openSUSE:Testing meeting 20100628

Jump to: navigation, search


  1. Own experience fo openSUSE RC1 + Factory


(18:59:45) LWFinger: Hi (19:00:04) bmwiedemann: Hi. I wonder if Holgi will be late. or rather how much. he left 18:06 (19:00:36) bmwiedemann: (CEST - so 54 mins ago) (19:01:17) bmwiedemann: but we could just start (19:01:30) LWFinger: OK (19:01:52) bmwiedemann: only 1 topic on today (19:01:59) bmwiedemann: "own experience of openSUSE 11.3 RC1+ Factory" (19:02:28) LWFinger: For me, I did not find any new bugs. (19:03:29) LWFinger: That is with 3 systems - all KDE. Two zypper dup upgrades and 1 full installation from the NET install CD (19:04:03) bmwiedemann: I have been running my automated tests again. And saw zypper complain about an outdated repo, failing to install updates. but only sometimes. (19:04:33) bmwiedemann: plus the LVM bug was still there yesterday (19:04:50) LWFinger: What is the LVM bug? (19:05:04) bmwiedemann: #1 on most annoying list. (19:05:05) elm: yes, I have discovered it too, but did not report about it,yet (19:05:16) bmwiedemann: I did open a bug (19:05:22) elm: bugno? (19:05:30) bmwiedemann: (19:06:11) bmwiedemann: it appears that it is not really bad, but it could be better. (19:06:48) Holgi: jo all (19:06:51) elm: oh very few bugs for 11.3 rc1 ! (19:07:03) LWFinger: Holgi: Hi (19:07:09) elm: will have to emit some as soon as my RC1-inst goes online. (19:07:09) bmwiedemann: elm: I have some more on my personal watch list ;) (19:07:37) Holgi heißt jetzt HolgiDE (19:07:43) HolgiDE hat den Raum verlassen (quit: Changing host). (19:07:43) HolgiDE [~holgi@CAcert/Assurer/HolgiDE] hat den Raum betreten. (19:07:43) Modus (+o HolgiDE) von ChanServ (19:08:13) LWFinger: The LVM bug is listed as fixed. How long does it take from acceptance into Factory and the fix in the iso? (19:08:20) bmwiedemann: 610199 612143 613824 (19:08:35) bmwiedemann: LWFinger: usually around 2 days (19:09:46) bmwiedemann: repo still has version from 2010-06-10 (19:11:52) bmwiedemann: HolgiDE: is it true that only bugs>=Major are fixed since MS7? (19:12:23) LWFinger: Based on the Bugzilla date, the fix for 598193 went to Factory today. (19:12:40) HolgiDE: bmwiedemann: general yes, but it's also possible that some others gets fixed if there is a new package being submitted (19:13:01) elm: has anyone tested apparmor? (19:13:12) bmwiedemann: elm: not me (19:13:14) elm: there had been some unresolved bugs with it at 11.2. (19:13:35) elm: will do this soon (possibly this evening) (19:13:50) LWFinger: I have not as I do not run a distributed. openSUSE patched, kernel. (19:14:10) elm: you run vanilla? (19:14:26) LWFinger: elm: yes. (19:14:55) elm: why? is this for more direct kernel bug reporting? (19:15:17) elm: even the jengelh -multimedia kernel has apparmor. (19:15:20) LWFinger: elm: I do wireless development and mainline testing. (19:15:55) LWFinger: elm: For both I use git updated kernels. (19:16:06) elm: mainline testing - vanilla? (19:16:14) elm: mhm; thx. (19:16:38) LWFinger: Yes, I'm running 2.6.35-rc3. (19:17:19) elm: so, then;: Have we already tested all main areas of RC1 ? - wlan, bluetooth, apparmor, nvidia, ati, (lvm, kde, gnome, pkgmgm: sure) (19:17:26) elm: any other area to test? (19:18:27) LWFinger: I have tested wlan enough to know that NetworkManager did not get broken. (19:18:54) elm: Some broad user area testing before a release (RCx) can never harm, I believe. (19:18:56) HolgiDE: elm: let's do the other way around - what did you test so far? (19:19:10) elm: I have just installed RC1. (19:19:23) HolgiDE: elm: so real RC1 or a later build? (19:19:36) elm: graphics so far only works with nomodeset on my external monitor - still with some bugs. (19:19:50) LWFinger: elm: What adapter? (19:20:40) elm: ATI Mobility Radeon HD-2700, 256MB (19:21:05) elm: Nonetheless it already works much better than in 11.2 times. (19:21:42) LWFinger: In the Fora, there are the usual complaints with the new graphics drivers. For nVidia