| odinho | Hey. So, I installed 12.10 on my laptop, and it was not a very nice experience. I had made ~32gb of space ready for it. And in the partition-place, I pressed "use the existing". At that point it instantly started doing stuff without telling me what it was going to do. | 15:32 |
|---|---|---|
| odinho | And it ended up with making a 13GB / partition, and 17GB swap. | 15:33 |
| odinho | That's... just evil :-) | 15:33 |
| xnox | do you happen to have 16GB of RAM? | 15:38 |
| stgraber | I had the same just happen to me actually but I expected it when I told the installer to just wipe the disk clean ;) | 15:38 |
| stgraber | (11GB / and 17GB swap in my case) | 15:38 |
| odinho | xnox: Yeah. :-) It might be a bit smarter about those "sane defaults". | 15:38 |
| xnox | odinho: for bonus points with a 19GB partition it whould have failed to install =) | 15:39 |
| xnox | (and 16GB swap) | 15:39 |
| odinho | xnox: ... Yeah. That'd be fun :D :P | 15:39 |
| odinho | xnox: I actually thought about making it 20GB. Because my SSD is only 256 GB, and I'm supposed to use OS X on the machine. | 15:40 |
| odinho | I'm only sneaking in Ubuntu to use that when I'm home. | 15:40 |
| antarus | xnox: wait, you are still doing swap = ram ? :) | 15:40 |
| antarus | (I can't really poke fun, we had swap = 3x ram for a while...) | 15:41 |
| antarus | which means on our HP Z620, you had 192G of swap :X | 15:41 |
| antarus | engineers were *pissed* | 15:41 |
| cjwatson | We should probably (a) act differently on laptops (b) on giant RAM use rather less swap but warn that hibernation will be impossible | 15:42 |
| odinho | Swap should really be capped at 4 GB anyway. Or something like that. -- Or does anyone actually want swap? | 15:42 |
| cjwatson | Just rather hard to integrate all the pieces there | 15:42 |
| stgraber | antarus: yep, just had that on a new VM machine I asked someone to setup ;) they installed stock Ubuntu using the "wipe the disk and don't ask me anything" option and they got a 192GB swap partition ;) what a waste of a nice 15K RPM SAS drive | 15:42 |
| cjwatson | You really don't want to run a Linux system without swap | 15:42 |
| antarus | cjwatson: I thought suspend to disk didn't work anyway | 15:42 |
| cjwatson | Lies | 15:42 |
| cjwatson | People keep trying to remove the UI for it and such | 15:42 |
| antarus | I'll give you a hint, it doesn't work ;p | 15:42 |
| cjwatson | But lots of people still use it successfully AIUI | 15:42 |
| cjwatson | ... for you | 15:43 |
| antarus | (reliably) | 15:43 |
| antarus | ;) | 15:43 |
| cjwatson | It's only very recently that Linux has been able to cope with defragmenting memory sanely when entirely swapless | 15:43 |
| cjwatson | And I still would definitely not trust it in such a moded | 15:43 |
| antarus | yeah there are some people at work that claim swapless machines are superior | 15:43 |
| stgraber | cjwatson: I just use zram-config ;) | 15:43 |
| antarus | I mostly call those people stupid | 15:43 |
| odinho | There's lots of memory-pressure things in the kernel now, -- because of virtual machines. | 15:44 |
| stgraber | cjwatson: that way both the kernel and my SSD are happy | 15:44 |
| cjwatson | Swapless means (a) absolutely no tolerance of overcommit (b) you have to have absolutely everything resident even if it's a process that hasn't been scheduled in months | 15:44 |
| odinho | So maybe it'd be possible to pressure mem usage down to swap-level if you still wanted to hibernate. | 15:45 |
| cjwatson | When somebody does it, let us know :) | 15:45 |
| odinho | cjwatson: When I have done it manually, I have done about 1 or 2G of swap. Anything over that seems rather ... too much :-) | 15:45 |
| odinho | At least allocating more ram than disk space should've triggered an assert somewhere :D | 15:46 |
| cjwatson | On a laptop I'd still go with >=ram, but on other systems it makes less sense | 15:46 |
| odinho | Heuristic could still be, yeah I'm on a laptop, but having ram=swap would mean I'm using over 10% of available disk space, woops, that's no good --> revert to using 2G. :] | 15:47 |
| mpt | xnox, for bug 1047384, I think we should go with your option 2. Delaying the passphrase question for more than one step (the keyboard layout step) would seem weird, almost like we'd forgotten about it. | 16:21 |
| ubot2 | Launchpad bug 1047384 in ubiquity (Ubuntu) "System Encryption Password set before setting keyboard locale" [Medium,In progress] https://launchpad.net/bugs/1047384 | 16:21 |
| ogra_ | use zram :) | 16:21 |
| ogra_ | makes the kernel happy and keeps your disk free | 16:21 |
| ogra_ | (wrt swap) | 16:22 |
