An improved "hosts" completion handling

["Christopher Browne" <cbbrowne@xxxxxxxxx>]

I had been using some code I'd seen "out there" to pull known_hosts as
a way of setting up a $hosts value so that ssh/telnet could do tab
completion on host names.

I think this was the source, more or less...
http://www.2701.org/archive/200405030000.html

With the changes in openssh, of late, ~/.ssh/known_hosts is now
littered with encoded host names that were getting pulled in such that
I'd get tab completion looking like the following:

cbbrowne@dba2:~/Slony-I/slony1-HEAD-pristine> ssh rg  ssh
rg570-da-logdbs
 Wednesday 17:20:40
rg570-da-logdbs
rg570-da-logdbs1
rg570-da-logdbs1.nj.site\ ssh-rsa\
AAAAB3NzaC1yc2EAAAABIwAAAIEAn7BPZtUqqjJzpVszaI71KUCVyvGicrOvUTFO4EiGzCHm7lYtJIbN0MblFo4CGwq3Z7cgzMwn/qdSxIXFnlIHIpPHojcZAnbSFXrJMIKB45VjK6XgFOB7ZQtpKg252F/nXV7vaTa/xgI6H+tza4w5UI5AZnrD/knq+l3qtSyuBDU=
rg570-da-logdbs2
rg570-da-logdbs2.nj.site\ ssh-rsa\
AAAAB3NzaC1yc2EAAAABIwAAAIEAn7BPZtUqqjJzpVszaI71KUCVyvGicrOvUTFO4EiGzCHm7lYtJIbN0MblFo4CGwq3Z7cgzMwn/qdSxIXFnlIHIpPHojcZAnbSFXrJMIKB45VjK6XgFOB7ZQtpKg252F/nXV7vaTa/xgI6H+tza4w5UI5AZnrD/knq+l3qtSyuBDU=

Bleah.  That's awful.

The following showed an alternative that does some better parsing, as
well as reading IP addresses.

http://www.zsh.org/mla/workers/2006/msg00280.html

Here's my adaption of this:

for file in $HOME/.ssh/known_hosts /etc/ssh/known_hosts; do
 if [[ -r ${file} ]] ; then
   _cache_hosts+=(${${${(u)${(f)"$(<${file})"}%%[
,|#]*}:#(#s)[0-9]##.[0-9]##.[0-9]##.[0-9]##(#e)}:#(#s)[0-9a-f:]##(#e)})
 fi
done
zstyle ':completion:*:hosts' hosts $_cache_hosts

Hope that is of assistance to others...
--
http://linuxfinances.info/info/linuxdistributions.html
"...  memory leaks  are  quite acceptable  in  many applications  ..."
(Bjarne Stroustrup, The Design and Evolution of C++, page 220)