to use the partx utilities. As for 'write', I think its such a niche utility that those that know they need it can build/install it under "your distro, your rules". Thoughts? Matt. -- http://linuxfromscratch.org/mailman/listinfo/lfs-dev FAQ: http://www.linuxfromscratch.org/faq/ Unsubscribe: See the above information page
