| --- Log opened Fri Oct 04 00:00:39 2013 |
| 00:11 | | Derakon[AFK] is now known as Derakon |
| 00:17 | | Vornicus [vorn@ServerAdministrator.Nightstar.Net] has joined #code |
| 00:17 | | mode/#code [+qo Vornicus Vornicus] by ChanServ |
| 00:22 | | You're now known as TheWatcher[T-2] |
| 00:24 | | himi [fow035@Nightstar-5d05bada.internode.on.net] has quit [Ping timeout: 121 seconds] |
| 00:24 | | Turaiel[Offline] is now known as Turaiel |
| 00:28 | | Derakon is now known as Derakon[AFK] |
| 00:33 | | You're now known as TheWatcher[zZzZ] |
| 00:36 | | ktemkin[awal] is now known as ktemkin[] |
| 00:53 | | Vornicus [vorn@ServerAdministrator.Nightstar.Net] has quit [[NS] Quit: Leaving] |
| 01:19 | < RichyB> | Derakon[AFK], stuff comes off via PCI, not PCI-E? |
| 01:21 | < RichyB> | Derakon[AFK], silly suggestion, but AIUI PCI-E is point-to-point unlike PCI which is a bus, so perhaps having a PCI image input and a PCI-E network card in the same machine, might allow the image to be put on a network without waiting for the camera to release the PCI bus. |
| 01:45 | | Derakon[AFK] is now known as Derakon |
| 01:46 | <&Derakon> | RichyB: the camera card is non-negotiable; it only works in x8 or x16 slots. |
| 01:47 | < RichyB> | aw, I thought from your description that the camera card was plain old PCI, not PCI-E. |
| 01:47 | <&Derakon> | Sorry, I got sloppy with terminology. |
| 01:48 | < RichyB> | So you don't get blocked if you throw the images away instead of trying to dump them to disk or network |
| 01:49 | <&Derakon> | Correct. |
| 01:49 | <&Derakon> | But if there is any disk or network activity, even if it's only nominal amounts per image, then I get blocked. |
| 01:49 | < RichyB> | what happens if you just keep the image in RAM for a bit? Or run md5 over it or something? |
| 01:49 | <&Derakon> | Like, I can write a single byte to a file, per image, and still have trouble. |
| 01:50 | <&Derakon> | Haven't tried wasting time yet, but ultimately we have to be able to keep up with the incoming image flow. |
| 01:50 | <&Derakon> | So anything that amounts to "Go slower" is a non-starter. |
| 01:53 | < RichyB> | I'm not telling you that you want to go slower, I'm asking you what happens if you try stressing things other than the network or disk controllers, such as memory bandwidth or the CPU. |
| 01:53 | < RichyB> | also, what happens if you try to put it up on screen? |
| 01:53 | | himi [fow035@Nightstar-36db723c.ffp.csiro.au] has joined #code |
| 01:54 | | mode/#code [+o himi] by ChanServ |
| 01:54 | <&Derakon> | Haven't tried display; this is raw image data. I do have code to put it into a format that can be displayed, but that's being done on a separate computer (on the other side of the network connection). |
| 01:54 | < RichyB> | Surely the manufacturer will have seen this before. Can you ask them about it? |
| 01:55 | <&Derakon> | I've been talking back and forth with the vendor for months, now. They think there's something wrong with the computer. |
| 01:55 | <&Derakon> | (Even though I've tried it on two computers now and they both fail in exactly the same way) |
| 01:55 | <&Derakon> | Unfortunately, testing your suggestions is going to have to wait until Monday. |
| 01:55 | < RichyB> | Those weren't suggestions for fixes, just things I'm curious about. |
| 01:56 | | * Derakon nods. |
| 01:56 | <&Derakon> | "Diagnostic tests", that's what we call those. :) |
| 01:56 | < RichyB> | Actual suggestion for a fix: perhaps the manufacturer is using computers/motherboards/chipsets with aggregate bandwidth that yours doesn't have? |
| 01:56 | <&Derakon> | We meet the specs they've specified, but those are just CPU/RAM/HD speed requirements. |
| 01:56 | <&Derakon> | And OS reqs. |
| 01:56 | < RichyB> | Like maybe they use a 2-socket server motherboard which, surprise, just happens to have two sets of PCI-E chips. |
| 01:57 | < RichyB> | Could be a requirement that they don't know they have because they've never tried a machine that doesn't meet it. |
