Hi guys,
I talked to Seth and he directed me here as somebody could be interested in my
ideas.
My idea in general is to make yum much more package-system independant (i.e.
so it could be used for rpms as well as debs and tgzs). Plus Make it possible
to run it on one machine and redistribute results on the rest of machines.
And final but still importand make it possible to collect data not only from
FS or rpmdb but also from RDBMS.
I walked through the code and started my own modification of yum which is just
a draft of what I would like it to be, and I haven't got any functionality
yet but initial ideas are as follow:
Need abstracting packageDB so we'll have different types of package
databases: rpmDB (localy installed rpm packages), FTPrpmDB (packages on
remote FTP server without header information in separate file), HdrFTPrpmDB
(packages on remote FTP server, headers available), DIRrpmDB (rpm's are in
local directory), SSHrpmDB (packages installed on remote server available via
SSH) (and we can sure derive debDB, emergeDB from package providing same
functionality so we can use it later).
class PackageDB:
def Load(self):
""" Load database into local variable """
pass
def get(self):
""" get local copy of database (variable) """
pass
def set(self, nevral):
""" set local copy of database (variable) """
pass
pass
Plus I'm trying to make nevral more abstract than it is right now so it can
manage info about non-rpm packages.
So is anybody interested?
P.S.
I'm using yum-1.0.1 as a base so if something like that is already done in
1.98 - sorry. For now 1.0.1 is my best bet as we are running rpm-4.0.x around
our machines and I don't want to backport changes from 1.98 to 1.0.1 :(
--
Dmitry Makovey
Web Systems Administrator
Athabasca University
(780) 675-6245
