Not logged in
Log in now
Create an account
Subscribe to LWN
LWN.net Weekly Edition for May 23, 2013
An "enum" for Python 3
An unexpected perf feature
LWN.net Weekly Edition for May 16, 2013
A look at the PyPy 2.0 release
Ubuntu 12.04 LTS "Precise Pangolin" released
Posted Apr 27, 2012 8:18 UTC (Fri) by boudewijn (subscriber, #14185)
There must be something really right about this release :-)
Posted Apr 27, 2012 8:38 UTC (Fri) by alecs1 (guest, #46699)
Report the God Damn bug upstream on behalf of your user.
Some guy from Launchpad reacted very quickly and talked on IRC with a KDE developer about that bug, but came back and told the reporter: "go report the bug upstream" and closed the bug as invalid, when he could have spent 3 more minutes to report upstream. I don't want to offend the guy who tried to be helpful, but I find this sort of behaviour is irresponsible and amounts to shoving crap on your users' throat. The original reporter never reported back, and I would do the same if I would get this kind of crap.
An upgrading from Oneiric to Precise made the system unbootable. I think the room for improvement is huge, Ubuntu is a far cry from getting their act together.
Posted Apr 27, 2012 8:48 UTC (Fri) by Mithrandir (subscriber, #3031)
Posted Apr 27, 2012 9:57 UTC (Fri) by alecs1 (guest, #46699)
The point stands: he spent some time to work on this, with 5 more minutes he could have achieved a lot more: give trackability to the bug, give the "customer" a sense that the vendor cares. None of my former employers would ever dream ignoring a client like that, even if bugs would linger for 10 years.
Posted Apr 27, 2012 12:06 UTC (Fri) by pboddie (subscriber, #50784)
As for the "entitlement" jibe, I guess the only response to that kind of narrow-minded thinking is for everyone to not bother reporting bugs at all and to just use something else instead. Yes, users can withdraw value from the interaction, too.
Posted Apr 27, 2012 8:51 UTC (Fri) by boudewijn (subscriber, #14185)
Btw, your launchpad guy, Apachelogger, is also a KDE guy :-).
Posted Apr 27, 2012 10:04 UTC (Fri) by alankila (subscriber, #47141)
Posted May 1, 2012 10:38 UTC (Tue) by nye (guest, #51576)
This is a good way to ensure that your software is only ever 90% done though. There will always be issues that affect only a few users since no two users are exactly the same, and usually so many of them that almost *all* users see a few of them.
Posted May 1, 2012 16:11 UTC (Tue) by raven667 (subscriber, #5198)
Posted Apr 27, 2012 9:58 UTC (Fri) by Alterego (subscriber, #55989)
For me this behavior is frequent on ubuntu, i got several bug reports closed that way, without anything done, even when i provided test cases.
So i switched back to debian and RHEL and am happy worker again :)
Posted Apr 27, 2012 11:19 UTC (Fri) by tpo (subscriber, #25713)
(this is here because +1 doesn't qualify as a valid comment)
Posted Apr 27, 2012 12:55 UTC (Fri) by sb (subscriber, #191)
Posted May 3, 2012 7:02 UTC (Thu) by ersi (subscriber, #64521)
I do however recognize the feeling of strongly agreeing with a comment from time to time. So you're not alone in that, I'll say. I do however see the comment section more like a mailing list, as in a reply-to-all way - and it's generally frowned upon to send messages without content :-)
Posted Apr 27, 2012 17:34 UTC (Fri) by mathstuf (subscriber, #69389)
What if the maintainer can't reproduce the issue or doesn't personally care about the issue nearly as much as you do? The maintainer is not always the best person to forward things upstream. *You* care and have the data and are more useful on the upstream bug that the aforementioned maintainer. With the inability to tack other users onto the CC list in most bug trackers I've seen, both the maintainer and the user have to go to the bug anyways. A maintainer's time is not well spent playing messenger between upstream and users that have encountered some bug just because the user cares about a bug, yet not enough to talk directly with upstream.
That said, if the issue is a common one or the maintainer can provide more information because they either know the code well enough to tell or dug a little deeper, then they might be a better fit for filing upstream.
They should care, but they don't have to care as much as users would like if it's some corner case.
Posted Apr 27, 2012 18:55 UTC (Fri) by alecs1 (guest, #46699)
I don't use Ubuntu (mainly because of poor upgradability, but while testing I did give up Unity precisely because of Klipper), but should I get treated like the reporter, I'll just leave, and not necessarily to Debian. And guess what, Ditto is a better replacement to Klipper and works OK on all Windows versions I tried. And Lyx will accept my bug report and donations from Windows too. Same will stand for Pidgin, Marble, Okular and QtCreator etc.
