01:06:30 <ehird> [[The one number it gave that seemed like it might be the number of infant deaths per year was actually the total number of deaths per year, multiplied by one child, i.e. pure nonsense.]]
01:10:13 <ehird> <W|A> $993,258.23 per person
01:10:25 <ehird> using estimates from 1993-2008 too
01:11:48 <ehird> ais523-in-future: i think they're improving it. It handles "gdp/average human height in centimeters"
01:20:55 <ehird> x!=0, log(x)!=0, x^(3/x^3+1/x^2+1)!=0, y = (x^3 (-W_n(-(e^((log(x))/x^2-(3 log(x))/x^3) log(x))/x^3))+x log(x)-3 log(x))/(x^3 log(x)), n element Z
01:21:06 <ehird> x=1, y-1; x=2, y=0
01:21:08 <ehird> x^y - x^3y + x - 3 = 0
01:32:12 -!- FireFly has quit ("Later").
01:56:02 -!- Sgeo has joined.
02:02:36 * Sgeo goes to set up a Paypal account while on Windows
02:03:57 <Sgeo> Unless a hacker's thinking of using a 2-minute Window between me transfering funds to the account and when I purchase stuff with it, I think I'm safe
02:06:35 <pikhq> They've already hacked in.
02:07:22 <Sgeo> I'm not going to hook it up to any "real" stuff
02:07:54 <Sgeo> "Once you link your bank account or credit card, you'll speed through online checkout without exposing your financial information." No thanks
02:11:05 -!- coppro has quit (Remote closed the connection).
02:11:06 <pikhq> I wouldn't trust Paypal to be hooked to a credit card or bank account if I had Knuth do a formal analysis of the OpenBSD operating system and *then* installed OpenBSD.
02:11:41 <pikhq> Now, if I could replace Paypal itself with OpenBSD, then maybe we'd be talking.
02:12:56 <Sgeo> I'm just going to use some SL money
02:13:05 <Sgeo> I should probably start working on my SL product
02:13:21 <Sgeo> And figure out how to receive Paypal payments
02:22:10 -!- Gracenotes has quit ("Leaving").
03:15:32 -!- Gracenotes has joined.
03:25:38 -!- bsmntbombdood has quit (Remote closed the connection).
03:44:13 -!- bsmntbombdood has joined.
04:28:00 -!- coppro has joined.
05:14:56 -!- GreaseMonkey has joined.
05:38:34 -!- bsmntbombdood has quit (Read error: 60 (Operation timed out)).
05:51:53 -!- bsmntbombdood has joined.
06:16:13 -!- psygnisfive has joined.
06:50:37 -!- evincar has joined.
06:56:11 <evincar> Pretty dead tonight, then, eh?
07:02:41 <Sgeo> I think I may accidentally have done something that my dad will kill me for
07:03:19 <Sgeo> Turned L$ into USD and deposited into a paypal account
07:03:28 <Sgeo> I have no clue about the tax implications
07:04:48 <Sgeo> L$ is the currency used in Second Life
07:05:24 <psygnisfive> http://news.ycombinator.com/item?id=612999
07:05:24 <psygnisfive> http://www.reddit.com/r/programming/comments/8l10s/a_brief_history_of_grammar/
07:05:25 <psygnisfive> http://digg.com/general_sciences/A_Brief_History_of_Grammar/
07:40:56 -!- evincar has quit ("ChatZilla 0.9.84 [Firefox 3.0.10/2009042316]").
07:52:38 -!- Sgeo has quit (Read error: 110 (Connection timed out)).
07:59:59 -!- clog has quit (ended).
08:00:00 -!- clog has joined.
08:40:01 -!- kar8nga has joined.
08:42:24 -!- oerjan has joined.
08:42:58 -!- BeholdMyGlory has joined.
08:50:50 <oerjan> <ehird> People who strip the fat of bacon are evil agents of the End Times.
08:52:17 <oerjan> lean, mean, evil agents who can easily outrun you
09:14:26 <GregorR-L> Wow. I hear this car zoom up behind me, so I start bleeding off speed (which is my automatic being-tailgated reaction), and look in the rear-view mirror to see ... a cop ... tailgating me ... without his lights on. He veers off of me and into the next lane, then zooms up and starts tailgating the next guy, whose automatic reaction was to speed up, presumably to faster than the speed limit, at which point the cop pulled him over.
09:18:47 -!- oerjan has quit ("Reboot").
09:22:15 -!- oerjan has joined.
09:26:30 <oerjan> _is_ it really the case that a single /// substitution can only loop if the destination contains the source?
09:38:24 -!- kar8nga has quit (Remote closed the connection).
09:47:41 * oerjan paranoidly checks for unicode
09:51:55 <Gracenotes> APL FUNCTIONAL SYMBOL CIRCLE DIAERESIS
09:56:34 <oerjan> /abcabc/bcabcabca/abcabc fails as the destination contains the source
09:57:02 <oerjan> even though two destinations can combine to create an extra copy
09:59:48 <oerjan> /abb/bbaab/abbbb -> bbaabbb -> bbbbaabb
10:02:26 <oerjan> /abb/bbaab/abbbb -> bbaabbb -> bbabbaabb -> bbbbaabaabb -> bbbbaababbaab
10:06:30 * oerjan just added a fourth debugging level to the interpreter just to check that :D
10:13:42 <oerjan> oh, also /ab/bbaa/abab, that's probably the simplest
10:25:50 <oerjan> make that /ab/bbaa/abb
10:43:26 -!- tombom has joined.
11:23:27 -!- GreaseMonkey has quit ("YES -> thor-ainor.it <- THIS IS *DELICIOUS*!").
11:50:12 -!- upyr[emacs] has joined.
12:02:13 -!- FireFly has joined.
12:05:56 -!- M0ny has joined.
12:07:03 -!- Slereah_ has joined.
12:10:43 -!- Slereah has quit (Read error: 110 (Connection timed out)).
12:21:47 -!- olsner has joined.
12:48:53 -!- tombom_ has joined.
13:01:14 -!- tombom has quit (Read error: 110 (Connection timed out)).
13:01:15 -!- tombom_ has changed nick to tombom.
13:11:03 -!- kar8nga has joined.
13:23:22 -!- kar8nga has quit (Remote closed the connection).
14:15:08 -!- oerjan has quit ("leaving").
14:25:52 <AnMaster> Interesting. GCC generates better code at -O2 for do{...}while(foo); loops than while(foo){...}
14:26:12 <AnMaster> that is when foo is assigned just a few lines before
14:26:21 <AnMaster> so it could easily figure out that the loop will always be run
14:32:21 <pikhq> GCC's optimisation is not exactly best-in-class.
14:32:30 <pikhq> It's just rather good.
14:32:45 <pikhq> Just like almost everything else about GCC.
14:33:03 <pikhq> So, it managed to be one of the best compilers by not being bad at much anything. ;)
14:34:41 <olsner> being the only generally available compiler certainly helps too
14:35:06 <pikhq> Well, yes. It *is* just about guaranteed to be available for your platform.
14:35:58 <pikhq> Unlike, say, icc, which is for i686-pc-linux-gnu, x86_64-pc-linux-gnu, i686-pc-windows-microsoft, and x86_64-pc-windows-microsoft...
