| --- Log opened Wed May 15 00:00:59 2013 |
| 00:01 | <@Reiv> | Next real question: I want to sum the absolute value of each cell in a column. |
| 00:01 | <@Alek> | yeah |
| 00:02 | <@Reiv> | So SUM(foo), for a foo of {2,-3,-1} is 6. |
| 00:02 | <@Reiv> | SUM(ABS(foo1:foo9)) gave me an error. |
| 00:16 | | * Reiv feasts apon Vornicus and his tasty brainmeats |
| 00:21 | | * McMartin nails a rather nasty bug they'd been dealing with, that turned out to be becuase of a slight difference in the behavior of the memory managers between VMware Players 4 and 5 |
| 00:21 | | * McMartin tees up the Frozen Synapse victory soundtracks |
| 00:21 | < Turaiel> | How'd you manage to debug that? |
| 00:22 | <&McMartin> | syscall tracing of them followed by auditing the code of ours that reacts to it and finding a bug in our teardown code that was inhibiting reinitialization in some circumstances |
| 00:23 | <&McMartin> | Ultimately a three-line fix, but the older versions didn't tickle the bug so it went unnoticed. |
| 00:23 | | You're now known as TheWatcher[T-2] |
| 00:29 | <@Azash> | https://www.rfc-editor.org/rfc/rfc6946.txt |
| 00:36 | | himi [fow035@D741F1.243F35.CADC30.81D435] has joined #code |
| 00:36 | | mode/#code [+o himi] by ChanServ |
| 00:38 | | You're now known as TheWatcher[zZzZ] |
| 01:03 | | gnolam [lenin@B19C04.C4B928.3477B4.1341ED] has quit [Client closed the connection] |
| 01:03 | | gnolam [lenin@B19C04.C4B928.3477B4.1341ED] has joined #code |
| 01:03 | | mode/#code [+o gnolam] by ChanServ |
| 01:15 | <@Reiv> | Nailed it. =SUMIF(Q2:Q34,">0")+ABS(SUMIF(Q2:Q34,"<0")) |
| 01:16 | <@Reiv> | I wonder if there isn't a cleaner way, but hell, it works |
| 01:17 | | VirusJTG_ [VirusJTG@Nightstar-09c31e7a.sta.comporium.net] has joined #code |
| 01:18 | | Turaiel is now known as Turaiel[Offline] |
| 01:21 | | VirusJTG [VirusJTG@Nightstar-09c31e7a.sta.comporium.net] has quit [Ping timeout: 121 seconds] |
| 01:21 | <@Azash> | 01:02 <@Reiv> So SUM(foo), for a foo of {2,-3,-1} is 6. < Wow |
| 01:25 | | ToxicFrog [ToxicFrog@ServerAdministrator.Nightstar.Net] has quit [[NS] Quit: ZNC - http://znc.in] |
| 01:26 | | ToxicFrog [ben@Nightstar-f9ce94e5.dsl.teksavvy.com] has joined #code |
| 01:26 | | mode/#code [+ao ToxicFrog ToxicFrog] by ChanServ |
| 01:46 | <@Reiv> | Azash: Nothing 'Wow' about it |
| 01:47 | <@Azash> | How come? |
| 01:47 | <@Reiv> | That was my desired outcome. The above code does it, ableit not entirely elegantly. |
| 01:47 | <@Azash> | Oh |
| 01:47 | <@Azash> | Sorry I misread |
| 01:47 | <@Azash> | I thought you meant that was the default behaviou |
| 01:47 | <@Azash> | r |
| 01:47 | <@Reiv> | Hell no. |
| 01:48 | <@Reiv> | Basically, I'm calculating both Gross and Turnover values for money. |
| 01:48 | <@Azash> | Right |
| 01:56 | <&McMartin> | ... this has to be a troll |
| 01:56 | <&McMartin> | https://www.elasticcobol.com/ |
| 01:58 | <@Reiv> | Wherein if you find $30k in gains and $20k in losses for a month, you may only have a Gross of $10k, but your /turnover/ was $50k and it's kinda importaint to note that to the boss when they ask how much you've done this month~ |
| 01:59 | <&McMartin> | Bzuh |
| 01:59 | <&McMartin> | Would you not have had 30k in gross gains there, with 10k net? |
| 01:59 | <&McMartin> | I get the impression I am missing something here |
| 02:01 | <@Reiv> | Possibly that I am using the wrong terms! |
| 02:01 | <@Reiv> | Net, not Gross, yes. |
| 02:01 | <@Reiv> | Net 10k, Turnover of 50k. |
| 02:02 | <@Reiv> | And the day the Net goes negative, it's not actually our departments fault, it means that the fuckups are starting to go against us and we really really want to find out why >_> |
| 02:09 | | Reiv [NSwebIRC@Nightstar-95746c1f.kinect.net.nz] has quit [Ping timeout: 121 seconds] |
| 02:11 | <~Vornicus> | Reiv: oh that's a nasty one |
| 02:12 | <~Vornicus> | Yours there is good but you could also get rid of the ABS: SUMIF(yadda,">0") - SUMIF(yadda, "<0") |
| 02:13 | | Reiv [NSwebIRC@Nightstar-95746c1f.kinect.net.nz] has joined #code |
| 02:13 | | mode/#code [+o Reiv] by ChanServ |
| 02:14 | <@Reiv> | McMartin: If it's a troll, it's a masterful one that's trying very hard to keep a straight face. |
| 02:16 | <~Vornicus> | (good lord, I've created a monster. |
| 02:17 | <~Vornicus> | (the 'partial' idea for excel is growing legs in my head) |
| 02:21 | <@Reiv> | partial? |
| 02:22 | <~Vornicus> | Like, I was thinking about countif and the difficulties we often have with it, namely that you only really get the right half of the inequality statement to configure. |
| 02:24 | <~Vornicus> | THere's no way to say "between 20 and 30", like you had above. |
| 02:26 | <@Reiv> | Interesting. |
| 02:26 | <@Reiv> | So my bodge job was the best way afterall! |
| 02:26 | <~Vornicus> | So I was thinking: what if it were possible to create a lambda-oid or partial or something that you can plug into your formula. Like, it'd look like =COUNTIF(foo, AND(20<_, _<30)) or something. |
| 02:27 | | Turaiel[Offline] is now known as Turaiel |
| 02:27 | <@Reiv> | Likewise, no way to SUM a column range while demanding each value be ABS? |
| 02:27 | <~Vornicus> | Similarly, but I think that'd require more support |
| 02:50 | <&ToxicFrog> | I think once you start having conversations like this it's a sign that you shouldn't be using Excel. |
| 02:56 | | Derakon is now known as Derakon[AFK] |
| 03:30 | | VirusJTG_ [VirusJTG@Nightstar-09c31e7a.sta.comporium.net] has quit [[NS] Quit: Program Shutting down] |
| 03:34 | <~Vornicus> | Yes, well |
| 03:38 | <@Reiv> | ToxicFrog: And yet the data I am given is in Excel, and the data I am returning to them is to be in Excel. |
| 03:38 | <@Reiv> | For bonus points, they'd quite like to be able to propagate my calculations forward, whilst using their Excel spreadsheet. |
| 03:39 | <@Reiv> | So yes, I could use something other than Excel. |
| 03:39 | <@Reiv> | So long as I still use Excel at every other point in the venture, not incidentally meaning that I'm the only one that can do the math in future too. ?? |
| 03:40 | <@Reiv> | So I'll just hammer together the slightly hacky trickery and call it done, because then as long as people can hit ctrl+D they can use the hacky bits too. :P |
| 03:49 | <~Vornicus> | Thing is. Excel could be truly great. As it is it's my go-to device for equation design. But it could be, with one or two features, a lot more powerful. |
| 03:54 | <@Reiv> | Not even especially onerous features! |
| 03:57 | | Kindamoody[zZz] is now known as Kindamoody |
| 03:57 | <~Vornicus> | They might be hard to program. |
| 04:00 | <@Reiv> | That's why we pay 'em the big bucks, right? |
| 04:00 | <@Reiv> | 'sides, all I really want is nesting. :p |
| 04:05 | | thalass [thalass@Nightstar-879f199f.bigpond.net.au] has joined #code |
| 04:06 | <~Vornicus> | well, no, what you really want is map and reduce. |
| 04:09 | < Syk> | so a cartographer on a diet? |
| 04:09 | < Syk> | :P |
| 04:17 | | * Vornicus ..patpats syk |
| 04:17 | < Syk> | :( okay so I know my pun wasn't that good |
| 04:25 | <&ToxicFrog> | zpool online \o/ |
| 04:27 | <~Vornicus> | what about ypool and xpool? |
| 04:34 | <&ToxicFrog> | obsolete |
| 04:34 | <@Reiv> | zpool? |
| 04:34 | <@Azash> | I assume zpools are related to zfs and are not just continental European terms for email systems |
| 04:51 | <&ToxicFrog> | Correct |
| 04:51 | <&ToxicFrog> | A zpool is a ZFS storage pool from which filesystems can be allocated |
| 04:51 | <~Vornicus> | zomg |
| 04:53 | | * ToxicFrog has just created an 8TB RAIDZ zpool on orias and is now mirroring /home over to it with triple replication |
| 05:07 | | JustBob [justbob@ServerAdministrator.Nightstar.Net] has quit [[NS] Quit: ] |
| 05:10 | | JustBob [justbob@ServerAdministrator.Nightstar.Net] has joined #code |
| 05:35 | | ToxicFrog [ben@Nightstar-f9ce94e5.dsl.teksavvy.com] has quit [[NS] Quit: Ex-Chat] |
| 05:35 | | ToxicFrog_ [ToxicFrog@ServerAdministrator.Nightstar.Net] has joined #code |
| 05:36 | | ToxicFrog_ is now known as ToxicFrog |
| 05:38 | | mode/#code [+ao ToxicFrog ToxicFrog] by ChanServ |
| 05:52 | | ErikMesoy|sleep is now known as ErikMesoy |
| 06:14 | | * Vornicus tries to remember how to generate e.g. [3,0,0], [2,1,0], [2,0,1], [1,2,0], [1,1,1], [1,0,2], [0,3,0], [0,2,1], [0,1,2], [0,0,3]. |
| 06:16 | | * Vornicus breaks out the knuth! yay! |
| 06:17 | <~Vornicus> | (because this one's on the twelvefold way, I know it) |
| 06:19 | < Syk> | I saw a hardbound box set of Knuth, while I was in perth |
| 06:19 | < Syk> | was going to get it, but... $180 |
| 06:20 | <~Vornicus> | Worth every penny. |
| 06:20 | < Syk> | uh |
| 06:20 | < Syk> | $280, rather |
| 06:20 | < Syk> | reduced from like $350 |
| 06:20 | <~Vornicus> | my statement stands |
| 06:20 | < Syk> | is it really? |
| 06:20 | | Xires [xires@Nightstar-c54fd8cd.feedthetrolls.net] has quit [Ping timeout: 121 seconds] |
| 06:20 | < Syk> | I know that they're good books and all |
| 06:21 | | Derakon[AFK] is now known as Derakon |
| 06:21 | | Xires [xires@C408D4.D7A475.EBB1B0.4FF4B9] has joined #code |
| 06:26 | < Syk> | I just hope Knuth gets to finish them all |
| 06:26 | <~Vornicus> | the number of times I've blown problems out of the water by merely opening knuth is larger than I can count now. |
| 06:28 | < Syk> | hmm |
| 06:29 | < Syk> | I guess I should read it someday |
| 06:29 | < Syk> | I'm assuming that it's better in hardback |
| 06:29 | <~Vornicus> | I'm not sure "read it" really applies here. |
| 06:30 | < Syk> | so it's more a reference? |
| 06:31 | <~Vornicus> | In a lot of ways, yes. |
| 06:40 | | Derakon is now known as Derakon[AFK] |
| 06:40 | | Turaiel is now known as Turaiel[Offline] |
