00:47:18 <elliott> what are the ways that have changed
00:47:29 <elliott> a disagreement between devs and the company
00:48:02 <elliott> involving companies in free OSes is always a bad idea i think.
00:48:14 <catseye> fuckin' crampin' our style
00:48:23 <elliott> catseye: oh shaddup it's true
00:48:37 <elliott> unless you were being serious
00:49:00 <catseye> elliott: neither serious nor in jest really
00:49:18 <elliott> Fuckin' rainbows, how do they work?
00:49:20 <Gregor> Though he is having sex with rainbows.
00:50:02 <elliott> "We got a theory" -- Insane Clown Posse, at the beginning
00:50:12 <elliott> "I HATE SCIENTISTS" -- Insane Clown Posse, near the end, paraphrased
00:52:04 <catseye> it IS true, but if you want to get *paid* for doing foss? guess what
00:52:13 <catseye> you must interface with the economy
00:52:44 <elliott> catseye: fuck getting paid :D
00:55:04 <catseye> anyway i liked that quote about the ibex
00:57:14 <catseye> Phantom_Hoover | GreaseMonkey, OK, so not being able to accurately store £0.10 is a tolerable error?
00:57:33 <catseye> if you are a large corporation with a valuation in the millions -- yes, totally.
00:57:45 <catseye> who gives a fuck about a few pence
00:57:56 <catseye> or whatever you call those things after decimalisation
00:58:02 <catseye> i think they're still pence...
00:58:34 <elliott> Peter Gibbons: [Explaining the plan] Alright so when the sub routine compounds the interest is uses all these extra decimal places that just get rounded off. So we simplified the whole thing, we rounded them all down, drop the remainder into an account we opened.
00:58:34 <elliott> Joanna: [Confused] So you're stealing?
00:58:34 <elliott> Peter Gibbons: Ah no, you don't understand. It's very complicated. It's uh it's aggregate, so I'm talking about fractions of a penny here. And over time they add up to a lot.
00:58:48 <elliott> catseye: Yes, they are still pence.
00:58:51 <elliott> Joanna: Oh okay. So you're gonna be making a lot of money, right?
00:58:52 <elliott> Joanna: Right. It's not yours?
00:58:52 <elliott> Peter Gibbons: Well it becomes ours.
00:58:52 <elliott> Joanna: How is that not stealing?
00:58:53 <elliott> Peter Gibbons: [pauses] I don't think I'm explaining this very well.
00:58:56 <elliott> Peter Gibbons: Um... the 7-11. You take a penny from the tray, right?
00:58:58 <elliott> Joanna: From the cripple children?
00:59:00 <elliott> Peter Gibbons: No that's the jar. I'm talking about the tray. You know the pennies that are for everybody?
00:59:04 <elliott> Joanna: Oh for everybody. Okay.
00:59:06 <elliott> Peter Gibbons: Well those are whole pennies, right? I'm just talking about fractions of a penny here. But we do it from a much bigger tray and we do it a couple a million times.
01:00:27 <elliott> I wonder if catseye has actually seen Office Space.
01:02:13 <catseye> elliott: I have but it was a long time ago.
01:02:30 <elliott> catseye: Yes, all those 11 -- almost 12 -- years ago, in 1999.
01:02:35 <elliott> [catseye feels UNSPEAKABLY OLD]
01:04:15 <catseye> also, if you are a large corporation: none of the developers you hire will be able to understand why you would not store currency in a float anyway. If you are lucky they will "understand" why it should be a double.
01:04:33 <elliott> catseye: dude, this place is a sanctuary. no talking about corporate development
01:04:51 <elliott> well okay but not all the time
01:04:57 <elliott> otherwise we'll just get #corporate-dev-sucks :P
01:07:10 <catseye> i got nothing. so don't let me stop you
01:07:27 <elliott> pikhq: You. Figure out a way to get a RAM-based filesystem to run a system from into RAM quickly.
01:07:56 <catseye> I think "John Freeman turned on off the computer" would make an awesome quit message.
01:08:00 <catseye> Hey, I did have something!
01:08:15 <elliott> catseye: But... but I thought it was stupid! :p
01:08:27 <elliott> Anyway it can't beat DEMOCRATIC PEER-REVIEW-BASED IRC DISCONNECTION.
01:09:03 -!- ais523 has quit (Remote host closed the connection).
01:09:20 <Gregor> As demonstrated above.
01:09:48 <elliott> Gregor: No, as demonstrated by this!
01:09:56 <elliott> (As a last final note, I ought to watch Idiocracy sometime.)
01:10:04 -!- elliott has quit (Quit: Connection reset by peer review).
01:11:12 -!- dbc has joined.
01:22:33 -!- catseye has quit (Quit: WeeChat 0.2.6).
01:26:23 -!- catseye has joined.
01:32:04 -!- augur has quit (Read error: Connection reset by peer).
01:32:27 -!- augur has joined.
01:53:26 <Gregor> adobe_acrobat_reader is a product that is a kind of map
01:53:32 <Gregor> president_elect is a U.S. person
01:54:02 <Gregor> Whoops, copied the wrong one, that one is true X-P
01:54:07 <Gregor> shanghai is the capital city of the country china
01:54:15 <Gregor> outdoor_air is a weather phenomenon
01:54:20 <Gregor> (Is that true, or just nonsense?)
01:54:58 <Gregor> outdoor_air is a weather phenomenon
01:55:07 <Gregor> john is a musician who is part of ben_folds
01:55:10 -!- augur has quit (Ping timeout: 240 seconds).
01:55:36 <Gregor> microsoft_sql_server is a consumer electronic device
01:56:22 <Gregor> envelope_llc_intelligent_nutrients is an office supply
01:56:30 <Gregor> barack is a politician who holds the office of secretary
02:13:35 -!- augur has joined.
02:18:08 -!- Mathnerd314 has quit (Ping timeout: 276 seconds).
02:19:16 -!- augur has quit (Ping timeout: 252 seconds).
02:28:06 -!- Mathnerd314 has joined.
02:28:45 <Quadlex> I'm about to call a loan shark about a job they are offering
02:31:43 <catseye> I almost took a job at a usurer last year :)
02:32:24 -!- augur has joined.
02:35:32 <catseye> Fancy word for loan shark.
02:36:32 <Quadlex> And they actually seem really cool
02:36:38 <Quadlex> http://www.cashdoctors.com.au
02:38:59 <catseye> My offer was kind of like that, too... programmers had their own offices, the pay was good, etc.
02:39:21 <catseye> I ended up going to a different place because their health insurance was kind of crap, though.
02:40:07 <Quadlex> We're socialist nazi commies
02:40:08 <catseye> I'm from Canada originally, but in this awful place called the United States right nw...
02:44:10 unexpected log event :(
04:46:56 <pikhq> So far, I've got to say Debian is less of a complete and utter pain than Gentoo.
04:47:55 <Sgeo> Why does GTK+ hate me using my trackpad to scroll?
04:48:45 <pikhq> ... And seems to perform better.
04:49:01 <pikhq> From inside of a VM.
04:53:05 -!- Sgeo has quit (Read error: Connection reset by peer).
04:53:52 -!- augur has quit (Ping timeout: 245 seconds).
04:59:32 <catseye> fungot: YOU APPEAR TO HAVE DIEDED
04:59:54 <catseye> fizzie: FUNGOT HAS APPEARED TO HAVE DIEDEDED
05:05:30 -!- Sgeo has joined.
05:07:47 * Sgeo allows Symantec to downgrade his beautiful new computer :(
05:13:07 <Gregor> calamari: Well, you only have two ears. With some physical simulation, should be totally feasible.
05:13:27 <Gregor> calamari: Reminds me of that barbershop recording ...
05:16:08 <pikhq> AAAARGH Debian (and clones) has one major, major breaker for me. Multiarch support is not quite finished.
05:17:03 <pikhq> calamari: Mplayer's "hrtf" plugin does that.
05:17:10 <Gregor> Yeah, multiarch sucks arse on Debian.
05:17:18 <pikhq> (abbreviation of "head-related transfer function", the method it uses for that)
05:17:20 <Gregor> I actually have a debootstrap-generated i386 chroot.
05:17:38 <pikhq> Gregor: Apparently they fell *just* short of getting multiarch into squeeze.
05:18:00 <pikhq> And here I thought it had the best chance of having decent multiarch.
05:18:14 <pikhq> Why is Gentoo's multiarch support *still* the only thing that even halfway works?
05:18:20 <Gregor> sid doesn't have useful multiarch.
05:18:28 <Gregor> pikhq: Because multiarch is (relatively) easy when you build everything.
05:18:40 <pikhq> Gregor: Except it's not even doing that.
05:18:50 <pikhq> Gregor: It's... Just got a handful of tarballs of x86_32 libraries.
05:19:07 <Gregor> Uhhh, that != multiarch support :P
05:19:17 <Gregor> Presumably Gentoo can build 32-bit libraries on x86_64?
05:19:31 <pikhq> Yeah, but it only does so for libc, IIRC.
05:20:10 <pikhq> The thing is, it also has packages for 32-bit only programs. Said programs on x86_64 depend on the tarballs instead of the actual libraries.
05:20:19 <pikhq> It's a damned hack, but *it fucking works*.
05:20:48 <Gregor> Debian has everything up to GTK+ for x86_32 on x86_64
05:21:15 <pikhq> Yeah, Gentoo's got... Basically all the dependencies for programs without x86_64 versions.
05:21:18 <Gregor> Plus apparently it has an installable x86_32 Java for some unimaginable reason.
05:21:41 * Sgeo is using a Toshiba Satellite T215
05:21:42 <pikhq> There's no official x86_64 Java binaries.
05:21:53 <pikhq> Builds just fine, though.
05:22:00 <Gregor> Ohhhhh, yeah, these are the sun packages, yeah.
05:22:07 <Sgeo> Oh, it's not considered a netbook
05:22:16 <Gregor> I of course have OpenJDK :P
05:22:30 -!- augur has joined.
05:22:50 <pikhq> Well, yeah. It's just a bit less of a pain to deal with.
05:23:30 <Gregor> I hate to say it, but Mandriva has great multiarch support :P
05:24:04 <pikhq> I have seen the dark side of its multiarch support.
05:24:31 <pikhq> RPM has far far too many limitations.
05:24:40 <Gregor> Oh, RPM is garbagesauce on rye.
05:24:42 <Sgeo> ATI Integrated Radeon stuff > Intel integrated?
05:24:55 <Gregor> Sgeo: Hard to imagine it wouldn't be.
05:25:02 <Gregor> Sgeo: Unless you want it to work on Linux.
05:25:13 <pikhq> Sgeo: Unless power usage is your consideration, yes.
05:25:34 <Sgeo> Now, why does SL hate this thing?
05:25:34 <pikhq> Gregor: Official drivers should just work, the X11 drivers should just have somewhat shitty OpenGL performance.
05:25:46 <Gregor> pikhq: That's what I mean by "work" :P
05:25:59 <Gregor> I'm not convinced that ATI+Linux will even outperform Intel+Linux.
05:25:59 <pikhq> Gregor: No, I mean, the ones that are *built into X11*.
05:26:14 <Sgeo> Figured out how to stop XChat from minimizing to tray
05:26:44 * Gregor continues to munch on infant bones while watching the conversation unfold.
05:26:52 <pikhq> Nvidia+Linux is definitely better than ATI+Linux, performance-wise.
05:27:25 * Sgeo tries SL on lowest settings
05:30:44 * Sgeo angrily blibbers at the "32-bit only" thing in SL's requirements
05:32:15 <pikhq> You can fix that by building it.
05:38:05 <Gregor> http://upload.wikimedia.org/wikipedia/commons/5/5a/Orchestra_layout.svg This orchestra layout lists both "double basses" and "contrebasses". ... fail.
05:38:07 -!- antivigilante has joined.
05:45:42 <Sgeo> That worked well
05:46:09 <catseye> Gregor: there should also be separate sections for baritones and euphoniums
05:46:24 <Gregor> catseye: Considering that the arrangement is almost entirely wrong, it's sort of irrelevant.
05:46:26 <catseye> (baritone horn, of course)
05:47:10 <catseye> Not that those ever get into an orchestra.
05:47:27 <Gregor> Screw your euphonium :P
05:51:55 <catseye> Gregor: Perhaps there should be a seating arrangement for the coronets.
06:09:35 -!- antivigilante has quit (Ping timeout: 252 seconds).
06:14:56 -!- realazthat has joined.
06:34:53 <coppro> can you do an upgrade without breaking my sound?
06:37:08 <Sgeo> VLC comes with FluidSynth stuff?
06:37:12 <Sgeo> I can just use a soundfont?
06:37:34 <Quadlex> Real Ubuntu users are happy with brown wallpaper and silence
06:37:35 <Sgeo> And unicorns will take flight?
06:38:09 <Quadlex> Unicorns don't need to fly to get around
06:38:25 <Sgeo> Like Omnipresent Man!
06:38:49 <Quadlex> Yeah, he's Omnipresent, but does the bastard ever do anything?
06:38:56 <Quadlex> No, he's too busy with his hand on his dick
06:47:02 -!- antivigilante has joined.
06:48:01 <Vorpal> <pikhq> So, looking at seeing if/how Debian's made an abomination of things, being seduced by Slackware, aaand... I think I'll try random things from there. <-- arch is pretty good IMO
06:50:11 <Vorpal> <pikhq> Nvidia+Linux is definitely better than ATI+Linux, performance-wise. <-- last I checked, also stability-wise.
06:50:15 <Vorpal> that was a while ago though
06:50:27 <Vorpal> though intel graphics is probably most stable
06:50:36 -!- quintopia has quit (Ping timeout: 265 seconds).
06:50:36 <Vorpal> worst when it comes to performance though
06:51:05 -!- quintopia has joined.
06:52:54 <Vorpal> <coppro> can you do an upgrade without breaking my sound? <-- that is why you stay on LTS if using ubuntu
06:56:06 -!- antivigilante_ has joined.
06:56:28 -!- antivigilante has quit (Ping timeout: 252 seconds).
06:56:47 -!- antivigilante_ has changed nick to antivigilante.
06:58:53 <Quadlex> Vorpal: you've got some borogrove on your face
07:00:32 <Vorpal> Quadlex, new here, nick-changer, previously idler or just someone who been away for some time?
07:02:13 <Vorpal> meh, don't have time to wait for answer, need to leave or I'll miss the bus
07:03:59 -!- antivigilante has quit (Ping timeout: 260 seconds).
07:04:43 -!- tombom has joined.
07:12:12 -!- Sgeo has quit (Ping timeout: 245 seconds).
07:30:50 -!- Sgeo has joined.
07:39:12 <fizzie> Re "brown wallpaper", also, the Meerkat wallpaper is pretty colorful and un-brown.
07:39:56 <fizzie> http://img3.imagebanana.com/img/x34gp6re/wartyfinalubuntu.jpg from a random google-image hit.
07:46:14 -!- Sgeo has quit (Read error: Connection reset by peer).
07:48:21 <olsner> if you don't have a word for purple, that *is* brown
07:49:58 <fizzie> And I guess it's not so dissimilar in color from the not-quite-as-brown Lucid wallpaper -- http://www.kilobitspersecond.com/stuff/warty-final-ubuntu.jpg -- too.
07:50:21 -!- tombom has quit (Quit: Leaving).
07:50:28 -!- oerjan has joined.
07:50:55 <fizzie> Karmic and Jaunty were pretty brown though.
07:51:50 <fizzie> (And Intrepid. And Hardy.)
07:58:59 <calamari> http://www.youtube.com/watch?v=LGkkyKZVzug&feature=player_embedded#!
07:59:59 -!- clog has quit (ended).
08:00:00 -!- clog has joined.
08:00:31 <fizzie> You people keep saying the same thing; that link was posted by alise in 2010-09-28 and nooga in 2010-09-29. And I don't think clog was even down then, so no excuse at all!
08:01:19 <fizzie> Speaking of excuses, was I supposed to do something to fungot?
08:24:34 -!- calamari has quit (Quit: Leaving).
08:24:49 <oerjan> ...and xkcd is back to just weird again.
08:29:37 <cheater99> linux aqualung with "metal" skin = awesome
09:09:39 -!- oerjan has quit (Quit: Later).
09:41:28 -!- nooga has joined.
10:06:12 -!- jcp has quit (Ping timeout: 240 seconds).
10:22:10 -!- jcp has joined.
10:49:17 -!- jcp has quit (Ping timeout: 245 seconds).
11:00:11 -!- jcp has joined.
11:30:17 -!- jcp has quit (Read error: Connection reset by peer).
11:33:49 -!- myndzi has joined.
11:36:23 -!- myndzi\ has quit (Ping timeout: 240 seconds).
11:38:40 -!- jcp has joined.
11:45:05 -!- jcp has quit (Ping timeout: 276 seconds).
11:51:41 -!- jcp has joined.
12:04:34 -!- jcp has quit (Read error: Connection reset by peer).
12:11:40 -!- jcp has joined.
12:21:40 -!- jcp has quit (Read error: Connection reset by peer).
12:29:11 -!- jcp has joined.
12:36:40 -!- sftp has joined.
13:29:53 -!- jcp has quit (Read error: Connection reset by peer).
13:40:40 -!- SgeoN1 has joined.
13:41:20 <SgeoN1> I just realized something horrible about Flatterland
13:42:36 <SgeoN1> It promotes the pseudoscientific bullshit involving taking scientific words, interpreting stuff as you please, and thinking it means something
13:43:10 -!- jcp has joined.
13:44:07 <SgeoN1> I feel no guilt about spoiling this: the last scene involves the protagonist basing arguments for gender equality on Flatland's supersymmetry
13:45:29 <SgeoN1> "Supersymmetric sister! ..."
13:46:03 <SgeoN1> I forgot the rest of her advertisement on Flatland's Internet
13:47:04 -!- jcp has quit (Read error: Connection reset by peer).
13:54:38 -!- Phantom_Hoover has joined.
13:55:11 -!- jcp has joined.
14:04:05 -!- jcp has quit (Read error: Connection reset by peer).
14:04:08 <SgeoN1> I don't want to criticise the book I loved as a kid, but I must!