| 01:57 | <&Derakon> | Remember that the problem happens even if we go straight to disk. |
| 01:57 | <&Derakon> | But I agree, there could be something like that. |
| 01:58 | < RichyB> | Also, I assume that you've eliminated the obvious sources of trouble inside your OS? No mad virus scanner or whatever. |
| 01:58 | <&Derakon> | Yeah, we nuke that stuff ASAP. |
| 01:58 | <&Derakon> | Rely on firewalls to keep us safe. |
| 02:02 | < RichyB> | What throughput does this data-firehose have? |
| 02:03 | <&Derakon> | The network link is only 1Gbps; we're going to have to upgrade that eventually. |
| 02:03 | <&Derakon> | The incoming data stream, currently, is 100FPS at 540x512, with 2 bytes/pixel. |
| 02:03 | <&Derakon> | Theoretically we ought to be able to manage 200FPS at 512x512. |
| 02:09 | < [R]> | Have you looked at the output of the machine in wireshark or the like? |
| 02:09 | | * [R] is wondering if you've coded it so that it's breaking the images into a bunch if really small packets. |
| 02:10 | < [R]> | Also, have you tried using a static TCP-window instead of the normal sliding TCP-windows? |
| 02:11 | <&Derakon> | I'm not a networking expert, but again, remember: it happens even when we write to disk instead of sending over the network. |
| 02:12 | < [R]> | Tried writting to a striped drive? |
| 02:12 | <&Derakon> | Yeah, the other computer has a RAID-1 drive composed of 4 3TB drives, IIRC. |
| 02:12 | < [R]> | Odd |
| 02:13 | < [R]> | Actually, have you tested the RAID array's max write speed? |
| 02:13 | <&Derakon> | No. |
| 02:13 | <&Derakon> | But again, remember, happens even if we only write 1 byte per image. |
| 02:14 | < [R]> | FS on the array? |
| 02:14 | <&Derakon> | Sorry, what? |
| 02:14 | < [R]> | What's the filesystem. |
| 02:14 | <&Derakon> | Uh, FAT32 I think. |
| 02:14 | < [R]> | ... |
| 02:14 | <&Derakon> | Whatever Windows likes. |
| 02:14 | < [R]> | Use something designed to store a fuckton of small files. |
| 02:15 | <&Derakon> | We're not writing a fuckton of small files; we're writing to a single file, one byte per image. |
| 02:15 | < [R]> | Oh. IIRC NTFS is actually pretty shitty for weird edge cases like this. |
| 02:15 | <&Derakon> | Basically I open a filehandle and then do handle.write('\x01') each time I process an image. |
| 02:15 | < [R]> | Is that file the only thing stored? |
| 02:16 | <&Derakon> | Correct. |
| 02:16 | < [R]> | Is it the only thing that will be stored? |
| 02:16 | <&Derakon> | The only other side-effect of the system is, once every hundred frames I write out to the console how long it took to process that batch of 100 frames. |
| 02:17 | < [R]> | Cause I'm thinking you might want to just raw-write to the array. |
| 02:17 | <&Derakon> | You mean, write the raw image data? That was the first "write to disk" thing I tried. |
| 02:19 | < [R]> | No |
| 02:20 | < [R]> | I mean write to the disk instead of the fs. |
| 02:20 | < [R]> | Well "disk" |
| 02:20 | <&Derakon> | I've a feeling that won't help. |
| 02:21 | < [R]> | ... actually |
| 02:21 | < [R]> | <Derakon> Basically I open a filehandle and then do handle.write('\x01') each time I process an image. <-- are you seriously doing: open(); write(); close()? |
| 02:22 | <&Derakon> | Gods, no. |
| 02:22 | < [R]> | If so. Don't. |
| 02:22 | <&Derakon> | The filehandle stays open for the duration of the program. |
| 02:22 | < [R]> | Then test your array. |
| 02:24 | <&Derakon> | I can't imagine that write speed is a problem, since we're only writing 100 bytes per second. |
| 02:25 | <&Derakon> | Do you have a hypothesis for why it could be failing? |
| 02:26 | <&McMartin> | If you aren't flushing after write, are you sure that stuff is hitting the disk? |
| 02:26 | <&Derakon> | No, I'm not. |
| 02:26 | < [R]> | Honestly, I'm thinking NTFS is being a twat, but you should really be testing the array regardless. |
| 02:52 | | Vorntastic [Vorn@Nightstar-ecdc617f.sub-70-211-16.myvzw.com] has joined #code |
| 03:14 | <&McMartin> | Wouldn't you know it, this program I wrote in VIC-20 BASIC is too slow. |
