| --- Log opened Tue Oct 30 00:00:16 2012 |
| 00:10 | | You're now known as TheWatcher[zZzZ] |
| 00:26 | | PinkFreud [WhyNot@NetworkAdministrator.Nightstar.Net] has quit [Ping timeout: 121 seconds] |
| 00:50 | | Netsplit *.net <-> *.split quits: @Vornicus, Orthia, iospace, @Derakon |
| 00:51 | | Netsplit over, joins: ~Vornicus, iospace, Orthia, &Derakon |
| 01:10 | | Derakon is now known as Derakon[AFK] |
| 02:20 | | Netsplit *.net <-> *.split quits: Orthia, @Vornicus, iospace, @Derakon[AFK] |
| 02:21 | | Netsplit over, joins: ~Vornicus, iospace, Orthia, &Derakon[AFK] |
| 02:31 | | Attilla [Obsolete@Nightstar-e04d82c1.as43234.net] has quit [Ping timeout: 121 seconds] |
| 02:35 | | syksleep is now known as Syk |
| 03:01 | | Kindamoody[zZz] is now known as Kindamoody |
| 03:11 | | Derakon[AFK] is now known as Derakon |
| 04:57 | | Kindamoody is now known as Kindamoody|breakfast |
| 05:17 | | ErikMesoy|sleep [Erik@Nightstar-5b5b3bc5.80-203-69.nextgentel.com] has quit [Connection closed] |
| 05:18 | | ErikMesoy [Erik@Nightstar-5b5b3bc5.80-203-69.nextgentel.com] has joined #code |
| 05:18 | | ErikMesoy [Erik@Nightstar-5b5b3bc5.80-203-69.nextgentel.com] has quit [Connection closed] |
| 05:19 | | ErikMesoy [Erik@Nightstar-5b5b3bc5.80-203-69.nextgentel.com] has joined #code |
| 05:19 | | ErikMesoy [Erik@Nightstar-5b5b3bc5.80-203-69.nextgentel.com] has quit [Client closed the connection] |
| 05:20 | | ErikMesoy [Erik@Nightstar-5b5b3bc5.80-203-69.nextgentel.com] has joined #code |
| 05:20 | | ErikMesoy [Erik@Nightstar-5b5b3bc5.80-203-69.nextgentel.com] has quit [Client closed the connection] |
| 05:21 | | ErikMesoy [Erik@Nightstar-5b5b3bc5.80-203-69.nextgentel.com] has joined #code |
| 05:21 | | ErikMesoy [Erik@Nightstar-5b5b3bc5.80-203-69.nextgentel.com] has quit [Client closed the connection] |
| 05:22 | | himi [fow035@Nightstar-5d05bada.internode.on.net] has quit [Ping timeout: 121 seconds] |
| 05:22 | | ErikMesoy [Erik@Nightstar-5b5b3bc5.80-203-69.nextgentel.com] has joined #code |
| 05:22 | | ErikMesoy [Erik@Nightstar-5b5b3bc5.80-203-69.nextgentel.com] has quit [Client closed the connection] |
| 05:23 | | ErikMesoy [Erik@Nightstar-5b5b3bc5.80-203-69.nextgentel.com] has joined #code |
| 05:23 | | ErikMesoy [Erik@Nightstar-5b5b3bc5.80-203-69.nextgentel.com] has quit [Client closed the connection] |
| 05:26 | | Derakon is now known as Derakon[AFK] |
| 05:35 | | himi [fow035@Nightstar-5d05bada.internode.on.net] has joined #code |
| 05:35 | | mode/#code [+o himi] by ChanServ |
| 05:46 | | Kindamoody|breakfast is now known as Kindamoody|out |
| 06:00 | | Syloq_Home [Syloq@NetworkAdministrator.Nightstar.Net] has quit [Client closed the connection] |
| 06:02 | | Netsplit *.net <-> *.split quits: Orthia, @Vornicus, iospace, @Derakon[AFK] |
| 06:03 | | Netsplit over, joins: Vornicus |
| 06:03 | | mode/#code [+qo Vornicus Vornicus] by ChanServ |
| 06:04 | | Netsplit over, joins: iospace, Orthia, &Derakon[AFK] |
| 06:33 | | ErikMesoy [Erik@Nightstar-5b5b3bc5.80-203-69.nextgentel.com] has joined #code |
| 08:23 | | Vash [Vash@Nightstar-b43e074a.wlfrct.sbcglobal.net] has quit [[NS] Quit: I lovecraft Vorn!] |
| 08:37 | | You're now known as TheWatcher |
| 09:42 | | Attilla [Obsolete@Nightstar-4c1bf761.as43234.net] has joined #code |
| 10:32 | | RichyB [richardb@Nightstar-3b2c2db2.bethere.co.uk] has joined #code |
| 10:39 | | You're now known as TheWatcher[afk] |
| 11:34 | | ErikMesoy [Erik@Nightstar-5b5b3bc5.80-203-69.nextgentel.com] has quit [Ping timeout: 121 seconds] |
| 11:36 | | You're now known as TheWatcher |
| 11:54 | | ErikMesoy [Erik@Nightstar-e8459361.80-203-16.nextgentel.com] has joined #code |
| 12:30 | | celticminstrel [celticminst@Nightstar-05d23b97.cable.rogers.com] has joined #code |
| 13:50 | | gnolam is now known as gnolapostate |
| 13:56 | | Pandemic [VirusJTG@Nightstar-09c31e7a.sta.comporium.net] has joined #code |
| 13:56 | < Pandemic> | hello all, are there any exchange gurus around I have an odd issue I haven't seen before |
| 13:57 | <@Tamber> | "It's... actually /working/." |
| 13:57 | < gnolapostate> | Hee |
| 13:57 | < Pandemic> | I've got one account, and only one account, that even though it has been assigned permissions to send on be half of another account cannot do so, because every time, that user does so, the address is resolved to an x400 address. At which point the transport service rejects it and then sends the user an NDR |
