| --- Log opened Fri Nov 28 00:00:38 2008 |
| 00:05 | | You're now known as TheWatcher[T-2] |
| 00:30 | | You're now known as TheWatcher[zZzZ] |
| 03:06 | <@gnolam> | Hmm. Maybe I should start using commit messages that /aren't/ from S. Neverov's "English-Russian Phrase-book". |
| 03:08 | | Consul [~consul@Nightstar-5023.dsl.sfldmi.ameritech.net] has quit [Quit: Leaving] |
| 03:34 | | Consul [~consul@Nightstar-5023.dsl.sfldmi.ameritech.net] has joined #code |
| 03:34 | | mode/#code [+o Consul] by ChanServ |
| 04:15 | <@MyCatVerbs> | gnolam: why. have you already a beet-harvesting combine? |
| 04:15 | <@MyCatVerbs> | s/\./,/ |
| 04:18 | <@gnolam> | People might get confused some day. |
| 04:22 | | Consul [~consul@Nightstar-5023.dsl.sfldmi.ameritech.net] has quit [Quit: Leaving] |
| 04:23 | <@gnolam> | > svn log |
| 04:23 | <@gnolam> | Tell us what provision is made for the collective farmers' recreation? |
| 04:23 | <@gnolam> | ------------------------------------------------------------------------ |
| 04:23 | <@gnolam> | r35 | gnolam | 2008-11-25 19:05:06 +0100 (Tue, 25 Nov 2008) | 1 line |
| 04:23 | <@GeekSoldier> | nice. |
| 04:23 | <@GeekSoldier> | I bet your other committers love those. |
| 04:25 | | gnolam [lenin@79.136.60.ns-4387] has quit [Quit: Z?] |
| 04:31 | | Consul [~consul@Nightstar-5023.dsl.sfldmi.ameritech.net] has joined #code |
| 04:31 | | mode/#code [+o Consul] by ChanServ |
| 04:47 | | Syloqs-AFH [~Syloq@ServicesAdmin.Nightstar.Net] has quit [Connection reset by peer] |
| 04:48 | | Syloqs_AFH [~Syloq@Admin.Nightstar.Net] has joined #code |
| 04:49 | | Syloqs_AFH is now known as Syloqs-AFH |
| 04:52 | <@Consul> | Okay, I have got to get SELinux sorted out. The Flash plugin is throwing thousands of errors. |
| 04:55 | | Consul [~consul@Nightstar-5023.dsl.sfldmi.ameritech.net] has quit [Quit: Leaving] |
| 04:59 | | Serah is now known as Mr_Wizard |
| 05:02 | | Consul [~consul@Nightstar-5023.dsl.sfldmi.ameritech.net] has joined #code |
| 05:02 | | mode/#code [+o Consul] by ChanServ |
| 05:03 | <@Consul> | Hrm, looks like that did the trick. |
| 05:09 | <@MyCatVerbs> | GeekSoldier: at least he's not filling their hovercraft with eels. |
| 05:10 | <@Consul> | Oh man, I thought for sure the size of my music collection would crash Rhythmbox. |
| 05:56 | | * ToxicFrog hugs JEdit's project manager plugin |
| 05:56 | <@ToxicFrog> | Consul: how big? |
| 06:20 | | * jerith has about 50k widely scattered tracks. |
| 06:20 | <@jerith> | Took forever to import, but ran fine thereafter. |
| 06:26 | <@ToxicFrog> | 17k, 60GB-ish here. Likewise. |
| 06:45 | | * ToxicFrog gnaws with irritation on the window control API |
| 06:51 | | Vornicus is now known as Vornicus-Latens |
| 07:12 | | AnnoDomini [~farkoff@Nightstar-29646.neoplus.adsl.tpnet.pl] has joined #Code |
| 07:12 | | mode/#code [+o AnnoDomini] by ChanServ |
| 07:23 | <@ToxicFrog> | Done: window title, resolution, fullscreen toggle, lines, points, ticks, delay |
| 07:24 | <@ToxicFrog> | Next up: mouse control and a shitload of drawing primitives |
| 07:25 | <@ToxicFrog> | The mouse interface at least is easy |
| 07:25 | <@ToxicFrog> | mouse(x,y) => x,y,buttons |
| 07:26 | <@ToxicFrog> | With signature int -> int -> (int, int, bool[]) |
| 07:36 | <@ToxicFrog> | This is something I don't do often, but |
| 07:36 | | * ToxicFrog winds C's entrails out on a stick, then kicks it in the preprocessor |
| 07:37 | <@ToxicFrog> | gcc -Wall -std=c99 -g -o /tmp/slep SLEP.c && exec /tmp/slep |
| 07:52 | <@McMartin> | Consul: Did you get pulseaudio sorted? |
| 09:47 | | You're now known as TheWatcher |
