=== Lcawte is now known as Lcawte|Away [00:42] mapp: did you do it yet? :P [06:18] HI daftykins [07:34] I GUESS thats life [07:35] get a job on purley way live in old ken street [07:51] mornin everyone [07:51] jesus christ [07:52] i love my ex girlfriend [07:52] i hope she's ok though [07:56] morning all [07:57] morning [07:57] gotta go outsidr in a min [07:57] im smoking 30+ a day again now [07:58] oh dear, expensive [07:58] well [07:58] its 2.50 a pack here [07:58] an its lofe hey [07:58] cant blame noone [07:59] im petrified of getting old and smoking and drink obv dont help [07:59] but what can i do [08:00] hmm.. pi Musicbox project = not coöperating today [08:01] its a funny thing [08:02] my dad worked at Wandsworth then moved to Coldingley [08:03] wish i could ge rid of all the drunk idiots [08:08] oh god .. they are going to launch the apple watch tonight. Can i crawl under a rock somewhere for the next 3 weeks ? [08:09] :D [08:09] yep [08:09] apple watcg [08:09] iudiots will buy it [08:10] id rather have 20 packs of cigarettes than that garbage [08:11] remember ima big time smoker [08:11] smoking mave 30 a day [08:11] i dont know [08:13] Noobs will flock me and ask "why don't i have one". [08:13] I feel another blogpost coming up [08:14] "the 80-IQ Economy" How to get dumb people to pay way too much for their gadgets [08:17] i dont do blogposts;p [08:18] i liv in gibraltar [08:18] but wannna be in SE1 [08:18] im an SE1 boy [08:21] lol .. I'll buy it :) [08:22] christ [08:22] im an alcohllic [08:22] ive got enoiugh issus [08:22] :P [08:23] even more reason to be spending your beer money on toys [08:25] nah# [08:25] * knightwise things mapp is drunk [08:25] shauno [08:26] thanks knightwise:P [08:27] i already said i have a drink problem] [08:37] shauno# [08:37] im SE1 for life [08:38] :( [08:38] never get anyway good eh [08:58] morning boys and girls. [08:58] hey brobostigon [08:59] hey knightwise [09:05] Morning all [09:06] morning davmor2 [09:08] morning [09:13] morning [09:39] Morning [10:02] morning has anyone got any experience creating customised livecds? [10:29] I have experience cos I've done it once [10:29] can't actually remember what I did though [10:29] OK : so, if you drag an email or an attachment off Thunderbird, it makes a link [10:30] yup [10:30] But Nautilus doesn't follow those links back to Thunderbird [10:30] nope [10:30] it's mental [10:30] Would be great it it worked... [10:30] and because no one is bothering to update thunderbird any more, it ain't gonna change [10:30] Well, it's not entirely Thunderbirds problem [10:31] Dr_Robotnic: https://help.ubuntu.com/community/LiveCDCustomization follow this and all your prayers shall be answered [10:31] I'd love it if it worked [10:31] we all would [10:31] but I've given up with email clients [10:31] not so much :( [10:31] gmail and owa work, so I'm not changing any more [10:31] You could just shove an imap:// link in a ticket (in your personal tracking system) and go look at it [10:31] SuperMatt, OWA, urrgh [10:32] i'm trying to install the drivers for a printer and having trouble [10:32] * awilkins has Thunderbird set up on his Office 365 email [10:32] awilkins: the new version for outlook 201 is quite nice [10:32] I keep getting asked for root password in the chrooted envrioment [10:32] Dr_Robotnic: ah, do the drives work *outside* the live cd? [10:32] What I really want is a kind of unified ticket tracker / email / productivity client [10:33] not sure, i can install the drivers on a ubuntu install [10:33] Something that does Getting Things Done (has a tickler feature) [10:33] Where everything (unarchived email threads, everything) is treated as an open task [10:34] It has a wiki for the knowledge management side, integration to the common phoneses like the Exchange plugin for Android, etc === Myrtti_ is now known as Myrtti === Lcawte|Away is now known as Lcawte === alan_g is now known as alan_g|lunch === marcushaslam_ is now known as marcushaslam [14:08] hmm i have received word that 8 x 4TB disks have shown up, now to plan how to have them all connected at once [14:08] (power wise) [14:09] time to download the internet... [14:09] nah, it'll probably be over 24 hours to initialise a RAID6 across those to start with :) [14:10] in the olden golden days my ex