I'll stop, and also apologise Harald Sitter that I chose his bug to exemplify what I consider a costly mistake. I'd also report the bug and link it from Launchpad, but I have no idea if it reproduces.
Posted Apr 27, 2012 19:11 UTC (Fri) by mathstuf (subscriber, #69389)
Looking at it, it was indeed handled poorly (overall). Instead of being closed as Invalid, it should probably have been kept open and some indication that there was an upstream bug to be tracked in it (I'm a Fedora maintainer and an RHBZ bug can be associated with other BZs). That way, when the upstream bug is fixed, it can be asked to be backported or whatever on the distro side. Of course, I don't know the Ubuntu bug life cycle, so maybe this was "valid" under that, but I'd say if that was the case, the lifecycle needs fixed.
Posted Apr 29, 2012 10:41 UTC (Sun) by jond (subscriber, #37669)
Posted Apr 29, 2012 14:46 UTC (Sun) by mathstuf (subscriber, #69389)
Posted Apr 27, 2012 16:08 UTC (Fri) by tjc (subscriber, #137)
It's been out one day!
Maybe in two weeks we'll be able to tell if it's a solid release or not.
Posted Apr 27, 2012 17:07 UTC (Fri) by Cyberax (✭ supporter ✭, #52523)
cgroup-bin breaks suspend on my computers (bug is reported and left without attention), radeon on dual-GPU notebooks has to be blacklisted, a lot of apps crash with 'ASSERT failure in : "Got an update for an invalid inteface. Investigate this.", file atspiadaptor.cpp, line 899' because of some changes in accessibility support. And so on.
In short, it's a usual Ubuntu release.
Posted Apr 27, 2012 19:54 UTC (Fri) by tcourbon (subscriber, #60669)
I have been also running it since Jannuary and find it quite stable. Also I second the assertion that Unity looks more like a viable shell than before. My workflow still fit better in KDE but I'm kind of fond of the HUD ("look mommy I use the drop-down menus without touching the mouse").
I never got into the Ubuntu-hate mood and that's not the release that will push me into the hating mud.
Posted Apr 27, 2012 21:10 UTC (Fri) by tjc (subscriber, #137)
I've been running Ubuntu since Warty, and I've come to expect a certain amount of brokenness with a new release. Usually a subsequent release will fix problems with the prior, but introduce a new set of problems of its own. This has motivated me to skip about every other release. Unfortunately, I've had more problems with the LTS releases, but this may be that I'm just "unlucky at hardware" and not characteristic of other people's experiences.
I don't know if this is really avoidable, considering all the untested hardware that it runs on, or if other Linux distros are any better. I've also run Debian testing for many years, and while that is sometimes a bit rough around the edges, the problems are more incremental in nature, so I don't notice them that much (with a couple of alarming exceptions over the years).
Posted Apr 28, 2012 9:52 UTC (Sat) by Pawlerson (guest, #74136)
Posted Apr 28, 2012 15:02 UTC (Sat) by dlang (✭ supporter ✭, #313)
It was not a fun upgrade for a lot of people, especially ones who thought they had good hardware because they weren't running a AMD or Nvidia video card.
Posted May 3, 2012 14:20 UTC (Thu) by Cato (subscriber, #7643)
Ubuntu often works well on lots of hardware, but getting specific hardware to work can be really painful, as with other distros.
Posted Apr 28, 2012 16:09 UTC (Sat) by tjc (subscriber, #137)
Posted Apr 28, 2012 17:40 UTC (Sat) by dlang (✭ supporter ✭, #313)
Unfortunantly a couple of years ago (around the time of the last LTS release), the Intel drivers got really messed up for about 6 months and broke on a lot of people's systems.
Posted Apr 30, 2012 16:38 UTC (Mon) by nye (guest, #51576)
The nvidia blob has in my experience been the only solid, dependable video driver for Linux.
Okay, to be fair to the Intel driver, it is now basically functional and you can expect that your system won't crash these days, but it's hardly bug-free - for example, I've not yet met a system using Intel graphics which doesn't need the screen resolution to be set manually in xorg.conf.
If you actually look at all the cases where people are claiming problems with the nvidia blob, the vast majority don't stand up to scrutiny - typically a developer notices that the reporter is using the nvidia blob, then jumps straight to the assumption that it must be the cause of all problems because it doesn't have the magic OSS sauce.