14:36:20 <pikhq> Or MS VC++, which is available for half of those.
14:37:16 <AnMaster> pikhq, you forgot that icc handles Linux on Itanium too
14:37:20 <pikhq> Yes. That's the only compiler I see replacing GCC,
14:37:34 <pikhq> For about the same reasons that GCC got to be ubiquitous.
14:37:56 <pikhq> Except that it's a bit better.
14:38:41 <fizzie> There's ICC for OS X and QNX, though.
14:39:47 <fizzie> Well, it's listed on http://software.intel.com/en-us/intel-compilers/ at least; I don't know about up-to-date versions. And no Fortran compiler on QNX.
14:46:25 <pikhq> Still, ICC sucks for non-Intel processors.
15:05:59 -!- upyr[emacs] has quit (Remote closed the connection).
15:49:19 -!- Judofyr has joined.
15:53:05 <AnMaster> anyone know a good non-compiled language to use for output to test lostking? From GCC I always get "out of memory". Even when letting it use 1 GB.
15:57:46 -!- Slereah_ has quit (Read error: 60 (Operation timed out)).
15:57:53 -!- kar8nga has joined.
16:01:52 -!- Slereah has joined.
16:04:12 -!- Judofyr has quit (Remote closed the connection).
16:05:56 -!- asiekierka has joined.
16:17:12 -!- kar8nga has quit (Read error: 104 (Connection reset by peer)).
16:22:55 -!- kar8nga has joined.
16:23:53 <ehird> psygnisfive: dude, why di you submit that to proggit
16:23:57 <ehird> it has nothing to do with programming
16:24:11 <ehird> also, downvoted for news site whoring.
16:24:19 <psygnisfive> it was the closest thing i could think to add it to, given whats on the sites :P
16:24:34 <ehird> http://www.reddit.com/r/linguistics
16:24:57 <ehird> psygnisfive: the list lists the most popular reddits.
16:24:59 <ehird> there are thousands.
16:25:36 <ehird> i'm going to log into digg and slashdot and hn for the first time in ages and downvote them too because i hate "UPVOTE DIS" buttons :)
16:25:50 <psygnisfive> well ill remove those buttons just for you then :P
16:25:50 <ehird> ooh, I don't have to for HN, the submission's deleted
16:26:15 <ehird> http://news.ycombinator.com/item?id=612999
16:26:17 <ehird> psygnisfive: "[dead]"
16:26:19 <ehird> I beg to differ, sir.
16:26:36 <psygnisfive> A Brief History of Grammar (wellnowwhat.net)
16:26:36 <psygnisfive> 3 points by psygnisfive 10 hours ago | 1 comment
16:26:39 <psygnisfive> 1 point by cendrillonea 8 hours ago | link
16:26:41 <psygnisfive> I really enjoyed this, and thought it was a well-detailed account that more than adequately explained the evolution of the grammar we use today.
16:26:47 <ehird> Hacker Newsnew | comments | leaders | jobs | submitlogin
16:26:53 <ehird> 1 point by cendrillonea 8 hours ago | link
16:26:55 <ehird> I really enjoyed this, and thought it was a well-detailed account that more than adequately explained the evolution of the grammar we use today.
16:27:00 <ehird> psygnisfive: you can see it because you are the submitter, I infer.
16:27:06 <ehird> for everyone else...
16:27:12 <ehird> yer submission's dead
16:31:28 <ehird> i don't care about grammar, tbh
16:32:16 <psygnisfive> its not so much about grammar as in "what do languages do blah blah blah"
16:33:34 -!- sebbu2 has changed nick to sebbu.
16:33:48 <psygnisfive> you have a finite language, that produces some small set of "kernel sentences", and then you have a number of unrestricted production rules that modify a sentence in particular ways, including adding in recursive elements
16:34:29 <psygnisfive> the TG approach wanted to see what was a sufficient enough set of kernel sentences and transformational rules in order to account for all over natural language
16:36:37 <ehird> psygnisfive: because fhd
16:36:41 <Slereah> Isn't beta more of a /v/ sound?
16:36:48 <ehird> Durünglork toshmosh, psygnisfive!
16:38:05 -!- Judofyr has joined.
16:41:54 <psygnisfive> anyway, ehird, you should check it out. just because. :P
16:42:08 <ehird> rewrite it in tardglish and i will
16:45:25 <ehird> Download now a shareware DOS executable version of UUdecode. Or, download a zipped version of the same.
16:45:26 <ehird> You can find fancier implementations of uuencode and uudecode in various places, but they do basically the same processing that my versions do. But if you are a software developer, beware that none of those other implementations make available the source code so that you can incorporate uuencode and uudecode functions into your own projects.
16:45:33 <ehird> http://my.execpc.com/~adw/uu.html
16:45:40 <ehird> It's written in BASIC!
16:49:27 <psygnisfive> http://www.flickr.com/photos/xorsyst/3494215156/
16:52:47 -!- Judofyr has quit (Remote closed the connection).
16:56:57 <ehird> ISO 3103 is a standard published by the International Organization for Standardization (commonly referred to as ISO), specifying a standardized method for brewing tea. It was originally laid down in 1980 as BS 6008:1980. It was produced by ISO Technical Committee 34 (Food products), Sub-Committee 8 (Tea).
16:56:59 <ehird> The abstract states the following:
16:57:01 <ehird> The method consists in extracting of soluble substances in dried tea leaf, containing in a porcelain or earthenware pot, by means of freshly boiling water, pouring of the liquor into a white porcelain or earthenware bowl, examination of the organoleptic properties of the infused leaf, and of the liquid with or without milk, or both.
17:01:13 <ehird> "Just as liberal feminists are frequently content with a minimal agenda of legal and social equality for women and 'pro-choice', so liberal (and even some socialist) mathematicians are often content to work within the hegemonic Zermelo-Fraenkel framework (which, reflecting its nineteenth-century liberal origins, already incorporates the axiom of equality) supplemented only by the axiom of choice."
17:06:01 <psygnisfive> ehird: was this tea standard introduced by a brit?
17:06:27 <ehird> I have a feeling I'd like tea if I could be arsed to make it Properly(TM).
17:06:51 <psygnisfive> well are you following the ISO 3103 standardized method for bewing tea?
17:08:55 <fizzie> It also suggests the obvious (non-regexp) solution: building a ISO 3103 compliant tea-making robot.
17:09:17 <ehird> Letter O considered harmful
17:09:17 <ehird> During the same Fortran Standards Committee meeting at which the name "FORTRAN 77" was chosen, a technical proposal was somehow smuggled into the official distribution, bearing the title, "Letter O considered harmful". This deceptively simple proposal purported to address the confusion that sometimes arises between the letter "O" and the numeral zero, by eliminating the letter from allowable variable names. However, the method proposed was to eliminate t
17:09:22 <ehird> he letter from the character set entirely (thereby retaining 48 as the number of lexical characters, which the colon had increased to 49).
17:09:25 <ehird> Among the "PRO" arguments was the assertion that this would also promote structured programming, by making it impossible to use the notorious GO TO statement as before. (Troublesome FORMAT statements would be eliminated, as well.)