| 06:41 | | himi [fow035@D741F1.243F35.CADC30.81D435] has quit [Ping timeout: 121 seconds] |
| 07:00 | | celticminstrel [celticminst@Nightstar-e83b3651.cable.rogers.com] has quit [[NS] Quit: KABOOM! It seems that I have exploded. Please wait while I reinstall the universe.] |
| 07:30 | | Kindamoody is now known as Kindamoody|out |
| 07:33 | | You're now known as TheWatcher |
| 07:38 | <@froztbyte> | I still don't have any Knuth books here |
| 07:38 | <@froztbyte> | really do need to acquire some |
| 07:50 | <@Azash> | I have them in my bookshelf |
| 07:50 | <@Azash> | On D:\ |
| 07:52 | | * Vornicus determines he needs a more, um, sensible solution to the problem of having the program documenting the decisions it makes. |
| 07:57 | | Reiv [NSwebIRC@Nightstar-95746c1f.kinect.net.nz] has quit [Ping timeout: 121 seconds] |
| 08:04 | <@froztbyte> | Azash: ew windows |
| 08:08 | <@Azash> | froztbyte: I prefer not to fiddle around for everyday use~ |
| 08:08 | <@Azash> | If I do work I just fire up the VMs |
| 08:09 | | * froztbyte does not understand what is being referred to |
| 08:10 | <@Azash> | I commented on your distaste saying I prefer a casual OS for the everyday stuff like gaming and browsing and then when I do work such as coding I just start up a virtual machine with something more suitable (typically xubuntu) |
| 08:11 | | * TheWatcher readsup, notes that you can get TAOCP1-4A straight from the publisher for $208, amazon 'only' wants $114. Anyone trying to charge $350 for it is taking the piss. |
| 08:12 | < Syk> | TheWatcher: Australia. |
| 08:12 | < Syk> | Physical bookstore. |
| 08:12 | <@TheWatcher> | Hah, yeah |
| 08:12 | < Syk> | Physical *specialist* bookstore. |
| 08:13 | < Syk> | with goddamn *classical* playing in the background |
| 08:13 | <@TheWatcher> | Huh, £140 in the UK |
| 08:13 | <@froztbyte> | Azash: I still don't know what "fiddle around" refers to |
| 08:13 | < Syk> | Oracle's Java: A Beginner's Guide was like $80 |
| 08:13 | <@TheWatcher> | Guess I should stick that on my wishlist, it's only slightly less likely than the keyboard... |
| 08:13 | <@froztbyte> | but I need to get to the office and be bored out of my mind anyway |
| 08:13 | <@froztbyte> | so |
| 08:13 | <@froztbyte> | \o |
| 08:16 | <@Azash> | froztbyte: 8. to waste time; trifle; dally (often followed by around ): Stop fiddling around and get to work. |
| 08:23 | | * Vornicus thinks he's figured out that bit. |
| 08:33 | <~Vornicus> | "TAOCP1-4A" for a moment there I thought you were leetspeeking about pudding. |
| 08:42 | < Syk> | haha |
| 08:43 | < Syk> | so if I'm ever going to get it, hardback is best? |
| 08:50 | <~Vornicus> | I have the hardbacks. I don't know if it's even /possible/ to get it paperback. As for ebooks, well, I've never liked them in the first place. |
| 08:52 | <@Azash> | Syk: It seems like books that aren't cornerstones (like TAOCP) have prices inversely proportional to their usefulness |
| 08:52 | <@Azash> | I got Anderson's Security Engineering for the equivalent of 30 dollars, new |
| 08:53 | < Syk> | heh |
| 08:55 | | AverageJoe [evil1@Nightstar-4b668a07.ph.cox.net] has joined #code |
| 09:19 | <@TheWatcher> | Yeah, I don't actually think there are paperback versions of it. |
| 09:21 | <@TheWatcher> | (it does amuse me that Knuth's page for them has a "SPECIAL NOTE TO THE SPEAKERS OF FRENCH AND OTHER EXOTIC LANGUAGES" on it.) |
| 09:21 | <@TheWatcher> | (I'd hitherto not considred French to be particularly exotic) |
| 09:21 | <@TheWatcher> | (But it must be, obviously) |
| 09:22 | | * Vornicus tries to remember his (ridiculous) algorithm for this task. |
| 09:24 | <@Azash> | I suppose in computing it is |
| 09:25 | <@Azash> | Excluding that ridiculous pastebin of C code made French by liberal use of defines |
| 09:38 | <~Vornicus> | okay: sort inputs descending in (x, y+z). Keep track of two lists: the results, and the y-z frontier as (y,z) entries, sorted ascending in y (and by that fact, descending in z). For each item: find the location it would be inserted into the y-z frontier as if y were the important part, and same for z. if the z location <= the y location, add this item to the result list, and splice it into |
| 09:38 | <~Vornicus> | the frontier. |
| 09:45 | <~Vornicus> | This makes the result list only those items that are not dominated in all three axes. Performance estimate: O(n log n) time, O(n) auxiliary storage (for the frontier). Typical case will have much less storage, but the time is bounded on the bottom by the sort. |
| 10:00 | | AverageJoe [evil1@Nightstar-4b668a07.ph.cox.net] has quit [Client closed the connection] |
| 10:06 | | RichyB [richardb@58734C.5279B7.EA7DF8.107330] has joined #code |
| 10:11 | | gnolam [lenin@B19C04.C4B928.3477B4.1341ED] has quit [Client closed the connection] |
| 10:11 | | gnolam [lenin@Nightstar-b2aa51c5.cust.bredbandsbolaget.se] has joined #code |
| 10:11 | | mode/#code [+o gnolam] by ChanServ |
| 10:12 | | ErikMesoy [Erik@A08927.B4421D.403BAA.818DC8] has quit [Client closed the connection] |
| 10:13 | | himi [fow035@Nightstar-5d05bada.internode.on.net] has joined #code |
| 10:13 | | mode/#code [+o himi] by ChanServ |
| 10:13 | | ErikMesoy [Erik@A08927.B4421D.403BAA.818DC8] has joined #code |
| 10:13 | | Vornotron [vorn@31356A.68201E.EB0611.E0094F] has joined #code |
| 10:14 | | Vornicus [vorn@31356A.68201E.EB0611.E0094F] has quit [Ping timeout: 121 seconds] |
| 10:16 | | Vornotron is now known as Vornicus |
| 10:16 | | mode/#code [+qo Vornicus Vornicus] by ChanServ |
| 11:08 | | thalass [thalass@Nightstar-879f199f.bigpond.net.au] has quit [[NS] Quit: Leaving] |
| 11:30 | <~Vornicus> | woot. after some silly direction errors, I have that working, and quickly enough to be of use. Let's see if I can handle the rest now. |
| 11:38 | | VirusJTG [VirusJTG@Nightstar-09c31e7a.sta.comporium.net] has joined #code |
| 11:38 | <@TheWatcher> | ... fucking hell |
| 11:40 | <~Vornicus> | ??? |
| 11:41 | <@TheWatcher> | Something's wrong with the universe - just wrote this monster function to do queries on the database for articles in the system I'm writing at work. It works. First time. |
| 11:41 | <@TheWatcher> | Doesn't happen, ever. |
| 11:41 | <~Vornicus> | Shit |
| 11:41 | <@Tamber> | ...maybe it's just being very subtly broken? |
| 11:41 | <@TheWatcher> | Yeah, probably. |
| 11:42 | < Syk> | or |
| 11:42 | <@Tamber> | I hope so; otherwise you just uncovered a bug in the universe. |
| 11:42 | <@Tamber> | D: |
| 11:42 | < Syk> | the universe is saving up bugs |
| 11:42 | < Syk> | and will later unleash them all at once |
| 11:42 | <@Tamber> | Australia: The universe's bug repository |
| 11:42 | <~Vornicus> | I remember when I first started programming, and I would write functions and they'd work the first time. |
| 11:42 | < Syk> | Vornicus: and then you got better and lost that skill? :P |
| 11:43 | <~Vornicus> | Now, I look back at my ridiculously simplistic code and wonder what the hell I thought I was trying to do. |
| 11:43 | < Syk> | heh |
| 11:43 | <@Tamber> | Vornicus, "10 GOSUB 20\n 20 PRINT 'HELLO'\n30 RETURN" |
| 11:43 | <@Tamber> | ? |
| 11:43 | <~Vornicus> | Error: RETURN without GOSUB |
| 11:44 | | * Tamber boggles |
| 11:44 | <@Tamber> | "...but it's right *there*!" |
| 11:44 | <@Tamber> | :) |
| 11:44 | <~Vornicus> | Right, but the second time you hit it it's not on the stack. |
| 11:44 | <@Tamber> | ...aha, right you are. |
| 11:45 | < Syk> | wow |
| 11:45 | < Syk> | XBMC on a raspberry pi is... incredibly slow |
| 11:45 | <~Vornicus> | Good god I haven't written commodore basic in 20 years and I can still catch those. |
| 11:45 | <@Tamber> | hehehe |
| 11:45 | <~Vornicus> | I stopped writing basic before I learned to program. |
| 11:46 | < Syk> | you know when you see one of those PCs, that like, run Windows Vista |
| 11:47 | <@Tamber> | "run"? |
| 11:47 | < Syk> | but have the hardware equiv of an underclocked pentium M |
| 11:47 | <@Tamber> | more like "limp" |
| 11:47 | < Syk> | with 16MB of shared graphics memory |
| 11:47 | <@Tamber> | or "hobble" |
| 11:47 | < Syk> | that's what XBMC on the raspi is like |
| 11:50 | <~Vornicus> | Okay, next up. Write the system-wide merger, and then the galaxy-wide adder, and then we'll have ourselves a time. |
| 11:52 | <~Vornicus> | oh, and the, uh, data reader... |
| 12:35 | | VirusJTG_ [VirusJTG@2B12AA.572255.206A2A.901581] has joined #code |
| 12:36 | <@froztbyte> | hehe |
| 12:36 | <@froztbyte> | Vornicus: if the world ends, I'd blame you |
| 12:36 | <@froztbyte> | but I might not have the internet to do at that time |
| 12:36 | <@froztbyte> | so just be pre-aware of that |
| 12:38 | | VirusJTG [VirusJTG@Nightstar-09c31e7a.sta.comporium.net] has quit [Ping timeout: 121 seconds] |
| 12:58 | <~Vornicus> | how is it my fault now? |
| 13:35 | | ToxicFrog is now known as ToxicFrog|W`rkn |
| 13:37 | | Vornicus [vorn@31356A.68201E.EB0611.E0094F] has quit [[NS] Quit: ] |
| 13:43 | <@froztbyte> | I didn't say it's your fault |
| 13:43 | <@froztbyte> | I just said I'd blame you ;p |
| 13:43 | <@froztbyte> | (re your apparently bug-less code earlier) |
| 15:36 | <&ToxicFrog|W`rkn> | Hmm. Possibly I should clean up the old filesystem *before* copying it over to the new one. |
| 15:36 | <&ToxicFrog|W`rkn> | rsync has been copying the kernel source for the last god knows how many releases of linux for ages now~ |
| 15:36 | <@froztbyte> | combine the steps with rsync! |
| 15:36 | <@froztbyte> | (it has a --exclude) |
| 15:42 | <&ToxicFrog|W`rkn> | You can't incrementally build the exclude list, though |
| 15:42 | <&ToxicFrog|W`rkn> | It's easier to just rm stuff before invoking rsync |
