Hi Colin, what else can I say, but WOW! thank-you ever so much for taking the time out to have a look at these issues. As soon as the rebuilds hit, do please poke the lubuntu-qa team so that the ppc testers can get onto them.
@ Adam, as you know the 'X' issue and have access to kit would you be prepared to make an exception to not testing 12.10 and having a chat with the #ubuntu-x people? I can raise the issue, but with no kit to back up testing of any debug reports that they need I'm a bit stuck. Else can you give some guidance as to what our current ppc testers should be talking to them about? Thanks, Phill. On 24 September 2012 01:12, Colin Watson <cjwat...@ubuntu.com> wrote: > On Sun, Sep 23, 2012 at 04:22:21PM +0100, o jordan wrote: > > I don't see any value in rolling back to A-3 or even rolling back > > specific packages (e.g. the xserver radeon driver which is the change > > that did it) to A-3. > > It's entirely unrealistic anyway over such a long time period. > Individual packages over a relatively short time period, sure, if they > are known to be the cause; the entire X stack (given that alpha-3's > Radeon driver package was for a different version of the X server with > an incompatible ABI) across two months, not so much! > > Reverting changes that cause regressions is sensible in many cases. But > the point of it is to get us back to a more stable state, not to spend > ages undoing an enormous amount of work and take us into a new > substantially-different combination of software that nobody has ever > tested before. > > > PowerPC can install fine. It can install fine on radeon. It just > > needs a yaboot parameter applying. It is how to comunicate this > > information to the user that is the problem. People just don't read > > information that is given to them. > > It's long been a guiding principle of Ubuntu that users shouldn't have > to set boot parameters by hand. If you've got to the point of > identifying which specific package version change is responsible, it > might not be so much harder to narrow it down to a particular > (presumably upstream) change, so that the X team can arrange for this to > work by default. > > > But, honestly, it is a really easy > > problem to overcome and I don't think a big deal needs to be made of > > it. I'm much more interested in getting things like > > https://bugs.launchpad.net/ubuntu-cdimage/+bug/1051313 and > > Applied, thanks. > > > https://bugs.launchpad.net/ubuntu/+source/ubiquity/+bug/1043066 fixed. > > I've applied the patch to yaboot-installer and uploaded, but it may be > too late for beta-2 and it will need a separate ubiquity upload as well > in any event. > > > The problem is with the changes that have been made > > to the xserver-xorg-video-radeon package this config now drops you > > into 8 bit graphics (forcing you this time to use a yaboot parameter). > > What is happening is that the radeon xorg driver detects drm is > > unavailable and unloads itself and the fbdev diver is used instead > > (for some reason this defaults to 8 bit). This didn't used to happen. > > You could certainly ask the question if it is possible to have KMS > > without drm? Not being a graphics expert, I don't know if that is a > > stupid question or not. > > The right answer, then, is to ask a graphics expert. :-) Try #ubuntu-x. > > > I can't explain it any clearer. Unless you can have KMS without drm, > > then the only choice is do you remove radeonfb for 12.10 or not. > > Either way, the boot message on the CDs needs to be updated and the > > timer removed so that people have a chance to read the new message. > > As I tried to explain on IRC, I'm extremely wary of updating the boot > message until I have acknowledgement from the X team (or kernel team if > necessary) that there's no other way. Boot messages that explain that > people have to set boot parameters are not something we like to do if at > all possible. > > > Does non-PowerPC Lubuntu suffer from this bug? Basically a previously > installed Lubuntu is labelled as Ubuntu in grub/yaboot: > https://bugs.launchpad.net/ubuntu/+source/os-prober/+bug/950914 > > I've followed up to this bug. But, in short, it has existed in some > form since the very first Ubuntu flavour was created in 5.04, and there > is no good reason to consider it any more important now than it was > then. > > Cheers, > > -- > Colin Watson [cjwat...@ubuntu.com] > -- https://wiki.ubuntu.com/phillw
-- Mailing list: https://launchpad.net/~lubuntu-qa Post to : lubuntu-qa@lists.launchpad.net Unsubscribe : https://launchpad.net/~lubuntu-qa More help : https://help.launchpad.net/ListHelp
