[19:49] <ibeardslee> morning
[19:55] <ojwb> morning
[20:36] <chilts> morning
[20:37] <ajmitch> morning
[20:52] <ibeardslee> such a pity that Precise still has so many errors and crashes
[20:54] <ajmitch> part of the problem is that some of these crashes weren't terribly visible in previous versions, but the new error reporting tool pops up when something silently dies
[20:55] <ajmitch> so it's a mix of perception & of things just being buggy
[20:55] <ibeardslee> perception of broken things shouldn't exist in an LTS release .. esp at the x.xx.1 release
[20:56] <lifeless> well
[20:56] <lifeless> its not that simple really.
[20:56] <lifeless> Consider - a background thread in a GUI crashes.
[20:56] <ibeardslee> the non-LTS versions should make that more visible so they get elimintated
[20:56] <lifeless> Some folk don't notice at all.
[20:56] <hads> I imagine you need the reports to find out what's broken.
[20:56] <lifeless> Some folk notice, because e.g. their ipod doesn't sync properly.
[20:56] <lifeless> We can't just gather the stats without peoples consent.
[20:57] <lifeless> And we don't have good enough heuristics that a global consent from a user will be good enough to avoid all chance of snarfing up passwords etc.
[20:57] <lifeless> So we have a choice, of not knowing that there is a problem, or of our users also knowing that there is a problem.
[20:58] <ibeardslee> sure .. but for most people .. Precise crashes more than previous versions
[20:58] <lifeless> *Because* its an LTS, its important we know whether its still got a high failure rate or not, otherwise we can't tell if allocating engineering effort to fixing it makes sense.
[20:58] <ibeardslee> sorry .. but for most people .. "Precise crashes more than previous versions"
[20:58] <lifeless> ibeardslee: See, *maybe* it does, *maybe it doesn't*. What we know about previous versions is worse than anecdotal.
[20:59] <lifeless> ibeardslee: what is clear is that we can now go out and fix precise.
[20:59] <lifeless> (and we are)
[20:59] <ajmitch> there was a spirited discussion about whether to turn off whoopsie in 12.04.1. The data just isn't there for previous releases of ubuntu
[20:59] <hads> From my perspective, until I saw the explanation that crash reporting was turned on, I thought that precise was just crashy.
[20:59] <lifeless> ajmitch: btw speaking of error reporting.
[21:00] <lifeless> ajmitch: did you see the php oops client link ?
[21:00] <ajmitch> lifeless: yeah I did thanks
[21:00] <ibeardslee> I agree, but I'm here defending precise against people telling me it crashes more than previous versions
[21:00] <lifeless> ibeardslee: I certainly understand *that*.
[21:00] <lifeless> ibeardslee: Windows had exactly the same psychological issue when it introduced error reporting
[21:01] <ajmitch> ibeardslee: previous releases had apport turned off at release, so this is where perception of crashiness comes in, but the data just wasn't there to compare against
[21:01] <ajmitch> there
[21:01] <ibeardslee> and because of that, I have others telling me they don't want to upgrade to Precise because it keeps crashing for others
[21:01] <ajmitch> there have been developers allocated for stable release maintenance & this is what they use to schedule what to work on
[21:02] <lifeless> ibeardslee: so, thats fine - the primary goal isn't to get everyone on precise, its to finally fix this horribly crashing platform we have.
[21:02] <lifeless> Precise is probably the least crashy we've been.
[21:03] <hads> Cool
[21:03] <lifeless> ibeardslee: I don't /care/ if O users avoid precise like the plague.
[21:04] <lifeless> ibeardslee: if we don't fix the root cause, the *justified* reputation of Linux as a crashy horrible desktop OS will remain.
[21:08] <ibeardslee> Agreed .. but unless people are aware that this LTS is all about fixing existing underlying issues, Precise *looks like* the buggiest release yet.
[21:09] <lifeless> yeah
[21:09] <lifeless> its a rub, thats for sure
[21:09] <mwhudson> i'm not sure what the alternative course of action is
[21:09] <mwhudson> i guess whoopsie could have stayed off until quantal
[21:10] <mwhudson> but then that's another 6 months of waiting?