| 15:42 | <&ToxicFrog|W`rkn> | It would be nice if rsync had an interactive mode with a "skip this dir and delete anything from it you've already transferred" button. |
| 15:43 | <&ToxicFrog|W`rkn> | why do I still have a complete source tree for Puppy Linux |
| 15:43 | < Syk> | because puppy linux is great |
| 15:43 | | * froztbyte throws Syk over the edge |
| 15:45 | | * Syk falls into the abyss of niche linux distributions |
| 15:48 | | * TheWatcher hairpulls at automake/conf/gofuckingcrazy |
| 15:48 | <@TheWatcher> | Twenty million different fucking versions, and woe betide you if you don't have just the right one installed |
| 15:49 | <@TheWatcher> | How is this pile of festing, putrid excrement supposed to make software /more/ portable again? |
| 15:49 | <@froztbyte> | heee |
| 15:49 | <@TheWatcher> | Need a sodding autoconf script just to work out which version of fucking autoconf you need |
| 15:50 | <@froztbyte> | don't you dare speak badly of the tower of win that C programmers worship, TW |
| 15:50 | <@froztbyte> | they'll write WELL-OPTIMIZED loops to squish you |
| 15:50 | <@TheWatcher> | I am a fucking C programmer. I teach the godsforsaken language. |
| 15:50 | <@froztbyte> | but yeah, I pretty much agree |
| 15:50 | <@froztbyte> | it's a festering shitpile |
| 15:51 | <@froztbyte> | and any time anyone mentions it, someone goes "but it's pretty good at what it does, given the ecosystem" |
| 15:51 | <@froztbyte> | to which my response is only: "you have low standards" |
| 15:51 | <&ToxicFrog|W`rkn> | Syk: these days I favour TinyCore |
| 15:52 | < Syk> | I just keep an ubuntu 12.04 live USB with me at all times |
| 15:52 | < Syk> | comes in handy |
| 15:53 | < Syk> | eg. "my windows has diederated!" |
| 15:53 | <&ToxicFrog|W`rkn> | TinyCore has the advantage that I can fit it on anything. |
| 15:54 | <@froztbyte> | including a 1.44MB stiffy disk? |
| 15:54 | <&ToxicFrog|W`rkn> | I have a rescue image that's TinyCore (with full wireless stack), memtest86, x86test, freeDOS with various drive diagnostics, and a few other things I don't remember that fits in 32MB. |
| 15:54 | <&ToxicFrog|W`rkn> | Anything that the computers I interact with will actually read, then~ |
| 15:55 | <@froztbyte> | I've just been using sysreccd lately |
| 15:55 | <&ToxicFrog|W`rkn> | Anyways, at that size, I just put it on every form of storage I have and make it bootable. |
| 15:55 | <@froztbyte> | since I deal with a fairly wide variety of base systems |
| 15:55 | <@froztbyte> | xfs/ext3/mdadm/jffs/reiser, then various assortments of retarded stuff |
| 15:55 | <@froztbyte> | I need to build a new rescue image thing soon though |
| 16:11 | <&ToxicFrog|W`rkn> | ZFS is so tasty |
| 16:12 | < Syk> | yes it is <3 |
| 16:12 | <&ToxicFrog|W`rkn> | Although the procedure for creating a degraded vdev deliberately is a bit wacky |
| 16:13 | | * ToxicFrog|W`rkn COMPRESSES ALL THE THINGS |
| 16:14 | < Syk> | compress it until your entire fs can be expressed on this potato chip |
| 16:15 | <&ToxicFrog|W`rkn> | I don't think lzjb is that good |
| 16:15 | <@froztbyte> | lzjb? |
| 16:15 | <@froztbyte> | ah |
| 16:16 | <@froztbyte> | haven't had to use it yet |
| 16:16 | <&ToxicFrog|W`rkn> | It's the default zfs inline compression thinger |
| 16:16 | <&ToxicFrog|W`rkn> | Designed to give useful compression levels without significantly affecting performance. |
| 16:16 | | * ToxicFrog|W`rkn has it turned on for /home, /orias/src, and /orias/media/books |
| 16:17 | <@froztbyte> | yeah, I'm familiar with its states of working, just not it itself |
| 16:17 | | * froztbyte has still not ever set up a zfs himself |
| 16:17 | <@froztbyte> | in a bit of a funny twist of fate |
| 16:17 | <@froztbyte> | even though I am familiar with the zfs zdb structs |
| 16:18 | <&ToxicFrog|W`rkn> | ...hmm |
| 16:18 | <&ToxicFrog|W`rkn> | extN doesn't support uid remapping on mount |
| 16:18 | <&ToxicFrog|W`rkn> | This is going to be super unpleasant |
| 16:19 | <@froztbyte> | extN only has good support for sucking |
| 16:20 | <&ToxicFrog|W`rkn> | (the old system and the new system use a different uid for me) |
| 16:20 | <&ToxicFrog|W`rkn> | (I have 3TB of files to chown) |
| 16:27 | | * gnolam adds some alpha. |
| 16:27 | <@gnolam> | Now let's see: will it blend? |
| 16:38 | <&ToxicFrog|W`rkn> | froztbyte: I do recommend setting a zfs at some point, so far it's pretty sweet |
| 16:39 | <&ToxicFrog|W`rkn> | My only real complaint is that you can't reshape a raidz |
| 16:39 | <@froztbyte> | at this stage I'll probably pass Go and roll directly to btrfs |
| 16:39 | <&ToxicFrog|W`rkn> | It kicks the shit out of mdadm and btrfs in every other respect |
| 16:39 | <@froztbyte> | at least, based on the timelines of when I'll have home raid-capable storage available again |
| 16:40 | < Syk> | i have two operating zfs things |
| 16:40 | < Syk> | my hp microserver has a pair of mirrored wd red 1tbs, and a client ibm server with mirrored wd red 2tbs |
| 16:40 | < Syk> | both operate beautifully |
| 16:41 | < Syk> | also, I am very surprised at how fast this little 1.4GHz turion dual core can kick data around |
| 16:41 | <&ToxicFrog|W`rkn> | This is 4x2TB RAIDZ of WD Reds with one disk missing (it's part of the array I'm copying the data from) |
| 16:41 | < Syk> | right now my zfs scrub is going at 107MB/s |
| 16:41 | <&ToxicFrog|W`rkn> | I've got /home with copies=3 and everything else straight on the raidz |
| 16:42 | <@froztbyte> | at present my home server is still lightning-dead from february |
| 16:42 | <@froztbyte> | only at the end of this month will I finally be able to fix that |
| 16:43 | <&ToxicFrog|W`rkn> | This is stage1 of a lengthy series of upgrades to Orias that will ultimately retire its current mdadm-based setup built on discarded gaming hardware and replace it with a small mITX system eSATA'd to the raidz, with an SSD for / /boot and ZFS ZIL and cache |
| 16:52 | | celticminstrel [celticminst@Nightstar-e83b3651.cable.rogers.com] has joined #code |
| 16:52 | | mode/#code [+o celticminstrel] by ChanServ |
| 17:07 | | VirusJTG_ [VirusJTG@2B12AA.572255.206A2A.901581] has quit [Client closed the connection] |
| 17:08 | | VirusJTG_ [VirusJTG@Nightstar-09c31e7a.sta.comporium.net] has joined #code |
| 17:08 | | Turaiel[Offline] is now known as Turaiel |
| 17:20 | | Derakon [chriswei@Nightstar-a3b183ae.ca.comcast.net] has joined #code |
| 17:20 | | mode/#code [+ao Derakon Derakon] by ChanServ |
| 17:30 | < RichyB> | ToxicFrog|W`rkn: "eSATA'd to the raidz" - what? |
| 17:31 | < RichyB> | You can have a Solaris or FreeBSD machine present a block file on a ZFS partition as an eSATA disk? |
| 17:32 | < RichyB> | Or did you mean eSATA'd to each of the set-of-disks which make up the raidz? |
| 17:33 | < RichyB> | (The former would be really, really awesome. I would *love* to be able to easily present disk images on a home-made NAS to a Windows machine as real SATA hard disks. ^_^) |
| 17:54 | | * Derakon mutters at this problem. |
| 17:55 | <&Derakon> | I have a bunch of axis-aligned tiles splatted down on a 2D plane. I want to draw a box on the plane and get an array of pixel data. |
| 17:55 | <&Derakon> | It's doable, just the offsets and intersections and stuff are painful to calculate. |
| 17:59 | <@gnolam> | Sounds like a classic tilemap. :) |
| 17:59 | <&Derakon> | The tiles aren't on a grid, though. |
| 17:59 | <&Derakon> | Their positions and sizes are arbitrary. |
| 17:59 | <&Derakon> | Fortunately they aren't rotated! |
| 18:00 | < Syk> | SIGH, so far in google IO |
| 18:00 | < Syk> | they showed a google version of the GS4 |
| 18:00 | < Syk> | with stock andorid |
| 18:00 | < Syk> | $649 |
| 18:00 | < Syk> | awkward silence, then a whistle |
| 18:08 | | * Derakon ponders going to lunch two hours early. |
| 18:51 | < RichyB> | Syk: is that more or less than the GS4 costs? |
| 18:51 | <&ToxicFrog|W`rkn> | RichyB: I mean there is an external drive enclosure containing the disks that make up the raidz, connected to the server via eSATA |
| 18:51 | < Syk> | RichyB: less |
| 18:51 | < Syk> | but still more than the N4 |
| 18:51 | < RichyB> | I would pay a small price premium for "stock Andoid" over the Samsung/HTC/whatever customised versions because they're always negatively-differentiated. |
| 18:52 | <@froztbyte> | heh |
| 18:52 | < RichyB> | ToxicFrog|W`rkn: ahhh. It would be so cool if you could easily set up "eSATA targets" like you can set up iSCSI targets. :) |
| 18:52 | | * froztbyte actually prefers the HTC-enabled ROMs, although he does take them in LeeDroiD flavour |
| 18:52 | | * froztbyte does not like blandroid. at all. |
| 18:53 | < RichyB> | I'm serious about that. You stick stock Android on, you get latest stable stock Android. |
| 18:53 | <@froztbyte> | but then you also have the latest blandroid |
| 18:53 | <@froztbyte> | :( |
| 18:53 | < RichyB> | You stick a vendor-differentiated version on, you get to wait an extra six months at least for small unimportant things like whichever fucking security vulnerability fixes are in upstream AOSP |
| 18:53 | < Syk> | cyanogenmod <3 |
| 18:54 | < RichyB> | Yes, Cyanogen FTW. |
| 18:54 | <@froztbyte> | LeeDroiD is the CM version of HTC stuff |
| 18:55 | <@froztbyte> | and it is pimp as hell |
| 18:55 | <@froztbyte> | Lee has clue, and spends it wisely |
| 18:55 | <&ToxicFrog|W`rkn> | HTC? |
| 18:55 | <@froztbyte> | so it's to the HTC default as CM is to the Samsung roms |
| 18:55 | < Syk> | froztbyte: uh |
| 18:55 | <&ToxicFrog|W`rkn> | RichyB: to be honest I have no idea what iSCSI is or how it works |
| 18:55 | <@froztbyte> | Syk: we've had this discussion |
| 18:55 | < Syk> | lee isnt in the scene anymore |
| 18:55 | < Syk> | for like |
| 18:55 | < Syk> | a year |