tried to download a webBBS to be available offline on his W98 or something. [14:10] people joined the effort, crashed the BBS and trashed the database. [14:10] it was a Very Bad Idea (tm) [14:15] hehe [14:15] i mean, oh dear === Lcawte is now known as Lcawte|Away [14:38] * Laney gets paranoid that laney@ubuntu.com has started getting double glazing spam at the same time that he actually started looking for it [14:39] lmao [14:41] yeah, I got a Mick George spam on mine. [14:55] browserstack has a new interface - funky [14:56] wow dead or alive had a number 1 30 years ago today now I feel old [14:56] davmor2, you _are_ old :-p [14:56] davmor2, we all are [14:57] diddledan: you're only old if you remember the video and not from youtube [14:57] damn it I'm old [14:57] but a year younger than popey and MooDoo so I don't care :D [14:58] thankyou google for giving me an extra 2GB :-p [14:59] I think the bigger o_O is "wow, dead or alive had a number 1" [15:00] wow, google do storage plans up to 200TB [15:00] err. 20TB [15:00] err. 30TB (get it right diddledan !) [15:01] 30TB is 299.99$ per month [15:01] 10TB is 99.99 [15:01] 1TB is 9.99 [15:02] pretty linear by the looks [15:02] diddledan: pick a number, any number, add a 0 to the end, divide it by 2, minus 50, multiple the final number by 2 and you should get to the right number ;) [15:03] teehee [15:03] is /this/ the card you were thinking of? [15:04] shauno: most of the time a woman will pick the queen of hearts and a man will pick the ace of spades interesting bit a trivia for you :) [15:11] I was thinking of my updated credit card... [15:13] Myrtti: that's a different trick when you can make a credit card disappear and reappear in the deck of cards :) [15:15] warning, graham clueless: http://www.hotforsecurity.com/blog/a-bad-week-for-uk-cybercriminals-11533.html [15:15] apparently I'm not well-known enough among law enforcement. must raise my profile [15:16] but that'd mean standing up? [15:16] good point [15:16] hmm [15:18] Hi, in ubuntu is it possible to have a laptop stay awake when the lid is closed when an external monitor is present? [15:20] * davmor2 blocks diddledan from the interwebz [15:21] :-o [15:21] but... but... [15:21] I NEED teh intertubes! [15:22] diddledan: okay have them back but at a slower speed but only cause I'm feeling generous :) [15:23] dankee === Lcawte|Away is now known as Lcawte [15:37] https://www.kickstarter.com/projects/1426448868/cuberox-six-screen-waterproof-linux-powered-comput [15:46] https://www.kickstarter.com/projects/1426448868/cuberox-six-screen-waterproof-linux-powered-comput [15:46] neat [15:47] popey, gonna buy? :-p [15:47] dunno. I'd rather make one [15:47] yeah, a pi would probably be able to power one of those [15:47] pi + box + 6 LED displays [15:48] battery pack and qi charger [15:48] or even a usb connector on the bottom [15:48] i am not convinced it needs all 6 sides with a display [15:48] I don't see the need for waterproofing either [15:56] andyc, It's possible to turn off the lid-closed-suspend thing completely [15:57] yes and no [15:57] it will still lock the screen [16:00] $299 retail? [16:01] is it apple announcement time yet? [16:01] it needs 6 faces because the only input is it's orientation [16:02] It has touchscreens too [16:02] I think they might be one pixel per face though [16:03] pi wouldn't fit inside it [16:03] also, it uses far too much power for a device like that [16:04] if for example you used a pi2 so you could run ubuntu, you'd get about 4 hours battery life if you filled the entire of the inside with lithium ion batteries [16:05] that would also explain the extremely high cost, i guess [16:07] I like their dubious claim that you can install Office on it [16:08] I'm guessing they mean "LibreOffice" [16:08] http://www.quickmeme.com/img/bf/bfc8800a2540c5274ebce706b295ed3102600ff4f50a934426edeb165c5fb49d.jpg [16:08] colleague just sent that to me [16:08] And the only way you'd actually be able to use it is over remote X11, because 6x16x16 just isn't enough for more than a couple of buttons... [16:09] diddledan, Isn't the point of Git that you CAN commit your code without pulling? [16:09] You can do a merge later? [16:09] awilkins, yeah, you'd think [16:09] This kind of just illustrates why SVN was rubbish [16:09] Or