It has been a few years since I followed/tested the Fedora development cycle. I am pleasantly surprised and pleased with the quality of the process, the responsiveness of the maintainers/developers, and the quality of the resulting product ... a really good job! I am pleased enough with F12 that I plan to upgrade all of my F11 systems to F12 when F12 is released and I have not updated like this in years (I recently upgraded from F9 to F11).
I am hoping that the recently described problems with the X server and firstboot are not a show-stopper for the Beta and that the Beta does appear next week so that I can give installing it a try. For all intents and purposes, I am running the beta now since I have kept my real (baremetal) F12 systems updated and well as installing the F12beta release candidates as qemu- kvm guests. But, when the beta is out, I will re-install using the beta DVD to make sure that works.
As I said, it has been a few years since I followed the development process. I assume that the current lack of additional rawhide updates is so that a "good" beta release could be done and that updates will resume once the beta DVD, etc. have been cut ... is this correct?
I also assume that the addition of function is now frozen with the beta and that only bug fix updates will be added to rawhide ... correct?
Gene
On 10/15/2009 04:28 AM, Gene Czarcinski wrote:
As I said, it has been a few years since I followed the development process. I assume that the current lack of additional rawhide updates is so that a "good" beta release could be done and that updates will resume once the beta DVD, etc. have been cut ... is this correct?
I also assume that the addition of function is now frozen with the beta and that only bug fix updates will be added to rawhide ... correct?
Correct. We are going through a Beta freeze. Maintainers aren't allowed to push updates to Rawhide directly anymore and would have to submit a request to release engineering to tag the update explaining why the fix is important enough (typically it is small enough as well).
The links from the schedule describe the freeze policies in more detail
https://fedoraproject.org/wiki/Releases/12
https://fedoraproject.org/wiki/Beta_Freeze_Policy
If you have more questions, feel free to ask.
Rahul
On Wed, 2009-10-14 at 18:58 -0400, Gene Czarcinski wrote:
I am hoping that the recently described problems with the X server and firstboot are not a show-stopper for the Beta and that the Beta does appear next week so that I can give installing it a try.
The problem only affects systems with multiple displays (or, more rarely, systems with one display but a driver bug which makes X _think_ they have multiple displays, as in Robert's case). It's not an absolute showstopper, you can blindly tab-enter your way through the wizard, or you can just boot the first time with the second display not connected. We won't be delaying the beta for it, is the final decision, it'll be fixed post-beta.
I hope to have time to do another install soon. I apologize for not being more active and filing bugs as I should. This fall is really busy for me.
Bob Cochran
On 10/14/2009 07:10 PM, Adam Williamson wrote:
On Wed, 2009-10-14 at 18:58 -0400, Gene Czarcinski wrote:
I am hoping that the recently described problems with the X server and firstboot are not a show-stopper for the Beta and that the Beta does appear next week so that I can give installing it a try.
The problem only affects systems with multiple displays (or, more rarely, systems with one display but a driver bug which makes X _think_ they have multiple displays, as in Robert's case). It's not an absolute showstopper, you can blindly tab-enter your way through the wizard, or you can just boot the first time with the second display not connected. We won't be delaying the beta for it, is the final decision, it'll be fixed post-beta.
On Wed, 2009-10-14 at 21:05 -0400, Robert L Cochran wrote:
I hope to have time to do another install soon. I apologize for not being more active and filing bugs as I should. This fall is really busy for me.
I was referring to Robert Day, who reported the bug in question this monrning :)
On 10/14/2009 09:50 PM, Adam Williamson wrote:
On Wed, 2009-10-14 at 21:05 -0400, Robert L Cochran wrote:
I hope to have time to do another install soon. I apologize for not being more active and filing bugs as I should. This fall is really busy for me.
I was referring to Robert Day, who reported the bug in question this monrning :)
I got my just desserts for not reading the entire thread.
Bob
On Wed, Oct 14, 2009 at 09:05:02PM -0400, Robert L Cochran wrote:
I hope to have time to do another install soon. I apologize for not being more active and filing bugs as I should. This fall is really busy for me.