| 09:58 | <@MyCatVerbs> | McMartin: does "sorted" mean "set on fire"? |
| 10:09 | <@McMartin> | It can! |
| 10:09 | <@McMartin> | It can also mean that it actually works. |
| 10:17 | | * MyCatVerbs still thinks fire is a roughly ideal solution to people who think that the ideal solution to audio layers on Linux forming a bewilderingly overcomplicated almost-acyclic graph of dependencies and pointless indirection is to add -one more- damn layer of indirection. |
| 10:22 | <@McMartin> | As long as it's The Last One? |
| 10:22 | <@McMartin> | Sound should fucking live in userland, goddamnit. |
| 10:22 | <@MyCatVerbs> | Nah. |
| 10:23 | <@MyCatVerbs> | Sound should imitate /dev/dsp, even if you have to hack things a little to make that into a magic filename that gives back a new, unique endpoint every time you open(2) it. |
| 10:23 | <@McMartin> | And /dev/dsp should live in fucking userland~ |
| 10:23 | <@McMartin> | Also, insufficient for sequencers. |
| 10:23 | <@MyCatVerbs> | I'll go with that. |
| 10:24 | <@MyCatVerbs> | Sequencers? Midi? Whatev. Put something else on top of those - another magic file that you can write MIDI streams down would be the obvious choice. |
| 10:24 | <@McMartin> | I'm not sufficiently up on The Philosophy Of Udev to know whether or not your hack is Pure Evil or not. |
| 10:24 | <@McMartin> | (Yes; /dev/sequencer, as hit happens.) |
| 10:25 | <@McMartin> | If it is - if devices are not to clone themselves when you open(2) them - then you have the raw device and then while still in user space, ONLY THE MIXER APP GETS TO TOUCH IT, and everyone else talks to it. |
| 10:25 | <@MyCatVerbs> | Either way. The point is that /dev/dsp gives you OSS compatibility, which means that you're suddenly compatible with, oh, Solaris, all of the BSDs... instead of creating yet another Linux-only layer. |
| 10:26 | <@MyCatVerbs> | If you make it magic, you get mixing for "free", at least from an API perspective. |
| 10:26 | <@McMartin> | Right, but if you make it magic and the magic isn't everywhere, you can no longer detect when the magic is there or not, which is Bad (tm) |
| 10:26 | <@McMartin> | This may be a matter they've already considered, and if it's a mixer app that can talk to /dev/dsp, it would seem that everybody wins |
| 10:27 | <@MyCatVerbs> | If that mixer app is going to be portable by virtue of talking to /dev/dsp, that's close enough to fine. |
| 10:55 | | Attilla [~The.Attil@Nightstar-9469.cdif.cable.ntl.com] has quit [Ping Timeout] |
| 11:02 | | * TheWatcher eyes this DProf output... 590MB >.< |
| 11:06 | <@Alek|gone> | ow |
| 11:07 | <@Alek|gone> | that's worse than my 10MB compilation files for a 2kb source and 100b exe in C++ |
| 11:08 | <@Alek|gone> | what IS DProf? |
| 11:18 | <@TheWatcher> | The Perl profiler. |
| 11:22 | <@Alek|gone> | ok -_- |
| 11:23 | <@Consul> | McMartin: PulseAudio seems to be less retarded in the latest Fedora than in the latest Ubuntu. |
| 11:35 | <@Consul> | Which is perhaps a little like saying Sudoku is less retarded than happy-slapping, but there you go. |
| 11:37 | | * TheWatcher eyes this code, bleghs, decides that while Perl makes for easiler implementation, this bugger needs c or c++ for speed |
| 12:29 | | Attilla [~The.Attil@Nightstar-9469.cdif.cable.ntl.com] has joined #code |
| 12:29 | | mode/#code [+o Attilla] by ChanServ |
| 14:17 | | Attilla [~The.Attil@Nightstar-9469.cdif.cable.ntl.com] has quit [Connection reset by peer] |
| 14:17 | | Attilla [~The.Attil@Nightstar-9469.cdif.cable.ntl.com] has joined #code |
| 14:17 | | mode/#code [+o Attilla] by ChanServ |
| 14:24 | | gnolam [lenin@79.136.60.ns-4387] has joined #Code |