| 13:58 | < Pandemic> | all other accounts work fine |
| 13:58 | < Pandemic> | nothing unusual in the x400 or address priority for the account |
| 13:58 | < Pandemic> | others can send on behalf of that account |
| 13:59 | < Pandemic> | and the user in question and send as herself normaly |
| 13:59 | | * Pandemic is more perplexed |
| 13:59 | < Pandemic> | most* |
| 13:59 | | * iospace hats Pandemic |
| 13:59 | <@Syk> | oh god exchange |
| 13:59 | | * Syk cries |
| 13:59 | < iospace> | hi new person ^_^ |
| 13:59 | < Pandemic> | lol |
| 14:00 | < Pandemic> | hello! |
| 14:00 | <@Syk> | >meeting room |
| 14:00 | | * iospace laughs at Syk |
| 14:00 | <@Syk> | >emails people who haven't worked here for 5 years |
| 14:00 | <@Syk> | EXCHANGE LOGIC |
| 14:00 | < Pandemic> | a rouge distribution entry check the list config on the Domain controlers |
| 14:00 | < Pandemic> | speficicly the GAL boxes |
| 14:01 | | * Pandemic has seen that one |
| 14:01 | <@Syk> | the domain controllers are a barrel of fuck too |
| 14:01 | < Pandemic> | yeah that would do it then |
| 14:01 | <@Syk> | i love how if we turn off one DC |
| 14:01 | <@Syk> | authentication for an app goes down |
| 14:02 | <@Syk> | said DC isn't even a PDC or anything |
| 14:02 | < Pandemic> | app is tired to use a static referance rather than a DNS authority look up |
| 14:02 | <@Syk> | it never /was/ a PDC |
| 14:02 | <@Syk> | and had no roles |
| 14:02 | <@Syk> | any domain role |
| 14:02 | <@Syk> | things |
| 14:02 | | * Pandemic has also seen that one |
| 14:02 | <@Syk> | you know what i mean |
| 14:02 | < Pandemic> | tied |
| 14:02 | < Pandemic> | loose the r from that word |
| 14:02 | < Pandemic> | will work better :) |
| 14:03 | | * Pandemic notes he is a dyslexic dysgraphic, spelling and typing may at times be questionable |
| 14:03 | <@Syk> | I was going to start the whole domain from scratch |
| 14:03 | <@Syk> | but then I quit a week ago |
| 14:03 | < Pandemic> | lol |
| 14:03 | <@Syk> | so I guess it doesn't matter :> |
| 14:03 | < Pandemic> | NYP is also an acceptable solution |
| 14:04 | < Pandemic> | (Not Your Problem) |
| 14:04 | <@Syk> | heh |
| 14:04 | <@Syk> | yeah |
| 14:04 | <@Syk> | i have 3 1/2 weeks to go |
| 14:04 | <@Syk> | uh |
| 14:04 | <@Syk> | 2 1/2 |
| 14:05 | < Pandemic> | sounds like an unplesent place |
| 14:05 | < Pandemic> | congrats for getting out |
| 14:05 | < iospace> | Pandemic: we all make typos here :P |
| 14:05 | <@Syk> | yups |
| 14:05 | <@Syk> | there's no director of corp, no finance manager in 2 months, no IT manager on friday, and no experienced IT staff on the 16th of Nov |
| 14:05 | <@Syk> | it's going to be a barrel 'o fun for them |
| 14:06 | < Pandemic> | I project their death in about.... 9 months or less |
| 14:06 | < Pandemic> | depending on how hard headed the owner is |
| 14:06 | <@Syk> | it's a council |
| 14:06 | <@Syk> | it legally can't die |
| 14:07 | <@Syk> | which is the most irritating bit |
| 14:07 | < Pandemic> | it looks like this /could/ be a propogation error that can happen in Exchange 2010, and EX2010 SP1 when an account is migrated from Ex2007 |
| 14:07 | < Pandemic> | word press says to "wait a few hours or restart the transport service |
| 14:07 | | Syloq_Home [Syloq@NetworkAdministrator.Nightstar.Net] has joined #code |
| 14:07 | < Pandemic> | well its like most government though |
| 14:07 | < Pandemic> | incompitent boobs that can't get any thing done |
| 14:07 | < Pandemic> | :-P |
| 14:08 | <@Syk> | unfortunately yeah |
| 14:08 | <@Syk> | :/ |
| 14:09 | < Pandemic> | hey! |
| 14:10 | < Pandemic> | I found something after an hour of googling |
| 14:10 | < Pandemic> | I'll check back at 2pm on the issue |
| 14:10 | <@Syk> | lol |
| 14:10 | | * Pandemic stabs the voodoo dall of steve balmer |
| 14:10 | <@Syk> | Pandemic: you need to get the doll |
| 14:11 | < Pandemic> | yeah..... |
| 14:11 | <@Syk> | put it in a pentagram made in virgin oak |
| 14:11 | <@Tamber> | And throw a chair at it. |
| 14:11 | <@Syk> | engraved with a WP8 device |
| 14:11 | < Pandemic> | lol |
| 14:11 | <@Syk> | put the doll in the middle |
| 14:11 | <@Syk> | and get all sweaty |
| 14:11 | <@Syk> | and then dance around the wood, yelling "DEVELOPERS DEVELOPERS DEVELOPERS DEVELOPERS" |
| 14:11 | <@Syk> | and hope that the mighty ballmer looks upon your wish favourably |
| 14:12 | <@Syk> | otherwise he'll remove it in Windows 9 |
| 14:12 | <@Tamber> | "get all sweaty" "dance around the wood" ~.~ |
