Jeremy,
I brought this question up at fudcon, and it relates a bit to this yum patch, and a bit to your kickstart patch. Really its a bit more like 2 questions.
1) Do you forsee the system installation code within anaconda and livecd creator merging at any point, or staying divergent? Phrased another way, do you ever see something like anaconda --rootpath replacing the installer code in livecd-creator? (or just generally coming to a point where there is a single code path to install generated system images)
2) Where do the spin differentiation / configurations exist in relation to pungi, and how do you see these relating to _livecd_ spin differentiation/configurations (which I visualize currently as differing kickstart files).
I suppose one issue that would clarify both of these, would be the theoretical possibility of something like the current livecd installer replacing the traditional anaconda install mechanism. Were that to happen, then it seems like the spin differentiation / configuration for the major spins, would just be different kickstart files. Do you see things going in this direction, or some other direction?
-dmc/jdog
____________________________________________________________________________________ 8:00? 8:25? 8:40? Find a flick in no time with the Yahoo! Search movie showtime shortcut. http://tools.search.yahoo.com/shortcuts/#news
On Fri, 2007-03-02 at 14:43 -0800, Jane Dogalt wrote:
- Do you forsee the system installation code within anaconda and
livecd creator merging at any point, or staying divergent? Phrased another way, do you ever see something like anaconda --rootpath replacing the installer code in livecd-creator? (or just generally coming to a point where there is a single code path to install generated system images)
I could see this going either way to be honest. It largely depends on what direction the people involved want to take. One advantage of the current livecd-creator code is that it's relatively small and thus easy to dig into if you need/want to. anaconda, on the other hand, requires a bit more of an investment.
That doesn't really answer your question, though :-/
- Where do the spin differentiation / configurations exist in relation
to pungi, and how do you see these relating to _livecd_ spin differentiation/configurations (which I visualize currently as differing kickstart files).
My personal goal is to have one sort of syntax that's use for defining these sorts of things. There was a (long-ish) discussion on anaconda-devel-list 3 or 4 months ago about how to define "distro specific features" (things like do I use NetworkManager or not, etc) with anaconda and the current answer is that you have to use an installclass. But the first steps have been taken so that we can use a kickstart config in anaconda to provide all of those sorts of defaults. At which point, we can hopefully use similar bits for everything. Not going to happen in the Fedora 7 timeframe, but it's the direction things are meandering in at least :)
I suppose one issue that would clarify both of these, would be the theoretical possibility of something like the current livecd installer replacing the traditional anaconda install mechanism. Were that to happen, then it seems like the spin differentiation / configuration for the major spins, would just be different kickstart files. Do you see things going in this direction, or some other direction?
I think the two have their own individual merit. A lot of people aren't going to want to have the overhead of creating their own spin and want to be able to do selection and configuration within the confines of something that's a bit more general purpose. But I want to try to make it so that the skills and what you learn doing things for one method are applicable for others as it will let you more easily pick the right tool for the job at hand.
Jeremy
livecd@lists.fedoraproject.org