| 03:29 | <&Derakon> | Throw more hardware at it~ |
| 03:29 | | * McMartin actually rewrites the core loop in machine language |
| 03:30 | <&McMartin> | It's just a simple program mapping keys to musical notes, but keyboard scanning was super-inconvenient in BASIC and had noticable stumbles. |
| 03:30 | <&Derakon> | We had a "piano" program on our C64. |
| 03:31 | <&McMartin> | Yeah, this is like those old programs, but simpler because the VIC's soundchip is shite~ |
| 03:31 | <&Derakon> | Heh. |
| 03:31 | <&McMartin> | If I wanted to make a C64 version I'd have to add a ton of extra things to tweak |
| 03:32 | <&McMartin> | On the plus side, though, I wouldn't have to migrate which voice is being used since I wouldn't have this ridiculous constraint regarding each voice being shifted an octave off from the others. |
| 03:32 | | VirusJTG [VirusJTG@Nightstar-09c31e7a.sta.comporium.net] has quit [[NS] Quit: Program Shutting down] |
| 03:46 | <&McMartin> | Really not a fan of the assembler I'm using for this project. |
| 03:46 | <&McMartin> | http://pastebin.com/EZxbpmR7 |
| 03:49 | | Vorntastic [Vorn@Nightstar-ecdc617f.sub-70-211-16.myvzw.com] has quit [Ping timeout: 121 seconds] |
| 03:51 | | Vorntastic [Vorn@Nightstar-ecdc617f.sub-70-211-16.myvzw.com] has joined #code |
| 04:08 | | Kindamoody[zZz] is now known as Kindamoody |
| 04:16 | | Vornicus [vorn@ServerAdministrator.Nightstar.Net] has joined #code |
| 04:16 | | mode/#code [+qo Vornicus Vornicus] by ChanServ |
| 04:55 | | Derakon is now known as Derakon[AFK] |
| 05:18 | | Orthia [orthianz@3CF3A5.E1CD01.B089B9.1E14D1] has quit [Ping timeout: 121 seconds] |
| 05:29 | | Kindamoody is now known as Kindamoody|afk |
| 05:46 | < [R]> | https://github.com/watsyurname529/linuxbeeppiano |
| 05:50 | | Turaiel is now known as Turaiel[Offline] |
| 05:55 | <&McMartin> | Hooray |
| 06:05 | | RichyB [RichyB@D553D1.68E9F7.02BB7C.3AF784] has quit [[NS] Quit: Gone.] |
| 06:06 | | ErikMesoy|sleep is now known as ErikMesoy |
| 06:08 | | RichyB [RichyB@D553D1.68E9F7.02BB7C.3AF784] has joined #code |
| 06:11 | | Kindamoody|afk is now known as Kindamoody |
| 06:22 | | himi-cat [qicruser@Nightstar-a933e808.optusnet.com.au] has joined #code |
| 06:22 | | mode/#code [+o himi-cat] by ChanServ |
| 06:25 | | Vorntastic [Vorn@Nightstar-ecdc617f.sub-70-211-16.myvzw.com] has quit [[NS] Quit: Bye] |
| 06:33 | | himi-cat [qicruser@Nightstar-a933e808.optusnet.com.au] has quit [Ping timeout: 121 seconds] |
| 07:15 | | You're now known as TheWatcher |
| 07:47 | | celticminstrel is now known as celmin|sleep |
| 08:13 | | Stalker [Z@Nightstar-b920a19c.cust.comxnet.dk] has joined #code |
| 08:32 | | Kindamoody is now known as Kindamoody|afk |
| 09:24 | | himi [fow035@Nightstar-36db723c.ffp.csiro.au] has quit [Ping timeout: 121 seconds] |
| 10:25 | | Orthia [orthianz@3CF3A5.E1CD01.B089B9.1E14D1] has joined #code |
| 10:25 | | mode/#code [+o Orthia] by ChanServ |
| 10:42 | | Syka_ [the@Nightstar-09306ea5.iinet.net.au] has joined #code |
| 10:44 | | Syka [the@Nightstar-1c1b171d.iinet.net.au] has quit [Ping timeout: 121 seconds] |
| 11:13 | | Reiv [orthianz@3CF3A5.E1CD01.B089B9.1E14D1] has joined #code |
| 11:13 | | mode/#code [+o Reiv] by ChanServ |
| 11:14 | | Orthia [orthianz@3CF3A5.E1CD01.B089B9.1E14D1] has quit [Ping timeout: 121 seconds] |
| 11:16 | | Reiv [orthianz@3CF3A5.E1CD01.B089B9.1E14D1] has quit [[NS] Quit: Going dooooown...] |
| 11:17 | | Orthia [orthianz@3CF3A5.E1CD01.B089B9.1E14D1] has joined #code |
| 11:17 | | mode/#code [+o Orthia] by ChanServ |
| 11:40 | | himi [fow035@Nightstar-5d05bada.internode.on.net] has joined #code |
| 11:40 | | mode/#code [+o himi] by ChanServ |
| 11:41 | | VirusJTG [VirusJTG@Nightstar-09c31e7a.sta.comporium.net] has joined #code |
| 13:36 | | Vornicus [vorn@ServerAdministrator.Nightstar.Net] has quit [[NS] Quit: Leaving] |
| 13:53 | | Stalker [Z@Nightstar-b920a19c.cust.comxnet.dk] has quit [Ping timeout: 121 seconds] |
| 14:14 | <@Tarinaky> | Anyone here have any experience with Mono? |
| 14:14 | <@Tarinaky> | tarinaky@Mirkwood:~/src/mono $ gmcs helloworld.cs -pkg:dotnet |
