The only time breaking up the list is necessary is if the list becomes longer then the max arg length. In that case you can put all of the packages you want to install/upgrade into a text file and pass the text filename to RPM.. (just note that you have to have at least two names listed or RPM doesn't want to work with the text file..) :|
Also on an erase option, the text file approach doesn't work so you have to place games if you want to try to do something like that for rpm -e.
(yes I have hit max argument length before.. real pain..)
--Mark
Aaron Hanson wrote:
Thanks. I'll look at the hash and percent options; maybe breaking up the list is really unnecessary effort.
-Aaron
-----Original Message-----
From: rpm-list-bounces@xxxxxxxxxx [mailto:rpm-list-bounces@xxxxxxxxxx] On Behalf Of Mark Hatle
Sent: Monday, April 12, 2004 12:42 PM
To: RPM Package Manager
Subject: Re: Install sequencing
You can add --percent if you want a _LOT_ of status information.. just be aware that is produce an insane amount of status. (Best if captured and interpretted by something else.)
--Mark
Aaron Hanson wrote:
Yes; all the run-time dependencies are correct. The reason I like breaking things up into multiple rpm transactions is because it provides more progress information to stdout and stderr. If
I remember
correctly, even 'rpm -Uv' is very quiet for a long time if it is working on a large set of packages....
-Aaron
-----Original Message----- From: rpm-list-bounces@xxxxxxxxxx [mailto:rpm-list-bounces@xxxxxxxxxx] On Behalf Of Mark Hatle Sent: Monday, April 12, 2004 12:26 PM To: RPM Package Manager Subject: Re: Install sequencing
If you have proper run-time dependencies, you can generally
bootstrap
the filesystem (provide a couple of necessary files and
directories)
and then do the initial RPM install in one big shot..
rpm -i <list of ALL of the packages>
(or -U) rpm will automatically reorder the install based on dependencies.. (but again this assumes you have proper run-time dependencies in your packages.)
--Mark
Aaron Hanson wrote:
Hi All-
I'm using RPM to install a complete embedded system of approximately
216 packages. Working from an alphbetized list of all the
packages,
I've created a very simple shell script that uses rpmcache,
followed
by a loop of 'rpm -i --aid' calls into a test install
area. The end
result is what I'd call a properly sequenced list of rpm
transactions
to install the complete system. But the resulting list is
not optimal;
it includes some VERY large groups of rpms that could actually be broken up into smaller sets.
Before I re-invent the wheel; has anyone got a
suggestion on the
best way to do this?
Thanks.
-Aaron
------------------------------------------------------------
----------
--
_______________________________________________ Rpm-list mailing list Rpm-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/rpm-list
_______________________________________________ Rpm-list mailing list Rpm-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/rpm-list
_______________________________________________ Rpm-list mailing list Rpm-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/rpm-list
_______________________________________________ Rpm-list mailing list Rpm-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/rpm-list
_______________________________________________ Rpm-list mailing list Rpm-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/rpm-list
_______________________________________________ Rpm-list mailing list Rpm-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/rpm-list