17:09:29 <ehird> The sole "CON" argument conceded that "this might invalidate some existing programs" but noted that most of these "probably were non-conforming, anyway".[16][17]
17:14:44 <ehird> http://img411.imageshack.us/img411/6782/hackerbymetallicbox.jpg
17:16:20 <fungot> Available: agora alice c64 darwin discworld europarl ff7 fisher* ic irc jargon lovecraft nethack pa speeches ss wp youtube
17:16:29 <fungot> Selected style: c64 (C64 programming material)
17:16:33 <fungot> asiekierka: two lines located on the serial bus secondary device address, the microprocessor from a low active chip select conditions have been moved up, or via dedicated handshaking lines.
17:17:22 <ehird> "Mitnick served five years in prison, of which four and a half years were pre-trial and eight months were in solitary confinement because the Judge was convinced he could start a nuclear war by whistling into a pay phone if he were near one."
17:17:23 <ehird> I WHISTLED FOR A NUCLEAR BOMB AND WHEN IT CAME NEAR THE LICENSE PLACE SAID "FRESH" AND HAD DICE IN THE MIRROR
17:44:10 <AnMaster> ehird, where is that quote from
17:44:26 <ehird> AnMaster: it's complicated.
17:44:32 <ehird> which quote are we referring to?
17:44:45 <AnMaster> the one about the nuclear war and the judge
17:45:27 <AnMaster> ehird, though I guess I should ask about the upper case line too.
17:45:45 <ehird> [[Kevin Mitnick]].
17:46:27 <ehird> And the UPPERCASE QUOTE is from The Fresh Prince of Bel-Air's theme song; if you don't have an attuned sense of post-hipster irony it may just be best to know that it's a *chan meme.
17:47:04 -!- tombom has quit (Read error: 104 (Connection reset by peer)).
17:47:36 <ehird> s/CAB/NUCLEAR BOMB/
17:47:38 <ehird> Just a minor adjustment.
17:47:39 <AnMaster> "In addition, as per the plea deal, Mitnick was prohibited from profiting from films or books that are based on his criminal activity for a period of seven years."
17:47:50 <ehird> What do you mean, what?
17:48:06 <ehird> http://www.imdb.com/title/tt0159784/
17:48:18 <ehird> There was a movie made ... very, very loosely based ... on Mitnick.
17:48:26 <ehird> Veeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeery loosely.
17:50:14 <AnMaster> ehird, why would he be forbidden from making money on such movies
17:50:33 <AnMaster> I can't see how a movie would harm anyone
17:50:36 <ehird> AnMaster: To stop him coming out on top?
17:50:44 <ehird> If he profits from selling his story, he's profiting from committing crimes. Indirectly.
17:50:50 <ehird> It's stupid, but I can see where it's coming from.
17:50:54 <AnMaster> ok, makes kind of sense I guess.
17:54:18 -!- kar8nga has quit (Remote closed the connection).
18:08:01 <ehird> http://lambda-the-ultimate.org/node/1562#comment-18623
18:09:48 <fizzie> The nuclear war quote is a "sourced statement" in the sense that it's from a newspiece saying "-- Mitnick was put in solitary confinement and prevented from using a phone after law enforcement officials convinced a judge that he had the ability to start a nuclear war by whistling into a pay phone, he [Mitnick] said." A bit funny that Wikipedia presents it as fact.
18:11:55 <pikhq> Mitnick able to start a nuclear war by whistling into a pay phone?
18:12:17 <pikhq> Could we *seriously* get some judges that know anything about the technology they maintain?
18:14:12 <fizzie> Given that it's only Mitnick himself who said they were worried about he being able to start a nuclear war, it is conceivable that he might have been a bit hyperbolic there.
18:14:44 <pikhq> By '95, inband signaling was outmoded.
18:15:10 <pikhq> So, he couldn't even conceivably do the Cap'n Crunch hack and get free long distance.
18:15:35 <fizzie> (Not that knowledge and judges exactly seem to meet often in, say, the EFF legal case descriptions.)
18:16:03 <pikhq> And legislators aren't much better.
18:16:27 <pikhq> (come on people: making bits uncopiable is like making water not wet. You're not doing it.)
18:17:06 <ehird> I nevercopybitslu1
18:17:15 <ehird> *I nevrcopybitslu1
18:17:45 <pikhq> I sense copied bits.
18:17:57 <ehird> well i was not copying bytes can I get off with a slap on he wrist
18:27:51 <asiekierka> I have getc/putc already. What else do I need?
18:28:19 <asiekierka> i have getc, putc/puts, hex2string, divmod (crappy), initalization. What do I need next for my C64 kernel?
18:31:14 <fizzie> Generally speaking, a process scheduler is often a nice thing to have in an operating system.
18:31:22 <pikhq> Possibly also a fork.
18:32:23 <fizzie> I guess the order is up to you; most people certainly would appreciate a file system too, yes.
18:33:13 <asiekierka> Well, the filesystem is Commodore default
18:33:39 <asiekierka> also, forking needs a whole multitasking infrastructure
18:34:44 <fizzie> It's more than what DOS had. :p
18:34:55 <asiekierka> I need to make the CIA programming code, the "state copier" code, and the "process selector" code
18:36:21 <fizzie> Anyway, with "file system" I was more talking about the kernel/user interface for accessing files than any on-disk formats or such.
18:36:50 <asiekierka> i do need the kernel interface to access files
18:36:58 <asiekierka> but yeah, I should get forking there first
18:37:17 <ehird> Forking operating system!
18:37:21 <ehird> Forking computer! Just work!
18:45:39 -!- oerjan has joined.
18:46:54 -!- jix has joined.
18:49:50 <ehird> % sudo port install xorg
18:49:55 <ehird> [giant lag as a slow tcl program calculates dependencies]
18:49:59 <ehird> it's been going for ~5 minutes now
18:51:06 <fizzie> What's wrong with X11.app, then?
18:51:43 <ehird> fizzie: This is the same Xquartz codebase, actually; it's just that MacPorts' matlab needs xorg-xextproto thing, and:
18:51:45 <ehird> Requested 'x11 >= 1.1.99.1' but version of X11 is 1.1.5
18:51:59 <ehird> and I looked for the new X11.app but it wants me to upgrade OS X to a newer minor release and I can't be bothered
18:52:13 <pikhq> ehird: What version of Tcl does that sucker run on?
18:52:15 <ehird> so i'm installing via this, which hopefully won't say thus
18:52:28 <ehird> % sudo port install xorg
18:52:29 <ehird> ---> Fetching xorg-applewmproto
18:52:32 <ehird> pikhq: how do I check
18:52:49 <ehird> as it tried to install xorg, the same error
18:52:49 <fizzie> There's a matlab in macports?
18:53:33 <pikhq> Hmm. I could've sworn 8.5 had a good bytecode compiler.
18:53:42 <ehird> X11, or X, is a vendor-neutral, system-architecture neutral \
18:53:44 <ehird> network-transparent window system and user interface standard. \
18:53:46 <ehird> In other words, it is a GUI for UNIX. X can use your network -- \
18:53:48 <ehird> you may run CPU-intensive programs on high powered workstations \
18:53:50 <ehird> and display the user interface, the windows, on inexpensive \
18:54:23 <ehird> How can I obliterate AppleX11 from my system, one wonders.
