[ATrpms-devel] Re: STSM/ATrpms FC1 DVD

Stefan Smietanowski stesmi at stesmi.com
Mon Feb 16 15:18:48 CET 2004


Hi Axel.

>>>alsa -- one of the selectable items is 
>>>alsa-kmdl-2.4.22-2166.nptl_35.rhfc1.at. Probably not the best label 
>>>there, since it has to be changed when the kernel changes, and it isn't 
>>>correct if I'm running an smp kernel.
>>
>>*point to Axel* Don't blame me :)
>>
>>The base name of the package is just that so that's why it needs to be
>>like that in the comps file.
> 
> Can the displayed string be altered, or does anaconda require showing
> the base name of the rpm?

It just displays the base name of the RPM. I guess with some hackery
it might be able to do other things BUT the problem is that there is
a number of other sub-programs of anaconda (used when making the ISO,
etc) that parse the comps.xml file so they would simply not find
what they're looking for meaning I'd have to find all instances that
reads the comps.xml file and change them and I don't think I should
do that as it means a likelyhood of bugs.

> Alternatively, all but the lirc kmdl rpms have a master packages that
> requires in the kmdl packages, e.g. alsa-driver for alsa, bttv for
> bttv-kmdl etc.

Yes, I know. But also note that someone might want to install
for instance lm_sensors but not the additional kernel modules and
just use what's in the stock kernel.

Either we find a generic solution or we leave it as is imo.

>>>lm_sensors -- I didn't choose to install it, but I was greeted at 
>>>startup by failures on the lm_sensors init script, so I guess it got 
>>>installed anyhow. I haven't looked at comps to see what might have made 
>>>it install...
> 
> 
> I am meaning to look into the automatic enabling of lm_sensors for
> quite some time now and am forgetting it. That's where a nagging
> bugzilla would have been helpful. I believe the bugzilla instance is
> ready to be launched.

Good. Tell me what you find out. Bugzilla will help alot in these
areas. Is it possible to add my DVD to it? After we find a suitable
name for it of course.

> Stefan, is it possible and does it make sense to test the DVD in NFS
> network install? Most parts of anaconda like the package selection
> algorithms would be the same, I suppose.

Is it possible: I don't know, I haven't tested those features :)

Does it make sense: Of course. Only thing you need is to burn
out the boot.iso to a CD-R(W). One for each spin of the installer.

If you're really interested I could set up an rsync that a selected
number of people can use to get the latest testing incarnation from.

Yes, it means an initial bandwidth hogging but it will get better
as people have 'it all'.

So I would in that case want a personal email (not to the list)
of every person that has an interest in rsyncing the whole
DVD and testing it. What I want in that case is the name
of the person, who that person is, the static IP or ip-ranges
that person might use and the level of interest of that person.

Jarod/Axel are given people of course, but I still want an email
from you two. Anyone else reading this are welcome to send me a
mail about it. I don't want too many people as testers right from
the start but a few would be handy.

> It would make testing easier, as one would only need to download a
> couple of files and us local mirrors of the packages. Shorter testing
> cycles and testers w/o DVD burners would be able to jump in.

Absolutely. I agree entirely.

// Stefan




More information about the atrpms-devel mailing list