14:04:37 <SgeoN1> Just like Friend Computer! Friend Computer never breaHSUDJEHDHDHDHHHU
14:05:25 -!- FireFly has joined.
14:09:13 <Phantom_Hoover> I remember you pondering why there wasn't a TV Tropes article on it a while ago...
14:09:40 -!- jcp has joined.
14:09:44 <Phantom_Hoover> And it doesn't seem to have a WP article, implying it's pretty obscure.
14:09:58 <SgeoN1> When it comes to books, I'm polygamous. Only with languages is it serial monogamy
14:11:38 <Phantom_Hoover> I thought the easy way to do calculus was from first principles...
14:13:01 <Phantom_Hoover> Although "easy" there depends entirely on one's point of view.
14:16:00 <SgeoN1> Going off irc to save battery
14:16:04 -!- SgeoN1 has quit (Quit: Bye).
14:17:06 <Phantom_Hoover> Heh: http://www.guardian.co.uk/world/2010/oct/11/chilean-miners-rival-churches-tussle
14:53:52 -!- augur has quit (Ping timeout: 245 seconds).
14:54:25 -!- elliott has joined.
14:55:02 <elliott> quintopia: A Golly ruleset sounds interesting enough!
14:55:17 -!- jcp has quit (Read error: Connection reset by peer).
14:56:11 <elliott> 17:16:34 <catseye> Gnome desktop.
14:56:12 <elliott> 17:16:37 <catseye> Trashcan.
14:56:46 <elliott> 17:53:26 <Gregor> adobe_acrobat_reader is a product that is a kind of map
14:57:39 <elliott> 20:46:56 <pikhq> So far, I've got to say Debian is less of a complete and utter pain than Gentoo.
14:57:45 <elliott> pikhq: stable, testing, sid?
15:00:43 <elliott> 22:37:34 <Quadlex> Real Ubuntu users are happy with brown wallpaper and silence
15:01:04 <elliott> 22:48:01 <Vorpal> <pikhq> So, looking at seeing if/how Debian's made an abomination of things, being seduced by Slackware, aaand... I think I'll try random things from there. <-- arch is pretty good IMO
15:01:10 -!- jcp has joined.
15:01:15 <elliott> It really isn't and pikhq is smart enough to have realised that already.
15:12:18 -!- jcp has quit (Read error: Connection reset by peer).
15:14:46 -!- Phantom_Hoover has quit (Ping timeout: 252 seconds).
15:22:41 -!- jcp has joined.
15:29:29 -!- jcp has quit (Read error: Connection reset by peer).
15:32:08 <fizzie> Purple is the new brown, eh?
15:32:48 -!- cpressey has joined.
15:35:19 <cpressey> < ais523> I'm amused at the way he called me a hacker with basically no evidence
15:35:31 <cpressey> wait, he... called you a hacker out of the blue?
15:35:44 <cpressey> bah, no ais523, and clog is probably still wack
15:36:54 <elliott> cpressey: ais523 has been "collaborating" with esr to get old c-intercal versions together because knuth wants it
15:40:10 -!- jcp has joined.
15:40:18 <cpressey> elliott: yes, i remembered that. but even in that context it's a bit creepy
15:44:45 -!- Sgeo has joined.
15:46:17 <Sgeo> Grrrrr at glare
15:46:31 -!- jcp has quit (Read error: Connection reset by peer).
15:52:41 -!- jcp has joined.
15:56:17 * Sgeo falls in love with WinSCP
16:01:24 <elliott> Urgh, pyFLTK's API is ugly.
16:01:54 <Gregor> The Fast Light ToolKit
16:02:05 <Gregor> And considering that FLTK's API is ugly itself, it's unsurprising that pyFLTK's API is ugly :P
16:03:37 -!- jcp has quit (Read error: Connection reset by peer).
16:04:20 <elliott> Gregor: Well, it's not that ugly.
16:04:52 <elliott> Gregor: If they just chopped the Fl_ prefixes off, and changed foo.prop(x) to "foo.prop = x" and foo.prop() into foo.prop, it'd be almost perfect.
16:04:59 <elliott> Change labelsize and the like into label_size and it'd be perfect.
16:05:55 <elliott> Oh, and making the constructor args be e.g.
16:06:03 <elliott> Button((110, 130), (100, 35), 'Okay')
16:06:08 <elliott> Button(110, 130, 100, 35, 'Okay')
16:06:12 <elliott> would be a nice finishing touch.
16:09:52 <Gregor> elliott: So, if they changed almost every aspect of the API, it'd be ALMOST perfect :
16:10:04 <elliott> Gregor: Well, firstly, those are all *Python* changes.
16:10:12 <elliott> I wouldn't change the C++ API.
16:10:39 <elliott> Gregor: A sed one the names, wrapping property-functions as properties -- not hard -- and unpacking a tuple in contructors.
16:10:52 <elliott> (Then, as a super-optional thing, maybe s/foobar/foo_bar/.)
16:11:09 <elliott> Not exactly major API changes, just aesthetics.
16:11:10 -!- jcp has joined.
16:11:18 <elliott> Gregor: Name something with a nicer API.
16:11:36 * Sgeo attempts to get used to this keyboard
16:14:58 <Gregor> elliott: Like all languages, all APIs suck.
16:15:07 <elliott> Gregor: Name one that sucks less.
16:15:37 <Gregor> Hell, even STL sucks less.
16:15:47 <Gregor> I don't use enough C++ to make relevant comparisons though :P
16:15:52 <Gregor> And that's suck a ridiculous challenge.
16:15:58 <Gregor> "Name an API that sucks less than <x>"
16:17:25 <elliott> Gregor: name a GUI API that sucks less than FLTK
16:18:26 <Gregor> The only GUI APIs I've used are FLTK and GTK+, and GTK+ sure as hell ain't it.
16:18:41 <elliott> Gregor: Lemme put it this way.
16:18:50 <elliott> If you think FLTK has a terrible API, you haven't seen every other GUI API.
16:19:06 <Gregor> Because GUIs are for pussies.
16:19:08 <elliott> Tk's API is better, sure. Hell, so is Shoes', although nobody uses that for serious stuff.
16:19:12 <elliott> AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
16:20:38 -!- jcp has quit (Read error: Connection reset by peer).
16:20:55 <Sgeo> Are delete and exists magical?
16:28:40 -!- jcp has joined.
16:31:30 <elliott> Gregor: Now you've got me mocking up half-code of a non-shitty GUI API...
16:31:59 * Sgeo wonders why, exactly, this thing came with a 64-bit version of Win7
16:32:20 <Sgeo> It only has 2gbs, are there any relevant advantages to 64-bit considering that?
16:32:29 <cpressey> curses, you must use curses so that your app works on all terminals. oh you want to colorize your bash prompt? here's some vt100 escape codes, enjoy
16:34:05 <cpressey> Sgeo: actually I have no idea what you mean. Syntactically they're slightly magical, because they take a whole hash ref as argument, but don't evaluate it. Semantically, the hash can be tied, so other things can happen when you delete or exsists. But that is, afaik, all.
16:34:39 * Sgeo meant the syntactical bit
16:34:53 <Sgeo> That is, in fact, magic that I can't reproduce in my own functions?
16:35:31 <elliott> Gregor: http://pastie.org/pastes/1218376/text?key=jzpmwvfrsp1qg51leacw
16:35:39 <elliott> Gregor: Look, it doesn't suck!
16:35:57 <elliott> Gregor: (_ is "arrange horizontally", | vertically)
16:37:40 -!- jcp has quit (Read error: Connection reset by peer).
16:38:37 <elliott> Actually, that switch's canecl should actually be
16:41:28 <elliott> Gregor: Psht, so unappreciative.
16:41:46 <elliott> Actually it should be case $cancel \cancel to cancel whatever event we're handling.
16:41:56 <elliott> Which implies to me that $cancel should have a default action of cancelling.
16:42:07 <elliott> Making the case perhaps unnecessary.
16:42:26 <elliott> cpressey: http://pastie.org/pastes/1218391/text?key=p2lt6cdvyvt1bgy7u9pj2q
16:44:11 -!- jcp has joined.
16:47:23 <cpressey> Sgeo: Not by any way I'm aware of, but my awareness essentially ends around 5.6 -- after that it's all a blur
16:47:31 <elliott> <elliott> cpressey: http://pastie.org/pastes/1218391/text?key=p2lt6cdvyvt1bgy7u9pj2q
16:47:31 <elliott> <elliott> cpressey: ADMIRE
16:47:49 <cpressey> elliott: attack of the mutant yamls.
16:48:16 <elliott> cpressey: As I said, _ means horizontal arrangement, and | means vertical
16:48:19 <elliott> The rest is somewhat obvious
16:48:32 <elliott> modified: [text contents != pages .(page_list selected) text]
16:48:32 <elliott> page_text: [pages .(page_list selected) text]
16:48:41 <elliott> page_text: [pages .(page_list selected) text]
16:48:42 <elliott> modified: [text contents != page_text]
16:49:32 <elliott> http://pastie.org/pastes/1218407/text?key=rvfz2k1dho97nel635zhg
16:54:51 -!- jcp has quit (Read error: Connection reset by peer).
16:55:23 -!- atrapado has joined.
17:00:19 <elliott> cpressey does not even acknowledge my design POWAH
17:02:41 -!- jcp has joined.
17:04:00 <elliott> http://pastie.org/pastes/1218407/text?key=rvfz2k1dho97nel635zhg You made me do this, it's *your* responsibility to comment!
17:07:21 <elliott> Wirth's law: "Software gets slower faster than hardware gets faster."
17:07:43 <elliott> Heh, didn't realise anyone had called it "Wirth's law" before.
17:08:15 <elliott> "Wirth attributed the saying to Martin Reiser" -- thus continuing to fulfil -- what's that law's name again?
17:08:25 <elliott> Where things are never attributed to the first person to think of them, instead the second.
17:10:32 -!- Phantom_Hoover has joined.
17:11:56 -!- jcp has quit (Read error: Connection reset by peer).
17:14:03 <elliott> Phantom_Hoover: can you link me to that site that has a shitload of spaceships catalogued in various rules?
17:14:09 <elliott> and it tells you which are inverses of others, etc.
17:15:00 <elliott> Phantom_Hoover: it's not affiliated with golly
17:15:34 <elliott> http://fano.ics.uci.edu/ca/
17:16:21 <elliott> Phantom_Hoover: the inverted versions, presumably
17:16:47 <Phantom_Hoover> Even then, there are definitely more than 40 spaceships in Life,
17:17:09 <Phantom_Hoover> And it's not that dated, either: it has the Caterpillar in it.
17:18:33 <Phantom_Hoover> Would it be that hard to replace give_obnoxious_warning_about_not_being_able_to_do_something_while_generating(); with stop(); do_thing();?
17:19:31 <Phantom_Hoover> It's... tolerable, if you zoom out when using Quicklife.
17:19:41 -!- jcp has joined.
17:19:52 -!- Sgeo has quit (Ping timeout: 264 seconds).
17:21:24 <elliott> its editing is so limitedly-designed as to be useless
17:21:46 <elliott> simple operations like clearing a 50,000 x 50,000 square somehow take ten minutes
17:22:04 <elliott> okay so maybe that's more reasonable :)
17:22:18 <elliott> you can't save colours per-pattern
17:22:46 <elliott> Gregor: Wow. Tiny Core Linux has a KDE package.
17:22:57 <Gregor> Tiny Core is awesome :)
17:22:57 <elliott> Gregor: Never has KDE been in a less appropriate setting X-D
17:22:58 <Phantom_Hoover> Well, clearing a square is going to take a long time unless you're clever about it.
17:23:08 <elliott> "I totally want a ten megabyte micro-distro... with KDE 4."
17:23:11 <Gregor> It's Tiny CORE Linux, not Tiny EVERYTHING Linux
17:24:07 <elliott> Tiny Core may be cool but it's not as cool as KITTEN
17:25:59 -!- augur has joined.
17:28:06 <HackEgo> 155|<Gregor> Well yeah, but furthermore unlike, oh, say, an Apple product, you don't have to sign their "we own your sperm" license agreement to GET that SDK and the requisite libraries. ... <Gregor> pikhq: Sure, but it's the only way Apple could get a first-born-son clause into a modern licensing agreement without infringing
17:28:27 <HackEgo> 215|<fungot> Sgeo: hahaah, and i love when they announced it i dare u to press alt f4 and your house ( acts 16:31 your bible)
17:28:59 <HackEgo> 127|<Ami> Discrimination fields ACTIVATE.
17:29:01 -!- jcp has quit (Read error: Connection reset by peer).
17:29:07 <HackEgo> 63|<fizzie> The thing is just to exist
17:29:17 <HackEgo> 126|<Warrigal> Ah, vulva. <Warrigal> What is that, anyway?
17:29:21 <HackEgo> 98|<fungot> ehird: every set can be well-ordered. corollary: every set s has the same diagram used from famous program talisman with fnord windows to cascade, someone i would never capitalize " i"
17:29:24 <HackEgo> 42|<ais523> after all, what are DVD players for?
17:29:27 <HackEgo> 19|<Warrigal> "You're at that stage in your life where you're going to want to do some things in private." --my mom
17:29:34 <HackEgo> 2|<Aftran> I used computational linguistics to kill her.
17:30:09 <HackEgo> 77|<ehird> no Deewiant <Deewiant> No?! <Deewiant> I've been living a lie <ehird> yep. <Deewiant> Excuse me while I jump out of the window ->
17:30:33 <HackEgo> 6|<Keiya> I think the freemasons are actually a cover for homosexual men.
17:31:34 <HackEgo> 88|<Madelon> both of you, quit it with the f-bombs. <Madelon> kaelis: what's the matter? something censoring stuff you're interested in?
17:31:37 <HackEgo> 128|<Slereah> I can do everything a Turing machine can do, except love
17:31:39 <elliott> Phantom_Hoover: http://pastie.org/pastes/1218391/text?key=p2lt6cdvyvt1bgy7u9pj2q BEHOLD (because nobody else will)
17:32:04 <elliott> Phantom_Hoover: http://pastie.org/pastes/1218501/text?key=kgj5a2hgnlvgoqqntjkg
17:32:10 <elliott> http://pastie.org/pastes/1218501/text?key=kgj5a2hgnlvgoqqntjkg
17:32:21 <elliott> Phantom_Hoover: I complained about pyFLTK's API, Gregor complained about FLTK's API
17:32:25 <elliott> I noted that all GUI libs sucked
17:32:30 <elliott> I wrote my own fakecode for something that sucked a bit less
17:32:39 <elliott> And the previous one sucks, only look at the latest one
17:36:38 <fungot> cpressey: riastradh oh, indeed. i'd prefer to just call fnord or whatever.)
17:37:40 -!- jcp has joined.
17:41:11 <elliott> Gregor: UNFORTUNATELY KDE4-DESKTOP.TCZ DOESN'T INSTALL.
17:41:21 <elliott> Should have gone with attilaspell.tcz
17:43:15 <Gregor> elliott: Only Atilla can install it.
17:43:31 <elliott> Gregor: Totally preempted you there.
17:44:15 <elliott> IT MIGHT BE WORKING THIS TIME
17:46:07 -!- jcp has quit (Read error: Connection reset by peer).
17:48:13 -!- flippo has joined.
17:48:49 <Vorpal> elliott, what distro is that?
17:50:06 <elliott> Vorpal: "I see no reason to listen to you"r questions.
17:50:41 -!- jcp has joined.
17:53:19 <cpressey> select the function you want to call from the dropdown
17:54:03 <elliott> cpressey: Sounds like Excel.
17:54:48 -!- flippo has quit (Changing host).
17:54:48 -!- flippo has joined.
17:59:49 -!- atrapado has quit (Quit: Abandonando).
18:09:40 -!- Phantom_Hoover has quit (Ping timeout: 252 seconds).
18:14:33 <nooga> someone just told me that i'm Roy from IT Crowd
18:14:59 <nooga> actually i heard it twice
18:15:06 <nooga> from my boss and from my date
18:15:28 -!- jcp has quit (Read error: Connection reset by peer).
18:23:00 -!- tombom has joined.
18:24:11 -!- jcp has joined.
18:26:31 -!- augur has quit (Ping timeout: 240 seconds).
18:28:22 <elliott> Gregor: you're right gui development sucks
18:36:27 <Vorpal> elliott, not giving me credit for saying that several times before? Oh well *shrug*
18:37:06 <elliott> Vorpal: You can't do GUI development because of some sort of cognitive deformity. I was merely referring to the current crop of *common* toolkits.
18:37:14 <elliott> There are a few I like, but they are mostly dead projects.
18:37:14 -!- jcp has quit (Read error: Connection reset by peer).
18:38:00 <elliott> And please either start ignoring me entirely, or talk in more than passive-aggressive slights.
18:41:18 -!- augur has joined.
18:42:41 -!- jcp has joined.
18:43:55 <nooga> elliot sound okay to me
18:44:16 <nooga> what's the difference between t and tt anyway?
18:48:28 <elliott> (OK, so it's not pronounced identically, but my point is made.)
18:54:24 -!- jcp has quit (Read error: Connection reset by peer).
18:55:41 <Vorpal> <elliott> Vorpal: You can't do GUI development because of some sort of cognitive deformity. I was merely referring to the current crop of *common* toolkits. <-- hm, no, I never claimed I wasn't able to do it at all in theory
18:55:55 <Vorpal> I said I disliked it, and implicitly referred to any GUI toolkits I used
18:56:44 <Vorpal> <elliott> And please either start ignoring me entirely, or talk in more than passive-aggressive slights. <-- do you know what "the golden rule" means?
18:57:20 <elliott> I am perfectly willing to not talk to you if you stop highlighting me with questions that ask for a response.
18:57:50 -!- cpressey_ has joined.
18:57:53 <fizzie> elliott: fi:aita = en:fence, fi:aitta = en:granary/storehouse/whatever. We have these delightful meaning-changes when it comes to "t"/"tt".
18:58:01 <Vorpal> elliott, I have nothing against talking to you in general.
19:00:57 <cpressey_> Korean also pronounces t/tt (and d/dd) differently.