18:54:59 <ehird> pikhq: I'm pretty sure the usual X11 files are shat in their usual place.
18:55:08 <ehird> pikhq: Evidence: I have a /usr/X11.
18:55:11 <ehird> Guess I'll obliterate that too.
18:55:17 <pikhq> That's vomit-worthy.
18:55:25 <ehird> pikhq: Xorg's organizational system is vomit-worthy.
18:55:33 <ehird> I mean, /usr/X11? C'mon, that's not very unix of you.
18:55:44 <ehird> /opt/X11 is FHS, /usr is modern-unix-OS
18:55:55 <ehird> Although Xorg uses /usr in recent versions, I think.
18:56:03 <ehird> And also uses HAL instead of xorg.conf by default, so perhaps we're seeing some sanity.
18:56:12 <pikhq> I don't have a /usr/X11.
18:56:18 <ehird> There you go then.
18:56:23 <ehird> Do you have /usr/X11R6?
18:56:35 <pikhq> I do have an xorg.conf, but that's because my monitor needs a specific modeline that can't be autodetected.
18:56:47 <fizzie> I have an /usr/X11R6, but there's only a symlink "bin -> ../bin" in it.
18:56:55 <pikhq> (the thing came with freaking Windows drivers, because it's signalling was broken)
18:56:59 <pikhq> I don't have a /usr/X11R6.
18:57:12 <pikhq> Which is good, considering I don't have X11R6 installed.
18:57:21 <ehird> pikhq: can't you configure HAL for it?
18:57:48 <pikhq> Also, when I set up X, I didn't have HAL installed.
19:00:03 <ehird> % port installed|grep xorg|awk '{print $1}'|xargs sudo port uninstall
19:00:11 <ehird> hmm, "% port installed|grep xorg|awk '{print $1}'|xargs sudo port uninstall xrender xpm " first actually
19:00:12 <fizzie> Heh, Debian's "x11-common" contains (in addition to the /usr/X11R6/bin -> ../bin symlink) also an /usr/bin/X11 which is a symlink to ".". I'm guessing compatibility-to-broken-programs related reasons.
19:00:15 <ehird> oh, Xft2 as well, heh.
19:02:02 <ehird> Let's try this now then.
19:02:05 <ehird> % sudo port install xorg
19:10:26 <ehird> pikhq: speed it up
19:13:26 <ehird> bsmntbombdood: yes. and you're ruining it with hyperthreading ;-)
19:14:27 <ehird> bsmntbombdood: "cache thrashing not not cache thrashing"
19:14:30 <ehird> "hack not not hack'
19:16:00 <fungot> asiekierka: when an nmi interrupt. the table that starts at 6. bit map are arranged in 25 rows of 40 columns, the
19:17:42 <fizzie> fungot: To be honest, that wasn't so cool.
19:17:42 <fungot> fizzie: in other words, the
19:18:21 <bsmntbombdood> ehird: think i'm ruining performance on this ssd with encryption?
19:18:38 <ehird> bsmntbombdood: encrypting an OS drive?
19:18:49 <ehird> that would most likely be bad
19:20:50 <bsmntbombdood> echo 3 > /proc/sys/vm/drop_caches; dd if=data bs=1M count=2K of=/dev/null
19:20:54 <bsmntbombdood> 2147483648 bytes (2.1 GB) copied, 18.8341 s, 114 MB/s
19:20:56 <ehird> bsmntbombdood: but if you're not doing the lvm alignment you're already showing RECKLESS ABANDON for performance, so ;-)
19:21:07 <ehird> also, sequential operations for the fail
19:21:13 <ehird> most OS work = random read/writes
19:21:27 <bsmntbombdood> well, it's supposed to get 200mb/s sequential reads is what i'm saying
19:21:56 <ehird> bsmntbombdood: if you disabled hyperthreading, SECURE ERASE'd that drive, and set up lvm2 with the alignment incantations, you could probably get away with it
19:22:47 <ehird> bsmntbombdood: yes, but writes are common
19:22:50 <ehird> think about firefox
19:22:56 <ehird> visit a page? 345734859345 gajillion small writes
19:24:53 -!- ais523 has joined.
19:25:32 <ehird> bsmntbombdood: btw, you are mounting /tmp as tmpfs right?
19:25:41 <ehird> loadsa speed benefits, less ssd writs
19:26:53 <AnMaster> I somehow suspect I need to work on that optimiser a bit!
19:26:55 <ehird> AnMaster: an extension of Ayn Rand's A=A?
19:27:16 <AnMaster> ehird: Partly implemented copy propagation:
19:27:25 <ehird> Slereah: JOHN GALT=JOHN GALT QED
19:27:39 <AnMaster> needs to be cleaned up to remove that dead node
19:28:25 <fungot> ^<lang> <code>; ^def <command> <lang> <code>; ^show [command]; lang=bf/ul, code=text/str:N; ^str 0-9 get/set/add [text]; ^style [style]; ^bool
19:28:53 -!- ais523 has quit (Remote closed the connection).
19:29:25 <ehird> bsmntbombdood: do you still have hyperthreading enabled? >:)
19:30:03 -!- ais523 has joined.
19:30:25 <ehird> bsmntbombdood: then no :-P
19:30:29 <ehird> [i'm just kidding around]
19:31:14 <ehird> bsmntbombdood: that's cores
19:31:22 <ais523> bsmntbombdood: it measures in erlangs
19:31:24 <ehird> 400% is max, or maybe 800% for hyperthreading
19:31:37 <ais523> 100% means full use of one core
19:31:44 <ais523> so if you have dual-core, it can go up to 200%
19:32:07 <ehird> bsmntbombdood: threads, foo
19:32:16 <ehird> bsmntbombdood: two pthraeds can max out two cores
19:32:21 <bsmntbombdood> ehird: things that i know are singlethreaded went to 102%
19:32:34 <ehird> bsmntbombdood: aha, I know why
19:32:38 <ais523> and it can go higher still under certain setups, like quad-core processors
19:32:43 <ehird> bsmntbombdood: the i7 transfers power from idle cores
19:32:48 <ehird> ais523: bsmntbombdood has a quad-core
19:32:52 <ehird> bsmntbombdood: to nonidle ones
19:32:59 <ehird> bsmntbombdood: iirc something like 30% faster clockspeed sometimes
19:33:02 <ehird> presumably it doesn't tell linux this
19:33:07 <ehird> so it can exceed 100% of stock clock speed
19:33:41 <ehird> bsmntbombdood: you think i'm wrong?
19:34:09 <fizzie> It's a bit confusing that the "%CPU" does that 100% == full use of a single core, but in the summary "Cpu(s): 3.8%us --" it's 100% == all processors in use.
19:35:05 <bsmntbombdood> and sometimes top says that top or xterm is using like 75%, which is weird
19:42:09 <ehird> bsmntbombdood: you'd better; it made up a large chunk of the price :P
19:42:14 <fizzie> I can't believe it's not butter.
