Jeremy Huntwork wrote: > > According to the link Dan posted, http://www.technovelty.org/linux/strip.html > (I > don't know if anyone in the community has verified yet personally), > --strip-all is the > one to worry about. --strip-unneeded will do the right thing for static libs. > So the > wording in chapter 5 may be wrong because it says --strip-unneeded will > destroy > static libs, when in fact it may not. > > I say 'may' because I am uncertain if there are caveats that exist when > dealing with the > difference of environments between chapter 5 and 6. I rather doubt it because > by that > point you have fully native libs, but still, it's always good to verify with > actual data >
I was looking at the binutils bugzilla, and in 2009 there were some patches applied to strip that effect --strip-unneeded and global symbols. This could be what has caused it to become safer to use on static libraries. I'm away for the weekend but when I get back I can set up a test environment with a backup of my tools and system running --strip-unneeded on the libraries. Then I'll try to do a bunch of normal things like run some servers and programs as well as build some packages with massive dependencies statically and just generally try to make it freak out. I usually only strip my tools because I like having the debugging symbols in my actual finished system. That being said though, I'm interested to see if this will go without a hitch. From a few forums I glanced over it looks like some distributions have been using this flag now for their stripping needs these days. Jonathan -- http://linuxfromscratch.org/mailman/listinfo/lfs-dev FAQ: http://www.linuxfromscratch.org/faq/ Unsubscribe: See the above information page