| 14:12 | < Pandemic> | lol Tamber |
| 14:12 | < Pandemic> | microsoft won't remove the bug in windows 9, they will just call it a feature |
| 14:12 | < Pandemic> | >.< |
| 14:14 | < Pandemic> | http://social.technet.microsoft.com/Forums/hu-HU/exchange2010/thread/bf68b7f4-90 2f-4e75-be2e-ed8fe571f15b |
| 14:14 | | * Pandemic head desk |
| 14:14 | <@Syk> | >new feature in windows 9 |
| 14:14 | <@Syk> | >the windows 9 filesystem is volatile, meaning you never have to worry about full hard drives again! |
| 14:14 | < Pandemic> | All I want from windows 9 is a UI that doesn't make me want to stab something..... |
| 14:17 | <@Tamber> | "All new Windows 9 user interface! No sharp edges or corners, so you can't hurt yourself on it!" |
| 14:23 | <&ToxicFrog> | Syk: what does "volatile" mean in the context of a filesystem? o.O |
| 14:23 | <@Syk> | ToxicFrog: volatile like RAM |
| 14:23 | < celticminstrel> | ...there's a Windows 9 already? |
| 14:23 | <@Tamber> | No. |
| 14:23 | <@Syk> | celticminstrel: nah |
| 14:23 | < Pandemic> | no |
| 14:23 | <@Syk> | we're making fun of what theyll do in windows 9 |
| 14:23 | < celticminstrel> | Oh. |
| 14:24 | <@Syk> | they've screwed the UI |
| 14:24 | < celticminstrel> | Well, that would be useful, provided you don't care about your data! :P |
| 14:24 | <@Syk> | so the kernel is the next thing |
| 14:24 | < Pandemic> | well at least windows 8 has held the pattern |
| 14:24 | < Pandemic> | ever other version sucks |
| 14:24 | < Pandemic> | I dunno |
| 14:24 | < Pandemic> | I always thought they screw the HAL before they screwed the kernel |
| 14:28 | <@Tamber> | Ah, so all the drivers won't work properly; and your printer will spew blood when you click print? |
| 14:28 | <@Tamber> | ...admittedly, on the printer front, that's not terribly surprising. |
| 14:29 | < Pandemic> | exactly! |
| 14:29 | < Pandemic> | it will be like Russian roulette with drivers. You never know which device will cause the next blue screen. |
| 14:29 | < Pandemic> | wait... |
| 14:29 | < Pandemic> | how is that any diffrent than now.... |
| 14:30 | < Pandemic> | ok looks like the kernal is next |
| 14:30 | <@Tamber> | With my, admittedly limited, experience with Win7; it's ...pretty decent, once I tracked down the dodgy memory. |
| 14:31 | < Pandemic> | Windows 7 I've been very happy with actually. |
| 14:31 | <@Tamber> | It still annoys the everloving shit out of me; but it's pretty stable. |
| 14:31 | < Pandemic> | lol |
| 14:36 | | * Syk hugs Windows 7 |
| 14:36 | <@Syk> | It's Not Vista(TM) |
| 14:39 | | You're now known as TheWatcher[afk] |
| 14:47 | | celticminstrel [celticminst@Nightstar-05d23b97.cable.rogers.com] has quit [[NS] Quit: And lo! The computer falls into a deep sleep, to awake again some other day!] |
| 14:49 | | himi [fow035@Nightstar-5d05bada.internode.on.net] has quit [Ping timeout: 121 seconds] |
| 14:53 | | PinkFreud [WhyNot@NetworkAdministrator.Nightstar.Net] has joined #code |
| 14:53 | | mode/#code [+o PinkFreud] by ChanServ |
| 14:55 | <@froztbyte> | Derakon[AFK]: heh, no stress |
| 14:55 | <@froztbyte> | that's just typically what I start doing though |
| 14:55 | <@froztbyte> | look from the outside of the loop, start stepping in |
| 14:56 | <@froztbyte> | if none of those show issues, I start poking at the variable updates |
| 14:56 | <@froztbyte> | if none of those show issues, I start dumping the fetch values |
| 14:56 | <@froztbyte> | (as a general rule for anything I didn't write myself) |
| 15:02 | | himi [fow035@Nightstar-5d05bada.internode.on.net] has joined #code |
| 15:02 | | mode/#code [+o himi] by ChanServ |
| 15:25 | < iospace> | yay methods of cleaning up redundant code :P |
| 15:53 | | * Tarinaky glares at (fuck the) group project group. |
| 15:53 | < Tarinaky> | What is the point of us having version control if you don't fucking use it. |
| 15:53 | < Tarinaky> | Much better to just upload a zip file containing some disparate code to Facebook >.> |
| 15:56 | < gnolapostate> | ... what |
| 15:56 | < gnolapostate> | Take a clue-by-four to them. >:E |
| 15:57 | <@Syk> | and then thwap them with the packing tube of +2 clue and nails |
| 15:58 | <&ToxicFrog> | Tarinaky: fashion a new case for the VCS server. |
| 15:58 | <&ToxicFrog> | From their bones. |
| 15:58 | <@Syk> | ToxicFrog: what if it's a virtual machine |
| 15:58 | <@Syk> | oh god we're going to need more killing for this VM cluster |