CVS [16:10] indeed, having to pull before commit just kills the whole benefit of decentralisation [16:10] that's what rebase is for [16:10] Ah, the rebase / merge dichotomy [16:10] I don't actually know what the idea of rebase is [16:10] I'm still fairly novice at git [16:11] rebase is the best feature of git [16:11] rebase lets you re-do your revisions as if they came from the top of the remote history instead of where you actually branched from [16:11] rebase -i is useful IMHO [16:11] Lets you make your own revision history less messy and more coherent [16:11] I can use it solidly, but anything more esoteric than the standard pull push commit and such is unknown to me [16:12] For (personal) example of the insane power : http://strangeowl.blogspot.com/2013/04/grafting-history-with-git.html [16:14] yeah, I'm intrigued by bzr [16:15] I love the full-on integration with launchpad and that ubuntu uses it extensively, but it seems outside the ubuntu community it's a bit of a black sheep [16:16] I liked Bazaar a lot. I used it for an internal project at a time when Git for Windows was i) not mature ii) not remotely comprehensible by normal users [16:17] git seems to be taking over all the things [16:17] Mercurial choked on their tree, SVN was too slow [16:17] even visual studio online from microsoft uses it [16:17] Jeepers [16:17] (you get the choice when you create a new project between MS' TFS or git) [16:18] VSS was horrible [16:18] I never played with that [16:18] I know that TFS is a different thing, but the taint doesn't wear off [16:19] VSS was utterly, utterly terrible. Basically the worse features of CVS combined with an obfuscated and fragile repository model that you shared publicly on your network via SMBFS [16:19] opengl-next is now known as "vulcan" [16:19] ouch [16:19] It came with tools that could analyze the repo to see if it was broken [16:20] anything on smbfs should die a slow and painful death [16:20] bzr doesn't have rebase [16:20] git pull --rebase [16:20] so good [16:20] They took an age and when they finished they usually told you, yes, it's broken, no, we can't fix it. [16:20] Best feature of VSS (and CVS, btw) : Purge. === Hornet- is now known as Hornet [16:21] Purge a file. Congrats, you have now erased it from all history. [16:21] dear god man [16:21] You can no longer build any revision of the software that contained that file. [16:21] that's counter to the whole idea of history of all the things [16:21] i think bzr's lack of bisect is worse though, makes it only suitable for developers who don't care about regressions [16:21] bzr has bisect [16:21] ali1234, no bisect? [16:21] no, it doesn't [16:21] There's a bisect plugin [16:21] wow that's a really awesome feature to be lacking [16:22] it has a plugin for it, which hasn't been maintained for 6 years and no longer works [16:22] and in fact it never worked, even when it was maintained [16:22] Hmm [16:22] git bisect is pretty sublime [16:22] it couldn't be implemented properly, because bzr has no way of telling the user what the current working dir revision is [16:22] The only thing that makes git bisect better is proper automated tests [16:22] so you could do your bisect, and then have no way of knowing which version you ended up on [16:23] Uh, no, bzr can tell you which revision you've got checked out [16:23] It will even stick a pretty flag on it in the pretty log viewer [16:23] it couldn't back when the bisect plugin still worked [16:24] another problem with bzr bisect was there was no "skip" option [16:24] * awilkins pulls his install of bzr-bisect [16:24] so if you hit a build that wouldn't compile, you had no way of continuing [16:25] and you couldn't change to another revision manually, because there's no way of knowing which revision you are currently on [16:25] ali1234, Believe you'd use bzr bisect move for that (but a bit manual) [16:25] move takes you directly to the numbered revision, but since there's no way to know what the current revision is, there's no way to know where you should go [16:26] bzr is pretty much a write-only VCS [16:26] it's great if you never want to look at the history or do anything with it and are just using VCS because your boss insists on it [16:27] TBH I think that encouraging rebase is counterproductive [16:27] it