| 14:14 | <@Tarinaky> | Package dotnet was not found in the pkg-config search path. |
| 14:15 | <@Namegduf> | Yeah, but not with building using gmcs directly. |
| 14:16 | | Syka [the@Nightstar-c1b1753b.iinet.net.au] has joined #code |
| 14:16 | <@Namegduf> | Is there a mono-dev package of some sort not installed? |
| 14:16 | <@Tarinaky> | I don't know. |
| 14:16 | <@Namegduf> | What distribution are you using? |
| 14:17 | <@Tarinaky> | Ubuntu. |
| 14:17 | | Syka_ [the@Nightstar-09306ea5.iinet.net.au] has quit [Ping timeout: 121 seconds] |
| 14:17 | <@Tarinaky> | E: Unable to locate package mono-dev |
| 14:17 | <@Namegduf> | Looks like you need to have mono-devel installed. |
| 14:17 | <@Namegduf> | http://www.mono-project.com/DistroPackages/Ubuntu |
| 14:18 | <@Namegduf> | Or mono-complete. |
| 14:18 | <@Namegduf> | If you already have that, maybe it's something else. |
| 14:19 | <@Tarinaky> | I tried mono-devel. |
| 14:19 | <@Tarinaky> | I'm trying mono-complete now. |
| 14:28 | <@Tarinaky> | helloworld.cs(6,29): error CS0117: `System.Windows.Forms.Application' does not contain a definition for `run' |
| 14:28 | <@Tarinaky> | I'm clearly doing something very silly. |
| 14:28 | <@Namegduf> | Sounds like a step forward. |
| 14:28 | <@Tarinaky> | Oh. Case. |
| 14:28 | <@Tarinaky> | Hurray! It worked. |
| 14:28 | <@Namegduf> | Awesome. |
| 14:30 | <@Tarinaky> | Mono stuff has silly dependencies. |
| 14:30 | <@Tarinaky> | But w/e. |
| 14:32 | <@Namegduf> | Well yeah. |
| 14:32 | <@Namegduf> | It was WINE-for-.NET, and now it's just a kind of buggy framework. |
| 14:33 | <@Namegduf> | It's not terrible but as stdlibs go it is one buggy, weird-API-design thing with a fair few sharp corners. |
| 14:33 | <@Namegduf> | Beats PHP by a mile, though. |
| 14:33 | <@Tarinaky> | Someone said there was money in C#. |
| 14:33 | <@Namegduf> | I suspect there is. |
| 14:33 | <@Tarinaky> | And as you know I am nothing if not mercenary. |
| 14:33 | <@Namegduf> | I mean, I get money from it. |
| 14:33 | <@TheWatcher> | yes, but there's also wirk involved in it ;P |
| 14:33 | <@TheWatcher> | *work too |
| 14:33 | <@Namegduf> | There's money involved in PHP, too. :P |
| 14:33 | <@Tarinaky> | I have yet to find a plan that involves being paid to do nothing. |
| 14:33 | <@Namegduf> | Although it is much better than that. |
| 14:34 | <@Tarinaky> | The best I can come up with is being paid to do very little. |
| 14:34 | <@Namegduf> | I'm not saying don't use it, just... silly dependencies are only a sort of secondary effect of its stdlib's weird design and corners, as a result of the history. |
| 14:35 | <@Tarinaky> | I've found ambiguity over whether Monodevelop does or does not support WinForms. |
| 14:35 | <@Tarinaky> | And Monodevelop crashes when I try to run it. |
| 14:35 | <@Tarinaky> | Likely because of a missing dependency. |
| 14:38 | <@Tarinaky> | Being able to bodge some gui boilerplate together using a RAD tool is more or less the main thing to attract me to C#. |
| 14:39 | <@Namegduf> | If you want portable RAD for GUI apps, Mono will not provide. |
| 14:39 | <@Namegduf> | Platform-specific issues are not unusual. |
| 14:39 | <@Namegduf> | Very common, in fact. |
| 14:40 | <@Namegduf> | You need to test somewhat carefully, notice where behaviour differs, etc. |
| 14:40 | <@Tarinaky> | I see. |
| 14:40 | <@Namegduf> | You also usually fall back on .NET on Windows, which has differing behaviour in a lot of places the documentation is kind of vague. |
| 14:40 | <@Tarinaky> | Better ideas? |
| 14:41 | <@Namegduf> | PyQt? |
| 14:41 | <@Tarinaky> | Ew. |
| 14:41 | <@Namegduf> | I don't think you'll find cross-platform Mono less ew. |
| 14:41 | <@Namegduf> | MonoGame is *much, much worse* and that's what I spend most of my time with, but when it comes to GUI stuff, Mono tends to do platform-specific things quite easily. |
| 14:42 | <@Namegduf> | Network code is fine aside Mono's thread pool not behaving like .NET's causing some "fun" problems. |