Bob Cochran
I don't even know if this is worth mentioning. I did a fairly quick test install at work, the way I usually do it. (This will, eventually be for a developer.) As it's work, I don't have time to really play around and see what's broken. But, at any rate, taking disk1 of the beta, unchecking everything, it installed without problem in VirtualBox. I then updated (nothing to update) and installed x-base, fluxbox and a few other things. Everything was smooth. I did the very minor test of changing default partitioning, just using a swap and a / of ext4 (no separate /boot.)
No issues. This is a very superficial test, and on a Virtual Machine.
However, an encouraging start.
On Wed, 2009-10-14 at 18:58 -0400, Gene Czarcinski wrote:
As I said, it has been a few years since I followed the development process. I assume that the current lack of additional rawhide updates is so that a "good" beta release could be done and that updates will resume once the beta DVD, etc. have been cut ... is this correct?
Partly. Some fixes still won't be accepted, though, and will be required to be 0-day updates or pushed to F-13.
I also assume that the addition of function is now frozen with the beta and that only bug fix updates will be added to rawhide ... correct?
Yes.
On Wed, 2009-10-14 at 18:58 -0400, Gene Czarcinski wrote:
It has been a few years since I followed/tested the Fedora development cycle. I am pleasantly surprised and pleased with the quality of the process, the responsiveness of the maintainers/developers, and the quality of the resulting product ... a really good job! I am pleased enough with F12 that I plan to upgrade all of my F11 systems to F12 when F12 is released and I have not updated like this in years (I recently upgraded from F9 to F11).
I'm glad you've had a great experience, but I'd be 'testing' each machine before just assuming f12 is good to go.
I'm having nightmares (which are slowly being dealt with). I'm pretty sure there a number of issues I will need to file bugs on (for example, pressing the eject button doesn't eject the disk in f11) but until I can run the desktop for more than a minute or so without a near total lockup then it's hard to really contribute in any other way.
I'm on a Dell Studio XPS so it's relatively modern hardware, but it's not exactly cutting edge. It's the ATI video card that seems to be causing most of the issues.
I've been wondering for some time whether a serious review of the way bugs are 'triaged' and resources are applied isn't needed. While I'm happy that other bugs for 'less vital[1]' parts of F12 are being addressed, I would have thought that basic system requirements like X, the kernel should just work, and if they don't, then extra effort should be put into addressing these issues.
As justification for this, I'm happy to contribute, but without a working X I can't really help out until the problem is resolved, which I'm sure everyone would admit is a big stumbling block. Hands up anyone who thinks it matters that <their favorite application> has a bug, if they can't get X running to use <their favorite application>. It's fair to say that <their favorite application>, working or otherwise, is pointless without a platform to run it on.
Anyhow, I'd running a live CD before upgrading lock, stock and barrel to f12. Not everyone's experiences have been so positive as yours.
Rodd
[1] I know less vital is vague, but I would class being able to run an alternative web browser as a less important bug than making sure that X works and is stable, as an example.
On Thu, Oct 15, 2009 at 10:47:06 +1100, Rodd Clarkson rodd@clarkson.id.au wrote:
I'm on a Dell Studio XPS so it's relatively modern hardware, but it's not exactly cutting edge. It's the ATI video card that seems to be causing most of the issues.
There have been some recent improvements from me when running with KMS enabled.
As justification for this, I'm happy to contribute, but without a working X I can't really help out until the problem is resolved, which I'm sure everyone would admit is a big stumbling block. Hands up anyone who thinks it matters that <their favorite application> has a bug, if they can't get X running to use <their favorite application>. It's fair to say that <their favorite application>, working or otherwise, is pointless without a platform to run it on.
As the maintainer of the games spin not having working graphics is a real pain. Because my monitors don't do EDID correctly, graphics hasn't worked right for a long time. Now I get 800x600 by default and 3d works enough on my ati 9200 that I can do some testing with a real games spin build. (I use xorg.conf files and had been using nomodeset to make my systems usable, but those are in the normal games spin.) Also testing glest (which I recently started maintaining) is hard with 3d not working very well.
On Thu, 2009-10-15 at 10:47 +1100, Rodd Clarkson wrote:
Hands up anyone who thinks it matters that <their favorite application> has a bug, if they can't get X running to use <their favorite application>.