is if you measure productivity by the size of the repo [16:28] I think rebase for your personal work branch is great [16:28] huge merge commits really blow up the history to the point of being unmanagable [16:28] I think for pushed revisions it sucks a big one [16:31] Huge merge commits are the product of huge patches, which is a separate topic - huge patches are what Git is designed to avoid, because it was the natural result of using SVN and similar, because they made merging hard [16:32] huge merge commits are what happens when the thing you are merging hasn't been rebased to the latest master version [16:32] A self-reinforcing loop ; SVN made merging hard, so people tried to do it less, which made their branch delta larger, which made merging hard... [16:32] if i had a choice i'd rather remove merge commits than remove rebasing [16:32] You couldn't do hotfixes that you then merged into the mainline then [16:33] of course you could [16:33] you just couldn't do it in the most lazy and ugly way possible [16:34] Erm, hello, I'm a programmer. The laziest way is by definition the way that requires the least work. [16:34] And the best. Work only leads to mistakes because a human is doing it. [16:35] Fix a bug ; i) Bisect to find it's introduction. ii) Work out which current supported releases it affects ii) Patch it on a branch from the lowest revision you care about (or straight after the bug) iii) Merge into all releases you care about iv) Merge into main branch [16:36] Now you can test for the fixed-ness of a given branch by checking for that revision ID in it's history [16:37] If you write this down in your bug database when people call with that bug you can look it up and go "Yup, that bug is still in that version, but you could upgrade to version X which is the closest hotfix, or upgrade to newer version X+1 for only £12.99! [16:37] merely having the patch doesn't guarantee it is fixed, if the merge commit isn't right [16:37] Well, you have that same problem if you rebase and cherry pick it everywhere [16:37] you do, but you have to actually think about it [16:38] Or write a unit test as the FIRST revision on your patch branch [16:39] Missed a step ia) Write a unit test and commit that first, make sure it breaks the build correctly as expected [16:39] yes, i can't argue with that [16:39] Doing it your way doesn't create any fewer objects, just some of the commits have fewer parents. [16:40] the problematic merges though, are when feature branches against a really old development version are merged directly in to mainline [16:40] that just creates a mess [16:40] Yes, but that's not rebase / merge's fault [16:40] You can prevent that by merging outward frequently [16:40] right, it's the fault of whoever decided rebase shouldn't be allowed [16:40] Which people with SVN habits don't do [16:41] Because "Merging is Hard" [16:41] and no, merging mainline into your feature branch all the time makes an even bigger mess, because now all the commits for a feature are scattered through history [16:42] * awilkins summons Obi-Wan [16:42] "That depends greatly on your point of view" [16:42] and if you need to bisect that feature you;ve got 10x as much work [16:42] my point of view is that of somebody who didn't write the code and is trying to fix it much later [16:42] because that's mainly what i do :) [16:43] The commits for that feature are still on that branch ; but there are some outward merges into it [16:44] but that branch is on some developer's machine, and all i've got is mainly [16:44] mainline* [16:45] How do you end up debugging it if it's only on HIS machine? [16:45] Those revisions have to have ended up in the build system to make a production bug [16:45] yes, scattered through the master branch in between hundreds of completely unrelated commits [16:46] If they are there the full history is there... the branch, maybe not, but a branch isn't any kind of continuity, it's a label stuck on a revision [16:46] exactly [16:46] Why would those revisions be scattered between the revisions in the master branch? [16:46] You only merge in when it's done [16:46] bisect has a beginning and an end [16:47] Yup [16:47] if you rebase your patches, they also have a beginning and an end [16:47] if you merge your patches you create a fork in history