19:01:02 -!- cpressey has quit (Ping timeout: 264 seconds).
19:01:19 -!- cpressey_ has changed nick to cpressey.
19:01:44 -!- Mathnerd314 has quit (Ping timeout: 276 seconds).
19:03:17 <fizzie> cpressey: Also, I didn't notice anything wrong with fungot; maybe you just hit the "I will only answer to the same guy four or so times in a row" limit?
19:03:17 <fungot> fizzie: then ask for more time to demi, irc-galleria, pouet, deviantart, newsgroups etc. it would be more fun
19:03:19 -!- Mathnerd314 has joined.
19:03:40 -!- jcp has joined.
19:04:18 <Vorpal> fizzie, does the t/tt thing change the quality of the t or the letter before it (the latter would be like in Swedish)
19:08:29 <fizzie> Vorpal: I think officially what it changes is the length of the silent part in /t/. It doesn't (much) change the neighbouring letters; Finnish has, or at least is approximated reasonably well by a simple mapping from text to phonemes.
19:09:08 <Vorpal> hm what do you mean with silent part though
19:10:00 <fizzie> Record something with a t and take a look if you want.
19:11:27 -!- jcp has quit (Read error: Connection reset by peer).
19:12:05 <fizzie> "In the articulation of the stop, three phases can be distinguished:
19:12:06 <fizzie> * Catch: The airway closes so that no air can escape through the mouth (hence the name stop). With nasal stops, the air escapes through the nose.
19:12:06 <fizzie> * Hold or occlusion: The airway stays closed, causing a pressure difference to build up (hence the name occlusive).
19:12:06 <fizzie> * Release or burst: The closure is opened. In the case of plosives, the released airflow produces a sudden impulse causing an audible sound (hence the name plosive)."
19:12:09 <fizzie> The middle part, then.
19:12:21 <Vorpal> of course, double consonants can change a lot in Swedish too, but in that case it changes the quality of the preceding vowel. Like sil (en:sieve) and sill (a kind of fish, seems like I forgot what that one is called in English)
19:13:28 <elliott> Grr, I hate it when TLDs and language codes don't match.
19:13:35 <elliott> se.wikipedia.org is not Swedish.
19:14:49 <fizzie> The 'sv' code is a bit annoying; I remember not noticing it when looking at a suitable keymap in NetBSD, since it only listed them by language code. (There wasn't one for fi, but sv is usable enough.)
19:15:25 <Vorpal> elliott, yes I agree it is annoying
19:15:26 <elliott> It would be nice to have a translate thing here that uses interwiki.
19:15:46 <Vorpal> elliott, I always have to think if I want se or sv when I select keymap on a livecd
19:15:48 <elliott> <HackEgo> Atlantic herring
19:16:02 <elliott> <HackEgo> Whoops no article lol
19:16:06 <Vorpal> elliott, herring sounds about right
19:16:22 <elliott> <HackEgo> ERROR REFUSE TO SHOW NYNORK BECAUSE AM BIGOT BOT
19:16:30 <elliott> nynork, what a spelling error
19:16:31 <Vorpal> elliott, what really? :D
19:16:52 <fizzie> Nynork sounds like some sort of a D&D monster.
19:16:53 <elliott> interestingly, the Swedish wiki links to the Nynorsk article for herring
19:17:02 <Gregor> tr has been borkleborked for a while.
19:17:05 <elliott> and that one links to the Bokm\oal one
19:17:11 <Gregor> IGNORE MY PAST THREE LINES.
19:17:21 <elliott> the Swedish one doesn't link to the bokmal one directly
19:17:29 <Vorpal> elliott, wait, tr is a bit bad name for it
19:17:34 <elliott> Vorpal: Yes yes, I didn't think.
19:17:43 <elliott> Instead ponder why interwiki is silly.
19:17:56 <elliott> What kind of interwiki doesn't catch second-level interwikis?
19:17:59 <Vorpal> elliott, trans or something like that might work (unless that is taken too, in which case I guess I didn't think)
19:18:02 <elliott> (i.e. "this listed interwiki links to this other interwiki")
19:18:09 <Vorpal> [think that is was] that is ;)
19:18:12 <elliott> Vorpal: maybe "iw" for interwiki
19:18:55 <Vorpal> elliott, iw says here: "Usage: iw [options] command" but then I wouldn't expect it on a server, since it deals with 802.11 stuff
19:19:11 <elliott> It goes in ~/bin, so who cares.
19:19:30 <Vorpal> elliott, tr is the kind of thing shell scripts in ~/bin might use
19:19:38 <Vorpal> but yeah, iw would work very well
19:19:43 <cpressey> Nynork nynork, it's a helluva town
19:19:58 <elliott> "<query-continue>" fuck you i want the whole query
19:19:58 <quintopia> cpressey thinks the way i do, apparently
19:20:41 -!- jcp has joined.
19:20:48 <elliott> olsner: I actually meant nynorsk
19:21:05 <quintopia> when you're blue and you don't know where to go to, why don't you go where fashion sits?
19:22:39 <elliott> http://redhatinbluesea.files.wordpress.com/2010/03/j4lup.jpeg
19:23:25 <elliott> Vorpal: Or even "interwiki" + "translate" = intertrans! The most vague thing ever.
19:23:35 <elliott> Except having connotations of an intersexed, transgender person.
19:23:42 <Vorpal> elliott, true, a bit tedious to type though
19:23:58 <Vorpal> I mean, there are some virtues in a short name
19:24:01 <elliott> "11 am? I'm afraid you're wikilate. You're wikifired."
19:24:32 <elliott> Vorpal: So do you have an wl(1)?
19:24:43 <Gregor> "wikileven wikiam? WikI'm wikfraid wiki're wikilate. Wiki're wikifired."
19:24:59 <Vorpal> elliott, not on any turned on computer at least
19:25:21 <Vorpal> elliott, that jpg (err, jpeg actually, how unusual), is it supposed to be on top of a communion bread?
19:25:33 <Vorpal> it kind of looks a bit like that
19:25:47 <Vorpal> elliott, yes it is obviously Putin :P
19:25:51 <elliott> http://en.wikipedia.org/wiki/Puttin'_on_the_Ritz
19:26:10 <elliott> So funny it has a capital SOOOOO?
19:28:37 -!- jcp has quit (Read error: Connection reset by peer).
19:28:43 <elliott> Gregor: HackEgo totally needs a way to take quoted arguments.
19:28:53 <elliott> By which I mean "please, god, don't make me feed every call through sh".
19:29:08 <elliott> Actually you can just use _.
19:29:37 <Vorpal> elliott, the `run thingy isn't that irritating IMO. Just 4 letters more
19:29:59 <Vorpal> but sure, it would be nice to avoid it
19:30:43 <Vorpal> elliott, still, there are more annoying things, like !haskell doing ghci then ghc, and not reporting any errors for some cases
19:30:52 <Vorpal> (instead just failing silently)
19:31:06 <elliott> Gregor: Wait, does HackEgo have network?
19:31:20 <Vorpal> elliott, I seem to remember he used the owner netfilter match
19:31:32 <Vorpal> to block it for all but a handful of things
19:31:35 <elliott> Gregor: tl;dr I need *.wikipedia.org access.
19:32:19 <Vorpal> elliott, that is, he was considering how to solve the network issue then I mentioned the owner match and he concluded it fitted perfectly, so unless he changed I presume he still uses that
19:33:10 <Vorpal> hm I just noticed how silly "so unless he changed I presume he still uses that" actually is if you read what it actually means
19:33:34 <Gregor> elliott: It has an http proxy, I can give extra access on request. Wikipedia seems reasonable, one sec.'
19:33:45 <elliott> Gregor: I need all language codes, mind :P
19:33:54 <Vorpal> (of course he might not be, in which case he isn't, and then doesn't use that)
19:33:58 <elliott> Can I just access normally or do I need to write my program for the proxy?
19:34:02 <Vorpal> okay, that was a bit too extreme to match zzo
19:34:14 <Gregor> elliott: It needs to be proxy-aware.
19:34:26 <Gregor> elliott: If you're writing something using e.g. wget to do requests, then you're fine.
19:34:29 <elliott> Gregor: Fucking fuckshit fuck you okay how do I do it?
19:34:42 <elliott> Gregor: Is it SOCKS or something horrible?
19:34:50 <Gregor> It's a standard HTTP proxy.
19:34:51 <Vorpal> elliott, that's python isn't it (urllib2 I mean)?
19:35:10 <Vorpal> elliott, why the 2, I never seem to remember seeing any urllib
19:35:21 <elliott> urllib2 uses it internally.
19:35:26 <elliott> In Python 3 it's a saner name.
19:35:39 <elliott> class urllib2.ProxyHandler([proxies])
19:35:40 <elliott> Cause requests to go through a proxy. If proxies is given, it must be a dictionary mapping protocol names to URLs of proxies. The default is to read the list of proxies from the environment variables . If no proxy environment variables are set, in a Windows environment, proxy settings are obtained from the registry’s Internet Settings section and in a Mac OS X environment, proxy information is retrie
19:35:40 <elliott> ved from the OS X System Configuration Framework.
19:35:52 <elliott> Gregor: What's the URoLogy?
19:36:04 <Vorpal> elliott, well that is only to be expected, they did use the opportunity of python 3 breaking things anyway to clear up some weirdness
19:36:51 <elliott> `run echo $HTTP_PROXY FUCK YEAH
19:36:56 <HackEgo> Runs arbitrary code in GNU/Linux. Type "`<command>", or "`run <command>" for full shell commands. "`fetch <URL>" downloads files. Files saved to $PWD are persistent, and $PWD/bin is in $PATH. $PWD is a mercurial repository, "`revert <rev>" can be used to revert to a revision. See http://codu.org/projects/hackbot/fshg/
19:37:18 <elliott> Gregor: Why couldn't you just IP firewall it ;__
19:37:26 <Vorpal> elliott, he is doing that as well
19:37:41 -!- jcp has joined.
19:37:41 <Vorpal> Gregor, you still use the owner match I presume?
19:38:42 <elliott> Gregor: "The default is to read the list of proxies from the environment variables"
19:38:45 <elliott> You should totally set that propertly
19:38:55 <elliott> *variables." (it's " ." in the source text but I'm sure that's a mistake.)
19:41:21 -!- Sgeo has joined.
19:41:32 <pikhq> elliott: I was testing Testing.
19:41:37 -!- cpressey_ has joined.
19:41:41 <elliott> pikhq: Testing? Testing testing?
19:41:55 <pikhq> elliott: Apparently Sid isn't going to get multiarch support until Squeeze is made stable.
19:41:59 <elliott> Testing testing testing? ("Testing to see if you want to test testing")
19:42:09 <elliott> MULTIARCH UP THE ANUS^WWAZOO
19:42:14 <Sgeo> Someone should write unit tests for the unit tests of the unit testing framework
19:42:25 <pikhq> So, basically, until then, Debian will suck more than Gentoo on amd64, regardless.
19:42:34 <Vorpal> pikhq, multiarch as in proper lib32?
19:42:40 <pikhq> elliott: So, how *is* Kitten going to handle multiple ABIs?
19:42:42 <Vorpal> that would be so awesome
19:43:17 <pikhq> Vorpal: Gentoo kinda sucks at it and has it very hacky.. (tarballs of libraries) Debian has that, but less comprehensive.
19:43:21 <elliott> pikhq: # pkg ins libc --arch=x86
19:43:28 <elliott> pikhq: # pkg ins libc --arch=x86-64
19:43:33 <elliott> pikhq: # pkg ins libc --arch=arm
19:43:38 <pikhq> elliott: And the filesystem layout?
19:43:39 <Vorpal> pikhq, most distros I know of has it very hacky
19:43:46 <pikhq> Vorpal: Quite true.
19:43:59 <pikhq> elliott: Per-ABI bin/lib/... dirs?
19:44:04 <Vorpal> pikhq, correction: all distros, except arch recently made it somewhat more sane
19:44:05 <elliott> pikhq: I'm open to suggestions on that. Probably something like /lib32, /lib64, /libarm.
19:44:20 <elliott> pikhq: Or, perhaps even: root is your native architecture, beyond that it's /arch/...
19:44:37 <elliott> But really, I'm open to suggestions.
19:44:46 <pikhq> elliott: Make it the actual ABI tuple.
19:45:02 <Vorpal> pikhq, arch still doesn't provide a complete set of 32-bit packages on 64-bit, but the repo and package layouts is quite a bit saner than in gentoo (one package matches one package, not one package matches a bunch of loosely related packages)
19:45:03 <elliott> pikhq: I might hide it away in an /arch/ directory to avoid the ugly.
19:45:09 <Gregor> Sorry, had an ad-hoc meeting.
19:45:14 <elliott> Wait, does NetBSD have ABI tuples?
19:45:26 -!- cpressey has quit (Ping timeout: 276 seconds).
19:45:30 <pikhq> It's an autotools convention that's gotten spread elsewhere.
19:45:51 <Vorpal> it is one of the few sane things coming from autotools
19:45:52 <elliott> pikhq: I'm not sure what the OS part of Kitten's would be though.
19:46:34 <pikhq> And it'd probably be best to have the same filesystem layout regardless of what the native ABI is; some packages hardcode paths in, and it's easier to not have to patch that.
19:46:43 <Vorpal> elliott, *-netbsd-<whatever sort of userland you use>
19:46:54 <Vorpal> bsd userland? I presume you wouldn't use gnu userland
19:46:56 <elliott> Vorpal: *-netbsd-kitten, then.
19:47:06 <elliott> I might use newlib libc with BSD coreutils. Who knows?
19:47:09 <pikhq> elliott: The convention is that the last bit is based on the libc.
19:47:25 <Vorpal> elliott, then *-netbsd-newlib
19:47:40 <elliott> Vorpal: But what if I change libc?! (Okay, yeah, everything will have to be rebuilt anyway :P)
19:47:49 <pikhq> For instance, i686-pc-linux-gnu for a "normal" Linux system, but i686-pc-linux-uclibc for a uclibc system.
19:47:59 <Vorpal> elliott, except you could now have them side by side when rebuilding!
19:48:07 <pikhq> elliott: If you have the filesystem layout based on the ABI tuple, you could have them both!
19:48:14 <pikhq> elliott: Even if you dynamically link!
19:48:22 <Vorpal> pikhq, the pc stuff, none of my systems have it
19:48:39 <elliott> As far as I'm concerned there's no reason not just to have an arch option when building packages.
19:48:42 <Vorpal> pikhq, wait, actually one does
19:48:53 <elliott> Easy enough for the package to have arch-specific dependencies, too.
19:48:58 <Gregor> Heh, Wikipedia was already whitelisted :P
19:48:58 <Gregor> `run wget http://en.wikipedia.org/wiki/ -O -
19:49:03 <HackEgo> <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"> \ <html xmlns="http://www.w3.org/1999/xhtml" lang="en" dir="ltr"> \ <head> \ <title>Wikipedia, the free encyclopedia</title> \ <meta http-equiv="Content-Type" content="text/html; charset=UTF-8" />
19:49:06 <Vorpal> pikhq, one is x86_64-unknown-linux-gnu, the other ones are x86_64-linux-gnu
19:49:10 <elliott> `run wget http://sv.wikipedia.org/wiki/ -O - | head
19:49:12 <HackEgo> <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"> \ <html xmlns="http://www.w3.org/1999/xhtml" lang="sv" dir="ltr"> \ <head> \ <title>Wikipedia, den fria encyklopedin</title> \ <meta http-equiv="Content-Type" content="text/html; charset=UTF-8" />
19:50:09 <Vorpal> elliott, nice coincidence, head getting (most of) <head>
19:51:08 <Vorpal> `run curl http://sv.wikipedia.org/wiki/ | head
19:51:10 <elliott> Gregor: What's the download-this command again?
19:51:28 <pikhq> So, it seems Mandriva is the only distro with actual multi-arch support.
19:51:47 <Vorpal> pikhq, mandriva is in all other aspects not a sane choice
19:51:58 <pikhq> Vorpal: Quite right.
19:52:06 <olsner> sounds pretty neat to have all architectures' packages available and installable
19:52:09 <Vorpal> pikhq, also I'm surprised it supports it if other RPM based distros don't
19:52:25 <elliott> olsner: Well, that's the thing; there isn't one pakage per architecture.
19:52:33 <Vorpal> yeah, mostly useful for ones you can run, for cross compiling it is a bit more work
19:52:34 <elliott> There's "newlib" and if it supports the given architecture, good for you.
19:52:44 <Vorpal> elliott, source based distro?
19:52:46 <elliott> Vorpal: It will treat it all as cross-compilation.
19:52:53 <olsner> wait, is this source-based?
19:53:00 <elliott> Just with the knowledge that "gcc -m32" is a valid 32-bit cross-compiler on 64-bit.
19:53:03 <Vorpal> elliott, ever done a canadian cross?
19:53:09 <elliott> Vorpal: olsner: Not source-based.
19:53:18 <elliott> Although you will be able to build-and-install a package with one command.
19:53:23 <elliott> If you really want to for some reason.
19:53:31 <pikhq> Vorpal: RPM itself actually has support for multiple ABIs. No other distro *actually uses it*.
19:53:36 <elliott> <Vorpal> elliott, ever done a canadian cross?
19:54:06 <Gregor> Canadian cross = just don't go there.
19:54:26 <elliott> Basically 32-bit and 64-bit support will just work... I'll have to tell the package manager about cross-compilers for --arch=arm on x86 or something, but that's not so difficult.
19:54:28 <Vorpal> elliott, I heard someone cross compiling gdb on OS X to run on linux (the resulting gdb binary that is). Why? Because it had to target debugging OS X. Remote kernel debugging.
19:54:31 <olsner> cross-compiling a cross-compiler? what's so hard about that? :D
19:54:36 <Vorpal> and it didn't want to compile on linux
19:54:44 <elliott> All it has to do is install the relevant cross-compiler.
19:54:51 <Vorpal> compared to that, a canadian cross seems like a stroll in the park!
19:54:57 <elliott> (There probably won't be one package per host and target architecture.)