| 18:55 | <@froztbyte> | Syk: we're not going to agree |
| 18:56 | < Syk> | CM to the samsung roms? |
| 18:56 | < Syk> | except CM support on samsung is shit |
| 18:56 | <@froztbyte> | yes |
| 18:56 | < RichyB> | ToxicFrog|W`rkn: iSCSI is is SCSI-commands-over-IP. |
| 18:56 | <@froztbyte> | "good" |
| 18:56 | < Syk> | because Samsung cannot into exynos |
| 18:56 | <@froztbyte> | isn't CM basically rooted in Samsung? |
| 18:56 | < Syk> | no, it's not? |
| 18:56 | <@froztbyte> | or at least, that used to be the case |
| 18:56 | <@froztbyte> | has that stopped? |
| 18:57 | <@froztbyte> | that was kinda my mental understanding of why my CM-based ROM experience has been so shit |
| 18:57 | < RichyB> | ToxicFrog|W`rkn: so it lets you share a block device over any network that supports IPv4 or IPv6 - typically nice cheap gigabit ethernet. |
| 18:57 | <@celticminstrel> | Suddenly I am being needlessly pinged. :/ |
| 18:57 | <@froztbyte> | "my phone's not a samsung, that's probably it" |
| 18:57 | < Syk> | froztbyte: CM is on everything |
| 18:57 | < Syk> | it's just that HTC are idiots |
| 18:57 | < Syk> | and Samsung make terrible drivers |
| 18:57 | <@celticminstrel> | What is this CM? |
| 18:57 | <@froztbyte> | cyanogenmod |
| 18:57 | < RichyB> | celticminstrel: CyanogenMod. |
| 18:57 | <@froztbyte> | Syk: well, objectively |
| 18:57 | <@celticminstrel> | I think I've heard of that actually... |
| 18:58 | <@froztbyte> | beyond the leedroid ROMs, I really haven't used a good ROM yet |
| 18:58 | <@froztbyte> | and even the leedroid ones had some prickly bits |
| 18:58 | < Syk> | it's because HTC cannot into 3rd party support |
| 18:58 | < RichyB> | A homebrew Android ROM project that ships very-new versions of Android for various handsets. With some customisations and stuff, I think? |
| 18:58 | < Syk> | because they're absolutely horrible |
| 18:58 | <@froztbyte> | HTC themselves don't matter for this statement |
| 18:58 | <@froztbyte> | their support is (was) irrelevant to the quality of lee's work |
| 18:59 | < RichyB> | ToxicFrog|W`rkn: if you're building a whitebox or homebrew SAN, iSCSI is how the clients will end up talking to the server. |
| 19:00 | <@froztbyte> | RichyB: speaking of shitty network block device interaction |
| 19:00 | <@froztbyte> | RichyB: have you ever used ATAoE? |
| 19:00 | < RichyB> | No, haven't even considered trying it. |
| 19:00 | < RichyB> | ATA is a shitty protocol even before you try networking it. |
| 19:01 | <@froztbyte> | I haven't really found a reason to either |
| 19:01 | <@froztbyte> | but I keep wondering whether there's any possible upside |
| 19:01 | <@froztbyte> | iSCSI initiators on linux are also not all too wonderful |
| 19:01 | <@froztbyte> | they take a bit of beating at times |
| 19:01 | <@froztbyte> | (though I must admit my knowledge is around 2 years stale on it now) |
| 19:01 | < RichyB> | Oh interesting. |
| 19:02 | <@froztbyte> | well, the open-iscsi initiator was a retard |
| 19:02 | < RichyB> | So it ATAoE throws IP away. It's just raw ethernet. |
| 19:02 | <@froztbyte> | yeah |
| 19:02 | <@froztbyte> | directed frames |
| 19:02 | <@froztbyte> | which seems like it might have potential |
| 19:02 | <@froztbyte> | but I haven't used it, nor found anyone who has |
| 19:07 | <&ToxicFrog|W`rkn> | RichyB: so, this may be a stupid question, but why not just...use iSCSI? |
| 19:07 | <&ToxicFrog|W`rkn> | Nothing ZFS specific, just expose the disks/disk images as iSCSI devices. |
| 19:09 | <&ToxicFrog|W`rkn> | (although ZFS does have iSCSI support built in) |
| 19:10 | <&ToxicFrog|W`rkn> | (I'm not sure how your proposed eSATA approach would work - eSATA connects two computers, one pretends to be a set of block devices?) |
| 19:14 | <@froztbyte> | yes |
| 19:15 | <@froztbyte> | so you could treat the latter set of disks as a jbod or single block device |
| 19:15 | <@froztbyte> | and chain a couple of systems together that way |
| 19:15 | <@froztbyte> | there's a couple of drive chassis which work this way |
| 19:17 | <&ToxicFrog|W`rkn> | I'm not clear on how that use differs from standard RAID backplanes |
| 19:17 | <@froztbyte> | "standard"? |
| 19:17 | <@froztbyte> | oh, hw raid? |
| 19:18 | <&ToxicFrog|W`rkn> | Yeah, external enclosures that can be configured between jbod/concat/raid1 |
| 19:18 | <@froztbyte> | ToxicFrog|W`rkn: think of it as a bunch of drives, like you'd normally have in your machine, but in another chassis |
| 19:18 | <@froztbyte> | and with only a single cable connecting them all |
| 19:18 | <@froztbyte> | rather than a bunch of sata cables |
| 19:18 | <&ToxicFrog|W`rkn> | Er |
| 19:18 | <&ToxicFrog|W`rkn> | So, eSATA, then |
| 19:18 | <@froztbyte> | not always |
| 19:19 | <&ToxicFrog|W`rkn> | Well, yeah, you need a port multiplied SATA controller |
| 19:19 | <@froztbyte> | some of them are some propriety crap just built for the chassis vendor |
| 19:19 | < RichyB> | ToxicFrog|W`rkn: the only advantage that I'm interested in is that you could use dirt-cheap ordinary SATA controllers as the clients rather than having to work out how to configure an iSCSI initiator on $os_of_your_choice. |
| 19:19 | <@froztbyte> | but yes, one's own version would be an esata multiport controller |
| 19:20 | <@froztbyte> | I wanna revisit iscsi sometime |
| 19:20 | <@froztbyte> | but I'll probably just spend time with ceph first :/ |
| 19:22 | <&ToxicFrog|W`rkn> | I'm now totally lost |
| 19:22 | <@froztbyte> | common reaction to storage |
| 19:22 | <&ToxicFrog|W`rkn> | I mean, I can see other uses for having a system that can present fake SATA devices |
| 19:23 | <&ToxicFrog|W`rkn> | Like having it present disk images as actual disks |
| 19:23 | <@froztbyte> | that duplicates IO |
| 19:23 | <&ToxicFrog|W`rkn> | But if your goal is just "access a bunch of disks with one cable", get a $30 SATA port multiplier, done |
| 19:23 | <@froztbyte> | first to the image layer, then again to the disks |
| 19:24 | <@froztbyte> | which doesn't sound like a big deal, but at volume can seriously become a pain in the ass |
| 19:24 | <&ToxicFrog|W`rkn> | Like, I feel like I don't understand what you and RichyB are talking about, because it seems like a really convoluted way of doing stuff we can already do |
| 19:24 | <&ToxicFrog|W`rkn> | But if that were the case it wouldn't be appealing |
| 19:25 | <@froztbyte> | hmmmm |
| 19:25 | <@froztbyte> | well, are you considering all of this from the viewpoint of having the storage (whatever flavour) directly attached to the system itself? |
| 19:26 | <@froztbyte> | also, for reference, as to <ToxicFrog|W`rkn> Like having it present disk images as actual disks |
| 19:26 | <&ToxicFrog|W`rkn> | I kind of was, yeah, hence my question about "eSATA connecting two computers" |
| 19:26 | <@froztbyte> | iSCSI exports a LUN/target, which is usually implemented in that way |
| 19:26 | <@froztbyte> | ToxicFrog|W`rkn: it's mostly just a thing if you run out of physical places to put the disks in the one machine |
| 19:27 | <@froztbyte> | like a friend of mine has systems with 96TB to a system |
| 19:27 | <@froztbyte> | they occupy a full rack |
| 19:27 | <&ToxicFrog|W`rkn> | ...right, but we already have iSCSI for that |
| 19:27 | <@froztbyte> | bunch of L2 cascaded chassis things |
| 19:27 | <&ToxicFrog|W`rkn> | And the argument for "but then you get to use cheap SATA controllers" doesn't hold any water unless you're actually using the SATA controller" i.e. are physically connected |
| 19:28 | <@froztbyte> | but with some other cross-connect, which has the bandwidth for running the chassis at useful speed |
| 19:29 | <&ToxicFrog|W`rkn> | ...so the idea is that you would go something like server -- SATA controller -- SATA/something bridge -- interconnect -- something/SATA bridge -- SATA connectors -- box o'disks? |
| 19:30 | <@froztbyte> | usually more server -> [cablespam] -> [boxes of disks with SATA expanders to each] |
| 19:30 | <@froztbyte> | I mean, if you're ghetto-building this |
| 19:31 | <&ToxicFrog|W`rkn> | No, I mean, whatever you were talking about with the cross-connect |
| 19:31 | <@froztbyte> | then this is the sort of thing I'm talking about: http://www.supermicro.com/products/nfo/storage.cfm |
| 19:32 | <&ToxicFrog|W`rkn> | Ok, but how do you talk to those? iSCSI? |
| 19:32 | <@froztbyte> | that is a very varied question |
| 19:34 | <&ToxicFrog|W`rkn> | So recontextualize this, I'm still trying to understand RichyB's question about RAIDZ-on-eSATA and networked(?) eSATA |
| 19:34 | <&ToxicFrog|W`rkn> | s/So/To/ |
| 19:34 | <@froztbyte> | ah |
| 19:35 | < RichyB> | ToxicFrog|W`rkn: you answered that question ages ago by explaining what you actually meant by "raidz-on-eSATA" |
| 19:35 | < RichyB> | ToxicFrog|W`rkn: the latter question was actually just "wouldn't it be nice if you could easily white-box things something that looks like a Dell MD3200?" |
| 19:36 | < RichyB> | MD3200 is this thing: http://www.dell.com/Learn/us/en/555/shared-content~data-sheets~en/Documents~ss69 4-powervault-md3200-direct-attached-storage.pdf?c=us&l=en&s=biz |
| 19:36 | <&ToxicFrog|W`rkn> | Yeah, but I still don't actually understand what you're asking, or what you mean by "white-box" |
| 19:37 | < RichyB> | "White-box" means put a machine yourself from commodity parts rather than buying a black box, ideally without having to make up a bunch of PCBs or ASICs yourself. |
| 19:38 | < RichyB> | The wikipedia article is more or less spot-on. http://en.wikipedia.org/wiki/White_box_%28computer_hardware%29 |
| 19:38 | | Nighthawk [Helix@Nightstar-33970526.range86-155.btcentralplus.com] has joined #code |
| 19:39 | <&ToxicFrog|W`rkn> | Aah |