| 14:24 | | mode/#code [+o gnolam] by ChanServ |
| 14:30 | <@GeekSoldier> | I love reading about a framework and thinking "...and?" |
| 14:31 | <@GeekSoldier> | Of course, I suppose I don't have a need to serve up articles using Django. Looks pretty, though. |
| 14:46 | | Attilla [~The.Attil@Nightstar-9469.cdif.cable.ntl.com] has quit [Connection reset by peer] |
| 14:50 | | Attilla [~The.Attil@Nightstar-9469.cdif.cable.ntl.com] has joined #code |
| 14:50 | | mode/#code [+o Attilla] by ChanServ |
| 15:22 | | * jerith usually looks at frameworks and thinks "is that all?" or "are they *insane*!?" |
| 15:24 | <@GeekSoldier> | Many of them evoke a "battery-powered hammer" response from me. |
| 15:31 | <@Consul> | Query: One of Faust's claims (it's a functional language that compiles to C) is that it can write C code that is better optimized than *most* humans can do it. Delusions of grandeur, or entirely possible? |
| 15:35 | <@jerith> | Consul: Most modern optimising compilers do it better than humans can. |
| 15:36 | <@Consul> | There's a new group of plugin makers for Linux audio, and I've been trying to talk to them about combining forces with some of the stuff I've been doing. |
| 15:37 | <@Consul> | But once I got to use Faust for DSP blocks in the back end, they froze up, insisting that "coding is not supposed to be easy" as if somehow I could poison their codebase. |
| 15:37 | <@Consul> | So I think I'm better off continuing to fly solo. |
| 15:38 | <@jerith> | "coding is not supposed to be easy"!? |
| 15:38 | <@Consul> | He started going off on the demo scene. |
| 15:38 | <@jerith> | It generally isn't, but anything that makes it easier is a good thing. |
| 15:38 | <@Consul> | And how the purists in the demo scene resented the ones that used "demo construction kits", etc. |
| 15:38 | <@Consul> | I accepted it as a polite "go away". |
| 15:39 | <@Consul> | The thing is, they both admitted that neither one is skilled with the math behind DSP. |
| 15:39 | <@Consul> | Not that I'm any kind of an expert, but I think I can knock together decent stuff. I like Faust because it looks like math. |
| 15:39 | <@Consul> | Rather than code. |
| 15:40 | <@jerith> | The basics of DSP maths are actually pretty simple. |
| 15:40 | | Syloqs-AFH [~Syloq@ServicesAdmin.Nightstar.Net] has quit [Ping Timeout] |
| 15:40 | <@jerith> | The hard part is figuring out what you need to do to the signal. |
| 15:41 | <@Consul> | The foundational maths are not difficult, sure. |
| 15:41 | <@Consul> | It's when you start getting to bilinear transforms and FFTs that it gets interesting. |
| 15:42 | | Syloqs_AFH [~Syloq@Admin.Nightstar.Net] has joined #code |
| 15:42 | <@Consul> | It's easy to throw together an EQ routine from an online cookbook, sure. |
| 15:43 | | Syloqs_AFH is now known as Syloqs-AFH |
| 15:43 | <@Consul> | I was thinking I wanted some more ammunition for the argument, but I think I'll just let it go. |
| 15:47 | <@Consul> | In fact, I'm thinking about just writing the stuff I want and not bothering releasing it. |
| 15:47 | <@Consul> | Then it doesn't matter how efficient it is or how it looks. |
| 17:10 | | You're now known as TheWatcher[afk] |
| 17:21 | <@ToxicFrog> | <Consul> But once I got to use Faust for DSP blocks in the back end, they froze up, insisting that "coding is not supposed to be easy" as if somehow I could poison their codebase. |
| 17:21 | <@ToxicFrog> | In that case, they should man up and write all their stuff in Brainfuck or Tetrafunge. |
| 17:23 | <@gnolam> | INTERCAL. |