19:55:10 <elliott> (Instead, a cross-compiling gcc or whatever will be a single metapackage, taking host and target as argument/options.)
19:55:13 <Gregor> Vorpal: There effectively is no anything-to-OS-X cross-compiler, so that's unsurprising.
19:55:16 <olsner> well, if you do canadian cross of the compiler, you'll probably want to do it of the rest of the toolchain too, including gdb
19:55:38 <elliott> `fetch http://pastie.org/pastes/1218858/text?key=rmrta4udnevkdlwt0vgxg
19:55:42 <HackEgo> 2010-10-13 18:55:36 URL:http://pastie.org/pastes/1218858/text?key=rmrta4udnevkdlwt0vgxg [1486/1486] -> "text?key=rmrta4udnevkdlwt0vgxg" [1]
19:55:43 <Vorpal> Gregor, actually the guy got the cross compiling to OS X to kind of work, but not for gdb.
19:55:50 <olsner> so the osx-linux-osx gdb thing would pretty much be included in the work for the canadian cross
19:55:55 <elliott> that actually doesn't work without `run
19:56:00 <elliott> `run rm "text?key=rmrta4udnevkdlwt0vgxg"
19:56:14 <Vorpal> Gregor, as in, he successfully cross compiled "hell of a world" as a test (he refused to use "hello world" after all he had to go through to make it work)
19:58:33 <elliott> `fetch http://pastie.org/pastes/1218882/text?key=1grvzo3sho5cexf4rrmog
19:58:35 <HackEgo> 2010-10-13 18:58:29 URL:http://pastie.org/pastes/1218882/text?key=1grvzo3sho5cexf4rrmog [1553/1553] -> "text?key=1grvzo3sho5cexf4rrmog" [1]
19:58:39 <elliott> `run mv text?key=1grvzo3sho5cexf4rrmog bin/wl
19:59:08 <olsner> if kitten is netbsd-based, will the freebsd opera builds work?
19:59:16 -!- oerjan has joined.
19:59:20 <elliott> olsner: maybe. who knows? if not, try the linux emulation ;)
19:59:25 <elliott> `fetch http://pastie.org/pastes/1218882/download?key=1grvzo3sho5cexf4rrmog
19:59:28 <HackEgo> 2010-10-13 18:59:22 URL:http://pastie.org/pastes/1218882/download?key=1grvzo3sho5cexf4rrmog [1124/1124] -> "download?key=1grvzo3sho5cexf4rrmog" [1]
19:59:41 <elliott> `run cat download?key=1grvzo3sho5cexf4rrmog >bin/wl; rm download?key=1grvzo3sho5cexf4rrmog
20:00:04 <HackEgo> Traceback (most recent call last): \ File "/tmp/hackenv.13346/bin/wl", line 5, in <module> \ import json \ ImportError: No module named json
20:00:42 <Vorpal> elliott, will kitten still run on toasters?
20:00:52 <elliott> Vorpal: Yes. If your toaster can run X11.
20:01:00 <elliott> (Okay, so X11 isn't required, but still.)
20:01:14 <Vorpal> elliott, hah. Will it be officially supported?
20:01:27 <elliott> Vorpal: Yes, if your toaster has a standard PC inside.
20:01:38 <Vorpal> elliott, ouch. That is not likely
20:01:48 <elliott> Neither is your toaster having any kind of computer inside :)
20:02:05 <Vorpal> elliott, well of course, but I meant, even ATX-mini is a bit large :P
20:02:07 <elliott> Vorpal: OTOH, all brands of coffee maker will be supported.
20:02:19 <elliott> Hopefully I will support RFC2324.
20:02:36 <elliott> Wow, RFC2324 is sucky. It isn't even REST!
20:02:51 <Vorpal> elliott, ah nice, I have one here that is completely devoid of anything like an integrated circuit
20:03:09 <elliott> Vorpal: I redefine "coffee maker" to mean "coffee maker with a computer hooked up to it".
20:03:12 <Vorpal> it is just a plain old electrical coffee maker, so simple inside that even your grandmum could understand it :P
20:03:26 <Gregor> What's all this about toasting kittens?
20:03:26 <Vorpal> elliott, and that computer must be a PC? :D
20:03:39 <Vorpal> Gregor, ...you did that intentionally right?
20:03:49 <Gregor> Toasted kittens? NEVER!
20:04:03 <Vorpal> <elliott> Wow, RFC2324 is sucky. It isn't even REST! <-- the cofee makers never rest!
20:05:00 <elliott> alias coffee="curl -d '' http://kitchenpc:999/brew"
20:05:03 <Gregor> elliott: How about 2.7
20:05:07 <pikhq> elliott: BTW, one of the major reasons to actually have per-ABI directories is so that you actually have the flexibility to do something like install x86 programs on ARM and have them "just work" with qemu (and Linux's arbitrary ABI support).
20:05:18 <elliott> alias coffee-status="curl -q http://kitchenpc:999/status"
20:05:19 <Vorpal> elliott, um that should be the BREW method :P
20:05:30 <elliott> $ coffee; coffee; coffee; coffee; coffee; coffee; coffee; coffee; coffee; coffee; coffee; coffee; coffee
20:05:50 <Vorpal> (as in, a flood of coffee)
20:05:52 <elliott> Vorpal: No, this is HTCPCPCP
20:06:01 <elliott> Hyper Text Coffee Pot Control Protocol Crappy. Perfect!
20:06:06 <Vorpal> elliott, yes indeed! it flooded your kitchen with cofee
20:06:09 <elliott> It is the RESTful alternatiev to HTCPCP.
20:06:35 <olsner> eugh, "BREW method" ... don't write brew with all-caps or I'll be reminded of Binary Runtime Environment for Wireless :/
20:06:55 <Vorpal> olsner, what was binary runtime environment for wireless?
20:07:03 <olsner> http://en.wikipedia.org/wiki/BREW
20:07:08 <elliott> Gregor: You can apt upgrade the stuff inside, right?
20:07:13 <elliott> Gregor: Plz just upgrade Python to 2.6
20:07:19 <Gregor> elliott: But I can install from source.
20:07:23 <elliott> Gregor: Okay... well... do that please :P
20:07:28 <elliott> Python 2.5.5 is ancientish.
20:07:46 <elliott> So I didn't think it'd have 2.7.
20:08:01 <Gregor> Worse yet, HackBot is on Debian Lenny.
20:08:04 <Gregor> Which is why it has 2.5
20:08:30 <Vorpal> Gregor, is lenny not like previous stable?
20:08:31 <elliott> Didn't they promise to start upgrading regularly?
20:08:40 <elliott> but it's had point-releases
20:08:49 <elliott> coppro: Get on to fucking that shit, stat.
20:09:01 <Vorpal> elliott, which was previous stable? There was something in between woody and lenny right?
20:09:14 <elliott> Gregor: I mean, ideally buzz of course, but hamm at the very least.
20:09:17 <Vorpal> elliott, why not python 3 btw?
20:09:31 <Gregor> elliott: I used potato once :P
20:09:33 <elliott> Vorpal: Python 3 is completely useless as it has approximately 0 library support and approximately 0 things written in it.
20:09:39 <quintopia> anyone want to hear a TERRIBLE math class homework question?
20:09:41 <elliott> Gregor: Potato is .2 versions newer than hamm! Can't have that.
20:10:04 -!- oklopol has joined.
20:10:06 <elliott> Man, what will Debian do when they run out of Toy Story characters?
20:10:11 <elliott> Start naming them after the individual soldiers?
20:10:21 <quintopia> Which of the following are continuous functions? (Select all that apply.)
20:10:24 <Gregor> elliott: Use Toy Story {2,3} characters, presumably.
20:10:30 <quintopia> 1 The temperature at a specific location as a function of time. 2 The temperature at a specific time as a function of the distance due west from New York City. 3 The altitude above
20:10:34 <quintopia> sea level as a function of the distance due west from New York City. 4 The cost of a taxi ride as a function of the distance traveled. 5 The current in the circuit for the lights in a
20:10:47 <elliott> Gregor: Even that plan has... flaws.
20:10:53 <elliott> Gregor: "Boy, I can't wait for Debian barbie!"
20:10:58 <Vorpal> quintopia, any other ones?
20:11:00 <Gregor> I was just about to say that :P
20:11:01 <Vorpal> quintopia, or was that all?
20:11:08 <pikhq> elliott: http://lackof.org/taggart/hacking/multiarch/ Oh, here's a thought on how to do multiarch "right".
20:11:13 <Gregor> elliott: I preordered the Barbie Computer Engineer :P
20:11:31 <Vorpal> quintopia, well I don't know about taxi, but all the other ones are discrete, due to quantum mechanics
20:11:38 <elliott> Gregor: Debian lotsohugginbear
20:11:38 <quintopia> my best guess at reading the asker's mind says "the first three" while the actual correct answer i believe to be "none of the above"
20:12:00 <elliott> Gregor: (Yes yes, it'd just be lotso, shut up.)
20:12:11 <quintopia> however, the problem is compounded by not knowing whether the asker wants time zones to be taken into account in number 2.
20:12:22 <elliott> pikhq: OK, well, it looks relatively good, except:
20:12:35 <elliott> pikhq: The -os prefix is stupid, Kitten is not going to run on the same FS as any other OS.
20:12:47 <Gregor> elliott: Maybe ... JUST MAYBE ... at some point they'll come to their senses and use numbers.
20:12:48 <Vorpal> quintopia, anyway, current in a circuit might be continuous if there is a dimmer connected (well not really due to quantum mechanics9
20:12:49 <elliott> pikhq: And /*/foo directories are silly and /foo/* is a far better structure for things like that.
20:13:16 <pikhq> elliott: But you might want to support another OS's ABI someday.
20:13:19 <quintopia> vorpal: ignoring quantum mechanical effects, 5 would be continuous with or without a dimmer.
20:13:21 <elliott> Gregor: Debian chattertelephone
20:13:26 <elliott> Gregor: Debian jackinthebox
20:13:41 <elliott> ("What are you running?" "Debian pricklepants.")
20:13:42 <Vorpal> quintopia, well yes, since current doesn't change instantly
20:13:48 <elliott> pikhq: Well, okay, NetBSD does support Linux emulation.
20:13:54 <quintopia> but i suspect the asker expects to also ignore the latency in change of current in a wire
20:14:00 <elliott> Gregor: Debian totoro (I <3 THAT CAMEO FOREVER)
20:14:01 <Vorpal> quintopia, so why not write a proper answer to it!
20:14:15 <pikhq> elliott: Just use it so that you can have multiple kernel ABIs "just work".
20:14:32 <pikhq> elliott: Not having to change things in the future is GOOD.
20:14:38 <elliott> pikhq: I'LL THINK ABOUT IT
20:14:39 <Vorpal> quintopia, surely you can't get in trouble from writing a paragraph answering it properly
20:14:50 <elliott> Gregor: Oh man Debian zurg.
20:14:55 <elliott> That had better be the best release EVER.
20:15:10 <quintopia> Vorpal: it's not my homework. i have no need to. i just wanted to post its awfulness here so that i could say that the person whose homework it is, after getting a wave of complaints that the question is ambiguous, said "i'll go ask somewhere where people will actually want to help me."
20:15:14 <Gregor> elliott: They can use that as a transitionary release by spelling it both "zurg" and "zerg"
20:15:22 <Gregor> elliott: Then they can have Debian Protoss and Debian Terran
20:15:27 <elliott> <Vorpal> quintopia, well I don't know about taxi, but all the other ones are discrete, due to quantum mechanics ;; oh come on
20:15:29 <Gregor> Admittedly that doesn't get them too much farther ...
20:15:37 <Vorpal> quintopia, okay then the person who asked for help is stupid too
20:15:39 <pikhq> Gregor: Unit names.
20:15:51 <Vorpal> elliott, wasn't it obvious it was a joke?
20:16:06 <elliott> Vorpal: No, because it wasn't funny :P
20:16:26 <elliott> Gregor: Wait. With Totoro they can "reasonably" claim that My Neighbour Totoro took place entirely within the Toy Story universe.
20:16:42 <Vorpal> elliott, okay, lets rephrase that: wasn't it obvious it couldn't have been meant to be taken seriously?
20:16:42 <elliott> Gregor: I'm talkin' Debian catbus here.
20:16:42 <pikhq> Gregor: Alternately, they could *commission* more Toy Story films just to have more namespace.
20:16:46 <quintopia> elliott: i also stated that the correct answer would be "none of the above" due to the discrete nature of the universe. want to bitch me out for pedanticism too?
20:16:49 <Gregor> elliott: There ya go, lots of new names.
20:16:52 -!- Sgeo has quit (Ping timeout: 264 seconds).
20:17:14 <elliott> No OS can ever possibly hope to live up to that name.
20:17:16 <pikhq> "Toy Story 4: Debian Versioning"
20:17:52 <elliott> pikhq: But Disney/Pixar are jerks, so they give all the characters names in Klingon.
20:18:18 <Gregor> But Debian wouldn't use the number-names in order.
20:18:27 <Gregor> They'd have Debian 5.5, then Debian 4.7, then Debian 8.2
20:19:04 <Gregor> Or maybe Debian 8.0 "Four Point Seven"
20:19:09 <Vorpal> that would be quite fun
20:19:31 <elliott> Oh man, the Tux would have to be a character in Toy Story 4: Debian Versioning.
20:19:54 <Vorpal> elliott, they could reuse the cast from previous movies in the newly commissioned one. That would be even more jerky
20:20:06 <elliott> deb http://http.us.debian/org/debian debian main
20:20:42 <Vorpal> elliott, oh and of course the Gnu gnu, just to irritate everyone but FSF. Oh wait, they would probably irritated too, since that is the natural state of them.
20:21:15 <elliott> Have rms appear as the main antagonist
20:21:17 <quintopia> they should have the cast be: woody, buzz, bo peep, the main bug from bug's life, the little boy from Up, Mike Waczowski, and Tux
20:21:28 <elliott> But they won't use outright evil characters for release names will they :(
20:21:38 <Vorpal> elliott, no. Debian GNU/Linux 12.0 "gnu"
20:21:50 -!- augur has quit (Remote host closed the connection).
20:22:14 -!- augur has joined.
20:22:22 <Vorpal> pikhq, sid is always fixed to unstable though afaik
20:22:24 <elliott> pikhq: ...is the unstable release.
20:22:30 <olsner> oh, are the debian releases named after toy story characters?
20:22:33 <elliott> They'll never codename experimental because that's giving it too much legitimacy :P
20:22:37 <pikhq> elliott: And an outright evil character.
20:22:44 <Vorpal> olsner, or actually no
20:22:47 <Vorpal> olsner, it is the other way around
20:23:04 <Vorpal> olsner, Toy story characters are named after Debian releases
20:23:13 <elliott> pikhq: sid is never released.
20:23:51 <Vorpal> elliott, experimental, is that like unstable, just more so?
20:23:58 <pikhq> Oh, the next release codename is actually coming from Toy Story 2.
20:24:19 <Vorpal> elliott, hm when did they run out of Toy Story characters?
20:24:23 <Vorpal> or when are they going to maybe
20:24:31 <pikhq> Vorpal: Staging area for things that have yet to hit unstable.
20:24:40 <quintopia> there are plenty of characters left
20:24:59 <quintopia> I would definitely use Debian T-Rex
20:25:07 <pikhq> I wonder what'll happen with Ubuntu.
20:25:14 <elliott> <Vorpal> elliott, hm when did they run out of Toy Story characters?
20:25:17 <elliott> They'll be fine for years.
20:25:30 <quintopia> they will start over from the beginning of the alphabet iirc
20:25:30 <Gregor> Since they have a release once every two years or so :P
20:25:31 <elliott> Vorpal: when ubuntu run out of alphabet
20:25:47 <elliott> So in something like 2017, OH GOD
20:25:52 <elliott> Zygotic Zebra will come out
20:26:18 <elliott> Gregor: Yes, even I'm amazed at how awesome that was.
20:26:19 <pikhq> And then Amorphous Aardvark. They have yet to use A.
20:26:34 <Vorpal> elliott, well, they will just continue incrementing the unicode code point
20:26:45 <Gregor> What letter did they start on?
20:26:46 <Vorpal> and when that runs out we are probably not using *nix based systems any more
20:26:48 <quintopia> or, more likely, zippy, "CAUSE ITS FAST LOL"
20:26:52 <elliott> They only started the alphabetical thing recently.
20:26:58 <pikhq> Gregor: W, then H, then B, then they went alphabetical.
20:27:00 -!- augur has quit (Remote host closed the connection).
20:27:03 <elliott> W, H, B, D, E, F, G, H, I, J, K, L, M, N is planned.
20:27:30 <Vorpal> elliott, which one was the first H?
20:27:32 <elliott> So they have A, B and C after 2017.
20:27:36 <pikhq> They reüsed H. Darnit.
20:27:36 <Gregor> OK, so next would be Þ
20:27:38 <Vorpal> the second would be hedgehog
20:27:49 <Vorpal> elliott, was that not the second one?
20:28:47 <Vorpal> elliott, on the other hand, think of a suitable name starting with c
20:28:50 <elliott> Anyway after Cashmere Civet they'll have run out.
20:29:02 <Vorpal> elliott, best I managed was "chaotic camel"
20:29:05 <Gregor> How about Cheshire Cat
20:29:08 <Vorpal> which does not sound very good
20:29:11 <pikhq> And move to Toy Story characters.
20:29:15 <quintopia> there's a whole list of planned future names for ubuntu versions already...put your suggestions there
20:29:21 <Vorpal> Gregor, Cheshire is not an adjective
20:29:34 <Vorpal> elliott, must it be first *two* letters?
20:29:43 <pikhq> Vorpal: Yes it is. "Of or relating to Cheshire."
20:29:58 <oerjan> <elliott> Have rms appear as the main antagonist <-- this made me think of the bathtub monster from Rose is Rose (both eat toes)
20:30:00 <Vorpal> elliott, they need to get aardvark
20:30:29 <Vorpal> elliott, "hardy heron", "karmic koala"?
20:30:32 <pikhq> Gregor: Anthropomorphic Aardvark. For the furries.