every time you do it [16:48] that fork has to be flattened somehow [16:48] Are we saying your master is just one big line of revisions and you have no branches off it? [16:48] it effectively becomes that once you try to bisect it, yes [16:48] i'm saying that it is best if master is one big line of revisions, yes [16:49] it often isn't [16:49] bisect when you merge feature branches becomes i) a search for the revision where the feature branch merged and then ii) A search for the revision in the feature branch which introduced the bug [16:49] and with bzr it never is if you have more than one developer [16:49] I prefer a master that only consists of tested completed feature branch merge revisions [16:51] that's a mess to deal with when you don't know which feature introduced the bug [16:51] That's what bisect is for [16:52] and bisect turns history into a single line of commits [16:52] except that in such a system, most of the commits are merge commits [16:52] AFAIK bisect navigates around the graph [16:53] it descends in to branches [16:53] Yup [16:53] But it will be quicker to find which branch to descend into if it first bisects master [16:53] which means instead of n steps, your bisect is now n * number of branches steps [16:54] What, it checks every revision for every branch? Think not. [16:54] it bisects every branch [16:54] https://github.com/git/git/blob/master/Documentation/git-bisect-lk2009.txt#L407 [16:54] it has to, because you can merge master in to your branch, and then merge your branch back in to master [16:55] which means the bug can come and go from the point of view of master [16:55] often there's no way to produce a canonical linear history when you have this type of structure so it can't just "bisect master" because master has loops [16:56] If you merge master out to your branch, that revision is not part of the POV of master, unless you merge it back into master, in which case it's part of master [16:56] And no, you can't have loops in git [16:56] It's a directed acyclic graph of revisions [16:56] *acyclic* [16:56] it still has parallel branches that later merge [16:57] Yes, it does ; and they are either an ancestor of the bad revision or not [16:57] yes, and often they all are [16:57] which means each one has to be bisected indepentently [16:57] Apple watch event live stream, song currently playing is "Busy Earnin'" :) [16:58] You bisect the shortest path through the graph (master, made of all those merge revisions) and prune most of them away [16:58] You can't claim that it's n * branches because the branches are MADE of n [16:59] With a good bisection algorithm you will have fewer steps bisecting a branched graph than a linear one [16:59] that makes no sense at all [17:00] If HEAD and HEAD^^ are both bad [17:00] And HEAD is a merge revision on top of a branch with n/2 revisions in it [17:00] You can just ignore n/2 of the effort because you know that the bug revision is an ancestor of HEAD^^ [17:02] that's not what ends up happening though [17:03] what actually happens is that because your branches are regularly merged with master, you've actually propogated the bug into multiple branches [17:03] so the merge commits are all bad, and the common ancestors are all good [17:03] Your branches are NOT regularly merged with master [17:03] and then you pick one and bisect it and just hit a merge commit, then continue down the rabbit hole [17:03] Master is merged to them [17:03] exactly "master is merged to them" [17:03] bugs and all [17:04] But you can still bisect the shortest path through the graph to revision zero through master and rapidly find the branch the bug was introduced on [17:05] Possible less explicitly on git than bazaar [17:05] Or other VCSs that actually have a notion of branch being a thing [17:05] all true except for the "rapid" part [17:05] Whereas in git a branch is a label [17:05] And there is no branch in the history, only where the label is now [17:15] gah] [17:15] another one of those days [17:15] passed out [17:15] wake up cant remember what time i wS UP TILL [17:16] you were still whining about being drunk near 9am [17:16] aha [17:16] so after 9am [17:17] clown behaviour again [17:17] * mapp should know better [17:18] weren't you meant to get up early to let someone in ? [17:18] yes [17:19] and