users, noureau seems to be the worst. (19:21:44) elm: The screen had been flashing black these days which had been extremly annoying (probably was one of these kernel issues). (19:21:51) HolgiDE: elm: maybe a manual xorg.conf would help (19:22:13) HolgiDE: elm: prevents Xorg to probe every time (19:22:14) elm: yes, I have tested with different manual xorg.confs. (19:23:03) elm: will report this soon (tomorrow or so). (19:23:32) elm: The graphics adapter is not well supported (no 3D up to now as far as I know.): (19:23:39) elm: now other bugs: (19:23:45) bmwiedemann: (just had to help with a 11.2 install on a laptop with ATI 125x something graphics which not only showed a black screen but did not accept input beyond sysrq. probably bad driver thre) (19:23:56) HolgiDE: elm: problem is: just (real) "critical"+"blocker" bug will be fixed during RC time (19:24:20) LWFinger: I finally had to dig around in /etc/X11/xorg.conf.d to fix a touchpad problem. Anything there for you? (19:24:22) HolgiDE: elm: so maybe yours are too late for GM (19:24:45) bmwiedemann: RC2 is due in 3 days. (19:25:07) elm: I hope that as many bugs as possible will become resolved until the final release?! (19:25:43) LWFinger: elm: We all do. (19:25:57) elm: glad to hear. (19:26:12) elm: too late for GM? (19:26:37) elm: the graphics driver issue may in deed be too laborious. (19:26:38) LWFinger: We are still fighting the two networking bugs introduced late in the 11.2 RC cycle. (19:27:09) HolgiDE: LWFinger: which? (19:27:13) LWFinger: The graphics driver problems are upstream. Until fixed there, no joy here. (19:27:45) LWFinger: HolgiDE: The routing and DNS problems. (19:28:08) HolgiDE: LWFinger: with or without NM? (19:29:02) LWFinger: HolgiDE: Both - the problem is that those bugs are on the 11.2 GM. Both were quickly fixed, but without network, no update. (19:29:08) elm: shouldn`t I report none of the graphics bugs at (19:29:33) HolgiDE: elm: no, please report all, but don't expect them all to be fixed (19:29:49) LWFinger: elm: Yes, but they will likely need to be pushed to the kernel devs. (19:30:47) elm: nonetheless I have found some minor issues that should be very easy to correct: MANPATH, fsck output on boot lost, a usability issue. (19:30:56) elm: . (19:31:07) HolgiDE: LWFinger: and there bugs for 11.3 for this issue? what severity? (19:31:22) elm: minor. (19:31:54) LWFinger: HolgiDE: No, these are fixed and no new ones found in networking. 11.3 looks good in this area. (19:32:21) HolgiDE: LWFinger: and those for 11.2? (19:33:02) LWFinger: IHolgiDE: Long fixed, but the GM was not remastered. As a result, any new user gets them. (19:33:51) HolgiDE: LWFinger: ahh, those are "just" in install systems kernel (19:35:04) LWFinger: Actually, one is in NetworkManager and one in name serving - not kernel problems at all. (19:35:31) HolgiDE: LWFinger: ok (19:36:12) elm: what about the following idea: (19:36:18) elm: Let us note who has already tested the following main areas: (19:36:22) bmwiedemann: is still open in 11.2, too - but was good in 11.3. (19:36:53) elm: wlan, bluetooth, apparmor, nvidia, ati, (lvm, kde, gnome, pkgmgm: sure) (19:37:13) elm: bmwiedmann: ok then? (19:37:23) elm: and attach all bugs found to the resp. area (19:37:29) elm: -> quick overview what is sure to work and what needs testing (19:37:38) LWFinger: bmwiedemann: That sounds like a kernel regression. I wonder if the kernel Bugzilla has it? (19:38:07) bmwiedemann: LWFinger: that is just the raid10.ko module not loaded. fix is rather trivial. (19:38:46) LWFinger: bmwiedemann: OK, a configuration problem. (19:38:46) bmwiedemann: elm: I run wlan on 11.3-RC1. no problems there. (19:39:21) LWFinger: elm: wlan has been good since M4. (19:39:35) bmwiedemann: only rfkill bug, but that is kernel/driver-related: 595586 (19:39:38) elm: will test bluetooth and DUNs, soon (there had been issues with 11.2) (19:39:57) elm: :) very well then (19:40:14) elm: 11.3 seems to become the best release for