| 14:42 | <@Tarinaky> | Well, the alternative is to abandon Mono and run back to java. |
| 14:42 | <@Namegduf> | I'd honestly suggest that. RAD GUIs are going to look like shit, may as well just use Java's stuff. |
| 14:43 | <@Tarinaky> | No RAD tools for Java though. |
| 14:43 | <@Namegduf> | Huh. |
| 14:43 | <@Tarinaky> | At least, so far as I am aware. |
| 14:43 | <@Tarinaky> | I really didn't like PyQt when I played with it though. |
| 14:44 | <@Namegduf> | Might be worth looking into any. |
| 14:49 | | * ErikMesoy is reading more of the crank material on Urbit. It uses "gonadic combinators". Where gloriously advanced, professional, theoretical and academic computer scientists have monads, the peasant language Urbit just has gonads. :D |
| 14:50 | < ErikMesoy> | Urbit also has the best haters: "This must be what Chapter 8 of the Punishment Book in Anathem is like." :D |
| 14:50 | < Syka> | what is this urbit |
| 14:50 | <@Tarinaky> | That's the martial language where software is right nfirst time isn't it? |
| 14:51 | < ErikMesoy> | Syka: it's an immutable combined operation system/programming language/social network written by an absolute crank who thinks that most Computer Science has been DOING IT WRONG for decades. |
| 14:51 | < ErikMesoy> | http://www.urbit.org/ |
| 14:51 | <@Namegduf> | The syntax is on drugs, and names are at most four characters. |
| 14:51 | <@Namegduf> | The theory is that people will learn it like they learn a new spoken language. |
| 14:51 | <@TheWatcher> | Sounds like some student projects I've marked. |
| 14:52 | <@Namegduf> | Based on their expert understanding of linguistics. |
| 14:52 | < ErikMesoy> | In addition to being OS, language and social network, it will also solve spam :D |
| 14:52 | <@Tarinaky> | TheWatcher: That's Abject Oriented Programming. |
| 14:53 | <@Namegduf> | The nautical terms remind me of seasteading. |
| 14:53 | <@Namegduf> | I suspect this is partially deliberate. |
| 14:53 | <@Namegduf> | Isn't this guy also Moldbug? |
| 14:53 | <@Namegduf> | Maybe that was a joke. |
| 14:54 | < ErikMesoy> | Definitely high class crank though. He has a working demo and a sense of humor. |
| 14:55 | <@Namegduf> | It's a pity they saddled their "social network" communications system which solves spam to their rather insane language that, well, no one will ever use for anything. |
| 14:56 | <@Namegduf> | Oh yeah, you could flag your political leaning, couldn't you? |
| 14:56 | < ErikMesoy> | People use Perl >_> |
| 14:56 | <@Namegduf> | People don't, however, use brainfuck. |
| 14:56 | <@TheWatcher> | Perl isn't insane. |
| 14:56 | <@Namegduf> | Which is a closer comparison in terms of readability. |
| 14:56 | <@TheWatcher> | PHP? /That's/ insane |
| 14:57 | < ErikMesoy> | This language has ruuuuuunes :D (digraphs) |
| 15:00 | <@TheWatcher> | (I still contend that if you print the source for PHP out with the correct font and characters per row, and glue the sheets of paper together, what you'll find is that it forms the largest trollface.jpg in the world) |
| 15:00 | <@Namegduf> | XD |
| 15:05 | <@Tarinaky> | As opposed to Perl which is meerely indestinguishable from line noise. |
| 15:05 | < ErikMesoy> | http://www.popehat.com/2013/10/03/what-is-urbit/#comment-1124800 |
| 15:11 | < RichyB> | TheWatcher, well |
| 15:12 | < RichyB> | I think that Perl and PHP are both mad, but Perl is an eccentric science-related-memetic-disease kind of mad; it makes bold, possibly wildly wrong choices, on purpose. |
| 15:13 | < RichyB> | PHP is just mad because it washes the methanol down with petrol and has no remaining functional brain cells. |
| 15:15 | <@TheWatcher> | Yeah, that's probably fair,a ctually |
| 15:24 | | Turaiel[Offline] is now known as Turaiel |
| 15:27 | | thalass [thalass@Nightstar-4ab061cc.bigpond.net.au] has joined #code |
| 15:27 | <&jerith> | RichyB: https://twitter.com/firxen/status/386103490293952512 |
| 15:28 | < RichyB> | I would fall over laughing if I'd heard that in person. |
| 15:28 | < thalass> | *snerk* |