i over sleep [17:19] :D === alan_g|lunch is now known as alan_g [17:54] http://open.spotify.com/track/4fC7KhiZuHg3qZ1rj0fDWT just leave this here and move on === alan_g is now known as alan_g|EOD [18:13] MooDoo, popey: you'll like this http://open.spotify.com/album/1NILZoyUbK0pPWWWAnudJ1 [18:20] my face when 4TB disks show up as 1.6TB on the controller card :) [18:21] then my sigh of relief when a firmware update has them show up correctly \o/ [19:06] hrmm, any thoughts on 64KB stripe size versus 256KB on an 8 x 4TB RAID6 volume? [19:08] doesn't that depend on available bandwidth per disk? [19:08] What's the use case? [19:08] Lots of dinky files? [19:08] http://www.tomshardware.com/reviews/RAID-SCALING-CHARTS,1735-4.html [19:09] Exactly where my browser landed @:-) [19:10] I'd probably go for the 64kB or even 32kB [19:10] 256kB won't do much to the multi-meg media files [19:11] But it will waste a lot of room on the dinky little source code files (assuming the file system doesn't do neat tricks like smush them into the metadata blocks) [19:16] awilkins: huge media [19:18] it's a clients music and film storage alone typically [19:29] hmm, i went with XFS on the last one too - but we're double the size now. wonder if i should look into alternatives [19:41] xfs should be ok with huge volumes I think (IANAL) [19:42] btrfs is too unstable still methinks [19:42] so the alternative would be zfs via fuse [19:43] I don't know if JFS is still supported? [19:51] ZFS is the way to go [19:51] * m0nkey_ likes ZFS [19:53] ZFS wants to manage the disks directly though, no? [19:53] i'm using a proper hardware RAID controller here. [20:20] Correct, ZFS doesn't like being on a hardware controller === aquarius__ is now known as aquarius [20:47] https://twitter.com/EmrgencyKittens/status/575034682930233344/photo/1 [20:48] so yeah no ZFS for me :) [20:49] Heh, it's not for everyone. Just us FreeNAS nuts. [20:49] ^_^ [20:50] this is a dedicated PC so i'd rather use the 3ware controller yeah [20:50] so is my FreeNAS box :) [20:51] oh yeah, the funky OS thing [20:51] nevermind, i shouldn't IRC and xbox night [20:52] que? [20:55] diddledan: Any idea what this cat has been licking http://poolhouse.s3.amazonaws.com/blog-assets-two/2015/03/disgusted5.jpg [20:55] DJones, one dreads to think [20:57] ugh.. the new macbook air, it's basically an ipad with a keyboard now [20:57] it's not actually an air [20:57] it's a "macbook" [20:57] afaict that is [20:57] which means the macbook is slimmer than the macbook air [20:58] I think it'll make sense eventually [20:59] Which doesn't make sense. Might as well drop the MacBook Air line [20:59] I quite like the look of those 5k imac hair removers [20:59] step 1, kill the non-retina pro (which feels like the not-too-distant future) [20:59] then yeah, step 2, once they have cost under control on this one, kill the air. go back to just having mb / mb pro [21:00] having such a cluttered 'offer' really isn't their style. it really feels like this is just a transition until they can make retina cost-effective enough to warrant killing the non-retina models [21:01] they've got a proliferation of imac too [21:02] 21inch 27inch and 27inch 5k [21:02] that's almost exactly the same issue though? they can't kill the 27" until they can get the cost down on the 5k [21:02] aye [21:03] i did see a cat pull disgusted face after she stuck her face in a bag she thought was groceries but was in fact used to contain used nappies [21:03] I think they need to cull one of the ipads too, and just go back to having "the big one" and "the little one" in each category [21:03] * m0nkey_ wont be buying another Mac [21:05] * m0nkey_ will buy a motherboard (http://www.newegg.ca/Product/Product.aspx?Item=N82E16813157504) next week [21:08] heh, newegg are one of the few thing I miss about the US [21:09] newegg, me ma, and not needing a drivers licence because there's only two cops in town - one's on a bicycle, and the other was on my bowling team [21:09] lol [21:13] ooh, just found a bar o chocolate I didn't know I had [21:13] Anyone know how to get Ubuntu or even Fedora to boot solo on the latest MBA firmware? I don't get bootwait anymore, just a flashing question mark. [21:23] m0nkey_: bios mode or efi? [21:23] both similar but vary slightly [21:23]