| 15:59 | <@froztbyte> | then you give it a virtual frame |
| 15:59 | <@Syk> | synergise the bones |
| 15:59 | <@Syk> | (???) |
| 15:59 | <@froztbyte> | and that virtual frame can become an art expo in your office |
| 15:59 | <@froztbyte> | but it does justify a killing spree |
| 15:59 | <@froztbyte> | (having to go after all the possible VM metal, that is) |
| 15:59 | <@Namegduf> | Tarinaky: If you're using git, you can copy the .git directory from another checkout into the non-version-controlled code's directory |
| 15:59 | <@Syk> | no court would convict |
| 15:59 | <@Syk> | unless you're in the UK |
| 15:59 | <@Namegduf> | Tarinaky: And it becomes a fully functional git checkout |
| 16:00 | <@Syk> | in which case you're be jailed for a year for hate speech after calling them blockheaded idiots |
| 16:00 | <@Namegduf> | Tarinaky: git status will show you the differences since last checkout, you can commit, reset, etc. |
| 16:00 | <&ToxicFrog> | Namegduf: that is still a lot more effort than the developers using version control in the first place. |
| 16:01 | <@Namegduf> | ToxicFrog: Oh, no disagreement at all. |
| 16:10 | | Syk is now known as syksleep |
| 16:15 | < Tarinaky> | Namegduf: I'll keep that in mind. FOrtunately this doesn't appear to use any resources in the existing git. |
| 16:16 | <@Namegduf> | Ah, okay. |
| 16:16 | < Tarinaky> | It's more that they just haven't bothered at all. |
| 16:23 | < Pandemic> | there, another 16 TB staged to become the third active DPM box |
| 16:23 | < Pandemic> | huza |
| 16:25 | | Kindamoody|out is now known as Kindamoody |
| 16:52 | < Pandemic> | http://www.engadget.com/2012/10/29/chrome-remote-desktop-out-of-beta/ |
| 16:52 | < Pandemic> | useful.... |
| 16:53 | < ErikMesoy> | And pretty good. I've played FTL over it. |
| 16:53 | | ErikMesoy is now known as Harrower |
| 16:53 | < Pandemic> | nice |
| 16:54 | < Pandemic> | and good to know |
| 16:54 | < Pandemic> | I have a friend that is trying to access her mac over the WAN with a windows box |
| 16:54 | < Pandemic> | that should work nicely |
| 16:57 | < Harrower> | I could play Dungeons&Dragons Online over Chrome Remote Desktop if I didn't mind lag measured in seconds. ;-) |
| 16:58 | < Pandemic> | yeah well, nothing is perfect.... |
| 16:59 | | EvilDarkLord is now known as Maze |
| 17:01 | < Tarinaky> | Fuck github for windows with a rusty fucking stick. |
| 17:01 | < Tarinaky> | I installed it to see what it looked like (so I could support it better) |
| 17:01 | < Tarinaky> | And it's dicked with the settings for all my local git repos. |
| 17:02 | <&ToxicFrog> | "installed it"? Wha? |
| 17:02 | < Tarinaky> | Github have a 'speshul' client for Windows. |
| 17:02 | < Tarinaky> | With a GUI. |
| 17:04 | <&ToxicFrog> | Aah. |
| 17:05 | <&ToxicFrog> | It's probably edited your ~/.gitconfig, wherever the fuck that is on windows |
| 17:05 | < Tarinaky> | No. It's edited the .git/config file. |
| 17:05 | < Tarinaky> | To change the source of origin. |
| 17:05 | < Tarinaky> | Which is mental. |
| 17:05 | < Tarinaky> | (Changed it from ssh://git@... to https://...) |
| 17:07 | < Tarinaky> | Currently trying to get my head around Java's synchronized and volatile keywords. |
| 17:08 | <&ToxicFrog> | Um |
| 17:08 | <&ToxicFrog> | You mean it searched your computer for all of your local git repos and edited origin in all of them? |
| 17:09 | < Tarinaky> | Yes. |
| 17:10 | <~Vornicus> | Synchronized methods: when a thread enters a synchronized method on an object, it gains a lock on that object. |
| 17:10 | <&ToxicFrog> | They're pretty straightforward. A synchronized block uses the specified object as a mutex, which is acquired when entering the block and released when leaving it, and upon leaving the block, caches are synchronized. |
| 17:11 | < Tarinaky> | I get that. |
| 17:11 | < Tarinaky> | What does Volatile do? |
| 17:11 | <&ToxicFrog> | volatile is the equivalent for individual variables: access to it is made atomic (via mutex protection a la synchronized) and it is guaranteed to be cache coherent (typically by not cacheing or reordering accesses to it at all) |
| 17:12 | < Tarinaky> | See, that's what's confusing me. |
| 17:12 | <&ToxicFrog> | Actually, disregard that "via mutex protection" part, more recent JVMs use atomic instructions like CAS |
| 17:13 | < Tarinaky> | Half of what I'm reading says that multiple threads are allowed to access a volatile variable at the same time. |
