goals/status?

[Jane Dogalt]

While I have my own projects and goals, I was wondering what the
general status is regarding this list - "fedora live cd".

Specifically, I'm curious as to what the top 3 unfinished goals are for
this project.  And what the current plan for achieving those goals is,
and where the output will be hosted for testing that meets those goals
(when that happens).

It seems to me, that there should be a checklist/bugzilla of features,
that describe what we are aiming for, such that if someone uses
kadischi or pilgrim or some other tool to generate a prototype, that
beta testers can refer to and evaluate said prototype.

Obviously there are many desired features, but what I'd like to see, is
maybe a weekly/monthly listing of the top 3 "todo items", so that we
can start knocking them off, and seeing things progress.

So, to clarify/reiterate-

1) where are the alhpa/beta prototypes (give me urls/torrents that
work)

and

2) what are the top (3) priority features that are missing.

One of the key things I've been working on in my own project, and
mentioning here for some time now, is that I'd like to see a completely
automated livecd generation system.  Such that a cvs tree of extra
files/configuration might get nightly/weekly built into a testable
.iso.  I think if we had that infrastructure, it would drastically
increase the rate of progress on the project.  It would be like "hey,
feature XYZ on build fedora-livecd-20061020a.iso is broken", or "this
build just got this feature, please test."

agree/disagree?

-dmc/jdog


__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com

[J. Hartline]

Jane Dogalt wrote:
>  While I have my own projects and goals, I was wondering what the
>  general status is regarding this list - "fedora live cd".
>
>  Specifically, I'm curious as to what the top 3 unfinished goals are for
>  this project.  And what the current plan for achieving those goals is,
>  and where the output will be hosted for testing that meets those goals
>  (when that happens).
>
>  It seems to me, that there should be a checklist/bugzilla of features,
>  that describe what we are aiming for, such that if someone uses
>  kadischi or pilgrim or some other tool to generate a prototype, that
>  beta testers can refer to and evaluate said prototype.
>
>  Obviously there are many desired features, but what I'd like to see, is
>  maybe a weekly/monthly listing of the top 3 "todo items", so that we
>  can start knocking them off, and seeing things progress.
>
>  So, to clarify/reiterate-
>
>  1) where are the alhpa/beta prototypes (give me urls/torrents that
>  work)
>     
One of Kadischi's priorities are installing a system from a LiveCD.
The work has been done in Anaconda to allow Anaconda to run from the LiveCD
environement with the exception of Anaconda using utilities in 
/usr/sbin/ rather
than in /sbin where these utilites would be normally in a real system.
Such as mkswap, mke2fs, tune2fs, mkreiserfs, mkfs.xfs, and so on.

You can refer to bugzilla entry 205137 to download a prototype disc I 
have created
that demonstrates the potential of this being a reality sooner than later.
(Thus my frustration when told Pilgrim is closer to this goal than 
Kadischi and that
Kadischi patches to Anaconda were of low priority. The writer is simply 
uninformed, and
I have already stated I was not mentioning it since it wasn't actually 
working yet but being
worked on.)
Have a look: http://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=205137

As soon as this can be rectified the LiveCD to system install is complete.
Polishing is only neccessary afterwards.

The current Anaconda can, without patching be run from within a LiveCD 
environment.
Where we are at now, is during installling packages an exception occurs 
stating the header
files do not exist, upon inspection they do actually exist.
>  and
>
>  2) what are the top (3) priority features that are missing.
>
>  One of the key things I've been working on in my own project, and
>  mentioning here for some time now, is that I'd like to see a completely
>  automated livecd generation system.  Such that a cvs tree of extra
>  files/configuration might get nightly/weekly built into a testable
>  .iso.  I think if we had that infrastructure, it would drastically
>  increase the rate of progress on the project.  It would be like "hey,
>  feature XYZ on build fedora-livecd-20061020a.iso is broken", or "this
>  build just got this feature, please test."
>
>  agree/disagree?
>     
I agree to the fullest.
Another of Kadischi's priorities is integration of some sort of user 
interface that can be used
but is not required to run Kadischi, to keep compatibility with 
Kickstart LiveCD generation, etc.

If you would like to see the current progress the Fedora Unity project 
is doing with Kadischi
visit: http://torrent.fedoraunity.org/torrents

There you will find several different makes of Fedora LiveCD discs.
Recently we have been able to and successfully compile LiveCDs for 
x86_64 also.

I am still waiting on Chitlesh to post the Kadischi PyQT UI code he has, 
if it is newer than
the last copy I saw from him which I have here on disk.

J. Hartline

[David Zeuthen]

On Sun, 2006-10-22 at 01:26 -0500, Jasper Hartline wrote:
>  One of Kadischi's priorities are installing a system from a
> LiveCD. 
Sorry, but I don't think just running anaconda from a live cd
environment consuming RPM packages shipped side-by-side with the live cd
file system qualifies as "installable livecd". At least you can't do
anything interesting in 700MB without requiring network access. 