19:43:30 <bsmntbombdood> i couldn't even play a movie at anything but the original size
19:44:56 <fizzie> You can sort-of blame Nvidia for not releasing the hardware specs like ATI/AMD did.
19:46:11 <fizzie> That's probably the other reason people haven't been reverse-engineering nvidia stuff more.
19:46:26 <ehird> bsmntbombdood: beats ati drivers at least
19:47:22 <fizzie> What sort of card was it again?
19:47:22 <ehird> bsmntbombdood: how lowest-common-denominator.
19:47:30 <ehird> fizzie: really old passive Nvidia from sparkl
19:48:06 <ehird> well, okay, not old
19:48:08 <AnMaster> <ehird> AnMaster: an extension of Ayn Rand's A=A? <-- btw, what did you mean with that?
19:48:16 <ehird> AnMaster: it's kind of complicated
19:48:30 <ehird> bsmntbombdood: true. Anyway, you told me you had no gfx card requirements :P
19:48:40 <ehird> AnMaster: "Ayn Rand is a retard"
19:48:57 <AnMaster> ehird, http://en.wikipedia.org/wiki/Ayn_Rand ?
19:49:14 <Slereah> I actually once tried to make a language based on A=A
19:49:15 <ehird> she's dead now, which is a positive aspect.
19:49:20 <ehird> Slereah: Objectivist C
19:49:50 <bsmntbombdood> ehird: i should have gotten a GTX 295 or something
19:50:05 <Slereah> Forgot the specs, but here's a program : http://membres.lycos.fr/bewulf/Russell/Fibo.jpg
19:50:12 <ehird> bsmntbombdood: is my sarcasm-o-meter setting off correct?
19:51:20 <ehird> bsmntbombdood: DING! Sarcasm!
19:51:20 <fizzie> The nvidia binary driver also added that VDPAU thing recently; it was long so that it supported only a bit of xvmc, and only for geforce series <=7. (The readme says G96, which is in 9400, does IDCT+motion-compensation on the card for MPEG1/2, H.264 and VC1 (WMV), and bitstream decoding only for h.264; G98 would do also the bitstream decoding part on the card for mpeg1/2 and wmv.
19:52:53 <ehird> bsmntbombdood: so was the reading correct
19:58:26 <fizzie> My passive gf7600gt card, with it's crummy G73 gpu and no VDPAU support in the nvidia drivers, can't do 1080p video in this rather-less-than-high-end system. :/ (Admittedly I don't think I have any 1080p media either.)
20:00:52 -!- upyr[emacs] has joined.
20:01:45 <pikhq> bsmntbombdood: Your card is somewhat shitty for 3D stuff, but should handle most any video you throw at it.
20:02:16 <bsmntbombdood> i don't think it's worth it to download a 15gb full movie though
20:03:04 <pikhq> What resolution is your monitor?
20:04:07 <pikhq> Look for 720p videos.
20:04:20 <pikhq> That being the best that your monitor can display.
20:04:36 <pikhq> And also, your graphics card is overkill for 720p video.
20:04:49 <pikhq> My Geforce 6300 does 720p video just fine. :p
20:04:55 <fizzie> This 1920x1200 would suit 1080p rather well, if I had a habit of watching any moving pictures.
20:05:05 <ehird> 20:03 bsmntbombdood: only 1280*1024
20:05:15 <ehird> why are you using such a tiny monitor on such a powerful system :P
20:05:36 <pikhq> I'm going to agree with ehird on this one.
20:05:45 <ehird> it's like wiring a cray up to a 14" CRT and playing a card game, a few decades ago
20:05:49 <pikhq> I've got a 1400x900 monitor, and that's because I'm cheap. ;)
20:06:00 <ehird> well, 14" crt isn't quite cray era
20:06:04 <ehird> let's say a dumb greenscreen terminal
20:06:15 <pikhq> Let's say a teletype.
20:06:23 <pikhq> With paper and everything.
20:06:41 <pikhq> Home computers of the era had better output.
20:06:43 <bsmntbombdood> hopefully you are not wasting your cray on rendering eye candy
20:06:47 <fizzie> Funny, according to wp "G86 and G98 [quite different from the G86 -- features VC-1 video decoding completely in hardware] cards are both sold as "8400 GS", the difference can only be told from technical specifications". It's like they *wanted* to confuse people.
20:07:12 <ehird> bsmntbombdood: 'card game' being the operative word
20:08:51 <pikhq> bsmntbombdood: How many gigaflops does your computer offer?
20:09:02 <ehird> pikhq: fuck that, how many bogomips?
20:09:45 <pikhq> I don't know how many bogomips the Cray offers.
20:09:48 <fizzie> That's a lot of bogons generated.
20:10:04 <ehird> bsmntbombdood: that's not many at all
20:10:14 <ehird> turn of hyperthreading, bitch :-)
20:10:33 <pikhq> My crappy Sempron offers 3215.44 bogomips.
20:10:51 <ehird> bsmntbombdood: *8 because of STUPOTHREADING
20:11:12 <fizzie> I get only 2000 bogomips... though I think that number changes when the cpufreq machinery decides to go up from 1 GHz to whatever-the-normal-frequency-was.
20:12:18 <fizzie> Heh, yes. If I start a "while true; do X=y; done" loop in another terminal, the bogomips rating goes to 5628.88. (2.8 GHz seems to be the normal frequency of a X2-5600+.)
20:12:59 <ehird> Deewiant: what cpu?
20:13:00 <ais523> what's with the caps there in the variable assignment?
20:13:03 <ais523> and why that assignment at all?
20:13:05 <ehird> ofc, this is very ghz-dependent
20:13:18 <Deewiant> ehird: Q9550, overclocked to 3.something
20:13:32 <ehird> ew, overclockers. /me steps away from Deewiant
20:13:35 <ehird> dirty overclocker :|
20:13:35 <fizzie> ais523: I couldn't think of what a no-op would be.
20:13:37 <Deewiant> (Default being 2.[89]something)
20:13:43 <ehird> bsmntbombdood: et tu, brute‽‽‽‽‽‽‽‽
20:14:07 <ais523> fizzie: the null string usually works
20:14:18 <ais523> although I'm not sure offhand if it would be do; done or do done
20:14:27 <fizzie> ais523: Well, "while true; do; done" gets bash: syntax error near unexpected token `;'
20:14:33 <ehird> i have an allergy to "HARDKOR" gamers and the related ilk. which happens to encompass just about anyone who knows hardware without being able to program, and anyone under the discipline of such invented by and as.
20:14:43 <fizzie> And "while true; do done" => "bash: syntax error near unexpected token `done'"
20:14:45 <ehird> if the sentence got ungrammatical later on it's because it ended up not being tuned to say what i was saying
20:14:53 <ehird> fizzie: while true; do; done
20:15:00 <Deewiant> fizzie: Interestingly, that works in zsh.
20:15:08 <fizzie> Deewiant: Well, bash is... peculiar.
20:15:36 <ais523> "while true; do; done" fails in dash too
20:15:47 <ais523> I have it installed, but it isn't my default shell
20:16:07 <ais523> ~ $ while true; do; done
20:16:08 <ais523> sh: syntax error: ";" unexpected
20:16:55 <fizzie> Based on a very cursory glance of the bash(1) man page "shell grammar" section, I can't really construct a "null command"y thing.