The problem is that bugs in X tend to be device-specific (and hence only affect people who are unfortunate enough to own that exact device), but bugs in, say, Firefox affect everybody.
The other issue is that the developers of, say, Firefox know crap all about fixing bugs in X. And vice versa. Coders aren't interchangeable resources, you don't just pick them at random and point them at problems. (Well, some people do. This appears to be how Windows is developed...) Even more specifically...Ben Skeggs (whose area of expertise is nouveau) might be able to fix basic bugs in radeon, but probably nothing complex. Not, at least, without a lot of extra training, during which time no-one's fixing bugs in nouveau.
On Wed, 2009-10-14 at 17:13 -0700, Adam Williamson wrote:
On Thu, 2009-10-15 at 10:47 +1100, Rodd Clarkson wrote:
Hands up anyone who thinks it matters that <their favorite application> has a bug, if they can't get X running to use <their favorite application>.
The problem is that bugs in X tend to be device-specific (and hence only affect people who are unfortunate enough to own that exact device), but bugs in, say, Firefox affect everybody.
I don't know if Firefox works or not ;-]
But I get that X is complex, I'm just saying that it would be great if more resources could be put into making the core system work (which would 'free up' more 'affected' people to find other issues and address them.)
The other issue is that the developers of, say, Firefox know crap all about fixing bugs in X. And vice versa. Coders aren't interchangeable resources, you don't just pick them at random and point them at problems. (Well, some people do. This appears to be how Windows is developed...) Even more specifically...Ben Skeggs (whose area of expertise is nouveau) might be able to fix basic bugs in radeon, but probably nothing complex. Not, at least, without a lot of extra training, during which time no-one's fixing bugs in nouveau.
I also get that it's very complex, but to be honest, as someone who has had a series of sub-par performances on a variety of laptops due to pour chipset support, I'd really like to see Redhat/Fedora putting a bunch of effort into this area for a release cycle and getting things a whole bunch better. Having graphics that just works is a big win for everyone.
The best graphics experience I've had with Linux (which for me is Fedora and Centos) has been using the proprietary nvidia driver (and that has a times sucked badly too). And while I'm willing to go without 3D support on my work desktop, basics like having a stable desktop, and having suspend/resume support aren't in my opinion small issues to be ignored. Ironically, I chose my current laptop (with it's ATI chipset) because I thought the ATI support was supposed to be reasonable and because I admired Fedora stance on free software and avoiding unsupported, proprietary drivers. (I now recommend people using nvidia and and proprietary drivers, or intel if they have to and avoid ATI all together)
R.
-- Adam Williamson Fedora QA Community Monkey IRC: adamw | Fedora Talk: adamwill AT fedoraproject DOT org http://www.happyassassin.net
On Thu, 2009-10-15 at 14:50 +1100, Rodd Clarkson wrote:
The best graphics experience I've had with Linux (which for me is Fedora and Centos) has been using the proprietary nvidia driver (and that has a times sucked badly too). And while I'm willing to go without 3D support on my work desktop, basics like having a stable desktop, and having suspend/resume support aren't in my opinion small issues to be ignored.
No-one ever said they were.
nouveau bugs closed during F12 Rawhide cycle: 53 https://bugzilla.redhat.com/buglist.cgi?query_format=advanced&classifica...
radeon bugs closed during F12 Rawhide cycle: 80 https://bugzilla.redhat.com/buglist.cgi?query_format=advanced&classifica...
intel bugs closed during F12 Rawhide cycle: 58 https://bugzilla.redhat.com/buglist.cgi?query_format=advanced&classifica...
Fedora works hard on these components. They're priorities for triage, blocker review, everything else. Red Hat works hard on these components. We pay more X.org developers than any other distributor, as far as I know. We're the only distribution who pays lead upstream developers to work on all three drivers, as well as the rest of X.org, including the server itself.
I'm just trying to make the point that both RH and Fedora work extremely hard on these components. You seem to be implying that (both RH and Fedora) we don't, which is not the case.