[21:10] <ajmitch> but then you don't get info on what needed fixing for 12.04.1
[21:10] <ibeardslee> No no, leave it on quantal .. it's not the LTS release.
[21:10] <ibeardslee> err /me reads and comprehends better
[21:11] <ojwb> you should have sent the code back in time and turned it on for maverick
[21:11] <ajmitch> I think that errors.ubuntu.com has been a very useful tool
[21:13] <lifeless> We have some assumptions.
[21:13] <lifeless> We assume that LTS users are statistically different to current-release users are different to in-development users.
[21:14] <lifeless> We assume that we get different issues from these groups; likely with significant overlap, but specifically not subsets.
[21:15] <ojwb> that's likely true, though if you'd started in the post-LTS release you'd have shaken out at least some of the issues LTS users hit too
[21:15] <lifeless> indeed!
[21:16] <lifeless> There are things being done to reduce the impact
[21:16] <lifeless> IIRC:
[21:16] <lifeless>  - the UI is being polished
[21:16] <ojwb> my hindsight is 20-20
[21:16] <lifeless>  - a cap on reports per user is being put in place to avoid dossing folk
[21:17] <lifeless>  - there are plans (likely to not be ported to precise) for wider opt-in that might permit no-interaction crash report handling, specficially for known crashes.
[21:17] <ojwb> having used linux as a desktop OS heavily for close to 20 years now, I can't say I agree that it's a crashy horrible desktop OS
[21:18] <ibeardslee> I agree
[21:18] <ojwb> sound has been a rough edge for a long time, on and off
[21:18] <lifeless> I use it, and love it, as my desktop. However, compared to Mac OS X and Windows 7, we suffer more failed processes and drivers, per user per day.
[21:19]  * ajmitch won't mention those bloody nvidia drivers
[21:19] <lifeless> Personal anecdata - wifi dies for me every day or two, sometimes multiple times per day; graphic lockups once a week or so if I use unity.
[21:19] <ibeardslee> until precise, I have to admit that I hardly ever noticed things going wrong.
[21:19] <lifeless> Thats intel video, not nvidia.
[21:19] <ojwb> well, unity was very crashy, at least in the initial release
[21:19] <lifeless> suspend resume are still fragile
[21:19] <ibeardslee> or at least if I did, it was an insignificant event
[21:19] <ojwb> but I think that's not "linux desktop" really
[21:19] <lifeless> firefox and chromium processes dying are routine
[21:20] <lifeless> libreoffice going up its own backend is common
[21:20] <mwhudson> iwlwifi causes the majority of my pain in precise i think
[21:20] <lifeless> so common, they put specific crash handling code for recovery of documents...
[21:20] <mwhudson> firefox was very very crashy for a while, but that turned out to be firebug
[21:20] <ajmitch> mwhudson: I can believe that, I still have to disable 802.11n to get usable wifi
[21:20] <lifeless> gvfs gets stuck on windows interop so often its not funny
[21:21] <ojwb> the driver for the wifi chipset in this laptop is very poor, but it's not in the kernel which is probably an indicator of the quality of it
[21:21] <ojwb> doesn't ms office do document recovery too?
[21:22] <lifeless> point is, put a linux desktop beside a windows 7 or mac for a month, do the same things on them both, and its extremely unlikely that the user observable crash rate on the linux machine will be higher than that of either other OS
[21:22] <lifeless> Its important to celebrate how far we have come, but its also important not to be blind to the issues we do have.
[21:22] <mwhudson> lifeless: i think you mean 'lower' there?
[21:23] <lifeless> mwhudson: BWAH. yes. Thanks/
[21:59] <ajmitch> talking of crash reports, there goes one now :)
[22:04] <ibeardslee> heh
[22:05] <ibeardslee> tried to ugprade my home PC to Quantal on the weekend
[22:05] <ibeardslee> won't do it, despite the video card supporting earlier 3D efforts it isn't supported by the lastest version of unity
[22:05]  * ajmitch did a fresh quantal install on the new home pc without too many issues
[22:06] <ajmitch> that's interesting, what card?
[22:06] <ibeardslee> an onboard nvidia geforce 6somethingmumble
[22:06] <ajmitch> 6000 series?