Basically, it's just politics.
Posted Apr 30, 2012 17:44 UTC (Mon) by mathstuf (subscriber, #69389)
I've had lots of issues with the blob. Mainly when dealing with multiple monitors or WM compositing. I've not used it since around Fedora 10 or so on a day-to-day machine (I have it on a work laptop in an alternate install for OpenCL testing, but that hasn't been booted in months).
> I've not yet met a system using Intel graphics which doesn't need the screen resolution to be set manually in xorg.conf.
I have a desktop (19" 1440x900 monitor) and netbook (9" 1024x600) with Intel graphics that have never needed an xorg.conf for anything.
Posted May 1, 2012 10:57 UTC (Tue) by nye (guest, #51576)
> I have a desktop (19" 1440x900 monitor) and netbook (9" 1024x600) with Intel graphics that have never needed an xorg.conf for anything.
Interesting - does that also hold for GDM/KDM/WhateverDM, or do you not use one? Did it need any special kind of configuration?
In fact I can use KDE's system settings to configure the right resolution each time I log in, rather than writing an xorg.conf from scratch, but that doesn't help for the login manager so I wonder if there might be a way to configure that.
This was all so much easier in the days when all you had to do was uncomment the right line in the provided XF86Config.
Posted May 1, 2012 18:39 UTC (Tue) by dlang (✭ supporter ✭, #313)
Posted May 2, 2012 11:28 UTC (Wed) by nye (guest, #51576)
I can live with it since I don't re-log very often and I haven't motivated myself to write the xorg.conf yet. On the one system I have that's running Gnome though I couldn't figure out how to set the resolution (it comes with what seems to be a preconfigured list of useless options) so there I did have to resort to writing an xorg.conf.
Note that at least one machine detected the resolution correctly on 2.6.32, but no kernels I've tried before or since.
I haven't tried any kernel versions that are particularly recent since the rate of regressions in Linux has conditioned me to upgrade only when it's completely unavoidable.
(And as for monitor *hotplugging* in KDE...that's a pretty guaranteed way to render this laptop non-responsive and require a power cycle)
Posted May 2, 2012 13:51 UTC (Wed) by nye (guest, #51576)
Actually now that I think about it, I believe it was 2.6.26 which worked. Before that was probably the era of massive breakage, and from there I skipped to 2.6.29 which was apparently when KMS was introduced for Intel, which seems a likely candidate for the breakage.
Posted May 2, 2012 16:09 UTC (Wed) by mathstuf (subscriber, #69389)
Posted May 10, 2012 13:15 UTC (Thu) by nye (guest, #51576)
The two I currently have access to are an Atom integrated system with an 'N10 Family Integrated Graphics Controller', and a more standard laptop with a 'Mobile GM965/GL960 Integrated Graphics Controller'.
Those don't really mean a great deal to me, to be honest. I just know that they're both very low-performance and seemingly not brilliantly supported.
Posted May 1, 2012 19:05 UTC (Tue) by mathstuf (subscriber, #69389)
Posted May 1, 2012 19:29 UTC (Tue) by dlang (✭ supporter ✭, #313)
It doesn't require you to use KDM, it's a problem in the KDE screen management after KDE starts.
Posted May 1, 2012 19:56 UTC (Tue) by mathstuf (subscriber, #69389)
Posted May 1, 2012 16:19 UTC (Tue) by juliank (subscriber, #45896)
On all 4 laptops that I own or administrate, none needs an xorg.conf; and all of them use intel graphics (and 3 different generations of intel graphics).
On boot, the kernel detects the correct screen resolution for the panels and automatically sets the correct mode. When plugging in a screen, the screen is detected, and the correct mode is set automatically.
Posted May 2, 2012 11:29 UTC (Wed) by nye (guest, #51576)
Okay, you're making me feel motivated to take the risk of trying a new kernel version - what version are you using to get that?
Posted May 2, 2012 13:00 UTC (Wed) by juliank (subscriber, #45896)
Posted May 2, 2012 13:53 UTC (Wed) by nye (guest, #51576)
So the same versions that get it wrong on all my machines then :(. Oh well.
Posted Apr 30, 2012 14:21 UTC (Mon) by muwlgr (guest, #35359)
Unity shell is worth a try
Posted Apr 27, 2012 20:52 UTC (Fri) by Felix.Braun (subscriber, #3032)
Copyright © 2013, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds