On Wed, Mar 3, 2010 at 7:49 PM, Bill Nottingham <notting(a)redhat.com> wrote:
Thomas Janssen (thomasj(a)fedoraproject.org) said:
> > When there's no policy, and the user has to guess whether or not they
> > need to do this for every package on their system, however, you have
> > a mess.
>
> Well, except there's nothing to guess. The regular user should by all
> means know what a security fix is. As well as a bugfix. And i think
> the regular user knows as well what an enhancement is. So he can
> decide very well what he want.
> Except you expect from future users to be even more dumb than bread.
So, a user starts out with kdelibs-4.2.2 in Fedora 11, and decides to
only take security updates.
Their update path is now:
4.2.2 ->
4.2.4 ->
4.3.1 <end>
Say they take security and bugfix. Their upgrade path is now:
4.2.2 ->
4.2.3 ->
4.2.4 ->
4.3.1 ->
4.3.2 ->
4.3.3 ->
4.3.4 ->
4.3.5 ->
4.4.0
How is the user supposed to 'by all means' know that that is
sensible?
Ok, i know what you mean. He has to trust us packagers (not only
speaking for KDE, i have as well other software i maintain) that i do
the right thing. Especially if it's a complex software like KDE, where
bugfixes are not just 2 lines of code to backport.
It's still all 4.x series and not a 4.x to 5.x change.
And of course if you use KDE as example you cant have only bugfixes
without enhancements. I'm not sure if it's the same with GNOME. I dont
have any insight there.
--
LG Thomas
Dubium sapientiae initium