| 19:39 | <&ToxicFrog|W`rkn> | Ok, so what you're actually looking for is something that behaves like SAS~ |
| 19:40 | < RichyB> | In a minute or two I'm going to put a fucking fork through my laptop. |
| 19:40 | < Nighthawk> | Why RichyB? |
| 19:40 | <&ToxicFrog|W`rkn> | I mean, if you want to share ZFS storage over the network in a way that looks like local storage, it has iSCSI support built in, and for sharing image files and suchlike, you can do that without getting ZFS installed. |
| 19:41 | < Nighthawk> | Things are never that bad.. is it going slow want some I.T tips.. PM me and we'll speed that little f***er up in no time |
| 19:41 | < RichyB> | Faulty cooling. The laptop started regularly overheating so I took it apart to try to clean the fan a couple days ago. |
| 19:41 | <&ToxicFrog|W`rkn> | If you want to connect shitloads of drives directly, you probably want SAS, which is what the MD3200 uses. |
| 19:41 | <&ToxicFrog|W`rkn> | This will not be cheap, of course, but the price of the SAS controller is going be peanuts next to the cost of the actual storage in an array that size. |
| 19:41 | < Nighthawk> | Zune filing system, prefer Ext4 |
| 19:41 | < RichyB> | Turns out there actually wasn't any crud at all inside the fan... so I'm at a total loss for why the laptop's ability to cool itself has fallen off a cliff. |
| 19:42 | <@Azash> | RichyB: Replace the thermal goop |
| 19:42 | < Nighthawk> | Is it a Toshiba? |
| 19:42 | < RichyB> | No, it's a Dell. |
| 19:42 | <@Azash> | I don't really know why but laptops seem to have a very common issue with the thermal paste giving out |
| 19:42 | < RichyB> | Azash: it's not the chip->heatsink that's dead, it's the laptop-body->outside world heat exchange that isn't taking place efficiently enough. |
| 19:42 | < RichyB> | The damn thing is just failing to move enough volume of air to avoid broiling my crotch. |
| 19:43 | < Nighthawk> | @Azash depends on the laptop, I've got a Tablet & Keyboard combo running Android 4.2 JB and it doesn't have a fan |
| 19:43 | <@Azash> | RichyB: If you say so |
| 19:43 | <@Azash> | Nighthawk: Alright? |
| 19:44 | < Nighthawk> | Touch screen kicks ass, specially when its smaller than the Mac AirBook pro.. it rocks :) |
| 19:44 | < RichyB> | I still haven't ruled out crud in the fan intakes on the bottom, but I still can't physically reach those without doing more disassembly than I care for. |
| 19:44 | < RichyB> | Anyway. Shortly after attempting that and putting it back together, I found out that I had 2 screws left over. |
| 19:44 | < RichyB> | Unfortunately, those screws happen to be responsible for preventing the fan enclosure from vibrating. |
| 19:45 | < RichyB> | Haven't been brave enough to take a second go at tearing it apart to get them back in yet, because those fuckers are deep. |
| 19:45 | <&ToxicFrog|W`rkn> | Nighthawk: I've found that using a Pixel has already rewired my reflexes to the point that I find myself reaching for the screen on my normal laptop sometimes. It's more useful than I originally expected. |
| 19:45 | | * Nighthawk says "well when I say the touch screen kicks ass, so does the nVidia Tegra 3 GPU and all those OpenArena death matches online |
| 19:45 | < Nighthawk> | fan? What fan? |
| 19:46 | < RichyB> | Nighthawk: ffffff ;) |
| 19:46 | < Nighthawk> | ARM chips, if its good enough for a Guided Missle its good enough for Linux ;) |
| 19:46 | < RichyB> | I'm thinking that my next laptop might actually be something like an Asus Transformer TF700 running a load of SSH and rdesktop clients |
| 19:47 | < Nighthawk> | RichyB ah the model up from mine ;) |
| 19:47 | < RichyB> | plus one or machines in the corner running Xen. |
| 19:47 | < Nighthawk> | TF300T unlocked, rooted, running CM10.1 over here ;) |
| 19:48 | < Nighthawk> | 15hours battery life, whats not to love |
| 19:48 | < RichyB> | When I saw the specs on the TF700, I decided I had to have one. :) |
| 19:48 | < RichyB> | It's not readily available in the UK yet, though. |
| 19:48 | <@froztbyte> | <Nighthawk> Zune filing system, prefer Ext4 |
| 19:48 | <@froztbyte> | lawl. |
| 19:48 | < Nighthawk> | It is I've seen them at PC World |
| 19:48 | < Nighthawk> | where I got mine @ RichyB |
| 19:49 | < Nighthawk> | you have to do a search online for a store with them in Stock |
| 19:49 | < Nighthawk> | then travel to buy |
| 19:49 | | * froztbyte , in a twist of fate, is managing to get rid of the last big ext3 storage array today |
| 19:49 | <&ToxicFrog|W`rkn> | RichyB: that's how I rolled for years, these days I prefer to have more stuff at the edge though |
| 19:49 | | * ToxicFrog|W`rkn high-fives froztbyte |
| 19:49 | <&ToxicFrog|W`rkn> | I'm still moving stuff off mine, it'll be at least another day, I think. |
| 19:49 | <@froztbyte> | goodbye ext3 on md-raid5, helloooooo xfs on md-raid6 |
| 19:49 | <@froztbyte> | well |
| 19:49 | < Nighthawk> | nice |
| 19:50 | <@froztbyte> | as soon as this array that I managed to resurrect manages to finish copying all the multiple thousand small files |
| 19:50 | < RichyB> | Nighthawk: http://www.pcworld.co.uk/gbuk/laptops-netbooks/ipad-tablets-and-ereaders/tablets /asus-transformer-pad-infinity-tf700t-1i099a-10-1-convertible-tablet-64-gb-21146 535-pdt.html says outta stock. |
| 19:50 | < RichyB> | I'd probably buy one immediately if they weren't. |
| 19:50 | <@froztbyte> | ToxicFrog|W`rkn: I don't know if you've seen the seek behaviour of ext3? |
| 19:50 | <@froztbyte> | https://oss.oracle.com/~mason/seekwatcher/ext3_vs_btrfs_vs_xfs.png |
| 19:51 | <@froztbyte> | zfs is also fairly favourable |
| 19:51 | < Nighthawk> | that sucks, try for a TF300T instead.. ;) its not that much of a performance biggy, the TF700 just has a better display but the display on the TF700 is so crazy that there's no games that can take advantage of that level of display yet |
| 19:51 | <@froztbyte> | but that first graph should basically explain any part of "don't ever use ext*" to most people |
| 19:51 | < RichyB> | Nighthawk: no deal. The ultra-insane display on the TF700T is the reason why I want it. |
| 19:52 | < Nighthawk> | lol |
| 19:52 | <&ToxicFrog|W`rkn> | froztbyte: that's kind of awesome |
| 19:52 | < RichyB> | I don't want that display for high-resolution 3d games. |
| 19:52 | < RichyB> | I want that display for two reasons: |
| 19:52 | <@froztbyte> | indeed |
| 19:52 | < RichyB> | 1. fonts |
| 19:52 | < RichyB> | 2. to lick the fonts |
| 19:52 | <&ToxicFrog|W`rkn> | That said, I don't really care about the performance, because all this data is accessed over the network anyways |
| 19:52 | <@froztbyte> | ToxicFrog|W`rkn: the software is on the same site, just take the filename out |
| 19:52 | <@froztbyte> | and blktrace is in the debian repos, presumably ubuntu too |
| 19:52 | < Nighthawk> | lol, make sure you buy some lens cleaner and spray it every now and then before you let the licking commence |
| 19:52 | <&ToxicFrog|W`rkn> | I'm ditching ext3+mdadm because ZFS owns owns owns and is much easier to manage |
| 19:53 | <@froztbyte> | the main reason I'm not bothering with zfs on this other thing is I'm out of this company in 12 days and I'm the only person with a bit of zfs clue |
| 19:54 | <@froztbyte> | the other reason is I think zfs might be a bit too heavy on the CPU which is already pretty hard run |
| 19:54 | < Nighthawk> | Debian is the one @froztbyte |
| 19:55 | < Nighthawk> | Even google is migrating a lot of its backbone over to Debian and ditching what they've been using.. which they're tight lipped about.. I wonder if it was Windows 8 Server.. snicker |
| 19:55 | <@froztbyte> | I ... what |
| 19:55 | < RichyB> | froztbyte and TF were talking filesystems, not distros. |
| 19:56 | < RichyB> | froztbyte: AIUI ZFS is pretty civilised as long as you don't try to use transparent compression or deduplication. |
| 19:56 | < Nighthawk> | I read the google internal updates on Google Plus @froztbyte they advertise a lot about where they're vision is going |
| 19:56 | <@froztbyte> | RichyB: yeah, I don't have the RAM for dedup, or the CPU assist (and RAM) for compression |
| 19:56 | <@froztbyte> | RichyB: but I mean "this is an HP Microserver N40L, and runs 2x 1.4GHz atoms" |
| 19:57 | <@froztbyte> | or turion64 or whatever was in that model |
| 19:57 | <@froztbyte> | its function, mostly, is backuppc over rsync protocol (not encap'd in ssh) |
| 19:57 | <@froztbyte> | the new config is raid6 + xfs, so that should also clear up some CPU space |
| 19:58 | <@froztbyte> | but it doesn't have a lot of juice for multiplexing :) |
| 19:58 | < RichyB> | I didn't think compression took much RAM? GNU gzip has a working-set size of about a megabyte, doesn't it? |
| 19:59 | | * Derakon ponders his code, wonders how to procedurally generate 512x512 test patterns that still look distinctive and ordered when resized to ~100x100. |
| 19:59 | <&Derakon> | Right now I'm just generating range(512*512) and then stuffing it into a uint16 so it rolls over, which is fine but makes it hard to tell if I have the right portion of any given image. |
| 19:59 | < Nighthawk> | ASCII art @ Derakon = patterns that still look distinctive and ordered |
| 20:00 | <&Derakon> | I don't have any drawing utilities. |
| 20:00 | <&Derakon> | This is pure manipulate-numpy-arrays stuff. |
| 20:00 | < Nighthawk> | hmm, piping raw ASCII art into the array.. interesting concept |
| 20:01 | | * RichyB coughs |
| 20:01 | <@froztbyte> | RichyB: I'd be talking out of my ass if I tried referring to real numbes on this |
| 20:01 | <&Derakon> | It's the code that generates images for a simulated camera so I can test my microscope software without being hooked up to an actual microscope. |
| 20:01 | < RichyB> | Almost-pure numpy core. ;) |