20:30:32 <Gregor> elliott: You already said Abalone
20:30:38 <Vorpal> elliott, that is first one letter, not first two
20:30:58 <Gregor> pikhq: It has to be aa
20:31:21 <elliott> elliott@dinky:~/Code/wl$ grep '^aa' /usr/share/dict/words
20:31:24 <olsner> shouldn't it rather be something like aardvarkomorphic anthrope to refer to furries?
20:31:25 <quintopia> for the pirates fed up with ubuntu
20:31:41 <Vorpal> <elliott> you're right Vorpal <-- of course ;)
20:31:57 <pikhq> Gregor: A'a Aarvdvark?
20:32:36 <elliott> "andalism" -- first line of [[Abalone]]
20:32:57 -!- cpressey has joined.
20:33:15 <Vorpal> elliott, anyway, jaunty was the only one I remember where it is actually first two letters that are the same.
20:33:31 <elliott> Vorpal: But after Cashmere Civet, they've run out of alphabet.
20:33:38 <elliott> So we go to Aa, Ab, ..., Az, ..., Zz
20:33:41 <Gregor> Accidental Acupuncturist
20:33:48 <elliott> Gregor: THAT IS NOT AN ANIMAL
20:33:48 <Gregor> Yes, an acupuncturist is an animal.
20:33:52 <Vorpal> elliott, so they keep incrementing the unicode codepoint I told you!
20:33:53 -!- Phantom_Hoover has joined.
20:33:58 -!- tombom has quit (Ping timeout: 272 seconds).
20:34:10 -!- tombom has joined.
20:34:10 -!- tombom has quit (Changing host).
20:34:10 -!- tombom has joined.
20:34:21 <Vorpal> elliott, they need to get something starting with the del char into the dictionary quickly
20:34:47 <pikhq> I can't wait for いたい いぬ (itai inu).
20:34:54 <quintopia> I can't wait til Gnaught Gnu comes out.
20:35:04 <Vorpal> elliott, once they reach the Chinese letters it will be easy
20:35:26 <Vorpal> elliott, hm, which one was I?
20:35:58 -!- cpressey_ has quit (Ping timeout: 252 seconds).
20:36:14 <Vorpal> elliott, I swear they pick intentionally obscure adjectives. Okay, lucid wasn't, but many other ones were
20:36:28 <cpressey> Fucking Falcon, how does that work?
20:36:48 <cpressey> (By the principle of the Multiparadigm clusterfuck, apparently.)
20:37:01 <cpressey> quintopia: haven't checked recently
20:37:02 <fizzie> elliott: Wordnet says: http://p.zem.fi/ab-animals -- okay, so it counts pretty much anything that's alive by any definition, but...
20:37:24 <elliott> Is that the ... result of an abortion?
20:37:46 <fizzie> elliott: Well, http://p.zem.fi/b5og you see.
20:37:49 <Gregor> Accredited Achatellinus
20:37:53 <quintopia> cpressey: you can do it with one hand, while typing
20:38:42 <fizzie> Still, there's two quite viable ab-animals there:
20:38:47 <fizzie> 1. abalone, ear-shell -- (any of various large edible marine gastropods of the genus Haliotis having an ear-shaped shell with pearly interior)
20:38:48 <fizzie> 1. abrocome, chinchilla rat, rat chinchilla -- (ratlike rodent with soft fur and large ears of the Andes)
20:38:51 <Gregor> When do we get to Augmented Australopithecine?
20:38:58 <elliott> fizzie: Already said abalone.
20:39:29 <Gregor> Or are WE the Augmented Australopithecine?
20:39:38 <oerjan> Abortus isn't animal, it's a murdered human! </pope>
20:39:38 <elliott> What # is u in the alphabet?
20:40:13 <Gregor> elliott: A, B, C, D, E, F, G
20:40:24 <elliott> Gregor: Augmented Australopithecine is in about 2039.
20:40:25 <Gregor> elliott: H, I, J, K, L, M, N, O, P
20:40:27 <coppro> do re mi fa so la ti do
20:40:38 <elliott> M=Math;C=12;f=[];R=[];A='charCodeAt';S='slice';for(P=0;P<96;){k="/SN;__/NK;OL/QN;__/OL;NK4L@@_C4_G@OL4SO@__4QN@OL3NB?_G3_K?OL/QN;__/SK;__4OL@__4LC@_G4LC@_G4_C@_G"[A](P);D="\0\0";for(j=0;k<95&&j<1e4;){v=M.max(-1e4,M.min(1e4,1e6*M.sin(j*M.pow(2,k/C)/695)))/M.exp(j++/5e3);D+=String.fromCharCode(v&255,v>>8&255)}R[P++]=new Audio("data:audio/wav;base64,UklGRgAAAABXQVZFZm10IBAAAAABAAEAwF0AAIC7AAACABAAZGF0YSBO
20:40:39 <elliott> "+btoa(D))}for(e=i=252;i--;)f[i]=i%C&&i<240?(i+1)%C?r=0:'█<br>':'█';t=p=4;function d(c){for(q=p+[13,14,26,25][r%4],i=1;i<99;q+=((i*=2)==8?[9,-37,-9,37]:[1,C,-1,-C])[r%4])if('36cqrtx'[A](t)&i)if(-c){if(f[q])return 1}else f[q]=c}function m(e){Q=[-1,0,1,C][e?e.keyCode-37:3]||0;d(0);p+=Q;r+=!Q;s=d(1);if(s)p-=Q,r-=!Q;d('▒');document.body.innerHTML=f.join('').replace(/0/g,'░');return s}onkeydown=m;o=
20:40:39 <elliott> function(){P=P%96;for(_ in[1,2,3])R[P++].play();if(m()){t=~~(7*M.random()),p=r=4;e=d(1)?1e9:e;for(y=0;y<240;)if(f[S](y,y+=C).join().indexOf('0')<0)f=f[S](0,C).concat(f[S](0,y-C),f[S](y))}setTimeout(o,e*=0.997)};o()
20:40:43 <Gregor> *do re mi fa sol la si do
20:40:57 <fizzie> fis@eris:~$ wn animal -treen
20:40:57 <fizzie> Hyponyms of noun animal
20:40:57 <fizzie> Search too large. Narrow search and try again...
20:40:57 <fizzie> (What a silly limited thing.)
20:41:39 <elliott> Phantom_Hoover: javascript
20:41:43 <elliott> see http://js1k.com/demo/730 in firefox
20:42:14 <Gregor> coppro: Just correcting your solfege :P
20:42:24 <fizzie> Have to love the data:audio/wav dynamic audio-synthery.
20:42:27 <coppro> elliott: you pinged me earlier
20:42:43 <oerjan> Gregor: it's ti in some languages
20:42:44 <coppro> 15:08 < elliott> coppro: Get on to fucking that shit, stat.
20:42:46 <elliott> Phantom_Hoover: It's also computed.
20:42:57 <coppro> 15:08 < elliott> coppro: Get on to fucking that shit, stat.
20:42:57 <elliott> well you are the coprophiliac
20:43:14 <Gregor> oerjan: LOSER LANGUAGES.
20:43:14 <quintopia> elliott: tetris is popular as a programming project precisely because it can be written with so little code
20:43:55 <elliott> quintopia: Oh come on, Tetris with speeding-up and *the original Tetris theme generated in wav* in 1024 bytes is pretty damn impressive.
20:43:56 <elliott> Phantom_Hoover: http://js1k.com/demo/730
20:44:50 <elliott> Phantom_Hoover: oh shut up
20:45:24 <Phantom_Hoover> I was thinking they had some awesome procedural generator specially configured to belt out the full theme.
20:45:38 <elliott> Phantom_Hoover: IT'S AWESOME GO TO HELL
20:45:55 <elliott> Anyway it's approximately the whole theme.
20:46:06 <quintopia> elliott: i'm not hearing the music in firefox, and also, this version doesn't seem to let me hold down a key to move the block a long distance
20:46:07 <elliott> It's enough to sing I Am The Man... to it.
20:46:13 <elliott> quintopia: which Fx version?
20:46:15 * pikhq needs a decent pair of speakers, but is cheap. Dammit.
20:46:27 <elliott> also, i don't think you realise just how little code 1k is
20:46:53 <elliott> I'm on 3.6.10, so, try upgrading.
20:46:58 <Gregor> elliott: Apparently it's in the repos, but since I hadn't updated Hackbot's chroot in roughly all eternity, it's not installed :P
20:47:08 <elliott> Gregor: I thought you said you were gonna source it.
20:47:18 <Gregor> elliott: I was, until I noticed it was in the repo anyway.
20:47:23 <elliott> Gregor: So how's that going
20:47:41 <elliott> I bet it breaks everything.
20:47:50 <Gregor> But *eh*, these things happen.
20:48:09 <Gregor> I've recreated the Hcakiki chroot three times after breaking it horribly by trying to do stupid things.
20:48:13 <Gregor> I can do the same with Hackbot.
20:48:56 <quintopia> also, where is my preview of the next block >:
20:49:47 <Gregor> elliott: Then compile that to <1K of JS.
20:50:42 <Gregor> <EgoBot> I do not like you, elliott.
20:51:04 <Gregor> <EgoBot> I do not like green eggs and ham, either.
20:51:12 <elliott> I do not like them, Ego I am.
20:51:32 <Gregor> <EgoBot> I do not like them on the 'net.
20:51:48 <Gregor> I can't rhyme *shrugs*
20:53:26 <Vorpal> what an unexpected result!
20:53:48 <Vorpal> but I guess it is too lazy to compute the question
20:54:05 -!- augur has joined.
20:54:27 <oerjan> in lazytris the block isn't actually drawn until it hits bottom. you better _pray_ you hit the right buttons.
20:55:13 <Vorpal> oerjan, actually, the blocks doesn't hit the bottom until you lose.
20:55:16 <Gregor> http://codu.org/imgs/rc.png Images I've left floating around codu.org that are totally self-descriptive.
20:55:53 <Gregor> It is the nineties and there is time for JSKlax?
20:56:21 <Vorpal> Gregor, any idea what rc.png actually is. I'm genuinely interested
20:56:28 <Gregor> Not even the foggiest bit of a clue.
20:56:43 <Vorpal> Gregor, to tell the truth it looks pretty random. Which rules out a lot of stuff rc could stand for
20:56:48 <Gregor> Clearly something with poorly-correlated axes.
20:57:09 <Vorpal> Gregor, what about timestamp on file?
20:57:15 <Vorpal> Gregor, maybe you could grep logs
20:57:24 <Vorpal> if ever mentioned on irc
20:57:27 <Gregor> I doubt it was #esoteric-related.
20:57:33 <Gregor> Or otherwise IRC-related
20:57:33 * augur greps gregor's logs
20:57:52 <Gregor> augur: I hope you're using a -E!
20:57:54 <elliott> <Gregor> It is the nineties and there is time for JSKlax? ;; wat
20:59:01 <elliott> oerjan: sequence.repeat$randomIO
21:00:22 <oerjan> try randoms<$>newStdGen instead
21:01:06 <elliott> Prelude Random> g<-newStdGen
21:01:09 <elliott> Prelude Random> randoms g :: [Int]
21:02:29 <pikhq> elliott: Whoever writes copy for the Xubuntu installer should be shot.
21:02:50 <oerjan> elliott: erm did it hang at [ ?
21:02:54 <pikhq> elliott: It's got grammatical and spelling errors.
21:03:12 <pikhq> The one I remember most strong was "linux".
21:03:19 <elliott> pikhq: (Xubuntu doesn't get nearly as much love as (K)ubuntu)
21:04:06 <pikhq> Gregor: What differences does it have from Sid?
21:04:19 <elliott> pikhq: It's sid for people who think that pre-installed KDE is a grand idea.
21:04:26 <elliott> (Well, the LiveCD uses KDE.)
21:04:34 <Gregor> elliott: The XFCE live CD uses XFCE.
21:04:48 <elliott> Meanwhile, sid uses whatever the fuck you install :P
21:05:02 <Vorpal> Gregor, I would be surprised if the xfce livecd used, say, fluxbox
21:05:05 <elliott> There is basically no reason not to just use testing anyway.
21:05:24 <Gregor> elliott: aptosid is more up-to-date than testing, at the cost of being less stable.
21:05:32 <elliott> testing is not exactly out-of-date.
21:05:40 <Gregor> elliott: Depends on the packages.
21:05:55 <Gregor> elliott: Some packages are basically right up to latest, others trail behind.
21:05:55 <pikhq> Testing's usually pretty up-to-date. It only ceases to be even vaguely so when they do a freeze for a new stable.
21:06:10 <Gregor> (Yeah, the freezes are annoying to)
21:06:21 <oerjan> !haskell import System.Random; main = print.take 10.randoms=<<newStdGen
21:06:47 <elliott> main=do{g<-newStdGen;print(randoms g::[Int])}
21:06:51 <Gregor> Yeesh, I'm doing a system upgrade on Codu and NOW everybody's hammering it >_<
21:06:58 <elliott> elliott@dinky:~/Code/lazytris$ runhaskell lazytris.hs | head -c 50
21:06:58 <elliott> [-7911342775086485719,-3419649560986927097,-704380lazytris.hs: <stdout>: commitAndReleaseBuffer: resource vanished (Broken pipe)
21:07:26 <quintopia> this game is hard: http://js1k.com/demo/823 (my high score after 4 plays is only 11!)
21:07:33 <Vorpal> elliott, I'm not surprised at the broken pipe
21:07:40 <Gregor> quintopia: http://sibeli.us/ <-- this game is hard
21:07:47 <elliott> Vorpal: I meant the integers.
21:07:55 <Vorpal> elliott, they *are* integers
21:08:08 <elliott> Gregor: That game makes you listen to Sibelius, though.
21:08:10 <Vorpal> elliott, so just do abs() or something if that is the issue
21:08:20 <Vorpal> elliott, that is nasty
21:08:37 <Gregor> elliott: s/makes/allows/
21:08:57 <Vorpal> Gregor, what is the goal of the game?
21:09:07 <elliott> Vorpal: Welcome to Sibeli.us! This is a web based version of that insipid game-music genre that's popular lately, with a twist: The music is Sibelius' Finlandia. Sound easy? Feel free to try! Use the keys '1', '2', '3' and '4' or 's', 'w', 'b' and 'p' when an action hits the relevant letter (and music). And have fun!
21:09:24 <elliott> Gregor: It needs to penalise you for pressing swbp when there's nothing there.
21:09:36 <Gregor> elliott: It just doesn't let your score go below 0.
21:09:37 <elliott> Gregor: Doesn't go negative though.
21:09:56 <Vorpal> <quintopia> this game is hard: http://js1k.com/demo/823 (my high score after 4 plays is only 11!)
21:09:59 <elliott> I want to go for minimum possible score!
21:10:04 <Vorpal> what is the goal of that game
21:10:49 <Vorpal> http://sibeli.us/ <-- "meh, flash"
21:10:56 <Vorpal> but then, I'm no sibelius fan
21:11:17 <Gregor> Vorpal: Flash only for audio.
21:11:25 <Gregor> Vorpal: And don't say HTML5.
21:11:30 <Vorpal> Gregor, don't you um need that for the page?
21:11:32 <Gregor> Vorpal: If you say HTML5, I punch you in the face.
21:11:35 <pikhq> Oh, that was the installer for the last LTS version. Why would I want *that*?
21:11:40 <quintopia> Gregor: that game would be ridiculously easy if it were smooth and synced with the music properly
21:11:44 * Gregor punches elliott in the face.
21:11:55 <elliott> Gregor: It's punch-worthy because it works!
21:12:11 <elliott> Or I guess Gregor supports IE.
21:12:21 <quintopia> gregor: no. it would probably work in windows. adobe supports windows...
21:12:28 <Gregor> quintopia: No, it's SHIT in Windows.
21:12:32 <Gregor> quintopia: It's much better everywhere else.
21:13:13 <Vorpal> <elliott> Gregor: It's punch-worthy because it works! <-- he thinks it doesn't
21:13:31 <Gregor> It doesn't work for ZEE for different reasons ...
21:13:46 <Vorpal> well, except for zee it works nicely
21:13:54 <Gregor> It doesn't work for Sibeli.us because no implementation communicates time very well.
21:14:05 <elliott> Your mom doesn't communism.
21:14:07 * Vorpal forces elliott to use gopher5
21:14:13 <Gregor> For ZEE HTML5 can't loop properly. Neither can Flash, but it's a modicum better.
21:14:35 <Vorpal> Gregor, why not just use jsmips + proper code to loop! ;)
21:14:50 <Gregor> Vorpal: Because HTML5 doesn't support synth, only feeding it audio files.
21:14:56 <elliott> main=do{g<-newStdGen;let b=randomRs(0,6)g::[Int]}
21:14:59 <elliott> The beginnings of LAZYTRIS
21:15:21 <Gregor> Vorpal: Honestly, if HTML5 Audio had a function to just hand it bits to spit out, that would be ideal.
21:15:39 <Gregor> Vorpal: It can do streams (sometimes), but that's a waste of bandwidth if you're just looping.
21:15:39 <Vorpal> Gregor, with streams you could do it properly server-side
21:16:01 <Vorpal> Gregor, maybe you need to do it locally instead of in the browser then
21:16:07 <Vorpal> Gregor, using sdl or such
21:16:16 <Gregor> Vorpal: The SDL version of ZEE loops beautifully :P
21:16:24 <Gregor> Which is why I sort of don't care that the browser version is crapsiloo.
21:16:28 <Vorpal> Gregor, oh, you somehow combined sdl + js?
21:16:39 <Gregor> http://codu.org/projects/gjs/
21:17:08 <Gregor> Codu is doing a system upgrade right now, yeeeesh
21:17:51 * Gregor projects hatred at Vorpal
21:18:09 <Gregor> <Gregor> Codu is doing a system upgrade right now, yeeeesh
21:18:11 <olsner> hatred: makes stuff work
21:18:30 <Vorpal> Gregor, I just updated you on the current situation
21:18:39 <cpressey> !run wget http://everything | bzip2
21:18:44 <Gregor> THANK YOU FOR KEEPING ME APPRISED.