| 15:30 | <&jerith> | RichyB: His next bit was "and also because I like it when people give me money." |
| 15:31 | <@TheWatcher> | Well, the first part was right, anyway. |
| 15:39 | < RichyB> | jerith, I have perfect sympathy for that. |
| 15:45 | | Stalker [Z@Nightstar-5aa18eaf.balk.dk] has joined #code |
| 16:01 | | celmin|sleep [celticminst@Nightstar-1afc984c.dsl.bell.ca] has quit [[NS] Quit: And lo! The computer falls into a deep sleep, to awake again some other day!] |
| 16:37 | | Turaiel is now known as Turaiel[Offline] |
| 16:51 | | ErikMesoy1 [Erik@Nightstar-404c8d4f.80-203-17.nextgentel.com] has joined #code |
| 16:52 | | ErikMesoy [Erik@Nightstar-404c8d4f.80-203-17.nextgentel.com] has quit [Ping timeout: 121 seconds] |
| 17:05 | | ErikMesoy1 is now known as ErikMesoy |
| 17:44 | | celticminstrel [celticminst@Nightstar-1afc984c.dsl.bell.ca] has joined #code |
| 17:44 | | mode/#code [+o celticminstrel] by ChanServ |
| 17:58 | | thalass is now known as Thalasleep |
| 17:58 | <@celticminstrel> | Okay, so they got 800k apparently. |
| 17:58 | <@celticminstrel> | Whoops, wrong channel. XD |
| 17:58 | <&jerith> | RichyB: The backstory is that FaceBook paid him for a proof of concept and the disagreed about whether the performance improvements would scale up to a more complete implementation. |
| 18:02 | | Thalasleep [thalass@Nightstar-4ab061cc.bigpond.net.au] has quit [Ping timeout: 121 seconds] |
| 18:03 | < RichyB> | But they also paid him to complete the implementation anyway? |
| 18:04 | | Turaiel[Offline] is now known as Turaiel |
| 18:04 | <&jerith> | RichyB: N |
| 18:04 | <&jerith> | o. |
| 18:05 | <&jerith> | But he has a hosting startup based on high-performance VMs. |
| 18:05 | <&jerith> | So he's working on the PHP implementation for that. |
| 18:05 | < RichyB> | Is the background of this "HHVM" or something else? |
| 18:05 | <&jerith> | Well, *someone*'s working on it. |
| 18:06 | < RichyB> | I was assuming that his project was HHVM, but now I realise that I didn't know that. |
| 18:06 | <&jerith> | https://twitter.com/CompSciFact/status/386171493261926400 |
| 18:06 | <&jerith> | RichyB: No, this is different. |
| 18:06 | < RichyB> | heehee |
| 18:06 | < RichyB> | Ah I see. |
| 18:06 | <&jerith> | FB was realising that HipHop was a bit of a dead end. |
| 18:06 | < RichyB> | Well⦠aahhhhyes. |
| 18:06 | <&jerith> | So they were exploring various options. |
| 18:07 | <&jerith> | One of which was Hippy. |
| 18:07 | < RichyB> | Doesn't it produce, like, multi-gigabyte binaries? |
| 18:07 | <&jerith> | RichyB: Yes. That's part of the reason it's a dead end. |
| 18:08 | <&jerith> | Also, you need to change your C++ compiler flags to have template recursion 100+ levels deep or something crazy. |
| 18:08 | < RichyB> | I remember reading someone's slides where they were explaining that they used bittorrent internally to distribute the compiled binaries on their LANs because they were too big to deploy from fireservers. |
| 18:08 | < RichyB> | and also that, in production, they have some 3-stage compilation thing going on |
| 18:08 | <&jerith> | HHVM is their new thing that's actually a bytecode interpreter. |
| 18:09 | < RichyB> | compile PHP -> C++, compile C++ with instrumentation flags, then run that binary for an hour or two, then recompile the C++ with profile-guided optimisations turned on. |
| 18:10 | < RichyB> | Is HHVM actually just a sane JITing VM? |
| 18:10 | <&jerith> | Anyway, fijal's currently looking for people with big PHP deployments that might be interested in hippy. |
| 18:11 | <&jerith> | RichyB: I don't know. I think it's still mostly vapourware or something. |
| 18:13 | <&jerith> | http://www.hhvm.com/blog/ |
| 18:14 | < RichyB> | Yeah, "too bad it doesn't run my code." |
| 18:14 | <&jerith> | Looks like HHVM has a fancy JIT but isn't actually a complete PHP implementation. |
| 18:14 | < RichyB> | I have an enormously-fast incomplete PHP implementation that came with my copy of gnu coreutils. |
| 18:14 | < RichyB> | It's called "/bin/true". |