| 20:01 | < RichyB> | Oh, different project. :D |
| 20:01 | <&Derakon> | Yeah, this isn't Pyrel. I'm at work. :p |
| 20:01 | <@froztbyte> | RichyB: would be interesting to try get some numbers on it, though |
| 20:01 | < RichyB> | Commercialise Pyrel, then. ;) |
| 20:01 | <&Derakon> | Love to. It's open-source and I'm not the only dev. |
| 20:02 | <@EvilDarkLord> | Draw a bunch of wide sine curves and layer them on top of each other? |
| 20:02 | <&ToxicFrog|W`rkn> | froztbyte: compression in ZFS is pretty lightweight by default, doesn't need much RAM or CPU. |
| 20:02 | <&Derakon> | Best I could manage would be a commercial module on top of it or something similar. |
| 20:02 | <@froztbyte> | the people I know who rock extensive zfs rigs do believe far more in the compression than the dedup, though |
| 20:02 | <&ToxicFrog|W`rkn> | deduplication will completely murder you, though |
| 20:02 | <@froztbyte> | yeah, dedup needs a shitton of RAM |
| 20:02 | <@froztbyte> | and adds a lot of CPU too |
| 20:02 | <@froztbyte> | I'll probably have a ZFS system someday. |
| 20:02 | <&Derakon> | Hm...I could probably generate a sinewave pattern and rotate it for subsequent images. |
| 20:02 | < RichyB> | froztbyte: I've seen some favourable benchmarks of Postgres-on-ZFS, where the machine had gobs of CPU to spare but its storage was spinning-rust. Real numbers but not the same case at all. ? |
| 20:02 | < Nighthawk> | yes I found it murderously slow using ZFS |
| 20:02 | <@froztbyte> | RichyB: haha |
| 20:03 | <&ToxicFrog|W`rkn> | I've got it running a massive copy from raid5 with compression enabled, on an old Athlon 64 X2 system |
| 20:03 | <&ToxicFrog|W`rkn> | And, well, the CPUs are maxed out, but that's all mdadm and rsync's internal checksums |
| 20:03 | <&ToxicFrog|W`rkn> | And it's pushing about 50MB/s |
| 20:03 | <@froztbyte> | oh, well, that's still generous |
| 20:03 | < Nighthawk> | because its trying to remember everything for the roll back feature which is great to have.. but thats there for lazy people who dont do backups |
| 20:04 | <@froztbyte> | root@sysresccd /root % grep 'model name' /proc/cpuinfo |
| 20:04 | <@froztbyte> | model name : AMD Turion(tm) II Neo N40L Dual-Core Processor |
| 20:04 | <@froztbyte> | model name : AMD Turion(tm) II Neo N40L Dual-Core Processor |
| 20:04 | < RichyB> | froztbyte: the issue that they were solving in that benchmark was that if you do things like parse Apache logs into SQL rows, you get abysmal space usage because postgresql's built-in compression (TOAST) only operates on individual tuples, which is great for individually-big tuples but not useful for large numbers of mostly-redundant tuples. |
| 20:04 | | Kindamoody|out is now known as Kindamoody |
| 20:04 | <@froztbyte> | erf |
| 20:04 | <@froztbyte> | even just that task .... akldsjklajjsad |
| 20:04 | <@froztbyte> | (I have the same reaction about "php-syslog-ng") |
| 20:05 | < RichyB> | froztbyte: but if you transparently compress the file that is backing the postgresql table with your data in it, then you get to take advantage of lovely things like the fact that LZO both compresses and decompresses faster than 7.2kRPM hard disks can read or write data. |
| 20:05 | <@froztbyte> | Nighthawk: those two things are independent |
| 20:05 | <@froztbyte> | RichyB: yeah |
| 20:05 | <&ToxicFrog|W`rkn> | froztbyte: so...that's faster and more recent than what I have, then? |
| 20:05 | <&ToxicFrog|W`rkn> | Athlon X2, not Athlon II. The original dual-core version. |
| 20:05 | <@froztbyte> | RichyB: the one very nice thing about stuff like zfs/btrfs |
| 20:05 | <@froztbyte> | RichyB: ZIL, and the btrfs equivalent |
| 20:06 | <@froztbyte> | for the most part it far exceeds journalling mechanisms |
| 20:06 | < Nighthawk> | RichyB the benefits of Solid State storage, write 150mb of data in seconds |
| 20:06 | < RichyB> | Yeah, typically you have this thing called "a finite budget". :P |
| 20:06 | <&ToxicFrog|W`rkn> | Nighthawk: that's not how ZFS snapshots/rollbacks work. It's not "trying to remember everything" for them. |
| 20:07 | < RichyB> | "Use ZFS, ZIL and L2ARC so that you mostly get SSD speeds in practice on anything other than fully-random reads" is good advice, "Just use SSDs for everything" is, uh, cute. |
| 20:07 | <@froztbyte> | ToxicFrog|W`rkn: mm, sorry, I might've misunderstood you then |
| 20:07 | <@froztbyte> | ToxicFrog|W`rkn: but yeah, it's a bit of a pickle |
| 20:07 | < Nighthawk> | &ToxicFrog thats the way Solaris was marketing it on OpenSolaris hence I went with Hammer-FS and dragon fly BSD |
| 20:07 | < RichyB> | Dammit. Why won't Fedora boot on this MacBook? |
| 20:07 | <@froztbyte> | I can't easily drop ZFS on here now without committing to it |
| 20:07 | <@froztbyte> | and I can't commit to it because I'm not there in 2 weeks |
| 20:08 | | * froztbyte wouldn't like leaving a potential failure in his wake |
| 20:08 | < RichyB> | froztbyte: yeah, in your situation I'd still go with xfs on raid6. |
| 20:08 | <&ToxicFrog|W`rkn> | froztbyte: this is a ~six year old, 1.8GHz, dual-core processor. |
| 20:08 | < RichyB> | The main constraint here is "cow-orkers don't know/love/grok ZFS" moreso than "ZFS won't work". |
| 20:08 | < RichyB> | Possibly also "the xfs implementation has already been done, ZFS isn't compelling enough to go and redo it." |
| 20:09 | <&ToxicFrog|W`rkn> | Nighthawk: then either their marketing is shit or you misunderstood it, I think. It's not versioned filesystem thing, it's a COW filesystem, which means that snapshots if you make them take negligible time. |
| 20:09 | <@froztbyte> | RichyB: pretty much |
| 20:09 | <@froztbyte> | RichyB: I want to, in the near future, buy another couple of microservers for home |
| 20:09 | <@froztbyte> | and then use them for random things |
| 20:09 | < Nighthawk> | Maybe it was the Celeron @ToxicFrog, to be fair I was running it on a shitty laptop |
| 20:09 | <@froztbyte> | but they don't really have enough CPU juice to justify them :( |
| 20:09 | <@froztbyte> | they're such a nice little box though |
| 20:10 | < RichyB> | froztbyte: I wish I could buy pre-built Shuttle PCs. |
| 20:10 | <@froztbyte> | (they're *fantastic* for random storage setups) |
| 20:10 | < RichyB> | Well, probably can, but I haven't seen anyone selling them. |
| 20:10 | <@froztbyte> | RichyB: why can't you? wrong country? |
| 20:10 | < RichyB> | Haven't seen anyone selling fully-assembled machines online, just the cases. |
| 20:10 | <@froztbyte> | ah |
| 20:10 | <@froztbyte> | glyph recently found another place that does interesting prebuilt servers |
| 20:10 | <@froztbyte> | if you care, I'll dig the URL out of my logs |
| 20:11 | <@froztbyte> | well, s/servers/machines/ |
| 20:11 | < RichyB> | I could put them together myself, but I'd rather someone else did it because I've come to the conclusion that I am unacceptably clumsy. |
| 20:11 | <@froztbyte> | whitebox stuff with useful hardware configurations, at non-insane pricing |
| 20:11 | < RichyB> | froztbyte: in the UK? Please. |
| 20:11 | < RichyB> | When I get around to buying a flat, I'm going to fill the damn thing with computers. :) |
| 20:12 | <@TheWatcher> | Saves on heating bills |
| 20:12 | < Nighthawk> | What about those desktop PC's that are the size of a wireless Hub @ RichyB then you just need to buy a USB to SATA disk reader and buy terabytes of SATA storage and plug and play your HDD's |
| 20:13 | < RichyB> | Nighthawk: those things have atrocious thermal problems. |
| 20:13 | < Nighthawk> | I never had any thermal problems... well not with the disk reader.. But I've not tried the desktop sized hub-style PC |
| 20:14 | < Nighthawk> | and on that recommendation I dont think I would |
| 20:14 | < RichyB> | I think that what I want is a small Unix box as a router and jumping-off point, a shuttle PC as a Xen host, normally switched off but with wake-on-LAN enabled, and a gaming machine lying around somewhere. |
| 20:14 | <@froztbyte> | hmmm, xen |
| 20:14 | | * froztbyte hasn't booted a dom0 in quite some time |
| 20:15 | < RichyB> | I have spent just enough time playing with KVM to have decided that I really prefer Xen at this point in time. |
| 20:15 | | * froztbyte has lately been drinking the kvm koolaid, means to taste the lxc nectar sometime soonish |
| 20:15 | < RichyB> | KVM has... quirks. Meanwhile, there's a small cluster of XenServer boxes hiding in the corner in work and they're pretty much solid. |
| 20:17 | <@froztbyte> | yeah, it's not a 1-size-fits-all |
| 20:17 | <&ToxicFrog|W`rkn> | RichyB: that's kind of sort of what we're going with here, except the "small unix box" is a consumer wireless router with CFW and the shuttle PC is a custom mITX build. |
| 20:17 | <@froztbyte> | Xen also feels snappier in a lot of case |
| 20:17 | <@froztbyte> | probably because of the central control of resources |
| 20:17 | <@froztbyte> | rather than the pre-launch crappery |
| 20:17 | < Nighthawk> | lol @ pre-launch crappery |
| 20:17 | < Nighthawk> | tru |
| 20:18 | < RichyB> | KVM has things like, well, there's SCSI emulation but it doesn't seem to really work very well so the devs took the support for boot-from-SCSI out... |
| 20:18 | <@froztbyte> | RichyB: https://www.system76.com/ is the thing |
| 20:18 | <@froztbyte> | I *think* they offer stuff to the UK too |
| 20:18 | <@froztbyte> | but icbw |
| 20:19 | < RichyB> | OTOH Xen has things like the fact that Citrix have been backing it for years as a supported way to boot violently-shit fussy-as-Hell misoperating systems like Windows. |
| 20:19 | <@froztbyte> | (I didn't look too extensively because getting anything useful to .za is a fucking pain in the arse) |