| 17:26 | <@ToxicFrog> | Or go all out and write that fucker in straight machine code. |
| 17:26 | <@ToxicFrog> | Ideally for a processor that doesn't exist anymore. |
| 17:27 | <@ToxicFrog> | Nothing like a program starting short int main[] = { |
| 17:53 | | gnolam is now known as gnolam|Gig |
| 18:04 | | crem [~moo@Nightstar-28703.adsl.mgts.by] has quit [Ping Timeout] |
| 18:04 | | crem [~moo@Nightstar-28703.adsl.mgts.by] has joined #code |
| 18:21 | <@Alek|gone> | bwahahaha |
| 18:21 | <@Alek|gone> | no. |
| 18:21 | <@Alek|gone> | 010111010010101001001011011 |
| 18:36 | | Vornicus-Latens is now known as Vornicus |
| 18:36 | | You're now known as TheWatcher |
| 19:55 | <@gnolam|Gig> | Binary solo! |
| 19:55 | | gnolam|Gig is now known as gnolam |
| 19:59 | <@gnolam> | AnnoDomini: on the topic of self-diagnosis: http://www.nytimes.com/2008/11/25/technology/internet/25symptoms.html?_r=1 |
| 20:13 | | Vornicus [~vorn@ServicesOp.Nightstar.Net] has quit [Quit: This computer has gone to sleep] |
| 20:39 | <@AnnoDomini> | gnolam: Cool. |
| 20:39 | <@AnnoDomini> | Today, we were learning how to literally punch better. |
| 21:09 | | Attilla [~The.Attil@Nightstar-9469.cdif.cable.ntl.com] has quit [Ping Timeout] |
| 21:13 | <@gnolam> | Knuckle positioning, recoil, that sort of thing? |
| 21:15 | <@AnnoDomini> | More like how to go from a ready position to a punch with a forward step, then follow up with the other hand, then duck, then punch again twice in reverse order. |
| 21:23 | | You're now known as TheWatcher[afk] |
| 21:48 | <@Consul> | ToxicFrog: They keep going on about the Faust team making false claims of code optimization, that a human can always do better. |
| 21:49 | <@Consul> | To tell you the truth, efficiency is not the highest on my list of priorities. |
| 21:49 | <@Consul> | Good sound is. |
| 21:49 | <@gnolam> | In other words, it's the age-old "C vs ASM" argument again. |
| 21:49 | <@Consul> | Followed by "as easy as possible to write." |
| 21:56 | <@Consul> | Like I said, I'll just continue my own mission. |
| 21:56 | <@Consul> | Does anyone know if Fedora has a package that would install basically all of OpenOffice? |
| 21:57 | <@Consul> | A metapackage, I guess one might call it. |
| 22:03 | <@Consul> | Never mind, I think I'll just install piecemeal. I don't need the DB front-end anyway. |
| 22:09 | | Vornicus [~vorn@ServicesOp.Nightstar.Net] has joined #code |
| 22:09 | | mode/#code [+o Vornicus] by ChanServ |
| 22:13 | <@ToxicFrog> | "openoffice", typically |
| 22:13 | <@ToxicFrog> | Although you can always just go: # yum install 'openoffice-*' |
| 22:29 | <@McMartin> | 13:50 <@Consul> ToxicFrog: They keep going on about the Faust team making false claims of code optimization, that a human can always do better. |
| 22:29 | <@McMartin> | This is both technically true and a pathetic lie at the same time. |
| 22:29 | <@McMartin> | It's technically true because nothing stops a human from personally executing the logic of any optimizing compiler on their code. |
| 22:30 | <@McMartin> | It's a pathetic lie because, seriously, what the shit is that. Not to mention that you'd be having to do it for every target platform |
| 22:30 | <@McMartin> | And that include "different varieties of x86". |
| 22:30 | <@jerith> | Although a human sometimes /can/ do better by grokking the code and choosing a different algorithm. |
| 22:30 | <@jerith> | Or knowing more about the kinds of data it'll see. |
| 22:30 | <@McMartin> | Once you're doing that, you are no longer optimizing. |
| 22:31 | <@McMartin> | On the former case |
| 22:31 | <@jerith> | But you generally do that anyway, at a higher level. |