long. (19:40:34) HolgiDE: did anybody of you succeed in getting an openvpn connection run through network-manager? (19:40:42) LWFinger: bmwiedemann: I looked over the record on 595586. I think that needs to go the wireless MLO - I will do so today. (19:41:02) bmwiedemann: HolgiDE: I only did a manual openvpn setup. that worked. (19:41:36) HolgiDE: bmwiedemann: I did not get it working out of the box with gnome (19:41:46) HolgiDE: bmwiedemann: but did not have the time to debug (19:41:56) HolgiDE: bmwiedemann: now I'm with build 687 (19:42:05) HolgiDE: bmwiedemann: and did not try so far (19:43:27) bmwiedemann: could try it on my eeepc. (19:43:53) LWFinger: bmwiedemann: Is that a Netbook? (19:44:16) bmwiedemann: yes (19:44:59) LWFinger: bmwiedemann: There seem to be a lot of problems with the support chip set that Intel uses for Atom cpus. (19:45:02) elm: Is the LVM bug really resolved? (19:45:10) elm: it is not marked as fixed. (19:45:28) elm: would have something to add ... (19:45:43) bmwiedemann: elm: did not the fix for now. all LVM tests still fail. (19:45:54) bmwiedemann: LWFinger: (19:45:59) elm: ok; then happy reporting. (19:46:03) LWFinger: bmwiedemann: We are chasing a DMA problem with b43 on Netbooks with Phoenix BIOS - very specific. (19:50:39) bmwiedemann: I am still tuning my automated tests. e.g. did you know that KDE Plasma layout varies by 1 pixel depending on CPU (32/64 bit) ? ff32.png ff64.png (19:51:13) elm: strange. how can that be? (19:52:02) bmwiedemann: also some pixels are off by one or two bit in colour. might be rounding issues - when people do floating point calculations with long int values. (19:54:15) [daemon]_ [] hat den Raum betreten. (19:55:42) elm: what software do you use for auto-testing? (19:56:54) elm: no matter. (19:57:00) bmwiedemann: kvm (19:57:25) bmwiedemann: and my own scripts, which I wanted to upload to gitorious soon (already got me an account there) (19:57:28) elm: and the mouse events and snapshots?� how are they generated? (19:57:40) elm: very interesting. (19:57:51) elm: I always wanted to do similar stuff. (19:58:24) bmwiedemann: elm: you can see the commands I send and replies I get in (19:59:02) elm: ... but I lacked in an approach to remote-control and screen snapshot kvm. (20:00:30) bmwiedemann: elm: some docs (INSTALL) linked at (20:00:50) elm: thx. (20:03:00) bmwiedemann: elm: but mostly the real stuff is in my code. you do kvm -monitor tcp:... and then can issue screendump, sendkey and mouse_move command lines. (20:04:18) elm: thx. This is exactly what I have been asking for (unfortunately not booted under Linux - so I can`t have a look by now). (20:04:46) bmwiedemann: elm: there is even a qemu for the Redmond OS (tm) (20:05:46) elm: ... and the doku of this functions is in the man page, or not? (20:05:54) elm: or info page (20:06:13) elm: because the kvm-man is scarse of details as far as I can remember. (20:06:30) elm: regarding this. (20:07:25) bmwiedemann: elm: (20:07:50) elm: thx. (20:07:51) LWFinger: I will need to leave soon. Will we meet next Monday? (20:08:08) bmwiedemann: LWFinger: I would like to. (20:08:24) elm: willingly. (20:08:27) bmwiedemann: so 2010-07-05 17:00 (20:08:36) LWFinger: OK (20:09:06) LWFinger: Bye - my machine will be on, but I will not be here. (20:10:02) bmwiedemann: I think we can thereby conclude the official meeting (and go on chatting anyway) (20:10:59) bmwiedemann: elm: I must admit that both documentation and functionality is not what it could be. e.g. only last week I got a patch merged into qemu that allows to send :;[]{}'" and such (20:12:51) bmwiedemann: and mouse_button does not do the right thing in touchpad mode (20:14:57) bmwiedemann: got to leave now. I wish you a nice evening. And have a lot of fun... (20:15:16) elm: many thanks; good bye. (20:15:47) elm: may ask you if I shouldn`t get anything? (20:17:26) elm: seems to be very clear and straight forward to use. (20:17:55) elm: thxx. - & nice evening for you, too bmwiedmann.