I had sent my reply to centos-devel, but as there is discussion here - here it is ...
Lance
---------- Forwarded message ---------- Date: Tue, 14 Dec 2004 12:47:18 +0000 (GMT) From: Lance Davis lance@uklinux.net Reply-To: centos-devel@caosity.org To: John Newbigin jn@it.swin.edu.au Cc: centos-devel@caosity.org Subject: [CentOS-devel] Re: [Centos] Issues/thoughts re CentOS-4
(The beta isnt public yet - so I have moved this thread to centos-devel - where it should probably be anyway ... )
On Tue, 14 Dec 2004, John Newbigin wrote:
This is a list of issues for discussion for CentOS-4. Most of it is based on my experience from building/maintaining CentOS-2 and observing CentOS-3. I have not seen the CentOS-4 beta yet so I don't know if any of this is already done. If these have been discussed on IRC then perhaps someone could produce a shot summary of the outcomes....?
- Just call it CentOS-4 not 4.1 etc. Too confusing and serves no real
purpose
It will be called CentOS-4 but will have 4.0,4.1,4.2 as iso releases.
It means that the version of centos-release rpm will always be 4 , but the contents will probably say 4 (4.0) or somesuch.
This is necessary to know what version people have installed (or at least what version they have (attempted to) upgraded to.
- All modified rpm should use mezzanine or equiv. This makes it easier
to track changes, build updates and generate a report on what has changed.
That needs someone to sort it out ... and all developers to be up to speed with the packages ...
- Modified rpm should have .c4.n added to the release
- c4 = centos 4. (works well once the 4.1 is dropped)
- .n = a number starting at 1. The spec only 0 does not scale if you
need to make another spec only change. The changelog will tell you what has changed.
yes it does scale - you make it c0.1 ( or c4.0.1 )
I think it is useful to indicate a spec only change, but am happy to go with .c4.0 for this purpose if you wish, I still prefer .c0
Also there is no problem having .c0 rpms in different centos versions because they are only non changed redhat rpms with trademark edits in the specfile. If the specfile was edited so as to produce different output then that would have to be a .1
- Mission: To stay as close to RedHat as possible
Issues: some things need to be modified to support this. For example the cpu/memory restrictions will be removed(like in CentOS-3). Given that we are doing this, should we also lift other artificial constraints like installing on i386?
The restraint for i586 will be removed, as I am working on doing for U4 , however I'm not sure i386 is a good idea ... It is easy enough - just build the kernel and have it in the install tree.
If we start making changes, where do we draw the line? In CentOS-2 I have fixed some small 1 line bugs (mostly packaging issues) but this might be a contentious issue. What about a CentOS+ repo. for enhanced versions?
Bug fixes are not contentious IMHO, but they should be pushed upstream.
a CentOS+ repo sounds like a good idea if someone is prepared to do the work. I think it would have to work along the contrib lines and be experimental / not supported
- Issue: should the distro be self hosting? I think as long as the
requirements are available from the addon/extras repos. then I don't think this is needed out of the box. Obviously the build environment needs to be well documented and reproducible.
indeed - as long as the components to make it self hosting are there then they do not need to be installed/installable from anaconda.
- Mission: Remove RedHat trademarks.
Issues: should we replace all rh logos/icons or just the ones required by RedHat? I think if the file/words can be replaced then yes, but package names etc. should be left alone.
all of the rh/logos and icons are removed - to give a 'better user experience'. There is no need to change the package names that redhat have chosen to give to the linux community. (In rhel4 redhat-config -> system-config - and also a lot of work has been done by Fedora removing the word redhat where it is uneccesary - this has flowed through into Rhel4 - indeed even in U4 for 3.3 there are occasions where I have had to remove patches because they had been modified upstream by Fedora project and incorporated by Redhat) .
- Issue: yum > Don't ever auto-modify
yum.conf. Too confusing.
It is a shame that yum does not support multiple configuration directories ...
Invent a system for selecting/allocating a mirror. Users should always use a 3rd party mirror and not the main caosity mirrors. Issue: every mirror has different paths. We need to clean up all the symlinks and multi versions on the mirrors. (We need someone in change of the mirrors)
There is no issue here - all of the mirrors use the same paths within the cAos repo - the mirror is just defined as the url including the path to the cAos repo - that does not cause any problems that I can see and can be used in automated scripts.
Someone does need to volunteer to write the script to 'auto-modify yum.conf' to use a local mirror (ok not 'auto' maybe) - it would be done on install and manually thereafter - I am guessing a yum-config-centos rpm or some such containing a bash script that actually produces the first yum.conf on install (using the geolocate stuff) and thereafter would only write a .rpmnew unless run manually. How about the yum.conf in the actual yum rpm ?? what will happen to that ??
It is also a shame that yum cannot use iso image sets - although I wonder if with a dvd and all packages on a single disc the [base] repo could be local disc - probably not :(
We have Rick Graves managing the mirror database now - I had been doing it up until recently.
- Issue: gpg key.
This should be installed by default so the user does not have to find & import it. At a minimum it should be in /usr/share/doc/centos-release and work like a RedHat box.
agreed - I have been discussing this at length with Skvidal and Orc etc, the most recent proposal is to auto install the gpg key only if it is the same key that the yum rpm has been signed with - by checking fingerprints, otherwise to squawk.
In any event to tell the user that the key has bene installed and how to check whether the key has been compromised/is the right key.
(Skvidal - sorry - not had chance to discuss with you since discussiuon with Orc)
- Issue: release QA
Before public release, a QA procedure should make sure that there are no obvious problems (like the comps issue of 3.3).
agreed
- Idea: automatic patch building.
It should be possible to build most patches without intervention. I don't know what the current systems in place are but if there were autobuilt unsigned patches available it would allow
- testing of patches before they are 'officially' released
- checking the progress if a patch
agreed - there is no automatic system at present.
Thanks for your input - great
Regards Lance