20:17:00 <ais523> $ while true; do :; done
20:17:07 <ais523> : is the no-op command
20:17:21 <ais523> I'd say all commands set $?
20:17:27 <ais523> so it's correct for a nop to, too
20:17:39 <ais523> that's like complaining that NOP in asm isn't a nop because it changes the IP
20:17:42 <ehird> a nop is an uncommand!
20:17:59 <ais523> there are actually two sorts of nop
20:18:07 <ais523> which is something I'm thinking about a lot in Underlambda?
20:18:12 <ais523> what's more noppish; true, or cat?
20:18:32 <fizzie> : is described as "No effect; the command does nothing beyond expanding arguments and performing any specified redirections. A zero exit code is returned." so I guess it's a reasonably noppy thing.
20:18:36 <ehird> 20:18 ais523: which is something I'm thinking about a lot in Underlambda?
20:18:49 <fizzie> Okay, next time I need a do-nothing loop I'll certainly use :.
20:18:54 <ehird> ais523: in a pipeline, nop is cat
20:18:59 <ehird> ais523: in a sequence of commands, nop is... nothing
20:19:02 <ais523> whereas true would destroy data
20:19:18 <ais523> it's the difference between 0 and 1 in Underlambda
20:19:19 <ehird> so, the only thing you can REALLY call a nop in shell is the pipeline nop - cat
20:19:27 <ais523> as in, it destroys data
20:19:29 <ehird> cat :: String -> String; cat = id
20:19:35 <ais523> because it leaves everything the same
20:20:22 <ehird> tac = unlines . reverse . lines
20:20:40 <ehird> yay, x11 installed
20:22:13 <ehird> sudo port install maxima
20:22:18 -!- jix has quit (No route to host).
20:23:44 -!- asiekierka has quit.
20:25:00 <ais523> ehird: why wasn't x11 installed before?
20:25:15 <ais523> I thought you'd installed it earlier, though
20:25:16 <ehird> and I removed the x11 i had
20:25:21 <ehird> 'cuz dependency issues
20:25:25 <ehird> so i installed it via macports
20:26:04 <ehird> ... and I possibly don't even need x11 for this
20:26:27 * ehird installs gnuplot +no_x11 +wxwidgets; installs tk +quartz; installs maxima
20:26:36 <ehird> which will also install sbcl, despite me having my own sbcl
20:28:16 <ehird> bsmntbombdood: it depends on gnuplot and tk
20:28:24 <ehird> for the UI, I assume; it may not be mandatory, but I want it.
20:30:15 <ehird> But I want plots and shit too, mmkay?
20:30:45 -!- kar8nga has joined.
20:34:15 <ehird> bsmntbombdood: 56k?
20:34:32 <ehird> bsmntbombdood: why did you buy an expensive i7 pc
20:35:09 <GregorR-L> I plug my CPU directly into the intarwebs.
20:35:30 <ehird> bsmntbombdood: tiny monitor ... shitty internet... :P
20:35:55 <ehird> it just seems a bit ridiculous
20:37:42 <GregorR-L> bsmntbombdood uses it to calculate the trajectory of micromissiles to kill anyone who makes fun of his monitor and/or internet connection.
20:38:12 <ehird> Shiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiii
20:40:18 <pikhq> bsmntbombdood: I strongly suggest that you invest in bandwidth and a better monitor.
20:40:35 <bsmntbombdood> i will probably get a better monitor, the bandwidth is more difficult though
20:40:47 <pikhq> The best investment for the former, of course, is moving to Japan, but moving to any first-world nation other than the US is also a viable option.
20:40:54 <GregorR-L> bsmntbombdood's options are limited after he bombed Comcast's basement.
20:41:19 <pikhq> bsmntbombdood: Dude, you didn't ask anyone to help you bomb Comcast?
20:41:24 <pikhq> How very selfish of you!
20:41:59 <ehird> pikhq: sth korea has better bandwidth iirc
20:42:41 <pikhq> Japan has 100 Mbps for $20 a month.
20:42:42 <bsmntbombdood> japan and south korea are both places where i do not want to go
20:43:16 <pikhq> bsmntbombdood: Might I suggest somewhere more... Euro?
20:43:47 <ehird> avoid sweden, anmaster's there
20:44:05 <pikhq> Norway's got Oerjan, though.
20:44:13 <ehird> Go for Norway or Finland.
20:44:25 <ehird> Finland are more net-wise, iirc, and they invented IRC, and fizzie & oklopol & Deewiant
20:44:32 <fizzie> Finland has (110 Mbps down, 5 Mbps up) for $73.99/month (blame the $ exchange rate) if you don't mind the cable-modem technologistics.
20:44:38 <ehird> Norway has fjords and oerjan.
20:44:38 <fizzie> Yes, the Finns invented fizzie.
20:44:49 <ehird> fizzie: What, all of them
20:45:11 <GregorR-L> ehird: That's just how Finns breed.
20:45:17 <AnMaster> Sweden has nice Internet unless you are far out in the country.
20:45:25 <pikhq> fizzie: Where I'm at, I can get 1 Mbps down, 256k up for $69.99 a month.
20:45:27 <ehird> bsmntbombdood: 5 up isn't all that bad...
20:45:44 <ehird> unfortunately we can't get cable here
20:45:54 <ehird> moving to finland sounds nice
20:46:00 <fizzie> With careful picking of where to live, I guess you can get reasonable prices for "100 Mbps both ways" here too; but it's not very widely available yet.
20:46:15 <pikhq> fizzie: 100 Mbps *both ways*?
20:46:25 <ehird> And downhill. Uh, both ways.
20:46:26 <fizzie> Uphill both ways, yes.
20:46:35 -!- Sgeo has joined.
20:46:46 <pikhq> Dammit, why must *all* of your infrastructure be better than ours?
20:47:04 <ehird> fizzie: how cold is finland?
20:47:47 <pikhq> Sgeo: US. Dealing with infrastructure that was last touched when my grandmother was a child.
20:48:04 <ehird> Internet-over-phone is ridiculously stupid.
20:48:29 <Sgeo> I know at least one person still on dialup :/
20:48:31 <ehird> fizzie: so yeah, coldity?
20:48:34 <Sgeo> Well, not in person
20:48:38 <fizzie> I guess fibre/ethernet-based 100MBps-up-and-down is still only in some random places, though. Personally I'm paying something like $70/month for puny 24 Mbps down, 1 Mbps up. (I happened to pick an expensive ISP, since they do the whole static-IP, custom-reverse-DNS, native-IPv6, tech-support-that-knows-what-they're-doing thing.)
20:48:55 <fizzie> ehird: It's about +10 degrees celsius here now.
20:49:01 <pikhq> I want phone-over-Internet, TV-over-Internet, etc.-over-Internet.
20:49:01 <AnMaster> a dictionary of dictionaries of sets. Is it time to try to work out a better data representation yet?
20:49:06 <pikhq> fizzie: 24 Mbps? Puny?