| 18:14 | < ErikMesoy> | How do I force-expand a bit.ly shortened URL? It points to a PDF and I want to know where the PDF is. |
| 18:14 | <&jerith> | . |
| 18:15 | <@Alek> | this, only for any tiny url. |
| 18:15 | < Syka> | ErikMesoy: wget! |
| 18:15 | < Syka> | it'll print out where it's being 304'd to |
| 18:15 | <@Alek> | -_- |
| 18:15 | < RichyB> | ErikMesoy, curl -sD- -o /dev/null http://bit.ly/abcde | grep Location |
| 18:15 | < Syka> | also curl! |
| 18:15 | < RichyB> | make that |
| 18:16 | < RichyB> | er, make that "grep -i location" # because actually it's supposed to be case-insensitive. |
| 18:16 | < RichyB> | curl -sD- -o /dev/null http://t.co/hw51jyAYdJ | grep -i location |
| 18:17 | < RichyB> | Works for any URL shortener that uses HTTP redirects instead of doing anything really evil like <meta refresh> or serving <script>window.location="http://xyz.com/";</script> at your browser. |
| 18:17 | < RichyB> | ⥠curl |
| 18:23 | < ErikMesoy> | ah, hosted on amazonaws |
| 18:42 | | Turaiel is now known as Turaiel[Offline] |
| 18:57 | | Stalker [Z@Nightstar-5aa18eaf.balk.dk] has quit [Ping timeout: 121 seconds] |
| 19:57 | | Stalker [Z@Nightstar-b920a19c.cust.comxnet.dk] has joined #code |
| 20:13 | < RichyB> | Larry Wall, ladies and gentlemen, is an excellent speaker. On Java: «But managers kind of like it because it looks like you're getting a lot done, you know, if 100 lines of Java code accomplish a task, then it looks like you've written 100 lines, even though in a different language, it might only take 5 lines. You know, it's like, you know, you can eat a 1-pound steak or you can eat, you know, 100 pounds of shoe leather and you feel a greater s |
| 20:13 | < RichyB> | ense of accomplishment after the shoe leather, but, you know, maybe they're some downsides.» |
| 20:18 | <&McMartin> | Larry Wall's stated design aesthetic is really good |
| 20:18 | <&McMartin> | I have real trouble squaring his stated design aesthetic with the language he actually, you know, designde |
| 20:18 | <&McMartin> | *designed |
| 20:23 | <&ToxicFrog> | Snrk |
| 20:23 | <&ToxicFrog> | I'm using that quote next time I'm discussing Java~ |
| 20:24 | < ErikMesoy> | When I run a new game I have installed, here are some reasonable things it can do: 1) Run the game in windowed mode. 2) Run the game in fullscreen at my current resolution. 3) Open a dialog asking what resolution and windowing I want to run it at. |
| 20:25 | < ErikMesoy> | Then there's the acceptable option: 4) Run in fullscreen at a different resolution, give me everything back on exit/alt-tab out. |
| 20:25 | < ErikMesoy> | I notice a lot of games seem to pick the shit option 5) Run in fullscreen at a different resolution, and when I exit, my windows have been resized and moved around. |
| 20:25 | | * ErikMesoy nerdrages. |
| 20:26 | <&ToxicFrog> | That may be the desktop/WM's fault, not the game's. |
| 20:26 | < ErikMesoy> | Obviously it doesn't just apply to games, but games seem more prone to fullscreening than most applications IME |
| 20:27 | < ErikMesoy> | ToxicFrog: I'm a little curious how it would be the WM's fault when I have some games that do 4 and some games that do 5. |
| 20:27 | < ErikMesoy> | At least it seems like one is hooking into the WM rightly, and the other wrongly. |
| 20:27 | <&ToxicFrog> | Oh. In that case I have no idea. |
| 20:27 | | * ToxicFrog remembers having this problem a long time ago, hasn't had it in years even when playing older games, concluded that it was a desktop environment issue, not a game issue |
| 20:32 | <&McMartin> | Hrm |
| 20:32 | <&McMartin> | I wonder if it's a question of how the resolution is changed |
| 20:33 | <&McMartin> | Like, if you're using Direct3D or something and say "hey, I'm taking over the screen, use this resolution" it preserves everything, but if you instead say "hey, give me a window with a canvas widget on it that's the size of the whole screen, and make the desktop this size" then it juggles things |
| 20:34 | <&McMartin> | I've played some modern games where Alt-Tabbing out actually *changes the resolution back* |
