On Sat, Mar 11, 2017 at 10:59:52PM -0600, Bruce Dubbs wrote: > Ken Moffat wrote: > > > > On both the AMD machines, no problem. But on both of the intel > > machines, I ended up with a symlink pointing to libnssckbi.so.orig > > (still works, but rather defeats the point of it). What I don't > > understand is why. I've now manually fixed them, but I'm still > > struggling to work out what I've overlooked. > > > > My code is slightly different: > > > > if [ -h /usr/lib/libnssckbi.so ]; then > > echo "libnssckbi.so is already a symlink, nothing to do" > > else > > echo "making libnssckbi.so a symlink to libp11-kit.so" > > mv -v /usr/lib/libnssckbi.so /usr/lib/libnssckbi.so.orig > > ln -sfv libp11-kit.so /usr/lib/libnssckbi.so > > fi [---] > > It looks like the instructions work, but I don't like them. I'd prefer to > see: > > if [ ! -h /usr/lib/libnssckbi.so ]; then > mv -v /usr/lib/libnssckbi.so /usr/lib/libnssckbi.so.orig > ln -sfv libp11-kit.so /usr/lib/libnssckbi.so > fi > > Which is a slightly less verbose verion of what you have. >
Two reasons: 1. In the "do nothing" case (i.e. a later upgrade of certs on a system where I have already done this) I want to see something in my log to tell me if it did anything. Probably I could have put a test in the "driver" scripts which invokes this (upgrade firefox, update certs, build that part of my desktop where I build firefox) but I didn't. Of course, doing that would not give me a log at all. 2. My 'client' scripts (build a package, install a bootscript, in this case decide whether or not to make a symlink) fail on non-zero status. 'If' tests withut an 'else' tend to provoke such failures when the test is not met. For some things in firefox (conditional tests which can lead to me adding, or not adding, things to _my_ default build, e.g. I only add the current sed for libevent if the version is suitable) I _do_ turn off the 'set' whilst doing the tests, but it can get messy and I prefer not to do that. > I don't know why you would get a symlink pointing to the .orig file. The > instructions above certainly wouldn't do that. > I got the same, again, on my i3 BLFS-7.9 system. Unfortunately, that has other problems (https:// broken, no valid certs found). I'll be looking at that when I'm more advanced in moving my newer systems to 4.9.14 or 4.10.2 kernels - there were ext4 fixes in those, the problems they solve are clearly uncommon, but I'll feel safer with them. And I haven't looked to see _when_ the problems were introduced, maybe 4.9.early is ok, but maybe the no longer maintained 4.8 is also affected. Sorry if I haven't noticed any missing 'o' letters in what I typed, this is on my netbook, the 'o' needs to be keyed harder at the moment. ĸen -- `I shall take my mountains', said Lu-Tze. `The climate will be good for them.' -- Small Gods -- http://lists.linuxfromscratch.org/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