20:49:07 <ehird> fizzie: what isp was that again?
20:49:13 <GregorR-L> The US got the infrastructure early, so as a result it has the oldest infrastructure. It's really quite simple.
20:49:16 <ehird> pikhq: eh, you can get that in London
20:49:17 <pikhq> That's absurdly good for the US.
20:49:39 <ehird> AnMaster: ... and?
20:49:44 <ehird> He was commenting on "puny'.
20:49:44 <AnMaster> <pikhq> That's absurdly good for the US.
20:49:49 <AnMaster> that is because it isn't the US!
20:49:50 <ehird> Don't make yourself look like a jackass...
20:49:50 <fizzie> ehird: Nebula; nbl.fi or something. Here in the southernmost points it's rather rarely below -10 degrees in the winter, though occasionally it happens.
20:50:09 <pikhq> AnMaster: ... And I've been talking about how the US has crappy infrastructure.
20:50:11 <ehird> fizzie: going out in anything below 10 degrees makes me feel like death
20:50:29 <AnMaster> ehird, you need thicker clothes then
20:50:30 <pikhq> By "crappy", I mean "Ethiopia is set to surpass us in 10 years."
20:50:36 <ehird> it's 9 centigrade today apparently
20:50:36 -!- MizardX has quit ("reboot").
20:50:40 <ehird> and I can handle that just fine
20:50:48 <fizzie> This year we had sub-zero temperatures at night-time as recently as... some two-and-a-half weeks ago?
20:50:49 <ehird> so -10 i could probably manage at worst.
20:50:57 <ehird> night-time doesn't bother me
20:51:13 <ehird> fizzie: do you think these uphill-both-ways isp do the ipv6 stuffs too?
20:51:24 <ehird> i mean, it's not like moving to finland is possible or feasible, but IN MY MIND it could be
20:51:29 <ehird> i could move my mind to finland
20:51:32 <ehird> and keep my body here
20:51:48 <AnMaster> ehird, remember you will have to learn Finnish when moving there!
20:51:52 <pikhq> I could learn Finnish.
20:52:07 <GregorR-L> ehird: Unfortunately, the mind-Finland uplink is over 56k.
20:52:11 <ehird> AnMaster: I've tried learning languages before and failing, but I could probably get by to start with using basic vernacular...
20:52:18 <oerjan> pikhq: i recall reading that somalia has pretty kickass cell phone network
20:52:18 <ehird> GregorR-L: AIEEEEEEEEEEEE
20:52:42 <oerjan> because there was no state to regulate it
20:52:44 <fizzie> ehird: I don't think they do. I'm not quite sure where they are available, to be honest. Can't find any availability maps.
20:52:46 <pikhq> We're speaking one of the most commonly spoken languages right now. :)
20:53:06 <pikhq> oerjan: That's the first time I've heard of a lack of regulation creating competition.
20:53:07 <ehird> fizzie: hmm... what about the 100mbps both ways, it'd be supernice to have fast internet + ipv6 + reverse dns
20:53:25 * pikhq has been dropping words lately today. :(
20:53:29 -!- tombom has joined.
20:53:40 <fizzie> ehird: Well, the student housing at our university has that. :p
20:53:52 <ehird> fizzie: that's a rather unhelpful option :P
20:53:57 <ehird> what about the 100mbps one way
20:53:59 <pikhq> fizzie: ... Your university gets 100mbps?
20:54:00 <AnMaster> pikhq, I'm not sure that is true actually.
20:54:03 <ehird> is what i meant to say lately
20:54:12 <pikhq> I don't think mine has that on our *LAN*.
20:54:14 <ehird> (i don't really talk much outside anyway, so finnish probably wouldn't be a problem for a good amount of time)
20:54:18 <AnMaster> pikhq, French is harder, trust me. I took French in school some years ago.
20:54:21 <fizzie> pikhq: Yes, well... I think it's a shared 1 Gbps uplink to the interwebs, though.
20:54:41 <fizzie> pikhq: Or maybe they updated that to 10 Gbps already. I'm not quite sure.
20:54:59 * pikhq is jealous either way
20:55:22 <AnMaster> sure English has that too, but not to the same degree.
20:55:44 <ehird> lol— someone said that a password-removing thingy for linux/windows didn't work on OS X
20:55:49 <ehird> Power button Command-S 'nuff said.
20:55:53 <fizzie> pikhq: Yes, the student housing network people upgraded the main Funet (Finnish university network thing, through which their internet goes) link to 10-gigabit-Ethernet on April 2nd.
20:55:55 <AnMaster> and except for "be", a simple rule for I/you/he/she/it/we/you/they
20:55:56 -!- MizardX has joined.
20:55:59 <ehird> EFI-level backdoors FTW...
20:56:29 <ehird> AnMaster: boots into single-user mode
20:56:38 <AnMaster> ehird, you can set EFI password
20:56:51 <ehird> AnMaster: yeaah but i doubt this retard has
20:56:55 <AnMaster> that is needed for all non-normal boots
20:57:01 <ehird> AnMaster: you can encrypt your HD too
20:57:01 <fizzie> ehird: Oh, and the 110Mbps-down-5Mbps-up cable-TV-company/ISP is Welho, which is rather in the clueless group of ISPs, so I really don't think they do any IPv6 stuff or anything.
20:57:08 <ehird> so there's easy ways to prevent it; just not common ones
20:57:21 <ehird> fizzie: shit sucks.
20:57:33 <ehird> AnMaster: yes, it comes with truecrypt built in
20:57:35 <ehird> and a UI to use it
20:57:43 <pikhq> fizzie: Most cities in the US are doing well to get 5Mbps down.
20:57:44 <fizzie> Most of the ISPs here have some silly "you are forbidden to run any server applications in your pipe" things in their contracts, and the general level of service for home customers seems to be "you get 5 IPs over DHCP and that's it".
20:57:57 <AnMaster> I remember seeing some "encrypt home dir"
20:57:58 <ehird> AnMaster: yes. Although, caution in that truecrypt, while open source, is shady.
20:58:03 <pikhq> We've got monopolies everywhere.
20:58:09 <ehird> The mods delete questions about the changelogs etc on their forums
20:58:13 <ehird> and they just release source tarball
20:58:17 <ehird> but, it has it built in.
20:58:31 <pikhq> Well, except for dialup providers. Those still compete with each other.
20:58:33 <AnMaster> ehird, does it let you encrypt the system dir :D
20:58:35 <fizzie> pikhq: I think the 24 Mbps down (well, theoretical; I get ~19 Mbps here) ADSL2+ thing is pretty widely available here, though personally I only know about the situation around the capital city here.
20:58:38 <ehird> I think so, AnMaster.
20:58:55 <ehird> AnMaster: Well, presumably it'd leave the bootloader unencrypted.
20:58:58 <AnMaster> ehird, and have you ordered that linux PC yet btw :)
20:59:09 <ehird> I need to assemble a bag of parts, make sure the psu is fine, etc
20:59:13 <ehird> so it'll be a lil bit
20:59:20 <pikhq> fizzie: And you guys even get actual bandwidth close to what's advertised?
20:59:26 <pikhq> That's even more impressive.