| 20:34 | | celticminstrel [celticminst@Nightstar-1afc984c.dsl.bell.ca] has quit [Connection reset by peer] |
| 20:37 | | celticminstrel [celticminst@Nightstar-1afc984c.dsl.bell.ca] has joined #code |
| 20:37 | | mode/#code [+o celticminstrel] by ChanServ |
| 20:38 | <@Namegduf> | I was working on something which wanted full screen which alt-tabbed well without crashing. |
| 20:38 | <@Namegduf> | I eventually said "fuck it" and went with fake full screen. |
| 20:39 | <@Namegduf> | Because DirectX's real full screen does weird things when it loses focus. You can make it work but it does annoying things. |
| 20:39 | | celticminstrel [celticminst@Nightstar-1afc984c.dsl.bell.ca] has quit [[NS] Quit: KABOOM! It seems that I have exploded. Please wait while I reinstall the universe.] |
| 20:40 | | * McMartin nods |
| 20:40 | <&McMartin> | This was one of the things that Vista fixed, but that nobody noticed was fixed until Win7, because fuck Vista~ |
| 20:41 | <&McMartin> | (DX10 stores GPU stuff that might need to be reloaded in userspace and reloads it behind the scenes, AIUI. DX9 and earlier are allowed to just trash all your textures and shaders whenever you lose focus.) |
| 20:41 | | celticminstrel [celticminst@Nightstar-1afc984c.dsl.bell.ca] has joined #code |
| 20:41 | | mode/#code [+o celticminstrel] by ChanServ |
| 20:41 | | Kindamoody|afk is now known as Kindamoody |
| 20:42 | <&McMartin> | (Caveat: This is half-remembered from my graphics books, I've never directly worked with this) |
| 20:43 | | celticminstrel [celticminst@Nightstar-1afc984c.dsl.bell.ca] has quit [Client closed the connection] |
| 20:43 | <@Namegduf> | Sadly, no. |
| 20:43 | <@Namegduf> | Although I was working with XNA, it uses a virtualised graphics device of that sort. |
| 20:43 | <@Namegduf> | I think it could retain state entering fullscreen, and exiting fullscreen on purpose, but loss of focus which wasn't deliberately initiated by the program caused a full device reset. |
| 20:44 | <@Namegduf> | (Which, as it happens, we can catch and handle because another platform was doing something almost as bad on a regular basis) |
| 20:44 | <&McMartin> | Ah, OK |
| 20:44 | <&McMartin> | Well, that's still a step up from the warnings I remember for DX9. |
| 20:44 | <@Namegduf> | So multidesktop things sucked. |
| 20:44 | <&McMartin> | Speaking of coding, today is National Vodka Day |
| 20:47 | <@Namegduf> | XD |
| 20:58 | | celticminstrel [celticminst@Nightstar-1afc984c.dsl.bell.ca] has joined #code |
| 20:58 | | mode/#code [+o celticminstrel] by ChanServ |
| 21:29 | | Kindamoody is now known as Kindamoody[zZz] |
| 22:32 | | Turaiel[Offline] is now known as Turaiel |
| 22:42 | < ErikMesoy> | Heeee. I asked the Urbit people for some of their drugs. One of them responded "no drugs are necessary, unless you count 'using git'." |
| 22:45 | | celticminstrel [celticminst@Nightstar-1afc984c.dsl.bell.ca] has quit [[NS] Quit: And lo! The computer falls into a deep sleep, to awake again some other day!] |
| 23:19 | | Vornicus [vorn@ServerAdministrator.Nightstar.Net] has joined #code |
| 23:19 | | mode/#code [+qo Vornicus Vornicus] by ChanServ |
| 23:42 | <&McMartin> | Whoa hey, the Windows port of Emacs has stopped sucking horribly. When did that happen |
| 23:42 | <@Azash> | Maybe you accidentally launched vim? |
| 23:42 | <&McMartin> | No, because vim is awful. |
| 23:43 | < RichyB> | er |
| 23:43 | < RichyB> | I tried Emacs 23 for Windows a while back and it was okay. |
| 23:43 | <&McMartin> | OK, I think I last tried aroudn Emacs 21 or so |
| 23:43 | <&McMartin> | This is 24.2 and it's basically identical to what runs on my Linux lappy |
| 23:44 | < RichyB> | Good enough to be able to load and run the zenburn theme, and good enough that I was able to use hexl-mode to edit my GTA:San Andreas savegame. |
| 23:44 | < RichyB> | Yes. |
| 23:50 | <&McMartin> | Oops, well, that's one piece of hilarious failcannon. |
| 23:51 | <&McMartin> | It keeps failing to save the backup file. |
| --- Log closed Sat Oct 05 00:00:54 2013 |