For the record, I do have an "installable livecd" with pilgrim now
though the code needs to be cleaned up, see attached for current code
dump. It should be self-explanatory more or less. With this I'm able to
install from the LiveCD in ca. 5-6 minutes. It works pretty nice... but
then again, I don't have much time (nor the inclination but that's
another matter) to hack on this (let alone time to participate on this
list) these days but at least I was hoping someone would pick up this
train of thought as I do believe the current Kadischi livecd
installation is sub optimal.

So.. It's a little sad we can't seem to find agreement on the simple
definition of what a "installable livecd" is... and I still do think
that doing work which is clearly not on par with the distributions that
Fedora competes with... is not the path we want Fedora to be on.
Something for Greg and co. to think about I guess.

Anyway, enough flaming from me. Perfect is the enemy of the Good or
whatever it is they say. Good luck.

     David

[Chris Negus]

On Mon, 2006-10-23 at 15:27 -0400, David Zeuthen wrote:
>  On Sun, 2006-10-22 at 01:26 -0500, Jasper Hartline wrote:
>  > One of Kadischi's priorities are installing a system from a LiveCD.
>  Sorry, but I don't think just running anaconda from a live cd
>  environment consuming RPM packages shipped side-by-side with the live cd
>  file system qualifies as "installable livecd". At least you can't do
>  anything interesting in 700MB without requiring network access.  
I have a practical issue related to this discussion. I need to produce
media within a few days after FC6 is released to distribute with the
Fedora 6 Bible. I'm allowed to have a DVD (for that I'll use the Fedora
install DVD) and a CD that I want to double as a live/install CD.

I would prefer to have something similar to the Ubuntu or Gentoo live
CDs. They have icons on the desktop that install a workable system from
the CD. (I don't think you need the network for a basic install, but I
could be wrong.)

But, in lieu of a better idea, I'll simply create a CD that (from the
boot prompt) will either boot the live CD or start an "askmethod"
install. I agree it is not an elegant solution, but I would love to hear
if there are any better ones available right now.

David, if you are close, I wouldn't my trying out whatever you have.

-- Chris Negus

[J. Hartline]

David Zeuthen wrote:
>  On Sun, 2006-10-22 at 01:26 -0500, Jasper Hartline wrote:
>    
> > One of Kadischi's priorities are installing a system from a LiveCD.
> >     
>
>  Sorry, but I don't think just running anaconda from a live cd
>  environment consuming RPM packages shipped side-by-side with the live cd
>  file system qualifies as "installable livecd". At least you can't do
>  anything interesting in 700MB without requiring network access. 
>
>  For the record, I do have an "installable livecd" with pilgrim now
>  though the code needs to be cleaned up, see attached for current code
>  dump. It should be self-explanatory more or less. With this I'm able to
>  install from the LiveCD in ca. 5-6 minutes. It works pretty nice... but
>  then again, I don't have much time (nor the inclination but that's
>  another matter) to hack on this (let alone time to participate on this
>  list) these days but at least I was hoping someone would pick up this
>  train of thought as I do believe the current Kadischi livecd
>  installation is sub optimal.
>     
In 700MB you don't have room for any RPM packages, so network would be 
required.
The problem with your approach is that you do not include flexibility
with your installation method. You're simply copying the LiveCD system 
to the disk.
With the Kadischi/Anaconda approach a standard Fedora Core installer 
interface is
brought up.

With a DVD it is possible (I've already done it and tested it) to 
include 3.2GB of RPM
packages (The whole of Fedora Core) with 1.5 or so GB left for the 
LiveDVD system.

Your script doesn't seem to handle any sort of upgrade path either, 
which Anaconda does.
The method I am working on simply includes a partitioner, an upgrade 
path, and flexibility
in package selection in one tool, Anaconda. Plus it is the stock Fedora 
Core installer mechanism.
>  So.. It's a little sad we can't seem to find agreement on the
> simple
>  definition of what a "installable livecd" is... and I still do think
>  that doing work which is clearly not on par with the distributions that
>  Fedora competes with... is not the path we want Fedora to be on.
>  Something for Greg and co. to think about I guess.
>
>  Anyway, enough flaming from me. Perfect is the enemy of the Good or
>  whatever it is they say. Good luck.   
>     
We don't neccessarily have to agree.
I think what is best is that we both work on both projects
and let the user base and other members of the list decide what they like.

You don't have much mind for flexibility and it is proven by the fact that
you think only one project should be worked on, Greg or mether I do not 
remember which
has already stated this isn't just the Kadischi Fedora Core LiveCD list.

Also I disagree to the fullest that an Anaconda to system from LiveCD 
install is on par
with other distribution's progess also. Fedora Core is knows to be a 
full, robust operating system
not a drop on a dime oh just include this and this with no choice 
distribution, lke DSL or Puppy Linux.
Those are minimal systems, and from a fully features robust system like 
Fedora Core, I would expect
nothing less than package selection choices and partitioner, with viable 
upgrade path from the distribution's
LiveCD project. Anaconda is also familiar to alot of people, so new 
users getting guidance from an already
knowledgeable user, would have not to learn yet another installation 
tool and method, but could use the
same method presented with Anaconda.

J. Hartline