[email protected] wrote:
On Wed, Aug 26, 2009 at 09:10:16PM +0100, Peter Tribble wrote:
Are you willing to accept that pkg(5) does have serious problems
with performance, and needs improvements of an order of magnitude
or more?
Why would I accept such a claim? You haven't provided me any real data
that you have a performance problem. You're comparing two completely
different packaging systems, and claiming, erroneously I might add, that
the two are equivalent.
At the end of a day you can actually compare them as both of them are
tools for sysadmin so they can install/uninstall software, look for a
file or dependencies. These are going to be mostly used features of any
packaging system. And most sysadmins (your customers, users) won't care
that much about internals but rather how responsive and easy to use it
is. So comparing these main features of packaging tools for sysadmins
from a performance point of view definitely is a valid test.
Having said that from my experience I can't agree with Peter that is is
that bad - actually it is quite responsive. Definitely still using too
much memory but it hasn't matter that much to me from a practical point
of view (still it shouldn't be using hundreds of MBs sometimes...). Is
it slower than rpm or pkgadd? My impression is that it actually is but
again it is not that bad and we should all keep in mind that this is a
very much work in progress and I can understand why developers want to
mostly focus on features and correctness first and then on performance
tuning. Perhaps some effort should be put right now on improving the
performance of basic features so bad impressions won't stick to pkg for
years to come.
--
Robert Milkowski
http://milek.blogspot.com
_______________________________________________
pkg-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
