Em Sex, 2006-01-06 às 00:57 +0100, Christoph Thiel escreveu:
> be much better. One option could be to just call rpm instead of using
> python-rpm to perform the installation of packages (like YaST does it ;)).
That wouldn't solve the real issue: cache. This is what makes Smart slow
most of the time, but the cache is a necessary evil, since it's from
where Smart gathers all data in order to calculate transactions, fetch
files, show info, etc.
IMHO Smart is doing the right way: dealing directly with rpm libs and
not being a wrapper. In theory it's faster than calling an external
comand (since there is less forking, memory allocation, etc), and is
less susceptible to changes on the external command output (don't forget
it's intended to run regardless of distribution). But, who am I to
judge? :)
BTW, I work with a very slow laptop, and I realized that after I started
using Smart with YaST2 reps instead of APT it got much much faster. I
think that's because it doesn't have to deal with apt indexes, which are
bziped (and slow to uncompress) and bigger than yast indexes.
> I would rather like to stick with SMART and get a much testing on SMART as
> possible. Because at the end of the day, we will have a mature package
So very true. If you always have apt available, there would be less
people testing and helping Smart growing. Besides, there would always be
the 3rd party packagers, who could still make it available regardless of
the distro decision.
> to use SMART -- and we both know that apt-rpm doesn't have a bright
> future...
What future? :)
--
% Mauricio Teixeira (netmask)
% mteixeira{a}webset{d}net <> Maceio/AL/BR
% TI+Telecom Analyst <> Linux Specialist
% http://mteixeira.webset.net <> http://pmping.sf.net
% [D0CE 6BD4 526B B7D1 6F4E 85FA A7A0 1A6F B23A A9EE]
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]