| 17:13 | <&ToxicFrog> | ....yes and no. |
| 17:13 | < Tarinaky> | Just that the caches are flushed after a write. |
| 17:13 | <&ToxicFrog> | You won't end up with cache acoherency issues, but there are still other ways for things to go wrong. |
| 17:14 | <&ToxicFrog> | In particular, each individual access is atomic, which means something like this: |
| 17:14 | <&ToxicFrog> | volatile node foo; foo = foo.next(); |
| 17:15 | <&ToxicFrog> | Will not necessarily work as expected, because the thread may get preempted while calling next, even though the individual reads and writes of foo are atomic. |
| 17:15 | < RichyB> | ToxicFrog: damn, that's a good example. |
| 17:15 | | * ToxicFrog bows |
| 17:15 | < Tarinaky> | Pre-empted? |
| 17:15 | <&ToxicFrog> | Tarinaky: suspended in favour of something else and resumed later. |
| 17:16 | < Tarinaky> | Okay... I thought that not happening was what atomic meant >.< |
| 17:16 | <~Vornicus> | Well there's two accesses happening there |
| 17:16 | <&ToxicFrog> | It is. |
| 17:16 | <&ToxicFrog> | But there's multiple accesses to foo happening. |
| 17:17 | <&ToxicFrog> | Read foo, call next() method on the object retrieved by the read, update the variable with the value returned by the method call. |
| 17:17 | < Tarinaky> | Oh. I see. |
| 17:17 | <&ToxicFrog> | The first and last of those operations are atomic. The one in the middle is not. |
| 17:17 | < Tarinaky> | foo.next() is atomic |
| 17:17 | < Tarinaky> | And foo = /*something*/ is atomic. |
| 17:17 | <&ToxicFrog> | The "foo." part is. The "next()" part is not. |
| 17:17 | < Tarinaky> | Oh! |
| 17:17 | < Tarinaky> | I think... wait... |
| 17:17 | < Tarinaky> | Why is the "next()" part not? |
| 17:18 | <&ToxicFrog> | foo is volatile. That means or reads or updates on foo are atomic. Once you read the value to get an object, though, you can be interrupted while doing things to that object, like calling methods on it. |
| 17:19 | < Tarinaky> | Ah. So it's the handle that's volatile? |
| 17:19 | < Tarinaky> | Not the object. Makes sense. |
| 17:19 | < RichyB> | Yes. |
| 17:20 | <&ToxicFrog> | It's the variable. The storage location in memory named "foo". Reading that is atomic. Changing what object it points to is atomic. Futzing with the object itself? Nope. |
| 17:20 | <&ToxicFrog> | (have I mentioned how much I hate lock-modify-unlock parallel architectures?) |
| 17:21 | < RichyB> | ToxicFrog: seen any of the transactional memory stuff that Intel have been up to? :) |
| 17:22 | <&ToxicFrog> | No, but I've been playing with the STM features in Clojure |
| 17:22 | <&ToxicFrog> | And my research is in message-passing parallel programming on embedded systems |
| 17:22 | <&ToxicFrog> | (and my work history in high-performance megathreading on a Thoth-inspired custom OS) |
| 17:22 | < RichyB> | I've tried the STM support in Haskell a couple times. It's very pleasant. |
| 17:23 | <&ToxicFrog> | A while ago, I independently came up with TM, thinking along the lines of version control |
| 17:24 | <&ToxicFrog> | You have lock-modify-unlock (P4, pthreads), send-receive-reply (Thoth, some DVCS use patterns), but where's copy-modify-merge? |
| 17:24 | < Tarinaky> | Any suggestions on how to design a jUnit test for thread safety? |
| 17:24 | <&ToxicFrog> | Then I found out that Tom Knight had beaten me to it by 25 years |
| 17:24 | <&ToxicFrog> | again |
| 17:24 | < Tarinaky> | I -think- I've made this prototype thread safe. |
| 17:25 | < Tarinaky> | How do I find out? |
| 17:25 | < RichyB> | I don't think copy-modify-merge is a perfect characterisation of STM. |
| 17:25 | < RichyB> | It's more, snapshot then perform ff-only merge, or retry |
| 17:26 | <&ToxicFrog> | RichyB: yeah, but it looks the most like CMM when using it |
| 17:27 | < RichyB> | There are some databases that work more like CMM though! |
| 17:28 | < RichyB> | ZODB is basically a persistent implementation of STM... it has a "resolveConflict" feature, where you can define a special method on any class you write, which will merge the effects of transactions that conflict on objects of that type. |
| 17:28 | < RichyB> | um, I might be explaining this really badly? :( |
| 17:28 | <&ToxicFrog> | Tarinaky: well, here's my best advice for debugging shared-memory parallel systems. |
| 17:28 | < RichyB> | By default, collisions in ZODB's TM implementation cause all but one of the colliding transactions to be retried. |