| xnox | mpt: hmmm.... but then we need to bring geolocation forward (for a better guess at keyboard layout) | 17:57 |
| xnox | mpt: and then we delay start of installation by a long time, but hey the user wanted encryption =))))) | 17:58 |
| mpt | xnox, ugh. The reason to bring keyboard layout forward is to have only one step between choosing encryption and entering the encryption key. If we brought geolocation forward too, we'd then have two steps between them, defeating the purpose. (-: | 17:59 |
| xnox | mpt: with option 1, it's two steps with speedy install. with option 2, it's two steps away with delayed install. =)))) both are bad with one slightly worse. | 18:00 |
| mpt | bother | 18:01 |
| xnox | mpt: do you agree with bug 1118589 | 18:04 |
| ubot2 | Launchpad bug 1118589 in ubiquity (Ubuntu Raring) "Please remove WebCam step" [Medium,Triaged] https://launchpad.net/bugs/1118589 | 18:04 |
| mpt | xnox, no | 18:15 |
| mpt | of course not :-) | 18:16 |
| xnox | mpt: great, shared a google doc with you. | 18:17 |
| xnox | mpt: i'll make a merge proposal and my bit is done. | 18:17 |
| Laney | xnox: oh, so my patch was wasted :( | 18:18 |
| xnox | Laney: i found out yesterday. Shared a google doc with you. | 18:19 |
| xnox | Laney: personally, I don't typically look my best when the webcam thing pops up. | 18:19 |
| Laney | I'VE BECOME SO NUMB | 18:19 |
| xnox | and we only show the avatar in the fast user switcher. | 18:20 |
| xnox | and account settings. | 18:20 |
| xnox | we don't show it in lightdm. | 18:20 |
| xnox | Laney: your patch was accepted and it is in Raring currently =) \o/ | 18:21 |
| mpt | xnox, and PolicyKit in future. | 18:43 |
| xnox | hehe =)) | 18:44 |
| mpt | xnox, I'll discuss it with Ivanka on Monday. Okay to hold until then? | 18:44 |
| mpt | Actually, Tuesday. | 18:45 |
| mpt | hm hm | 18:45 |
| xnox | mpt: sure. I can split it into separate package such that adding/removing that step is a matter of changing what packages are seeded on to the CD. | 18:45 |
| mpt | I'll try to get this sorted tomorrow, actually | 18:45 |
| xnox | (cuase e.g. it makes little sense in Xubuntu/Mythbuntu/Studio as they have avatarless desktop) | 18:46 |
| mpt | oh, ok | 18:46 |
| mpt | I was about to say, splitting into separate packages sounds like work for the sake of delaying decisions :-) | 18:46 |
| mpt | but fair enough if it helps derivatives | 18:46 |
| stgraber | mpt: *flavours ;) | 18:54 |
| stgraber | we don't care about derivatives, but we do about flavours | 18:55 |
| mpt | right | 18:55 |
| mpt | I'd forgotten about flavours | 18:55 |
| * mpt ducks as stgraber throws something | 18:55 | |
| stgraber | ;) | 18:56 |
| bdmurray | I'm gonna set bug 1095684 to Low or Medium since making it happen (double-click) seems convoluted to me | 18:59 |
| ubot2 | Launchpad bug 1095684 in ubiquity (Ubuntu Raring) "Ubiquity crashes upon clicking 'New partition table' twice in the manual partitioning step during raring installation" [High,New] https://launchpad.net/bugs/1095684 | 18:59 |
| bdmurray | Seem reasonable? | 19:00 |
| cjwatson | lots of people are in the habit of double-clicking everything | 19:01 |
| bdmurray | okay | 19:03 |
| infinity | bdmurray: My parents double-click constantly. They get multiple web browsers for every link they click, etc. Totally not convoluted. | 19:55 |
| infinity | bdmurray: And also really hard to untrain. | 19:55 |
| luv | hi there, I'm sure you have got this question loads of times, anyway, would you guys be happy to accept a patch which would allow to enable/disable amazon (and possible other) lenses on install time? I am happy to do the actual code and this issue is quite important to me because i am not really happy that search i do after pressing <win_key> are send to the internet. | 22:09 |
| luv | im sure loads of people like the amazon lens but this is something that should be enabled/disabled on install time. again, im happy to do the coding. | 22:10 |
| === kentb is now known as kentb-out | ||
| cjwatson | luv: I'd take a patch that allowed preseeding the "Include online search results" setting in System Settings → Privacy, if it's possible to set that system-wide. Less sure about UI; that would have to go to the design team - but you'd need to do the first part of the work regardless | 23:54 |
| cjwatson | luv: Though, TBH, such a preseedable setting ought to be in whatever package implements the general "Include online search results" setting (i.e. more general than just the Amazon results in the shopping lens), rather than actually involving an installer patch as such | 23:55 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!