Ironically, I chose my current laptop (with it's ATI chipset) because I thought the ATI support was supposed to be reasonable and because I admired Fedora stance on free software and avoiding unsupported, proprietary drivers. (I now recommend people using nvidia and and proprietary drivers, or intel if they have to and avoid ATI all together)
I'm not sure why you'd have got that impression. ATI is getting a lot better recently, but it's still not really on a level with Intel as far as free driver support goes, or NVIDIA as far as proprietary driver support goes.
On Wed, 2009-10-14 at 21:46 -0700, Adam Williamson wrote:
I'm not sure why you'd have got that impression. ATI is getting a lot better recently, but it's still not really on a level with Intel as far as free driver support goes, or NVIDIA as far as proprietary driver support goes.
Adam
Ah, maybe I'm being a little over grumpy. If that's what Fedora is doing for X then I'm very impressed and more than a little in awe as to the contributions made to X by Fedora and Redhat.
It's easy to put on rose tinted glasses and as I've had nothing but issues with my ATI card, it's really easy to think that ATI support really sucks and that no-one is doing anything about it. So if I'm sounding a little glum, it's possibly slanted by my less than stellar experiences.
Here's hoping my X issues will be addressed soon, so I can get on with testing.
Thanks for taking the time to respond.
Rodd
On Thu, 2009-10-15 at 18:26 +1100, Rodd Clarkson wrote:
On Wed, 2009-10-14 at 21:46 -0700, Adam Williamson wrote:
I'm not sure why you'd have got that impression. ATI is getting a lot better recently, but it's still not really on a level with Intel as far as free driver support goes, or NVIDIA as far as proprietary driver support goes.
Adam
Ah, maybe I'm being a little over grumpy. If that's what Fedora is doing for X then I'm very impressed and more than a little in awe as to the contributions made to X by Fedora and Redhat.
It's easy to put on rose tinted glasses and as I've had nothing but issues with my ATI card, it's really easy to think that ATI support really sucks and that no-one is doing anything about it. So if I'm sounding a little glum, it's possibly slanted by my less than stellar experiences.
Here's hoping my X issues will be addressed soon, so I can get on with testing.
Thanks for taking the time to respond.
No problem. I understand how you feel - I get frustrated when my hardware's broken too. For the record, RH employs two radeon developers, Dave Airlie and Jerome Glisse, who both contribute actively to upstream development of the driver and to maintaining it in Fedora (though Dave is doing an increasing amount of work on the X server side, which I think is partly why we brought Jerome in).
I'm going to have another look at your bug(s) and see if there's anything we can look at that we didn't already...
On Thu, 2009-10-15 at 18:26 +1100, Rodd Clarkson wrote:
Here's hoping my X issues will be addressed soon, so I can get on with testing.
BTW, on the triage topic - we (Fedora this time :>) used to have Francois Cami in the Bugzappers, who more or less exclusively worked on triaging radeon bugs, which was a great help. Unfortunately he hasn't been able to do much on Fedora for a while due to other commitments. Nevertheless, it looks like your latest bug - https://bugzilla.redhat.com/show_bug.cgi?id=528593 - got picked up for triaging by Matej two days after being filed, which is pretty quick, I'd say.
On Thu, 2009-10-15 at 08:36 -0700, Adam Williamson wrote:
On Thu, 2009-10-15 at 18:26 +1100, Rodd Clarkson wrote:
Here's hoping my X issues will be addressed soon, so I can get on with testing.
Just for clarity, I didn't mean by this that they weren't being "addressed". I guess I was saying, here's hoping they are resolved soon.
BTW, on the triage topic - we (Fedora this time :>) used to have Francois Cami in the Bugzappers, who more or less exclusively worked on triaging radeon bugs, which was a great help. Unfortunately he hasn't been able to do much on Fedora for a while due to other commitments. Nevertheless, it looks like your latest bug - https://bugzilla.redhat.com/show_bug.cgi?id=528593 - got picked up for triaging by Matej two days after being filed, which is pretty quick, I'd say.
And yes, while I have had a couple of bugs of late that seemed to wallow in bugzilla (a few bigger ones I've nagged about on the list hoping for some action, which works some times) this particular problem was picked up quite quickly and seems to be moving forward in an reasonable manner (since it's kind of a huge stumbling block for me).
I should take a moment to thank you for your help both here, and with this bug too.
Rodd