| 20:19 | < Nighthawk> | you know your getting old when your still using PC based SCSI card converters |
| 20:19 | <&ToxicFrog|W`rkn> | Eurgh, I need to set up the mail stack |
| 20:19 | <@froztbyte> | ToxicFrog|W`rkn: ssmtp |
| 20:19 | < RichyB> | ssmtp FTW. |
| 20:19 | <@froztbyte> | (unless you need a real mail stack) |
| 20:19 | <&ToxicFrog|W`rkn> | I need MTA, MDA, and IMAP. |
| 20:20 | <&ToxicFrog|W`rkn> | Previous stack was postfix, maildrop, and dovecot. |
| 20:20 | < Nighthawk> | mutt mail, pine or sendmail... |
| 20:20 | | * RichyB shudders. |
| 20:20 | <@TheWatcher> | sendmail *twitch* |
| 20:21 | <&ToxicFrog|W`rkn> | MTA needs to be able to forward to an authenticated relay, MDA needs to understand user-defined filtering rules. |
| 20:21 | < Nighthawk> | sendmail, isnt as bad as it used to be.. |
| 20:21 | < RichyB> | Last time I tried to use mutt, I gave up and configured dovecot instead so that I could use Thunderbird. |
| 20:21 | < RichyB> | mutt's usability is worse than dovecot's. :P |
| 20:21 | <&ToxicFrog|W`rkn> | Nighthawk: are you sure you weren't using postfix's "sendmail" aliase~ |
| 20:21 | < Nighthawk> | they've plugged a lot of the sendmail rape bugs |
| 20:22 | < Nighthawk> | like using a sendmail server as a forwarding point for spam |
| 20:22 | < Nighthawk> | its down to the admin if its badly setup |
| 20:22 | <@TheWatcher> | sendmail configuration requires the consuption of at least three illicit substances just to appear sane. |
| 20:22 | < Nighthawk> | exactly |
| 20:22 | < Nighthawk> | lol |
| 20:23 | < RichyB> | That's generally considered to be a bad thing. |
| 20:23 | | * Nighthawk inhales from the crackpipe |
| 20:23 | | * Nighthawk drops some LSD |
| 20:23 | < RichyB> | "Yeah, it sucks balls, but if you spend a lot of time and energy on learning to cope with it, it can be made to behave in a barely-acceptable manner. Also, they've fixed a bunch of the security vulnerabilities that they happen to know about." |
| 20:23 | | * Nighthawk eats some mushrooms and says "ah, sendmail!" |
| 20:24 | <@TheWatcher> | RichyB: are we back to PHP again?~ |
| 20:24 | < Nighthawk> | they'll always be new vulnerabilities but thats just the end users Job find a problem, dont sit there trying to fix it yourself submit a fu**ing RFC and pull-request |
| 20:24 | < RichyB> | Is sendmail the PHP of MTAs? |
| 20:25 | < RichyB> | Or is it PHP that's the sendmail of programming languages? |
| 20:25 | <&Derakon> | sendmail came first, PHP is more notorious. |
| 20:26 | <&ToxicFrog|W`rkn> | froztbyte: based on a quick search sstmp is just an MTA, so it won't do what I need |
| 20:26 | < Nighthawk> | PHP is notorious because you can do more with one line of PHP code than you can with MySQL |
| 20:26 | <@froztbyte> | <Nighthawk> sendmail, isnt as bad as it used to be.. |
| 20:27 | <@froztbyte> | yes |
| 20:27 | <@froztbyte> | yes it is |
| 20:27 | < RichyB> | I prefer usage stories like, "The configuration files are both easy to write and to read and the system doesn't have gotchas or weird edge cases in places where you're plausibly going to run into them. The devs take steps to rule out entire categories of security issues by definition." |
| 20:27 | <@froztbyte> | ToxicFrog|W`rkn: yeah, it's just good for getting mail /off/ the box |
| 20:27 | <&ToxicFrog|W`rkn> | Oh boy oh boy oh boy, do we have a PHP apologist here? |
| 20:27 | <&ToxicFrog|W`rkn> | It's been ages! |
| 20:27 | <@froztbyte> | well, play it slow |
| 20:27 | <@froztbyte> | I need to finish this food that I'm making |
| 20:27 | <@froztbyte> | (thus the sporadic responses) |
| 20:27 | < RichyB> | ToxicFrog|W`rkn: yes, ssmtp is just for punting to an actual-smtpd that's running elsewhere. |
| 20:28 | < Nighthawk> | finish it as in eat it or finish it as in cooking it? |
| 20:28 | <&ToxicFrog|W`rkn> | So, ssmtp may be worth looking at to replace postfix, but not maildrop or dovecot |
| 20:28 | < Nighthawk> | lol @ dovecot how old is.. nevermind |
| 20:28 | < RichyB> | It's useful just because it puts a thing called "sendmail" on your PATH and supports the usual command-line interface. |
| 20:29 | < Nighthawk> | SMTP - Send Mail Transfer Protocol.. which came first the chicken or the egg? |
| 20:29 | <&ToxicFrog|W`rkn> | Nighthawk: ten years, which makes it younger than a lot of useful software. |
| 20:29 | < RichyB> | Main uses for ssmtp are "it's easier than configuring a fully-fledged MTA" and "it's a program that runs, does its stuff, and then quits, rather than a daemon that hangs around." |
| 20:29 | <&ToxicFrog|W`rkn> | And it's "Simple", actually, not "Send" |
| 20:30 | | Kindamoody is now known as Kindamoody[zZz] |
| 20:30 | <&ToxicFrog|W`rkn> | RichyB: yeah, the MTA side on this server basically needs to do two things: |
| 20:30 | < Nighthawk> | yes but daemons take up RAM and when your trying to juggle a workload is that a good thing |
| 20:30 | <&ToxicFrog|W`rkn> | - take local mail and hand it to maildrop |
| 20:30 | <@TheWatcher> | (Which makes it just about the most hilariously incorrectly named system ever...) |
| 20:30 | < RichyB> | e.g. if you want to get FreeBSD in <32MB, replacing the default localhost-only sendmail with a copy of ssmtp helps. |
| 20:30 | <&ToxicFrog|W`rkn> | - take remote mail and forward it to an authenticated relay |
| 20:30 | <&ToxicFrog|W`rkn> | If SSMTP can do both of those, is in the repos, and is easier to configure than postfix, it's in. |
| 20:30 | < Nighthawk> | forward it unauthenticated to a Relay |
| 20:31 | < Nighthawk> | thank god for PKI |
| 20:31 | < RichyB> | I don't know if it can do either of those. |
| 20:32 | | * Nighthawk fires up strongSwan and gpgcli |
| 20:34 | | * Nighthawk mutters "no I dont work at a bank, I just dont see the sense in sending a message in the clear when you can layer it at the packet layer first then armor it with a PKI and send it that way!" SSLsniff that fu**er |
| 20:34 | < RichyB> | Whoever thought up the idea of making all USB3 equipment be physically blue deserves a pat on the back. |
| 20:37 | <&Derakon> | Whyso? |
| 20:38 | < RichyB> | "Does this port I'm about to plug in support USB3?" "Is this a USB3 cable or a USB2?" "Is this a USB3 or USB2 hard disk?" can all be answered entirely by eyeballing the connectors. |
| 20:38 | < RichyB> | Much easier than hunting down information or looking for little logos or whatever. |
| 20:38 | < Nighthawk> | lol |
| 20:38 | < Nighthawk> | eyeballing the connectors.. classic |
| 20:39 | < RichyB> | Deadly serious; this is the kind of standardisation that saves a lot of time in small increments. |
| 20:39 | < RichyB> | Time is money. |
| 20:39 | <@froztbyte> | <Nighthawk> SMTP - Send Mail Transfer Protocol.. which came first the chicken or the egg? |
| 20:39 | < Nighthawk> | speaking of connectors I bought some with a USB female to male and tried connecting up a HP printer that didnt have the cable with it in the box |
| 20:39 | <@froztbyte> | either drunk, or trolling |
| 20:39 | < Nighthawk> | I swear manufacturers are getting tight |
| 20:40 | <@froztbyte> | ToxicFrog|W`rkn: yeah, my main server is postfix+dovecot+theirsql |
| 20:40 | <@froztbyte> | I need to get rid of that latter part |
| 20:40 | < Nighthawk> | it didt work, the printer was complaining of no connection, but it had worked fine previously |
| 20:40 | < Nighthawk> | it was the 6.00 connectors |
| 20:40 | <@froztbyte> | just haven't had sufficient round tuits to care yet |
| 20:40 | < Nighthawk> | one lead change and it worked like a charm |
| 20:41 | < Nighthawk> | drunk & troll.. nice.. way to stab a guy in the back, just trying to contribute my 2 cents worth arsehole |
| 20:42 | <@froztbyte> | just my 2c too, good sir |
| 20:42 | <@froztbyte> | feel free to go right off ignoring it |
| 20:43 | < Nighthawk> | this place blows, I'll go back to playing with Hammer-FS on Dragonfly-BSD and let it fall off me like water off a ducks back, you know about Hammer right, oh of course you do, you know its designed with Raid in mind |
| 20:43 | | Nighthawk [Helix@Nightstar-33970526.range86-155.btcentralplus.com] has left #code ["Leaving"] |
| 20:43 | <&Derakon> | ...okay then. |
| 20:43 | < RichyB> | Might not have been drunk. |
| 20:43 | < RichyB> | Could've just been a dick. |
| 20:43 | <@froztbyte> | rofls |
| 20:44 | <@froztbyte> | well |
| 20:44 | <&Derakon> | Either way, a bad fit. |
| 20:44 | <@froztbyte> | (and what I say next might sound offensive) |
| 20:44 | <@froztbyte> | the impression I get from them is that they're what we'd call in afrikaans an "ou ballie" |
| 20:44 | <@froztbyte> | aka old guy with just enough smarts to figure out these newfangled computer things |
| 20:45 | <@froztbyte> | but you're far more likely to hear about his helpful ability to fix your printer when you meet him in the paintshop or so |
| 20:45 | <@froztbyte> | there's certainly a place for those kinds of people in the world, and it's even how many get started |
| 20:45 | <@froztbyte> | but yeah, what Derakon said |
| 20:48 | <&ToxicFrog|W`rkn> | froztbyte: upon research, it looks like ssmtp won't do what I want, and neither will msmtp (which fills the same niche since ssmtp development stopped and is what's actually in the repos) |
| 20:49 | <@froztbyte> | ah |
| 20:49 | <@froztbyte> | guess you get postfix then |
| 20:50 | <&ToxicFrog|W`rkn> | yep |
| 20:50 | <&ToxicFrog|W`rkn> | Unless someone has a better suggestion |
| 20:50 | <@froztbyte> | you could go for exim~! |
| 20:51 | <@gnolam> | RichyB: didn't know that about USB3. |
| 20:51 | <@gnolam> | But yes. Color-coded ports are wonderful. PC 97 was a godsend for usability. |
| 20:52 | <@froztbyte> | that reminds me, I need to add some clarifying pictures to my latest blog post |
