terminology and the hierarchy of releases
Jim Cornette
redhat-jc at insight.rr.com
Mon Feb 9 23:06:58 UTC 2004
Robert P. J. Day wrote:
>
> On Mon, 9 Feb 2004, William Hooper wrote:
> ...
>
>>A better way to look at it:
>>
>>What value would there be to using the FC1 base and updates repos on a
>>rawhide install? I don't really think there is any because you want the
>>development packages to be installed, not the FC1 packages.
>
>
> just to be pedantic, it's not a rawhide "install" you'd be looking at
> here; it's an *update* based on rawhide. the question was, if you were
> looking at doing a massive update based on the rawhide repo, would
> having the updates or testing repos in your yum.conf potentially cause
> problems?
>
> and, based on some of the previous postings, technically, that should
> work, since upgrading based on rawhide is supposed to simulate upgrading
> to the next release, which should work.
>
> but, realistically, who knows?
>
> rday
>
>
I have upgraded an all development system and also have a system that is
an upgrade from FC1 with repositories for testing, updates and base. I
don't see too much differences between the systems.
There is an item that gets hidden with an upgraded system and probably
gets a lot of problems overlooked. This is for a default account and if
things will be created from the default skel or not. Xmms is one item
that had this problem on both of my systems. I believe creating and
deleting user accounts would be a good practice to thoroughly check out
programs for functionality. Having existing users and seeing what goes
well and what fails is also a good idea, in my view.
After listening to some discussions about some programs never making it
to development, but going straight into testing. (compilers, etc). I
think that at least these repositories should be investigated to see the
interaction if one was to compile software or things that bypass
development.
About the factor of upgrading from test1, test2, test3 then to the
release. I have a laptop with an Athlon processor. It is working fine
through this method. It all depends upon your interest in the testing.
Having different setups and situations is a positive move to test more
possibilities and possible interactions in hardware/setups.
A valid test would be to try as many different options and schemes out
and report the results.
The only pristine conditions That I could see is fedora 1 with no
updates applied to test release or straight to the test release. I doubt
that people will have either type of condition when upgrading from
Fedora 1 to Fedora 2. You need the spectrum in between to really test
out the quality of the next release.
Just my perception.
Jim
More information about the fedora-test-list
mailing list