21:19:02 <cpressey> `run wget http://everything | bzip2
21:19:18 <Vorpal> Gregor, IT WAS SO LITTLE. NOTHING TO THANK ME FOR.
21:19:26 <oerjan> olsner: Gregor is all about red hats
21:19:44 <Vorpal> Gregor, what js engine do you use?
21:19:45 <olsner> oerjan: ooh, that's clever
21:20:24 <Gregor> As a professional in the field of JavaScript behavior: Screw V8 :P
21:20:44 <Vorpal> Gregor, it is probably faster though
21:20:55 <olsner> oh, V8 is broken? nice :D
21:21:06 <Gregor> Didn't mean for that to be caps.
21:21:18 <Gregor> Vorpal: But since all of my time is spent in SDL, that's irrelevant.
21:21:32 <olsner> [from] whence Screw V8 then?
21:21:32 * elliott wonders if Gregor will bother qualifying his statements about v8
21:22:20 <Gregor> V8 has great performance in a possibly-general-case-but-maybe-not-it's-never-really-been-quantified, and EXTREMELY bad performance in other cases.
21:22:51 <Gregor> With semantically-innocuous changes to JS code, I can make v8 balloon up 30x.
21:23:01 <Gregor> I can't get SpiderMonkey to go more than 4x.
21:23:11 -!- GreaseMonkey has joined.
21:23:11 -!- GreaseMonkey has quit (Changing host).
21:23:11 -!- GreaseMonkey has joined.
21:23:16 <olsner> you should write a JS benchmark :D
21:23:26 <Gregor> olsner: See my upcoming PLDI paper (if it's accepted)
21:23:54 <Gregor> Vorpal: http://google.com/search?q=pldi
21:24:16 <Vorpal> Gregor, pldi.net or the sigplan link?
21:24:16 <quintopia> gregor: runs smoother in chromium, but letters are still about a half second ahead of music. got a 362174 somehow
21:24:22 <Gregor> Vorpal: The SIGPLAN link X_X
21:24:31 <quintopia> those fast sections should have the letters moving faster :/
21:24:32 <Vorpal> Gregor, there was also a stanford link
21:24:40 <Gregor> quintopia: A half second? That's mega-screwy, something weird in your audio.
21:24:46 <Gregor> Vorpal: That's just a particular instance of the conference.
21:25:12 <olsner> http://sss.cs.purdue.edu/projects/dynjs/ looks like a good link
21:25:14 <Vorpal> Gregor, ah, sigplan is acm, meaning I should be able to get it through the university proxy if it is acccepted
21:26:19 <elliott> Gregor: You shoulda put up a paper online, changed one wording in one paragraph and then submitted it :P
21:26:52 <Vorpal> elliott, fail indeed :P
21:26:54 <elliott> Is http://sss.cs.purdue.edu/projects/dynjs/ it?
21:27:15 <elliott> Gregor: Is that TeX with Times?
21:27:40 <Vorpal> elliott, slanted, where?
21:27:43 <Gregor> That is LaTeX with SIGPLAN's instituted templates.
21:27:54 <elliott> the text is slanted slightly
21:28:06 <Vorpal> I don't see it. Oh well
21:28:16 <Gregor> elliott: I think your brain is slanted slightly.
21:28:16 <Vorpal> elliott, TeX with Comic Sans would have been worse anyway
21:28:43 <Gregor> Also, at least it's not Word :P
21:28:49 <Vorpal> elliott, my pdf reader does not list times as being embedded
21:29:14 <Vorpal> there is NumbusRomNo9L and such
21:29:57 <Vorpal> elliott, I use qwerty yes :P
21:30:37 <Vorpal> elliott, as long as it isn't nubus I'm happy
21:30:37 <elliott> seriously? Nimbus is just an imitation of Times.
21:30:54 -!- augur has quit (Remote host closed the connection).
21:31:09 <Gregor> Comic Serif OH GOD WHY
21:31:17 <elliott> http://www.swiss-miss.com/2008/04/helvetica-serif.html
21:31:18 -!- augur has joined.
21:31:24 <Vorpal> elliott, Comic Antiqua! (spelling?)
21:33:07 -!- wareya has quit (Read error: Connection reset by peer).
21:33:32 -!- wareya has joined.
21:34:37 <Gregor> elliott: I totally lied.
21:34:39 <Gregor> `run /opt/python27/bin/python -c 'print "Hello"'
21:35:09 <Vorpal> Gregor, how did you lie?
21:35:23 <Gregor> Vorpal: I said I was going to just install a newer system Python. But I did not :P
21:35:36 <Vorpal> Gregor, so you compiled from source in the end?
21:35:49 <elliott> Gregor: Now replace the normal python with it :P
21:35:56 <Gregor> elliott: Not gonna happen.
21:36:17 <Gregor> elliott: Because that's part of the system, and since the upgrade went all kerplutz I've decided maybe I just want to leave that as-is.
21:36:37 <Gregor> If you want /usr/bin/python2.7, that can be arranged.
21:36:44 <elliott> Gregor: Okay, put /opt/python27/bin first in the PATH.
21:36:58 <Gregor> Actually, that most certainly can be arranged.
21:37:26 <cpressey> Yeah, I had some fun after moving my system python to 2.7
21:37:56 <cpressey> Just putting a Python 2.7 called 'python' in my ~/bin which is on my path
21:38:17 <cpressey> Suddenly #/usr/env/bin python has EXCITING NEW SEMANTICS
21:39:24 <Gregor> cpressey: Luckily, putting it in this PATH is more OK as it's only used for things within HackEgo :P
21:39:32 <Vorpal> cpressey, new semantics? how so?
21:39:44 <cpressey> Gregor: I'm sorry to hear that. (ok, not really)
21:39:46 <pikhq> Vorpal: Suddenly, it pulls up the python in his ~
21:39:57 <Vorpal> pikhq, yes that is obvious
21:40:04 <Vorpal> pikhq, I meant, "it isn't like it went to python3"
21:40:10 <Vorpal> so things should just continue to work
21:40:20 <pikhq> Point releases can still break things sometimes.
21:40:24 <cpressey> Vorpal: tell that to ubuntu system tools
21:41:25 <HackEgo> Traceback (most recent call last): \ File "/tmp/hackenv.15489/bin/wl", line 37, in <module> \ q = query(continue_id) \ File "/tmp/hackenv.15489/bin/wl", line 29, in query \ response = urllib2.urlopen(url).read() \ File "/opt/python27/lib/python2.7/urllib2.py", line 126, in urlopen \ return _opener.open(url,
21:41:30 <cpressey> we totally need #!/usr/bin/guess_format_and_interpret
21:42:09 <pikhq> We totally need an OS with sufficient metadata that the format of every file can be known.
21:42:21 <pikhq> Of course, I also totally need a pony.
21:42:45 <elliott> `run wl 'sv sill' 2>&1 | tr -d '\n'
21:42:47 <HackEgo> Traceback (most recent call last): File "/tmp/hackenv.15552/bin/wl", line 37, in <module> q = query(continue_id) File "/tmp/hackenv.15552/bin/wl", line 29, in query response = urllib2.urlopen(url).read() File "/opt/python27/lib/python2.7/urllib2.py", line 126, in urlopen return _opener.open(url, data, timeout)AttributeError:
21:42:57 <elliott> `run wl 'sv sill' 2>&1 | tr -d '\n' | tail -c +50
21:43:00 <HackEgo> ckenv.15603/bin/wl", line 37, in <module> q = query(continue_id) File "/tmp/hackenv.15603/bin/wl", line 29, in query response = urllib2.urlopen(url).read() File "/opt/python27/lib/python2.7/urllib2.py", line 126, in urlopen return _opener.open(url, data, timeout)AttributeError: ProxyHandler instance has no attribute 'open'
21:46:38 <Gregor> DISCOVERY: Browser that supports HTML5 Audio + PCM .wav + data: URLs = do client-side loop unrolling!
21:46:38 <Gregor> (More like OBSERVATION)
21:54:05 <Gregor> To loop oggs in a web browser: 1) download ogg via XHR, 2) decompress ogg with JS code, 3) duplicate decompressed audio a bunch of times, 4) write that into a data: URL as a PCM .wav, 5) load that into HTML5 Audio
21:55:34 <Gregor> It's only too bad you can't stream like that.
21:58:30 <pikhq> elliott: Well, I must say. Xubuntu > Debian when it comes to "just works".
21:58:37 <Gregor> He's a homeless, jobless layabout living off social security and drugs.
21:59:01 -!- augur has quit (Ping timeout: 252 seconds).
22:01:52 <Phantom_Hoover> Gregor, do you and him actually hate each other, or is it just a joke?
22:02:14 <oerjan> Phantom_Hoover: we all hate each other here. especially you.
22:02:38 <oerjan> hope that cleared things up
22:02:56 <Gregor> Fuck you oerjan, we don't hate each other!
22:03:12 <oerjan> ...i hate people who cannot even follow our hate policy properly
22:03:19 <olsner> Gregor: I have now read your paper
22:04:30 <Gregor> olsner: Congratulations on your literacy.
22:04:51 <olsner> thanks :D I'm quite proud of it myself
22:05:19 <oerjan> olsner is an ace at littering
22:05:53 <pikhq> WHAT THE HELL WHY IS FLASH IN XUBUNTU ACTUALLY USING CPU WELL
22:06:01 <pikhq> IT IS NOT RAPING MY CPU
22:06:09 <pikhq> UNLIKE ON REAL HARDWARE
22:06:32 <oerjan> this is virtually unheard of!
22:06:47 <pikhq> It's using like 30% CPU for Youtube. Note: Flash in Gentoo uses like 110% CPU.
22:07:04 -!- augur has joined.
22:09:13 <olsner> java is alright, but surely not *applets*?
22:10:20 <Gregor> Now that we're done hatin' on each other, anybody have an opinion on my crazy JS-data-URL-wtf plan? :P
22:11:36 <pikhq> Gregor: Delicious.
22:12:22 <oerjan> now embed that in MIPS and we're all set
22:13:06 <Gregor> Phantom_Hoover: Idonno, that's why I made up nonsense :P
22:13:15 <Gregor> No one but cpressey is privy to that information.
22:14:32 <oerjan> even cpressey doesn't know, the nsa wipe his memory whenever he leaves and reinstate it when he comes back
22:15:07 <Gregor> oerjan: Though leaving him with an imprint that will lead him to have an inexplicable desire to come back the next day.
22:15:25 <oerjan> well that part was obvious
22:15:51 <olsner> the other cpressey probably thinks he's discussing esoteric knitting patterns all night
22:16:01 <Gregor> The only way you can stop being an ex-anything is by becoming one again ...
22:16:35 <elliott> Gregor: That didn't work to get your ex-wife back now, did it?
22:16:53 <Gregor> elliott: No, I became my wife.
22:17:17 <elliott> Also illegal in 13 states.
22:18:16 <Phantom_Hoover> oerjan, so do you sit with a sign saying "will simplify for food"?
22:18:48 <oerjan> no no, it's "will make incomprehensible for food"
22:21:53 <cpressey> < Phantom_Hoover> Someone tell me what cpressey actually does.
22:23:06 <olsner> oh, you're a musician?
22:23:21 <Vorpal> Phantom_Hoover, you mean what he does for a living: well, I know he complained about python in a work-related context. I think he works as a programmer.
22:24:44 <elliott> <elliott> Gregor studies Ph.D.s
22:24:45 <elliott> <elliott> cpressey does Python
22:25:07 <Vorpal> elliott, he studies people with Ph.D.?
22:25:23 <elliott> Vorpal: No, he studies Ph.D.s themselves.
22:25:23 <cpressey> olsner: you haven't heard my MED songs!
22:25:37 <elliott> foetus is a hypercorrection
22:25:54 <Vorpal> elliott, ah. The abstract Ph.D.s are quite tricky to study I heard. Takes a lot of skill.
22:26:25 -!- GreaseMonkey has quit (Remote host closed the connection).
22:27:37 <Vorpal> elliott, that is indeed more tangible
22:27:44 <elliott> Someone appreciate this :(
22:27:48 <fizzie> Gregor: With a fast enough Javascript engine, you could have a single <audio> tag playing a long-enough constant-value audio file, and then just use the scripting to modify the volume property for each sample.
22:27:57 <Vorpal> elliott, I tried to figure it out and gave up
22:28:13 <elliott> [[Deja dit (pronounced dé-ja-di) Even though deja dit is French for "already said", it actually is used to describe the strange feeling you get when you're in a situation, and feel like you've been in the exact same situation before and spoken the same words as if you have studied them and just repeated them.]]
22:28:16 <elliott> The literal meaning works :P
22:28:19 <Gregor> fizzie: Also, HTML5 Audio is made of fail, that wouldn't work (in any current browser)
22:28:31 <cpressey> i never ever ever get that feeling
22:28:35 <cpressey> but i get deja vu all the time
22:28:40 <Vorpal> elliott, yes I know what deja vu is
22:28:43 <fizzie> Gregor: You could even record the constant beep from a PC speaker for authentic speaker-style music done that way.
22:29:09 <pikhq> Such a shame HTML5 audio has no API for generating audio.
22:29:20 <Gregor> pikhq: data URLs MIGHT BE IT!
22:29:21 <elliott> Often I get dream dejav u.
22:29:23 <pikhq> Except, of course, generating a data URI at runtime.
22:29:31 <Gregor> However, yes, an actual API would be nice :P
22:29:31 <elliott> As in, as I remember a dream, I swear I've had it before.
22:29:39 <elliott> Gregor: What happens when the <video> runs out?
22:29:43 <elliott> i.e. the unrolled loop finishes
22:29:50 <elliott> i.e. it is replayed all 9045378957389573495 timse
22:29:52 <Gregor> elliott: That's the no-streaming issue.
22:29:56 -!- antivigilante has joined.
22:30:04 <pikhq> Won't work on Internet Explorer, though.
22:30:12 <Gregor> pikhq: It should on IE9
22:30:15 <pikhq> They limit data URIs to 256 kb.
22:30:40 <fizzie> Gregor: There's the mozilla-specific extension for it, https://developer.mozilla.org/en/Introducing_the_Audio_API_Extension -- but that's cheating, of course.
22:30:45 <elliott> Gregor: Can't you just insert an identical <audio>, play it to 1 second (have the first second be silence), and pause it, ready to play when the other one finishes?
22:30:56 <elliott> Or does the API not support that?
22:31:19 <Gregor> elliott: You can, but its timing is garbage.
22:31:41 <Gregor> elliott: Getting the current time is godawfully broken, and there's usually a nondeterministic pause between calling the play function and the sound actually playing.
22:31:42 <elliott> Gregor: How about: Start playing the new one from zero seconds when there's one second left in yours.
22:31:47 <elliott> Then have the first second be silence.
22:31:48 <pikhq> "For security reasons, data URIs are restricted to downloaded resources. Data URIs cannot be used for navigation, for scripting, or to populate frame or iframe elements."
22:31:55 <pikhq> WHY YOU ASSHOLES WHY.
22:31:56 <Gregor> elliott: I have a library that does exactly that.
22:32:07 <elliott> pikhq: Like Xubuntu, huh? :P
22:32:10 <Gregor> elliott: Much better in Flash than in HTML5.
22:32:17 <elliott> Lame @ no data <iframe>s >:D
22:32:25 <elliott> (I have no idea what you'd do with a data: <iframe>.)
22:32:38 <pikhq> elliott: It seems to just work, which is a nice change.
22:33:08 <pikhq> elliott: Unfortunately, its multiarch support sucks, which means I *will* find it not working, and I *will* want to throw things at people after that.
22:33:10 <elliott> pikhq: If your VM is up to it, try Ubuntu proper; it sucks, but it's worth admiring the suck, right?
22:33:38 <elliott> Wow, Xubuntu 10.10's default wallpaper is hideous with those panels.
22:34:03 <elliott> pikhq: 9.10 is nicer: http://upload.wikimedia.org/wikipedia/commons/6/68/Xubuntu-karmic.png
22:34:10 <Vorpal> elliott, the hidious thingy
22:34:23 <elliott> http://www.google.co.uk/images?q=xubuntu+10.10&um=1&ie=UTF-8&source=og&sa=N&hl=en&tab=wi&biw=1366&bih=629
22:34:28 <pikhq> I had to switch to Mist to make it not painful.
22:34:36 <elliott> pikhq: As are the window borders.
22:34:45 <pikhq> But, of course, I'd have to switch to Mist anyways.
22:35:20 <elliott> (Actually I'll probably make a new theme for Kitten, like Grey Mist but... more polished.)
22:35:23 <pikhq> Yeah, yeah, I know.
22:35:45 <elliott> pikhq: Do you have any opinions on SERVICE MANAGEMENT
22:35:57 <Vorpal> pikhq, elliott: no love for themes like clearlooks or similar?
22:36:02 <elliott> Phantom_Hoover: http://www.google.co.uk/images?q=xubuntu+10.10&um=1&ie=UTF-8&source=og&sa=N&hl=en&tab=wi&biw=1366&bih=629
22:36:05 <elliott> Vorpal: clearlooks is unbearably cluttered.
22:36:17 <pikhq> Vorpal: I don't think you realise how nice the Mist engine is.
22:36:25 <elliott> It was clearly designed by someone who didn't raelise that putting LINES EVERYWHERE does not aid usability.
22:36:37 <Vorpal> pikhq, I really like the light and airy feeling of clearlooks
22:36:40 <elliott> One needs a warped definition of clear to call Clearlooks clear.
22:36:45 <elliott> Vorpal: Mist is "light and airy".
22:36:52 <elliott> It's the exact same colour scheme as Clearlooks.
22:36:55 <Vorpal> elliott, hm, okay, screenshot?
22:37:02 <elliott> Vorpal: System -> Preferences -> Appearance
22:37:30 <Vorpal> elliott, like clearlooks but more blocky in other words
22:37:54 <elliott> How about I stop talking about this and you stop talking about this otherwise I'll have to pstrangle you.
22:38:19 <elliott> pikhq: SERVICE MANAGERS DOOD
22:38:21 <Vorpal> elliott, how do you screenshot one window in gnome?
