My whole experience with sed and awk amount to cut, copy and paste, so any help is very nice.
The idea is that once we have this list we can then run an 'ls -al [item in list]' and get results like. >ls -al /lib/libhistory.so.5 lrwxrwxrwx 1 readline readline 17 Jan 9 06:13 /lib/libhistory.so.5 -> libhistory.so.5.1 now we know that readline is required to install bash. This kind of information can be added to a database dynamically. Then at some point in the future if we wish to update readline we can query the DB and list all items that require readline to install. For me I can track package versions because I added the package version as description to my package users. So, I can use /etc/passwd like [EMAIL PROTECTED]:/# cat /etc/passwd | grep readline readline:x:10015:10015:readline-5.1:/usr/src/readline:/bin/bash & [EMAIL PROTECTED]:/# cat /etc/passwd | grep bash: bash:x:10019:10019:bash-3.1:/usr/src/bash:/bin/bash then I'm sure some sed and awk magic can be worked here before placing things in the DB, but the results would look like bash-3.1 required readline-5.1 when installed or more importantly the following program(s) requre(d) readline-5.1 please test them after upgrading to ensure no dependencies were broken bash-3.1 ... ... etc. Finally, the result is that when going beyond BLFS you build and maintain your own package tree and can trouble shoot broken dependecies in a reasonable fashion, even as they occur. -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