| 22:31 | | * Consul munches his turkey salad sandwich. :-) |
| 22:31 | <@McMartin> | On the latter, it's either actually the former, or it's the kind of thing where you *really* want an adaptive JIT. |
| 22:31 | <@jerith> | Quite. |
| 22:32 | <@jerith> | For sufficiently small data, a JIT can cost more than it gains. |
| 22:32 | <@McMartin> | For sufficiently small data, efficiency is by definition not a concern~ |
| 22:32 | <@jerith> | But in that case, the difference is meaningless. |
| 22:33 | <@McMartin> | Demoscene sensibilities break down hard in the modern era. |
| 22:33 | <@McMartin> | My personal favorite of this came out of UQM. |
| 22:33 | <@McMartin> | It's actually more computationally expensive - by, like, orders of magnitude - to use the old x86 soundtrack than it is to do OGG decoding. |
| 22:33 | <@McMartin> | ER |
| 22:34 | <@Consul> | The whole idea in the demoscene was to push hardware to its limits. That's no longer a strict necessity unless you're doing a cutting-edge game (something some of the folks here are doing, I understand). |
| 22:34 | <@McMartin> | The old ProTracker soundtrack. |
| 22:34 | <@McMartin> | Consul: Right, but there are other bits that are kind of 'hanging on', with some limit pushing that comes later |
| 22:34 | <@ToxicFrog> | AIUI, in scene you use mods and similar not so that it costs less at runtime, but so that you can pack a 3d raytracer with particle effects and five distinct audio tracks into 50K. |
| 22:35 | <@jerith> | There /are/ places where that kind of thing is a feature. |
| 22:35 | | You're now known as TheWatcher |
| 22:35 | <@McMartin> | So OGG is straight decompression, which is easy, but ProTracker playback has evolved from "hand wavetable data to a chip" into a cubic interpolator, which is insanely expensive. |
| 22:35 | <@jerith> | Embedded stuff, etc. |
| 22:35 | <@McMartin> | Aye. |
| 22:36 | <@jerith> | Ugh. That might be my car alarm. |
| 22:40 | <@jerith> | Nope, not mine. |
| 22:41 | <@jerith> | Although I did get to run down and up four flights of stairs in my boxers. |
| 22:41 | <@jerith> | (A sight to terrify small children, that is.) |
| 22:41 | <@Consul> | Always a win. |
| 22:42 | <@McMartin> | It makes me sad that I can't find the spandex turtlenecks I like anymore. |
| 22:43 | <@jerith> | Last time I visited my mother's classroom (she teaches three year olds), two of them burst into tears and a third attached herself to my leg for the duration of my visit. |
| 22:43 | <@Consul> | McMartin: Oh, now if that's not a sig-worthy sentence, nothing is. |
| 22:43 | <@jerith> | I apparently manage to hit both extremes simultaneously. |
| 22:43 | <@McMartin> | It really isn't, though. I'm kind of spindly, so having cuffs that Always Fit is really, really nice. |
| 22:44 | <@Consul> | I'm so fat I have my own zip code. |
| 22:44 | <@Consul> | No spandex for me. |
| 22:44 | | * jerith is sad about his uniform. |
| 22:44 | <@McMartin> | Well, the actual bulk of the outfit is cotton. |
| 22:44 | <@jerith> | Having sleeves that actually reach more than halfway down my forearm would be nice. |
| 22:44 | <@McMartin> | However, normally this means I can get whipcrack sounds from the sleeves just by jogging. |
| 22:45 | <@McMartin> | So it's nice to not have that. |
| 22:45 | <@jerith> | Also, breathing would be handy. |
| 22:45 | <@jerith> | And it is now nearly 1am, which is /way/ past my bedtime. |
| 22:45 | <@jerith> | 'Night all. |
| 22:47 | <@Consul> | Night, jer |
| 23:09 | | Janus [~clairvoir@Nightstar-9198.vdsl.bright.net] has joined #code |
| --- Log closed Sat Nov 29 00:00:49 2008 |