| 17:28 | <&ToxicFrog> | Tarinaky: "Through me the way into the suffering city, Through me the way to the eternal pain, Through me the way that runs among the lost. Before me nothing was made but sequential things, and I execute concurrently. Abandon every hope, ye that enter." |
| 17:29 | < RichyB> | However, if *all* of the objects in which collisions occurred happen to succeed in merging from both transactions, then both transactions go ahead anyway. |
| 17:29 | <&ToxicFrog> | There is a reason Thoth kernels give you shared memory and then lots of ways to never have to use it~ |
| 17:29 | < Tarinaky> | ToxicFrog: I'm sharing that on Facebook. |
| 17:30 | < RichyB> | e.g. if you have a hit counter or something, you use a BTree.Length object |
| 17:30 | < RichyB> | if two different transactions call length.change(+1) on it, it's got a conflict resolver that'll act as though .change(+2) was called on it instead. |
| 17:31 | < RichyB> | I think that maybe Riak has something like this too, for resolving conflicts where there are disjoint sets of initiator names on the vector clocks? |
| 17:35 | <&ToxicFrog> | RichyB: that's pretty slick |
| 17:39 | < RichyB> | They're really easy to get wrong, apparently. :) |
| 17:39 | <&ToxicFrog> | Yeah, I'm not surprised |
| 17:39 | <&ToxicFrog> | Still cool |
| 17:39 | | * ToxicFrog cranks up the r0xx0r and returns to proofreading his thesis |
| 17:46 | < Tarinaky> | In hindsight my bluster about us not needing a database was misplaced. |
| 17:46 | < Tarinaky> | I can only conclude I am a masochist. |
| 17:47 | | * ToxicFrog cleans that modified quote up a bit, puts it in comments in his thesis >.> |
| 17:50 | < RichyB> | IMHO, for a lot of applications, it's pretty sane to go with: everything is internally sequential -> interaction between concurrently-running processes is performed entirely in a database -> lean on the DB's support for making concurrency work. |
| 17:52 | <&ToxicFrog> | Yeah. |
| 17:52 | <&ToxicFrog> | And in applications where this isn't the case, STM or message passing, for serious. |
| 18:05 | < RichyB> | Raw STM is a pretty okay "database" for applications that have no need of persistence ;) |
| 18:15 | < Tarinaky> | Eclipse appears to have had a major bard. |
| 18:15 | < Tarinaky> | *barf |
| 18:15 | < Tarinaky> | Halp ;/ |
| 18:15 | < Tarinaky> | Launching Eclipse spews out errors and none of the editors or anything open :/ |
| 18:16 | < Tarinaky> | I tried reverting my installed plugins, but that did nothing. |
| 18:25 | <&ToxicFrog> | I cannot help you. |
| 18:25 | | * ToxicFrog ported his development environment to another OS rather than use Eclipse. |
| 18:38 | < Tarinaky> | Now I get the joy of not only having to reinstall all my plugins. |
| 18:38 | < Tarinaky> | I have to convince Eclipse they aren't already installed. |
| 18:38 | < Tarinaky> | Fucking. Hell. |
| 19:02 | | Kindamoody is now known as Kindamoody[zZz] |
| 19:03 | | PinkFreud [WhyNot@NetworkAdministrator.Nightstar.Net] has quit [Ping timeout: 121 seconds] |
| 19:20 | < iospace> | oh semicolons D: |
| 19:31 | | PinkFreud [WhyNot@NetworkAdministrator.Nightstar.Net] has joined #code |
| 19:31 | | mode/#code [+o PinkFreud] by ChanServ |
| 19:39 | <&ToxicFrog> | Chapter 3: proofed. |
| 19:39 | <&ToxicFrog> | I am now 2/3rds of the way through the proofreading process. Yay! |
| 19:40 | < iospace> | we could use eclipse for our assembly class |
| 19:40 | < iospace> | i told eclipse to fuck off and just used core dumps |
| 19:40 | <&ToxicFrog> | Yeah, the devkit for this board uses Eclipse |
| 19:40 | <@Azash> | "we could use eclipse for our assembly class" wut? |
| 19:41 | <&ToxicFrog> | I crowbarred the underlying tools out and taught my build system to invoke them through wine instead |
| 19:51 | | Harrower is now known as ErikMesoy |
| 19:52 | | Maze is now known as EvilDarkLord |
| 19:56 | <&ToxicFrog> | Chapter 4 done. |
| 19:56 | | RichyB [richardb@Nightstar-3b2c2db2.bethere.co.uk] has quit [[NS] Quit: Leaving] |
| 19:57 | <&ToxicFrog> | TODO: link the remaining tables and figures into chapter 5, re-polish the section on the case study, proofread the conclusions and future work...and that's t. |
| 20:04 | <@Azash> | \o/ |
| 20:04 | <@Azash> | So close! |
| 20:23 | | celticminstrel [celticminst@Nightstar-05d23b97.cable.rogers.com] has joined #code |
| 20:40 | | Derakon [chriswei@Nightstar-a3b183ae.ca.comcast.net] has joined #code |
| 20:40 | | mode/#code [+ao Derakon Derakon] by ChanServ |