22:38:41 <elliott> Vorpal: use the Take Screenshot application
22:39:02 <Vorpal> I pressed prtsc, is that a different app? Since there is no option for it there
22:39:40 <elliott> It is the post-screenshot screen of Take Screenshot.
22:39:45 <elliott> Try my advice before questioning it.
22:40:56 <Vorpal> elliott, okay what about calling mist more rectangular then? Would you be okay with that?
22:41:03 <elliott> http://www.ultimatepp.org/L$www$uppweb$chss$en-us.html_12.png Mist
22:41:29 <elliott> It does need a bit of polish.
22:41:36 <Vorpal> I prefer the "rounder" feeling of clearlooks to be honest
22:41:39 <elliott> pikhq: SEEERVIIICE MAAANAAGEEEEERRRRSSS
22:41:55 <pikhq> elliott: MAKE IT WORK MAGICALLY
22:42:00 <elliott> pikhq: THAT IS LESS HELPFUL
22:42:04 <Vorpal> elliott, upstart is actually damn nice
22:42:08 <Vorpal> elliott, did that help?
22:42:14 <pikhq> Vorpal: I like the complete and utter lack of flashiness of Mist, myself.
22:42:15 <Vorpal> I looked at how the init scripts work and such
22:42:17 <elliott> Vorpal: No. Upstart is nowhere near as nice as Kitten's.
22:42:35 <elliott> And the sysv init directory style has always been and will always be an abomination.
22:42:38 <Vorpal> elliott, okay, I give it the benefit of not-having-seen-it
22:42:47 <elliott> As is the stupid set of runlevels.
22:42:50 <Vorpal> elliott, and upstart doesn't use the sysv init style
22:42:57 <elliott> A bunch of numbers ... and S!
22:43:15 <elliott> Vorpal: Besides, service managers don't just initialise.
22:43:25 <Vorpal> elliott, upstart does that :P
22:43:39 <elliott> Vorpal: /etc/init.d isn't even close to what I mean.
22:43:49 <Vorpal> elliott, I don't mean /etc/init.d either
22:43:56 <Vorpal> which is *completely* different
22:44:10 <Vorpal> elliott, upstart uses /etc/init, sysv uses /etc/init.d
22:44:12 <elliott> That has nothing to do with managing services post-boot.
22:44:19 <elliott> That's a bunch of .conf files.
22:44:34 <pikhq> Which upstart, in fact, does.
22:44:49 <elliott> Yes. But not with /etc/init.
22:44:58 <elliott> unless i'm sorely mistaken
22:45:04 <Vorpal> elliott, you do: stop foo
22:45:12 <Vorpal> and it magically looks at /etc/init/foo
22:45:19 <elliott> (Woot, polluting the global namespace.)
22:45:22 <Vorpal> actually stop is a symlink to initctl
22:45:27 <Vorpal> elliott, yeah I dislike that too
22:45:40 <elliott> Anyway it's simply not as advanced as my system.
22:45:45 <elliott> Mine is more like running svscan as process 1.
22:45:59 <pikhq> elliott: "stop foo"
22:46:13 <pikhq> Oh, what vorpal said.
22:46:13 <elliott> <elliott> Yes, I know of initctl.
22:46:19 <elliott> <elliott> Anyway it's simply not as advanced as my system.
22:46:20 <elliott> <elliott> Mine is more like running svscan as process 1.
22:46:38 <elliott> coppro: i'm taking a break to keep my sanity for a bit
22:46:59 <elliott> coppro: ideally, the new update that makes wine work perfectly will come out before I decide to try again :)
22:46:59 <Vorpal> Every five seconds, svscan checks for subdirectories again. If it sees a new subdirectory, it starts a new supervise process. If it sees an old subdirectory where a supervise process has exited, it restarts the supervise process. In the log case it reuses the same pipe so that no data is lost. <-- um, polling? ugh
22:47:13 <Vorpal> inotify and wait() should cover it
22:47:51 <elliott> Shut up about implementation details and respect djb's skills.
22:47:57 <Vorpal> elliott, okay wtf: "svscan starts one supervise process for each subdirectory of the current directory, up to a limit of 1000 subdirectories. "
22:48:02 <Vorpal> elliott, I'm a djb fan in general
22:48:11 <cpressey> daemontools was way before notification was anything like widespread
22:48:14 <Vorpal> elliott, but not in the specific case of svscan
22:48:28 <Vorpal> but what about the limit of 1000 subdirs?
22:48:28 <pikhq> It should use inotify now, though.
22:48:39 <elliott> djb doesn't really change his software if it works.
22:48:46 <elliott> especially since inotify is *linux specific*
22:48:58 <Vorpal> elliott, true, you want kqueue on *bsd
22:48:59 <elliott> which nobody has pointed out yet
22:49:11 <Vorpal> elliott, there are abstracting libraries
22:49:12 <elliott> Vorpal: i rather think djb's philosophy is to use something that works the same everywhere.
22:49:12 <cpressey> you are vehemently free to write and distribute a patch that makes it use inotify
22:49:20 <pikhq> elliott: Still, *polling*.
22:49:32 <elliott> pikhq: I don't get this pollophobia everyone has.
22:49:41 <elliott> How do you think your mouse works?
22:49:42 <Vorpal> elliott, "bad on laptops"
22:50:01 <pikhq> elliott: Depends on the mouse, actually.
22:50:03 <elliott> Anyway, I never said I would use svscan.
22:50:09 <elliott> pikhq: You have a continuous mouse?
22:50:09 <cpressey> the only cron i've seen that doesn't poll is an experimental one i wrote in erlang for fun
22:50:13 <Vorpal> elliott, there is a difference between waking up a small chipset in your mouse and waking up a bloody power hungry x86 CPU
22:50:21 <pikhq> elliott: A PS/2 mouse generates interrupts, IIRC.
22:50:30 <Vorpal> pikhq, so does USB mice, I checked
22:50:33 <elliott> pikhq: At certain intervals of movement.
22:50:46 <elliott> Anyway, those are implementation details and irrelevant to my description of Kitten's system.
22:50:47 <pikhq> elliott: ... Yeeessss...
22:50:57 <cpressey> (my cron in erlang (crone) computed the time to the next event due and slept exactly that much :)
22:51:03 <elliott> pikhq: Your mouse polls the mechanics inside...
22:51:22 <Vorpal> elliott, again: <Vorpal> elliott, there is a difference between waking up a small chipset in your mouse and waking up a bloody power hungry x86 CPU
22:51:24 <pikhq> elliott: Yes, you get polling from the single-purpose CPU. Hooray.
22:51:46 <pikhq> elliott: The one that doesn't even have power control.
22:52:04 <elliott> I'm not saying it should poll.
22:52:12 <elliott> I'm just saying that polling every five seconds isn't the worst damn thing a program can do.
22:52:20 <Vorpal> true, it could busy loop
22:52:21 <elliott> Especially if there was no other alternative at the time.
22:52:26 <Vorpal> (but that would be absurd)
22:52:40 <pikhq> Vorpal: Windows 95 does.
22:52:59 <Vorpal> pikhq, okay well so did DOS
22:56:20 <cpressey> i am quite interested in asynchronous (in the sense of clockless) processors, but they have their share of downsides
22:56:42 <elliott> pikhq: Question: why do all package description formats suck?
22:56:45 <Vorpal> cpressey, iirc ais523 is somewhat of an expert on those
22:57:10 <Vorpal> cpressey, well, as least having lectured about such things unless I completely misremember
22:57:16 <Vorpal> expert may be stretching it a bit
22:57:43 <Vorpal> cpressey, and I find them interesting too
22:58:15 <Vorpal> cpressey, but implementing anything like a pipeline in them seems near impossible
22:58:31 <Vorpal> (amongst other issues)
22:58:49 <cpressey> implementing all pipeline-like circuits in them is impossible
22:59:06 <cpressey> (i don't know that. i am just talking.)
22:59:28 <Vorpal> cpressey, well, don't know about impossible, very hard though
23:00:21 <cpressey> well, pipelines assume instructions, instructions assume a processor, a processor assumes a clock -- what we have here is a mismatch of expectations
23:01:01 <cpressey> if you go clockless you probably want to get away from all the stuff we've developed under the clocked model
23:01:25 <Vorpal> cpressey, you need some way to control the processor though, be it instructions or something else
23:01:27 <elliott> cpressey: Infinitely Long Instruction Word
23:01:35 <Vorpal> (otherwise it is bloodly useless)
23:02:01 <Vorpal> elliott, hm ILIW does have a certain ring to it
23:02:47 <cpressey> Vorpal: you "control" it by configuring it how to react
23:03:03 <Vorpal> cpressey, so an advanced FPGA basically?
23:03:06 <elliott> cpressey: event-based processor, i approve
23:04:24 <elliott> pikhq: Do you want to hear my crazy plan for integrating package management and configuration management? -- also because what?
23:04:28 <Vorpal> cpressey, needs to the self-programmable of course
23:04:53 <Vorpal> cpressey, never seen (or heard of) a self-programmable FPGA. I wonder why not
23:05:22 <elliott> > I am taking a holiday next week. For that time, I think cvs will be
23:05:30 <elliott> -- Theo de Raadt, August 2002
23:05:40 <cpressey> Vorpal: i think i've seen something similar to an FPGA but where the gates are set with static RAM -- that would be the basic idea. expensive to do densely, though...
23:05:45 <elliott> > I am locking all the trees until the development community decides
23:05:45 <elliott> > how future releases will be done.
23:05:49 <elliott> -- Theo de Raadt, August 2010
23:06:05 <elliott> > FUCKING DEVELOPERS WHY DO THEY CHANGE MY OPERATING SYSTEM *CRY*
23:06:12 <elliott> -- Theo de Raadt, internal monologue
23:06:39 <Vorpal> cpressey, expensive indeed
23:06:51 <cpressey> i haz the keez to de treez and i LOCKS DEM
23:06:56 <elliott> Note to self: Don't use OpenBSD for anything.
23:07:12 <cpressey> "Best to gaze askance, or use a mirror"
23:07:21 <Vorpal> elliott, yeah freebsd picked up the useful bits, such as pf
23:07:27 <pikhq> elliott: Jeeze, it's like he *wants* a fork.
23:07:40 <elliott> cpressey: For now we see in a mirror darkly!
23:08:21 <elliott> Vorpal: Well. I don't much like FreeBSD myself.
23:08:27 <elliott> cpressey basically summed it up with "the Linux of the BSDs" :)
23:08:42 <elliott> NetBSD kernel + Kitten pkg(1)! wooooo
23:08:44 <pikhq> Ports kinda sucks.
23:08:46 <Vorpal> elliott, it has however two excellent things: pf and the best network config ever
23:08:49 <elliott> i forget the historical /sbin manpage section
23:09:15 <elliott> pikhq: who would fork openbsd though
23:09:16 <Vorpal> I believe some sbin commands are (8) on linux
23:09:23 <elliott> it's been clear for a while that their security is basically just bluffing
23:09:35 <elliott> and the other BSDs should have basically the same
23:10:05 <Vorpal> elliott, that is fairly new, so not really historical at al
23:10:13 <cpressey> elliott: they all kind of steal from each other anyway, so the rivalries between them look silly
23:10:18 <cpressey> maybe that accts for the drama
23:10:29 <elliott> Five years without a remote hole in the default install!
23:10:30 <elliott> One remote hole in the default install, in nearly 6 years!
23:10:32 <elliott> Only two remote holes in the default install, in a heck of a long time!
23:10:38 <elliott> ...too bad the default install is useless!
23:11:33 <elliott> pikhq: I GUESS YOU DON'T WANT TO KNOW ABOUT PKG(8) ;___;
23:11:50 <Vorpal> elliott, pf is a nice firewall however, easier to use than netfilter by far
23:12:17 <elliott> http://mail-index.netbsd.org/netbsd-announce/2010/09/13/msg000110.html
23:12:19 <pikhq> elliott: Well, they do have a large number of things that actually make it more secure... Most of which are now in other *NIX, but hey.
23:12:19 <elliott> Introducing NPF, NetBSD's new packet filter
23:12:21 <Vorpal> elliott, and is the network config anything like rc.conf of freebesd?
23:12:45 -!- tombom has quit (Read error: Connection reset by peer).
23:12:47 <pikhq> I don't think anybody's done anything similar to their X sandboxing, though.
23:12:53 <elliott> "NPF is the third packet filter in NetBSD, after IP Filter and PF."
23:13:20 <elliott> The issue with NetBSD is that the default "distribution" is a bit out of date.
23:13:29 <elliott> iirc their X.Org was about three years old.
23:13:43 <elliott> that's why you use Kitten :)
23:14:32 <elliott> do the other BSDs have NetBSD's nice driver architecture?
23:14:42 <elliott> where the meat of the driver code is separated from platform-specific stuff
23:14:53 <elliott> so you can use the same hardware on different architectures and buses with minimal to no porting
23:15:03 <pikhq> No, that's pretty exclusive to NetBSD.
23:15:09 <elliott> see, NetBSD! it's awesome!
23:15:35 -!- cpressey has quit (Ping timeout: 260 seconds).
23:15:44 <pikhq> It's such a shame you won't be able to get the (irritating) binary drivers for 3D cards working there, though.
23:16:01 <elliott> "This package provides the oficial NVIDIA graphic driver for
23:16:07 <elliott> God knows if it's any good.
23:16:53 <elliott> "Quentin Garnier has made a loadable kernel module (LKM) version of the NVidia drivers on NetBSD. This release is very preliminary, rough and mostly meant to test the installation procedure. You will need a NetBSD-current system but the code itself should be quite backward compatible with some caveats. For example, you need 'options KVM86' in your kernel config. Known working hardware includes RIVA TNT
23:16:53 <elliott> 2 Model 64 (PCI), GeForce2 MX/MX 400, Vanta(AGP) and more!"
23:17:08 <elliott> pikhq: you can use nouveau right?
23:17:10 <pikhq> Actually, you could probably port it. How Nvidia rolls is that they've got a binary blob that's *OS independent*, and a stub to abstract the kernel.
23:17:19 <pikhq> elliott: I use AMD ATM.
23:17:33 <elliott> pikhq: well ati opened their specs recently :D
23:17:34 <pikhq> Which... Actually sucks ass on Linux.
23:17:51 <elliott> the open source one is too, but... sometimes actually less so
23:17:56 <pikhq> radeon is a bit worse ATM.
23:18:29 <elliott> pikhq: You could always run the driver in a VM! Wait, what?
23:18:54 <elliott> ...The Linux emulation in the BSDs is amazing, though.
23:19:34 <pikhq> And it'd be even better if you had multi-arch working right. :)
23:19:34 <elliott> http://www.gnu.org/manual/manual.html -> "Ferret" -> http://www.gnuferret.org/ -> LOL SQUATTED
23:19:50 <elliott> install linux debian on bsd :D
23:20:12 <pikhq> elliott: Not without a chroot.
23:20:21 <pikhq> Well. Or complete insanity.
23:20:40 <elliott> :( NetBSD doesn't have FreeBSD jails.
23:20:40 <Vorpal> elliott, where/when will kitten be available?
23:21:04 <pikhq> It does run as a Xen Dom0, though.
23:21:13 <Vorpal> elliott, those jails are kind of shitty anyway when it comes to network config
23:21:15 <elliott> Vorpal: http://kitten.st/ or something of that nature, and as soon as I get the time. Half-term is coming up, so that'll be a week of hackery if all goes well.
23:21:25 <elliott> Yeah, but the jails were the basis of my VPS hosting strategy!
23:21:34 <pikhq> And apparently its IO routines are sufficiently good that guests sometimes *run faster as a DomU than native*.
23:22:01 <Vorpal> pikhq, sounds like it's native IO routines are rather bad
23:22:26 <elliott> you do realise that Xen can run other OSes?
23:22:27 <pikhq> Vorpal: No, NetBSD as a Dom0 makes *other* DomUs run faster than native.
23:22:30 <elliott> the guest may be linux in that case
23:22:34 <elliott> pikhq: I wish you could get BFS for NetBSd :)
23:22:47 <Vorpal> pikhq, that explains things
23:22:56 <elliott> pikhq: Speaking of BFS, I'll probably have tweaks in my X.org distribution that gives it ridiculous scheduler priority.
23:23:01 <elliott> Like BFS' recommended configuration has.
23:23:34 <pikhq> cheater: I'm pretty sure NetBSD's scheduler is better-designed than Linux's.
23:23:58 <Vorpal> pikhq, c is nowhere near e on qwerty, what layout?
23:23:58 <pikhq> elliott: Oh, awesome. NetBSD doesn't *just* have Linux emulation.
23:24:20 <Vorpal> elliott, writing a package manager would take more time than "one week of hackery"
23:24:23 <pikhq> elliott: It emulates FreeBSD, Darwin, Solaris, SunOS 4, *and* Linux system calls.
23:24:37 <elliott> Vorpal: I never said it'd be ready then.
23:24:43 <elliott> Vorpal: Besides, a simple package manager can be done very quickly.
23:24:49 <Vorpal> pikhq, not any other system?
23:24:53 <elliott> <Vorpal> oh the brain sched?
23:25:03 <elliott> Vorpal: The *Linux emulation*-alike.
23:25:05 <elliott> Vorpal: That means: native.
23:25:15 <elliott> Just syscall implementations.
23:25:22 <elliott> This is why Quake II, for instance, works perfectly on NetBSD.
23:25:30 <Vorpal> pikhq, Why not Tru64 !? It is mission critical for me!
23:25:31 <elliott> Because it just... Has implementations of the Linux syscalls.
23:25:45 <Vorpal> elliott, also I know what it means
23:25:50 <pikhq> It supports different ones for different CPU architectures.
23:25:55 -!- Phantom_Hoover has quit (Ping timeout: 252 seconds).
23:25:56 <Vorpal> elliott, and it means "some stuff works, but others don't"
23:26:02 <elliott> Vorpal: Meh. Enough works.
23:26:08 <Vorpal> elliott, X would not for example
23:26:14 <elliott> pikhq: Just say that you want me to tell you how pkg(8) does stuff and we can get it over with :P
23:26:16 <pikhq> Vorpal: No, it's actually "complete up to a certain version".
23:26:18 <elliott> Vorpal: X is available on NetBSD.
