'su lfs' dropping into the background
bdubbs at swbell.net
Tue Jan 20 19:35:42 PST 2004
Greg Schafer wrote:
>On Mon, Jan 19, 2004 at 11:39:15PM -0600, Bruce Dubbs wrote:
>>.profile is never looked at if .bash_profile (or .bash_login) exists.
>Yeah, good point. I just RTFM'd for bash and confirmed. Even if it did, it
>still wouldn't matter coz the currently emplyed method is that nifty :-)
>>Its not a problem here. What could be a problem is /etc/profile on the
>No. Definitely not. That's why the new method was developed in the first
>place -- to overcome that very problem. The current book does try to explain
>it. Please review and see if we can improve the text.
Hmm. You're right of course. However there may be a problem here. If
a user exits from the lfs user and reenters, there is no assurance that
the 'source ~/.bash_profile' will be run again. If that happens, there
could be an extra CFLAGS or some such variable left in the environment.
Taking a critical look at the text as it is now written, I feel it does
explain the process, but the content is relatively dense. That is,
there is a lot of detail packed into a couple of paragraphs.
That said, about the only improvement I can thnk of is to add a note
after the 'source' command that if a user exits from the lfs user and
reenters, the command needs to be run again.
My wording would be:
"Caution: If you exit from the lfs user shell and reenter (or start an
additional instance), the .bash_profile command is not automatically
run. Be sure to run the above command each time you start up the lfs
More information about the lfs-dev