| 20:40 | <&Derakon> | operator = [min, max][shouldFillView]; zoom = operator(zoom, clientSize[1] / imageSize[1]) |
| 20:40 | <&Derakon> | Should I feel bad about this? |
| 20:41 | < ErikMesoy> | I don't see why |
| 20:41 | <@Tamber> | It's nowhere *near* enterpris-y enough! |
| 20:42 | <&ToxicFrog> | The use of [] for both list literal and list subscript threw me a bit, would it be inappropriate to use (min,max)[...] insttead? |
| 20:42 | <&Derakon> | TF: fair. |
| 20:43 | <&Derakon> | (This is for determining the scale and translate parameters needed to get an image to either fill or fit a given view area) |
| 20:43 | <&Derakon> | (Given the image is not allowed to be distorted) |
| 20:44 | | * Derakon cackles madly as he runs an experiment with one camera at 240x256 and the other at 512x256. |
| 20:44 | <&McMartin> | Derakon: Congratulations, you've just re-invented multiple dispatch~ |
| 20:44 | <&McMartin> | Which is badass, so no, you don't have to feel bad~ |
| 20:44 | <&Derakon> | McM: oh, I've done this kind of thing before. |
| 20:44 | <&Derakon> | Just, usually not with builtins. |
| 20:45 | <&Derakon> | Stylistically it felt a bit odd. |
| 20:45 | | Vash [Vash@Nightstar-b43e074a.wlfrct.sbcglobal.net] has joined #code |
| 20:45 | | mode/#code [+o Vash] by ChanServ |
| 20:47 | <&Derakon> | (Also, usually multiple dispatch is done based on the type of an argument, not its value, yes?) |
| 20:47 | <&McMartin> | (The distinction kind of blurs) |
| 20:47 | <&Derakon> | Though I guess if you step back a meta level, then the type of a parameter is a value itself, in a way. |
| 20:47 | <&McMartin> | (multiple dispatch means there's multiple things you're looking at) |
| 20:48 | <&ToxicFrog> | (pattern matching supremacy) |
| 20:48 | <&McMartin> | (I actually misread the code at first; I read it as parsing equivalent to x[a][b](c, d, e)( |
| 20:48 | <&McMartin> | (oops :())) |
| 20:48 | <&Derakon> | Augh, parens |
| 20:50 | <&Derakon> | I note that the Wikipedia article about multiple dispatch is largely about how to achieve it in languages that don't directly support it. |
| 20:50 | <&McMartin> | That's because Common Lisp is the only language that directly supports it |
| 20:50 | <&McMartin> | (Clojure claims to but its "support" looks a whole lot like what you do to implement it by hand in Python) |
| 20:54 | <&McMartin> | (Whether pattern matching counts as multiple dispatch is fraught enough that I tend to run screaming from the question - it depends on whether it even counts as OO.) |
| 21:02 | | * Derakon takes a moment to cheer the concept of publish/subscribe systems, such that he can insert debugging data into his image viewers just by publishing a fake event. |
| 21:03 | <&Derakon> | The image viewer is just waiting for new-image events; it doesn't care where they come from. |
| 21:03 | <&McMartin> | P/S is awesome |
| 21:04 | <&McMartin> | It's also a major part of why, for all its many flaws, Swing is one of the few GUI APIs I can program directly |
| 21:06 | <&Derakon> | Every GUI system I've worked with has had the concept of an "event queue" at some level. |
| 21:06 | <&McMartin> | Well yes |
| 21:06 | <&Derakon> | Though I admit that SDL's wasn't exactly tightly implemented. |
| 21:06 | <&McMartin> | That's one level down. |
| 21:06 | <&McMartin> | SDL's is not far from Win32's. |
| 21:29 | | You're now known as TheWatcher |
| 21:31 | | * Derakon mutters at mapping from screen coordinates to texture coordinates. |
| 21:39 | | himi [fow035@Nightstar-5d05bada.internode.on.net] has quit [Ping timeout: 121 seconds] |
| 22:14 | | ErikMesoy is now known as ErikMesoy|sleep |
| 22:27 | <&Derakon> | Cool thing I have to take into account: cosmic ray strikes on our cameras. |
| 22:28 | <&Derakon> | They show up as pixels that are much brighter than their neighbors, and occasionally cause problems. |
| 22:28 | <&McMartin> | Hee |
| 22:29 | | * McMartin merges his branch back into trunk! |
| 22:30 | | * Pandemic bonks the AC in the server room |
| 22:30 | < Pandemic> | I hate that stupid cooler |
| 22:49 | <~Vornicus> | I need to get games out of my head and into code, but every time I try I get thwarted by a lot of stupid stuff |
| 22:50 | <&Derakon> | Game development has lots of fiddly stuff you have to wrestle with before you can get to the meat, sadly. |
| 22:51 | <&Derakon> | Even if you use an engine, a) it won't do everything for you, and b) you have to know how to work with the engine. |