23:26:35 <elliott> Vorpal: Do you want to hear my pkg(8) chatter or should I /msg it to pikhq?
23:26:37 <Vorpal> pikhq, I'm sure it doesn't emulate stuff like /proc/mtrr or such
23:26:41 <pikhq> Vorpal: I believe they recently switched the syscall implementation from 2.4 emulation to 2.6 emulation.
23:26:51 <Vorpal> elliott, I'm okay with it
23:27:14 <pikhq> The x86 list appears to be: BSDI, FreeBSD, Interactive UNIX, SCO Unix, SCO Xenix, Linux, and Solaris.
23:27:15 <Vorpal> pikhq, syscalls alone does not make linux
23:27:30 <elliott> pikhq: Okay, so, packages. People like to think a package is one bit of software! THESE PEOPLE ARE NAZIS.
23:27:41 <Vorpal> pikhq, there is the file system hierarchy
23:27:51 <elliott> Vorpal: QUIT INTERRUPTING THE GENIUS
23:27:55 -!- antivigilante has quit (Quit: Ex-Chat).
23:27:57 <Vorpal> pikhq, /proc /sys and /dev look different
23:28:11 <Vorpal> elliott, *implements timesharing of #esoteric*
23:28:28 <pikhq> Vorpal: Which it supports by having an overlay of the filesystem so things look right.
23:28:32 <elliott> pikhq: Packages, right! A package can be one piece of software. But who says it can't handle system configuration too??
23:28:33 -!- antivigilante has joined.
23:28:49 <Vorpal> pikhq, fun, but some stuff would mean poking hardware, /proc/mtrr comes to mind again
23:28:55 <elliott> pikhq: FOR INSTANCE. There is no reason you can't have a package that implements, say, some specific configuration set for a piece of software.
23:29:28 <Vorpal> pikhq, and /dev/mem for doing stuff like messing with pci config space
23:29:36 <pikhq> Vorpal: ... Okay, and? That doesn't exactly prevent them from *implementing* that.
23:30:03 <Vorpal> pikhq, could be tricky preventing stuff acting up though
23:30:11 <elliott> pikhq: But then, consider this configuration *schema* (think axiom schemas, e.g. of ZFC): "Make the default window manager [WM]." As an example.
23:30:23 <elliott> pikhq: Do you want a custom package for every window manager? Of course not! That's ludicrous.
23:30:25 <Vorpal> pikhq, anyway, what about 3D support? Would nwn for linux run well under netbsd
23:30:41 <elliott> pikhq: Of course, you have an abstract x-window-manager package that all the WMs provide.
23:30:55 <elliott> pikhq: x-window-manager requires that the WMs provide, say, a start-executable line.
23:31:05 <elliott> dwm's start executable would be "dwm"
23:31:18 <elliott> pikhq: Then you have a "metapackage" called default-wm.
23:31:34 <Vorpal> elliott, sounds like virtual packages with a few extensions so far?
23:31:36 <elliott> default-wm is a set of packages, "default-wm X", for every WM X.
23:31:39 <pikhq> Vorpal: The 3D rendering would be done via X. And X has 3D support.
23:31:42 <elliott> It may be "default-wm --wm=X". Whatever. That's not important.
23:31:59 <elliott> pikhq: So, default-wm's install script writes X's start-executable plus other stuff to the right file, and whatnot.
23:32:12 <Vorpal> pikhq, yes obviously it would be done that way, the question was how well the thing would work. Stuff like ALSA and so on
23:32:18 <elliott> Or 'default-wm jwm', whatever.
23:32:23 <pikhq> elliott: So, what, basically metapackages can take options?
23:32:34 <elliott> pikhq: Pretty much. Yes. And every combination of options constitutes one package.
23:32:41 <Vorpal> pikhq, since it uses ALSA directly iirc, not OSS
23:32:45 -!- augur has quit (Remote host closed the connection).
23:33:04 <elliott> pikhq: And since we have a more extensive "provides x-window-manager" and other generic packages systems, that can require certain relevant properties to exist, we can write metapackages like this for many system configurations.
23:33:19 <pikhq> Vorpal: ALSA probably doesn't work. But, then, ALSA doesn't necessarily work on Linux.
23:33:21 <elliott> pikhq: You can imagine a sans-font metapackage.
23:33:24 -!- augur has joined.
23:33:33 <elliott> pikhq: Oh, and "default-wm X" would conflict with all other "default-wm Y"s.
23:33:44 <pikhq> elliott: Of course.
23:33:47 <Vorpal> pikhq, a bit unplayable then
23:33:48 <elliott> Whereas, imagine if you didn't even have a sans-font package, but a generic-font package.
23:33:53 <elliott> generic-font sans 'DejaVu Sans'
23:34:05 <elliott> generic-font X A would conflict with all other generic-font X Bs.
23:34:13 <elliott> But generic-font X A would not conflict with any generic-font Y Bs.
23:34:20 <elliott> Because "generic-font sans foo" and "generic-font sans bar" obviously conflict.
23:34:28 <elliott> But "generic-font sans foo" and "generic-font serif bar" don't.
23:34:33 -!- yorick has quit (Quit: Poef! Going to Rome, see you in 8-9 days.).
23:34:38 <Vorpal> pikhq, also I never had problems with alsa on any of my computers. (VIA on board, Intel HD Audio, Intel <pre HD thingy I forgot the name of>, SB Live! 5.1)
23:34:42 <elliott> pikhq: And of course, the package can contain as much logic as it needs to make this all work.
23:34:50 <elliott> pikhq: Or perhaps even a gcc package.
23:34:53 <Vorpal> pikhq, in all cases ALSA worked perfectly after perhaps adjusting some mixer levels
23:34:55 <pikhq> Vorpal: The point is that ALSA is *entirely optional* on Linux.
23:35:01 <elliott> pikhq: With "host" and "target" options.
23:35:06 <elliott> pikhq: One package for every cross-compiler.
23:35:09 <Vorpal> pikhq, only in theory, not in practise
23:35:15 <elliott> pikhq: But all in one metapackage.
23:35:23 <elliott> You have "cross-gcc x86 arm" installed.
23:35:24 <pikhq> Vorpal: It's one of two possible kernel-level sound systems.
23:35:26 <Vorpal> pikhq, well, there is OSSv4, but...
23:35:29 <elliott> But there's only one package behind it all, "cross-gcc".
23:35:36 <elliott> pikhq: Voila, cross-compilation is suddenly sane!
23:35:43 <pikhq> Hrm. Wait. ALSA's userspace stuff can be told to output to OSS, can't it? Victory!
23:35:56 <pikhq> elliott: Beautiful.
23:36:07 <elliott> And OSSv4 has its own ALSA-emulation library that basically-works most of the time.
23:36:15 <elliott> But it's easiest just to configure ALSA itself, yes.
23:36:26 <pikhq> elliott: Though you'd want cross-gcc i686-netbsd-kitten arm-netbsd-kitten
23:36:48 <elliott> cross-gcc --host=i686-netbsd-kitten --target=arm-netbsd-kitten.
23:36:55 <elliott> HOST is just what format the resulting gcc is, right?
23:37:03 <elliott> Don't need a specific option for it.
23:37:07 <elliott> That's already in the package manager.
23:37:19 <pikhq> elliott: Canadian cross.
23:37:39 <elliott> pikhq: That's where you compile a compiler for Z whose architecture is Y, on an X system, right?
23:37:54 <elliott> pikhq: Here's a Canadian cross for you:
23:37:54 <pikhq> You could support *Canadian* cross correctly. :)
23:38:17 <elliott> # pkg ins gcc-cross --arch=x86 --host=ppc-... --target=arm-...
23:38:25 <elliott> pikhq: (in --arch, "foo" is shorthand for "foo-netbsd-kitten")
23:38:29 <elliott> Or whatever the triple is.
23:38:34 <elliott> pikhq: Although do I have to say i686?
23:38:38 <elliott> I want a name for generic x86, dammit!
23:38:46 <elliott> And I refuse to play GNU's evil game!
23:38:54 <pikhq> elliott: Sorry, but i386, i486, i586, and i686 are different ABIs.
23:39:06 <olsner> hmm, of everything supports cross-compilation, then --arch=x would map to --host= rather than --build=
23:39:06 <elliott> YOUR MOM'S A DIFFERENT ABI
23:39:11 <elliott> pikhq: What's the triple element for x86-64?
23:39:16 <elliott> Please don't say x86_64 with an underscore.
23:39:29 <elliott> # pkg ins gcc-cross --arch=x86 --build=ppc-... --target=arm-...
23:39:32 <elliott> --arch being part of the package manager.
23:39:40 <elliott> pikhq: BUT I HATE THE UNDERSCORE :'(
23:39:47 <elliott> # pkg ins gcc-cross --arch=i686 --host=ppc-... --target=arm-...
23:39:58 <elliott> That installs the package "gcc-cross --arch=i686 --host=ppc-... --target=arm-...".
23:40:09 <elliott> Would install the package "gcc --arch=i686".
23:40:23 <elliott> pikhq: Technically I could consider all gccs cross-compilers, but I don't think gcc's build system agrees.
23:40:31 <elliott> Would install the package "gcc --arch=x86_64-netbsd-kitten".
23:40:34 <elliott> Assuming you're on x86-64.
23:40:37 <elliott> Whatever. You get the idea.
23:40:44 <pikhq> elliott: It only almost agrees.
23:41:10 <elliott> pikhq: You know, Go's solution to this is so much better.
23:41:23 -!- Sgeo has joined.
23:41:23 <elliott> pikhq: It uses the Plan 9 solution: All compilers are just regular programs that build on whatever architecture you want.
23:41:34 <elliott> They start with a number or letter denoting the architecture, followed by the one-letter program name.
23:41:41 <elliott> e.g. l for linker, c for C compiler, g for Go.
23:41:48 <elliott> pikhq: And so... you just... compile them.
23:41:51 <olsner> how do you select which gcc to use on a multiarch installation? can you choose which one to use of the x86-to-x86_64 cross compiler or the x86_64 non-cross compiler?
23:42:14 <elliott> olsner: the x86-to-x86_64 one is an x86 binary that compiles to x86_64 code, right?
23:42:37 <elliott> It'd be in /arch/i686-netbsd-kitten/bin/x86_64-netbsd-kitten-gcc.
23:42:40 <elliott> Or does the gcc part come first?
23:42:41 <olsner> I suppose it could even be arm-to-x86_64 running in qemu, if this is done right
23:43:02 <olsner> I mean, can you tell the package manager to do it?
23:43:03 <elliott> Although, it'd be *-netbsd or *-newlib really, right?
23:43:12 <elliott> olsner: oh, you mean, what to compile a package with?
23:43:38 <elliott> --arch=foo will just use a native compiler targeting foo if it can get away with it, i.e. /bin/foo-gcc.
23:43:42 <pikhq> olsner: Depends on the path; the prefix is only based on what the ABI tuple is.
23:43:57 <elliott> I guess you could do --cc=/arch/i686-netbsd-kitten/bin/...
23:44:13 <pikhq> elliott: What, is /bin/ a symlink to the native arch's bin?
23:44:21 * Sgeo vaguely wonders what's with elliott's new nick
23:44:29 <elliott> pikhq: I'm pretty sure not. Because I'm not *that* crazy about multilib.
23:44:37 <elliott> Sure, it's nice, but I refuse to design an OS around it.
23:44:42 <elliott> pikhq: Basically, you'll never see /arch/NATIVE.
23:44:50 <elliott> pikhq: /arch/NATIVE may be a symlink to /.
23:44:52 <pikhq> elliott: But then you'd need different binary packages for native or non-native installs.
23:44:57 <catseye> Why is it... n/m I'll google
23:45:08 <pikhq> elliott: But programs aren't.
23:45:15 <pikhq> elliott: They *really truly do hardcode paths*.
23:45:22 <olsner> well, let's say the 64-bit compiler is faster, then you'd want to use a 64-bit-hosted compiler for both 32-bit and 64-bit builds
23:45:24 <elliott> pikhq: Oh, *binary* packages.
23:45:34 <elliott> pikhq: Right, well, if they hardcode paths, that's a critical bug and we patch it.
23:45:43 <elliott> pikhq: Or maybe I just make the goddamn thing source-based.
23:45:49 <pikhq> elliott: This is very hard to root out sometimes.
23:45:55 <elliott> pikhq: Or maybe I just refuse to let people use such stupid programs.
23:46:03 <pikhq> Making /arch/NATIVE a symlink to / should fix it.
23:46:18 <elliott> But I won't tell the programs they're in /arch/NATIVE, that's ridiculous.
23:46:30 <elliott> pikhq: I like multilib, right? I just don't want to structure everything around it.
23:46:39 <pikhq> They will have to think they're in /arch/NATIVE or else things break.
23:46:48 <elliott> pikhq: Fine: Broken programs think they're in /arch/NATIVE.
23:46:49 <pikhq> ... Actually, the first thing to break is dynamic linking.
23:46:57 <pikhq> The path to the dynamic linker is hardcoded.
23:47:06 <catseye> "This is called a `Canadian Cross' because at the time a name was needed, Canada had three national parties. "
23:47:22 <elliott> pikhq: I'm going to commit suicide now.
23:47:36 <Sgeo> Radeon HD 4200
23:47:43 <elliott> pikhq: I think I've just tipped over the edge to "let's just have a tool to set up a chroot".
23:47:54 <pikhq> elliott: NO STOP IT THATS EVEN WORSE
23:48:16 <catseye> elliott: you need to stick the adjective "one-button" in there somewhere
23:48:45 <pikhq> elliott: With multilib set up right, programs can depend on binaries of *any* architecture but libraries of the same architecture, for one.
23:48:55 <olsner> hmm, on a multiarch system you could do a canadian cross where the compiler that compiles for the "build" ABI is actually a cross compiler... another level of cross compilation :)
23:49:07 <elliott> pikhq: Okay, I've tipped over to something else.
23:49:09 <pikhq> Say, some freakish x86-only emulator frontend.
23:49:22 <elliott> pikhq: "Just fucking have different binary packages for native and non-native because otherwise I'll end up drinking myself to death".
23:49:23 <Sgeo> Portal should work with that right?
23:49:49 <pikhq> elliott: That's somewhat nasty. Buuut not hard to do at all.
23:49:52 <elliott> pikhq: OR, maybe, make all non-native things source-only
23:50:03 <elliott> (source packages will be trivial to install. Like "same command as usual" trivial.)
23:50:17 -!- augur has quit (Ping timeout: 255 seconds).
23:50:32 <elliott> pikhq: That also stops me having to buy 1,000,000 TB of disk space as my supported platforms increase and I have to compile for every combination of them.
23:51:00 <pikhq> elliott: Well, it'd be pretty easy to just make the non-native binaries if you already have the packaging configuration set up.
23:51:09 <pikhq> pkg build --native foo;pkg build --non-native foo
23:51:30 <elliott> pikhq: But now imagine I have 10 supported platforms.
23:51:35 <elliott> pikhq: Every package gets compiled 11 times.
23:51:47 <elliott> Every package-for-architecture.
23:51:54 <elliott> pikhq: Yeah, I might just do that.
23:52:15 <elliott> With 10 supported platforms, and separate non-/native packages, that's 20 compilations of each package.
23:52:19 <elliott> Of which 10 are basically identical.
23:52:33 <elliott> pikhq: So I *might just say* that any non-native package on a system is source-based.
23:52:41 <Sgeo> WHY DOES AMD'S SITE TRY TO DENY THE EXISTENCE OF 4200????
23:52:44 <Sgeo> I WANT DRIVERS
23:52:50 <olsner> maybe you can delta-encode the non-native version cleverly
23:52:55 <elliott> Want 64-bit foo on 64-bit? Sure. 32-bit on 32-bit? Sure. 32-bit on 64-bit? Sure, but it'll install slowly.
23:52:59 <pikhq> elliott: Hrm. Another problem. I think you'd actually need a *native* install of each and every architecture to build these things if you distinguish between native and non-native.
23:53:06 <olsner> or at least share all the files that weren't dependent on installation location
23:53:07 <elliott> olsner: true, i could store it as a diff -- but now this is just getting insane
23:53:25 <elliott> pikhq: But at the same time I don't want to have everything thinking they're in /arch/FOO!
23:53:53 <pikhq> Whereas if you have no such distinction, you can just have your build system have all the cross-compilers.
23:54:20 <elliott> pikhq: Sure is a good thing I have weeks to think about this.
23:54:22 <pikhq> ... And, in case a build system tries to run native programs, the relevant qemu.
23:54:23 <elliott> Right now I just want to punch everyone.
23:54:34 <elliott> pikhq: Come to think of it.
23:54:47 <elliott> If there's a port for a system, it better bloody well be installed on at least one system that can fucking build packages.
23:54:53 <elliott> I don't want to support systems with three users.
23:55:12 <pikhq> elliott: Yeah, but you almost certainly want your packages to be built on a reasonably beefy system.
23:55:19 <pikhq> Rather than, say, a freaking ARM phone.
23:55:33 <olsner> beefy system with relevant qemu's :)
23:55:34 <catseye> my build farm is *five* ARM phones
23:55:50 <elliott> pikhq: I'd just like to know that the actual architectures I'm likely to support are i686 and x86_64. :P
23:56:35 <pikhq> elliott: Sooo, you want to make it hard to support embedded architectures? Okay, you suck more than Debian now. Kthx.
23:56:58 <elliott> pikhq: No no no I totally will support them in the sense that uh
23:57:06 <elliott> pikhq: I'm just saying that I'm not going to distribute LiveCDs for ARM
23:57:42 <pikhq> But you are going to have a very painful build experience.
23:57:59 <elliott> pikhq: I need to figure out how to do this :P
23:58:03 <catseye> i like the feeling of freedom you get driving a bus
23:58:14 <elliott> `addquote <catseye> i like the feeling of freedom you get driving a bus
23:58:21 <olsner> aaah, catseye is a bus driver
23:59:13 <HackEgo> 239|<catseye> i like the feeling of freedom you get driving a bus
23:59:25 <elliott> pikhq: Okay but seriously you're in charge of telling me how to do multilib.