[22:07] <ajmitch> nvidia has some weird numbering
[22:07] <ibeardslee> GeForce 6150
[22:07] <ajmitch> that should still be supported by the latest nvidia drivers
[22:07] <ajmitch> after that it'll only be supported on their legacy branch
[22:08] <ibeardslee> won't do the 3D for Precise either
[22:09] <ajmitch> odd
[22:10] <ibeardslee> http://www.asus.co.nz/Motherboards/AMD_AM2/M2NPVVM/#specifications
[22:10] <ibeardslee> that's the motherboard (and video)
[22:11] <ajmitch> could be that unity is expecting certain functionality that the card won't support
[22:12] <ajmitch> run /usr/lib/nux/unity_support_test -p
[22:14]  * ibeardslee makes a note for this evening
[22:15] <ajmitch> you can't ssh into your home machine? :)
[22:15] <ibeardslee> home PC is off .. this irc session is on my home server
[22:15] <ajmitch> ah right
[22:16] <ibeardslee> was getting an old shuttle up and running yesterday as well .. ahh edgy, the memories
[22:16] <hads> wakeonlan $DESKTOP :)
[22:16] <ibeardslee> that has been configured, but the network is still plugged into the shuttle
[22:18] <ibeardslee> been considering the power useage of that old PC, compared to my existing home server .. that could be a decent upgrade path
[22:26] <Atamira> morning
[22:27] <Atamira> waiting for the rain to hit
[22:27] <ibeardslee> it certainly hit here over the weekend
[22:28] <Atamira> it hits and stops and hits again here
[22:28] <Atamira> tho im sure we dont have all of it
[22:28] <Atamira> i know one of the posthops in hamilton had their roof cave in due to hail
[22:32]  * ojwb wonders if that's a typo for postops or postshops
[22:33] <ibeardslee> you mean you haven't seen the post hop?
[22:34] <ojwb> no, the bottle shop had sold out of it
[22:34] <ibeardslee> it's a variation of the lindy hop
[22:40] <lifeless> ibeardslee: not going to go the ARM route?
[22:41] <ibeardslee> lifeless: for the home server?  got a vagueplan to be able to run a couple of vms for things like calibre, koha etc
[22:41] <lifeless> ibeardslee: yeah, there is kvm for ARM nowadays
[22:49] <Atamira> ;p;
[22:49] <Atamira> oops
[22:49] <Atamira> im tired..
[22:49] <Atamira> post shop
[22:52] <ibeardslee> and now based on previous discussions starts letting all crashes get reported
[23:03] <ibeardslee> lifeless: that does change things a wee bit then
[23:03] <lifeless> ibeardslee: google has stuff on it
[23:06] <ibeardslee> although the advantage with the x64 server is just the one arch being cached by apt-cacher-ng
[23:07] <lifeless> have you seen the baserock ?
[23:07] <ibeardslee> no I haven't
[23:08] <lifeless> http://www.baserock.com/
[23:08] <lifeless> 8 x quad-core
[23:08] <ibeardslee> sounds overkill for a home server?
[23:09] <lifeless> probably uses a total of 50W at peak.
[23:09] <lifeless> ibeardslee: no such thing as overkill :)
[23:09] <ibeardslee> there is so when $$ are in short supply
[23:10] <ibeardslee> how could I get 4 x 3TB disks in those?
[23:11] <lifeless> you'd need a separate container for them, but each board has SATA on it
[23:11] <lifeless> anyhow, its just one of the servers out there doing this
[23:12] <ibeardslee> the HP 40L was something I'd considered
[23:12] <lifeless> theres a bunch of Ubuntu folk loving on the HP microservers
[23:13] <ibeardslee> but there is much to consider in terms of what I run at home and how live it is 24x7
[23:13] <lifeless> ... which the 40L appears to be
[23:13] <ibeardslee> yes
[23:13] <lifeless> the HP website has terrible google juice.
[23:13] <lifeless> first page of google for hp 40L has no links to the HP site :(
[23:14] <lifeless> but apparently the HP40L is a Ryobi 4 V Lithium-Ion Screwdriver
[23:15] <ibeardslee> well there you go
[23:15] <ibeardslee> try the HP N40L
[23:16] <lifeless> yeah
[23:16] <lifeless> just found it on their site
[23:16] <lifeless> thanks!