| 22:51 | <~Vornicus> | Like for instance pygame crashing terribly. |
| 22:51 | <&Derakon> | ...that's a new one on me. |
| 22:51 | <&Derakon> | What makes it crash? |
| 22:51 | <~Vornicus> | No idea. |
| 22:51 | <~Vornicus> | It just froze up after about 30 seconds. |
| 22:52 | <&Derakon> | You haven't managed to trace it down to "if I do X, it crashes, but if I don't then it survives, albeit uselessly"? |
| 22:52 | <~Vornicus> | it's not like I was doing anything crazy: just panning over a tile map. |
| 22:53 | <~Vornicus> | Hadn't even gotten around to adding controls. The event loop just waited for a quit |
| 22:53 | <&Derakon> | I'd be interested in taking a look, though I have to go offline shortly. Pastebin and I can take a look when I get home. |
| 22:54 | <&Derakon> | First guess is you somehow have a memory leak. |
| 22:54 | <&Derakon> | Despite being in Python. |
| 22:54 | | * TheWatcher stabs things like "result = (result/len/2 + 0.5);" |
| 22:55 | <&Derakon> | Use more parens! |
| 22:55 | <@TheWatcher> | Yeah, damn right |
| 22:55 | <&Derakon> | Or different parens, really, since the only ones there are kind of silly. |
| 22:56 | <@TheWatcher> | Silly, and add insult to injury |
| 23:00 | | * Derakon vanishes. |
| 23:00 | | Derakon [chriswei@Nightstar-a3b183ae.ca.comcast.net] has quit [[NS] Quit: leaving] |
| 23:10 | | himi [fow035@D741F1.243F35.CADC30.81D435] has joined #code |
| 23:11 | | mode/#code [+o himi] by ChanServ |
| 23:15 | | * Vornicus tries it now and it's pretty much working. |
| 23:29 | | Derakon[AFK] is now known as Derakon |
| 23:29 | <&Derakon> | Good to hear. |
| 23:29 | <&McMartin> | Short day? |
| 23:31 | | * Vornicus is unsure what went wrong last time |
| 23:32 | <~Vornicus> | Though this bit of code here always throws an error on exit (but it's a Windows error, not an exception because if it were an exception he'd have a stack trace saved off. |
| 23:34 | <~Vornicus> | oh, that might be it. |
| 23:34 | <&Derakon> | McM: not unusually; I got started at 8. |
| 23:35 | | * Vornicus moves the flip to before the event loop. |
| 23:35 | <&Derakon> | (I also keep a tight rein on my working hours) |
| 23:35 | <~Vornicus> | nope. Still errors out and I don't know why. |
| 23:36 | <&Derakon> | Pastebin / error message? |
| 23:36 | <~Vornicus> | Only difference between the two programs once it's in the main loop is that one's got a try-except around it, and the other one doesn't. |
| 23:36 | | * Vornicus pushes his directory to bitbucket |
| 23:38 | <~Vornicus> | think it was bitbucket anyway |
| 23:38 | <&Derakon> | You've forgotten where your remote was? |
| 23:40 | <~Vornicus> | Somewhat! |
| 23:40 | <~Vornicus> | Now I've forgotten my password. |
| 23:41 | <~Vornicus> | No, I had not, usernames are case sensitive -_- |
| 23:42 | <~Vornicus> | https://bitbucket.org/Vornicus/vornda okay. starting.py, when it exits, gives a Windows error - python.exe has encountered a problem and needs to close. We are sorry for the inconvenience. |
| 23:44 | <~Vornicus> | fullscreentest.py -- which by the way I seem to have forgotten to include the big image, just pick something bigger than 640x480 and it'll work fine -- does not give such an error. |
| 23:45 | <&Derakon> | Any difference between sys.exit() and sys.exit(0), maybe? |
| 23:45 | <~Vornicus> | it was originally sys.exit() and it still did it, I tried giving an explicit exit code to see if it would change anything and it did not. |
| 23:45 | <&Derakon> | I don't know what Windows' process-exit error code system is. |
| 23:47 | <&Derakon> | Hmph...I don't have pygame installed and its build system can't compile. |
| 23:51 | <&Derakon> | What happens if you don't specify HWSURFACE? |
| 23:52 | <&Derakon> | Otherwise, I'm out of ideas, sorry. |
| 23:53 | <~Vornicus> | that also loses me DOUBLEBUF, but it works now! Need to put a sleep in cuz it's a busyloop. |
| 23:53 | <&Derakon> | Best guess then is that the program's not properly releasing resources when it exits. |
| 23:54 | <&Derakon> | I thought there was an SDL cleanup function you were supposed to call on exit but I can't find it in my own code. |
| 23:54 | | * Vornicus pokes at jerith. |
| 23:54 | <~Vornicus> | it is weird though that fullscreentest didn't have that problem and starting did. |
| 23:57 | | gnolapostate is now known as gnolam |
| --- Log closed Wed Oct 31 00:00:31 2012 |