| 20:52 | <@gnolam> | (Especially with 3.5 mm jacks.) |
| 20:54 | <@gnolam> | Usability isn't something I normally associate with USB connectors though. |
| 20:55 | <@froztbyte> | http://blog.froztbyte.net/2013/04/usb-port-orientation-usability-idea/ |
| 20:55 | <@gnolam> | USB - the only connector you have to try 3 out of 2 possible orientations to successfully plug in. |
| 20:56 | <@froztbyte> | little bit of quantum tech in each of them! |
| 20:56 | <@froztbyte> | free 180 degree rotational symmetry with every purchase! |
| 20:56 | <@gnolam> | Non-integer spin! |
| 21:13 | <&ToxicFrog|W`rkn> | froztbyte: why isn't that a serious suggestion? |
| 21:15 | < RichyB> | A lot of Apple's non-standard connectors have 180-degree rotational symmetry for about this reason. |
| 21:15 | < RichyB> | As much as I hate non-standardness in connectors, they're often really quite good. |
| 21:16 | <@froztbyte> | ToxicFrog|W`rkn: because running exim is a bit of a pain in the ass |
| 21:16 | <&Derakon> | You mean, it doesn't matter if they're upside-down or not? |
| 21:16 | <@froztbyte> | well, no |
| 21:16 | <&ToxicFrog|W`rkn> | Aah. |
| 21:16 | <@froztbyte> | running it is fine |
| 21:16 | <&ToxicFrog|W`rkn> | More of a pain in the ass than postfix? |
| 21:16 | <@froztbyte> | configuring it is a bit of a shitty effort |
| 21:16 | <@froztbyte> | ToxicFrog|W`rkn: the config syntax is pretty much a spiritual derivative of m4 |
| 21:17 | <@froztbyte> | and there are a /lot/ of knobs |
| 21:17 | <&ToxicFrog|W`rkn> | Ok, getmail, despite its general shittyness at configuring multiple servers, still destroys fetchmail by virtue of handling stuff with malformed From: headers. |
| 21:17 | <&ToxicFrog|W`rkn> | oh |
| 21:17 | <@froztbyte> | some of which might change defaults between versions, and/or be new |
| 21:17 | <&ToxicFrog|W`rkn> | Also it supports mailboxes=ALL, which owns |
| 21:17 | <@froztbyte> | so you get to pick apart the pieces to find out what's what if you ever run into a showstopper |
| 21:18 | <@froztbyte> | postfix I've found somewhat easier to deal with in general |
| 21:18 | < RichyB> | Derakon: yeah. The Mac power connectors in particular are a work of art. |
| 21:18 | <@froztbyte> | exim is .... well, it makes an "application platform" out of email |
| 21:19 | <&Derakon> | Haha! Success! http://derakon.dyndns.org/~chriswei/temp2/composite.png |
| 21:19 | <@froztbyte> | there are certainly places for it, but I bet you probably don't want it |
| 21:19 | < RichyB> | They're held in only by a magnet, so if you smack them they fall off rather than damaging the power port, and they have 2-rotational symmetry so you can fumble them into place in either direction. |
| 21:19 | < RichyB> | Derakon: worst desktop background ever. |
| 21:19 | <&Derakon> | Heh. |
| 21:19 | <@froztbyte> | Derakon: I have flux running on my desktop |
| 21:19 | <@froztbyte> | currently the blue channel is probably about 30% down from normal |
| 21:19 | <@froztbyte> | the trip that does to my eyes... |
| 21:19 | <&Derakon> | Given an arrangement of tiles in arbitrary shapes, this managed to combine them into a single composite image. |
| 21:20 | <&Derakon> | (The arbitrary shapes happen to be a regular grid in this case, but still!) |
| 21:20 | < RichyB> | and I'm speaking as the man who had a drawing of a girl chasing down other children with a bloody seagull corpse atop her head as his background. |
| 21:20 | <@froztbyte> | rofl |
| 21:21 | <&McMartin> | My current desktop background is a screencap from BIT.TRIP CORE |
| 21:21 | <&McMartin> | I've been thinking of replacing it with an X-COM Geoscape |
| 21:21 | <&McMartin> | Maybe slightly edited |
| 21:21 | <&McMartin> | INDIA QA IS PANICKING |
| 21:21 | < RichyB> | Hee. |
| 21:21 | < RichyB> | My current is a picture of a nuclear reactor glowing due to Cherenkov. |
| 21:22 | <@froztbyte> | my background is on rotation |
| 21:23 | | Turaiel is now known as Turaiel[Offline] |
| 21:23 | <@froztbyte> | couple of pics from vladstudio, couple from interfacelift, couple from another site I've forgotten the name of now |
| 21:24 | < RichyB> | Previously it was this: http://www.reddit.com/r/FWEPP/comments/sixb5/the_download_timer_said_3_seconds/ |
| 21:24 | <@froztbyte> | nice |
| 21:28 | <&Derakon> | McM, have you seen this? http://www.youtube.com/watch?v=JxS0_ckSwqk |
| 21:28 | <&Derakon> | A raycaster written in 254 bytes for the C64. |
| 21:29 | <&McMartin> | I have not, but that sounds both interesting and in some sense about right~ |
| 21:35 | | RichyB [richardb@58734C.5279B7.EA7DF8.107330] has quit [[NS] Quit: >:3 This is BunThulhu. Copy him into your quit message to help him take over the Internet.] |
| 21:36 | | Turaiel[Offline] is now known as Turaiel |
| 21:42 | <&McMartin> | Oooh, he does the "load the executable into the load RAM itself to hijack the load process trick~ |
| 21:42 | | RichyB [richardb@58734C.5279B7.EA7DF8.107330] has joined #code |
| 21:55 | | Turaiel is now known as Turaiel[Offline] |
| 21:55 | | VirusJTG__ [VirusJTG@Nightstar-09c31e7a.sta.comporium.net] has joined #code |
| 21:58 | | VirusJTG_ [VirusJTG@Nightstar-09c31e7a.sta.comporium.net] has quit [Ping timeout: 121 seconds] |
| 22:15 | < RichyB> | Has anyone here ever tried to get a Linux liveCD or USB stick to boot on a MacBook? |
| 22:15 | | * RichyB is currently sitting amid a fairly hefty pile of yak fur here. |
| 22:16 | <@Tamber> | As long as it's not got to the point of sharks, yet |
| 22:16 | < RichyB> | The live DVD I burned freezes about eight seconds after booting the kernel. |
| 22:17 | < RichyB> | :| |
| 22:17 | <@Tamber> | :x |
| 22:18 | < RichyB> | I saw a random forum posting somewhere claiming that CD booting was a bit unreliable but USB booting works fine. |
| 22:18 | < RichyB> | The Macbook doesn't recognise the disk as prepared by the usual mechanism. :| |
| 22:22 | <&ToxicFrog|W`rkn> | good luck |
| 22:22 | <&ToxicFrog|W`rkn> | My current you-are-entering-a-world-of-pain project is reinstalling windows on Quetzalcoatl, since changing to a UEFI motherboard has confused it terribly |
| 22:23 | < RichyB> | Well |
| 22:23 | | * RichyB shrugs |
| 22:23 | < RichyB> | Thank you for the solidarity, at least. :D |
| 22:25 | <&ToxicFrog|W`rkn> | (I would offer help, but I know sweet fuck all about macs) |
| 22:30 | | * RichyB is trying to get Fedora onto it. |
| 22:30 | | * RichyB is not an OS X fan. |
| 22:30 | <@gnolam> | RichyB: I've had generally much more luck getting stuff to boot from USB than CDs/DVDs. |
| 22:30 | | * RichyB is quite enamoured of the hardware though. |
| 22:37 | < RichyB> | Where I last got to was that a USB disk written with "dd if=Fedora-18-x86_64-Live_Desktop.iso of=/dev/sdc" failed to be recognised by the MBP's boot loader |
| 22:37 | < RichyB> | hrmn |
| 22:38 | <&ToxicFrog|W`rkn> | Are fedora's isos multiboot capable? |
| 22:38 | < RichyB> | I've gone down the "shake a yak" route and tried installing Fedora in a virtualbox VM and using the USB passthrough to let the fedora VM write the live-USB thing with "livecd-iso-to-disk --format --reset-mbr --efi ..." |
| 22:38 | < RichyB> | In theory, yes. |
| 22:38 | < RichyB> | Bah, let's try another track. Maybe the bootloader just doesn't like the USB3 enclosure? |
| 22:39 | < RichyB> | This livecd-iso-to-disk thing is taking forever (wonder if it's mkfs'ing the whole 1TB disk I plugged in?) |
| 22:39 | | * Derakon blarghs, has run out of brain. |
| 22:39 | <@Azash> | Does mkfsing really work on the entire HDD though? |
| 22:39 | < RichyB> | Depends on the filesystem. |
| 22:39 | < RichyB> | I'll try the iso again but with a USB2 cable to downgrade it. |
| 22:40 | < RichyB> | e.g. ext2,3,4 want to write a number of inodes in advance that's proportional to the size of the disk. |
| 22:40 | < RichyB> | All the Windows filesystems are proportional to the size of the disk. |
| 22:41 | < RichyB> | The only filesystems I've seen that have O(1) mkfs are the fancy tree-based Unix ones like zfs, reiserfs, (xfs I think?) |
| 22:41 | <@Azash> | Right |
| 22:43 | < RichyB> | I think that maybe jfs (from IBM) has a constant-time mkfs too. |
| 22:43 | < RichyB> | ACK |
| 22:43 | < RichyB> | That was it. The damn bootloader just doesn't have USB3 support. >_< |
| 22:44 | | Reiv [NSwebIRC@Nightstar-95746c1f.kinect.net.nz] has joined #code |
| 22:44 | | mode/#code [+o Reiv] by ChanServ |
| 22:47 | | Vornicus [vorn@31356A.68201E.EB0611.E0094F] has joined #code |
| 22:47 | | mode/#code [+qo Vornicus Vornicus] by ChanServ |
| 22:51 | | himi [fow035@Nightstar-5d05bada.internode.on.net] has quit [Ping timeout: 121 seconds] |
| 23:00 | <&ToxicFrog|W`rkn> | RichyB: it's always the simple things. |
| 23:10 | < RichyB> | I am reminded at this point of the time I spent a day panicking about the HDD in a desktop machine that had nothing more than a frayed SATA cable. |
| 23:10 | | Alek [omegaboot@Nightstar-56dbba0f.in.comcast.net] has quit [[NS] Quit: bbl] |
| 23:23 | | VirusJTG__ [VirusJTG@Nightstar-09c31e7a.sta.comporium.net] has quit [Ping timeout: 121 seconds] |
| 23:23 | | ErikMesoy is now known as ErikMesoy|sleep |
| 23:32 | | VirusJTG [VirusJTG@2B12AA.572255.206A2A.901581] has joined #code |
| 23:38 | | Derakon [chriswei@Nightstar-a3b183ae.ca.comcast.net] has quit [[NS] Quit: leaving] |
| 23:43 | | VirusJTG_ [VirusJTG@Nightstar-09c31e7a.sta.comporium.net] has joined #code |
| 23:43 | | VirusJTG [VirusJTG@2B12AA.572255.206A2A.901581] has quit [Ping timeout: 121 seconds] |
| 23:48 | | Alek [omegaboot@31356A.619C16.B6C27D.BF0D0D] has joined #code |
| 23:48 | | mode/#code [+o Alek] by ChanServ |
| --- Log closed Thu May 16 00:00:15 2013 |