20:59:32 <AnMaster> ehird, and your befunge interpreter, how is it coming along :)
20:59:40 <ehird> AnMaster: I think I will stab you with a stick.
20:59:56 <pikhq> I expect next you'll be telling me that your cable service doesn't recompress HD streams to the point that they're unwatchable?
21:00:01 <ehird> For asking that :)
21:00:25 <ehird> pikhq: let's move to Finland; we can pool our resources to pay for residence and kick-ass internet
21:00:44 <AnMaster> pikhq, um, Why do you use cable instead of your own antenna.
21:00:53 <pikhq> ehird: We'll probably be better able to pay for university, too.
21:00:59 <Deewiant> pikhq: I measured my connection last weekend and it indeed managed something like 107/4.8 Mib/s
21:01:01 <ehird> AnMaster: Do you... are you...
21:01:05 <ehird> Do you even know what...
21:01:07 <pikhq> AnMaster: Because that gets us 2 stations.
21:01:20 <ehird> AnMaster: cable channels are not the same as analogue channels
21:01:29 <ehird> also, analogue is terrible quality
21:01:37 <AnMaster> ehird, there are no analogue channels in Sweden. They are digital, over antenna
21:01:56 <AnMaster> some kind of political motivation for switching to that.
21:02:02 <pikhq> ehird: Nearly 3/4ths of our analog broadcasters were shut down.
21:02:11 <ehird> pikhq: there are only 5 analogue channels in the uk
21:02:12 <pikhq> As of June, all OTA signals in the US will be digital.
21:02:27 <ehird> and digital = non-antenna in the UK
21:02:34 <ehird> I don't know about this digital-antenna shit but we have none of it
21:02:37 <pikhq> Oh, the UK. The *other* country with third-world infrastructure.
21:02:43 <AnMaster> ehird, why do you call it "shit"
21:02:56 <pikhq> Though I thought you guys had nearly 75% coverage with Freeview?
21:02:57 <ehird> AnMaster: shit does not mean rubbish
21:03:03 <ehird> unless it means rubbish
21:03:05 <ehird> in which case it means rubbish
21:03:32 <pikhq> ehird: At least you've got OTA stations worth watching.
21:03:33 <ehird> pikhq: Yes, and "freesat from sky", which has a marginally better channel choice, so we use it. We also have Sky, but $$$ and the channels are shit.
21:03:36 <ehird> Also bbc's HD freesat.
21:03:44 <ehird> pikhq: you sure about that?
21:03:48 <AnMaster> ehird, but no I can't say I'm able to detect the difference between shit::stuff() and shit::rubbish()
21:03:50 <ehird> channel 4 and some stuff on five, maybe
21:03:54 <ehird> and a few bbc documentaries
21:03:59 <pikhq> Which is more than we have.
21:04:00 <ehird> but most of BBC 1 stuff, and all of ITV, is rubbish
21:04:02 <AnMaster> err, sorry, wrong type description language
21:04:08 <pikhq> You know the channels you get on Sky?
21:04:21 <pikhq> That's US television.
21:04:35 <ehird> It's pretty awful.
21:04:49 <ehird> fatal error encountered in SBCL pid 88478(tid 2685568800):
21:04:50 <ehird> bad runtime option "--userinit"
21:05:04 <pikhq> US culture is pretty nice. When you remove the corporations from it.
21:05:17 <pikhq> ... That is, when you remove the last 50 to 100 years of it. :p
21:06:16 <ehird> Maxima 5.18.1 http://maxima.sourceforge.net
21:06:18 <ehird> Using Lisp SBCL 1.0.28
21:06:20 <ehird> Distributed under the GNU Public License. See the file COPYING.
21:06:22 <ehird> Dedicated to the memory of William Schelter.
21:06:24 <ehird> The function bug_report() provides bug reporting information.
21:07:17 <ehird> YAY xmaxima starts
21:07:42 <ehird> Terribly ugly, but!
21:09:27 <fizzie> Deewiant: It's that Welho 110/5?
21:09:43 <ehird> There's a wxmaxima.
21:09:53 <ehird> Deewiant: context?
21:10:32 <fizzie> ehird: <Deewiant> pikhq: I measured my connection last weekend and it indeed managed something like 107/4.8 Mib/s
21:10:47 <ehird> Deewiant: Do they do ipv6?
21:11:05 <Deewiant> I'm not on that connection now so I can't test.
21:11:40 <Deewiant> Most Finnish ISPs probably don't, only Nebula or whatever it was that fizzie hyped.
21:12:11 <ehird> Deewiant: But Nebula doesn't do the fast!
21:12:18 <ehird> Finns are internetty people :<
21:12:39 <pikhq> In 2012, the US will still have people with 56k modems.
21:12:47 <pikhq> ... As their only option.
21:13:14 <pikhq> I don't think you understand how US infrastructure works.
21:13:17 <AnMaster> <ehird> There's a wxmaxima. <-- yes I use it personally
21:13:26 <pikhq> The US will be the last holdout on IPv4.
21:13:30 <GregorR-L> pikhq: I'm being a happy optimist :P
21:13:36 <pikhq> I expect us to upgrade sometime after World War III.
21:13:38 <ehird> AnMaster: how do you get just a simple maxima console with it?
21:13:41 <ehird> where you can plot and stuff
21:13:57 <ehird> AnMaster: just a regular maxima REPL
21:13:58 <fizzie> Nebula's fastest one with a price that's not "ask us" is 48M/6M, which is basically two bonded ADSL2+ channels; and that's horribly expensive, $200/month or so.
21:14:06 <pikhq> ... Speaking of. Anyone want to play global thermonuclear war?
21:14:10 <ehird> fizzie: what would ask us imply
21:14:11 <AnMaster> ehird, that is what you get in the main text box of wxmaxima?
21:14:18 <ehird> AnMaster: oh, "Maxima process terminated."
21:14:26 <ehird> guess i have to done configure shit
21:14:36 <AnMaster> ehird, that sounds like what is technically called "a bug"
21:14:38 <fizzie> ehird: Probably at least more than that.
21:14:45 <AnMaster> ehird, it worked out of box for me on Gentoo :D
21:14:55 <ehird> fizzie: but also $$$
21:15:07 <ehird> AnMaster: seems my maxima install isn't workin'
21:15:24 <ehird> "Maxima program: /opt/local/bin/maxima"
21:15:28 <AnMaster> and I don't know, since it just worked out of box here
21:15:40 <fizzie> ehird: There's on their product list "SHDSL/EFM 2-100M" (and I think that technology implies the same speed both ways, so you could get 100M/100M) but it's probably a really large pile of $s for that.
21:15:48 <ehird> add "-s 4010" to additional parameters
21:16:07 <ehird> connect maxima to server on port N
21:16:13 <ehird> and 4010 was in the prefs as wxmaxima's port
21:16:29 <ehird> Maxima encountered a Lisp error:
21:16:29 <ehird> The function SETUP-SERVER is undefined.
21:16:31 <AnMaster> ehird, there are no extra parameters here
21:18:16 <AnMaster> ehird, wxmaxima preferences screenshot, in case it helps: http://omploader.org/vMW95dA
21:18:29 <ehird> ew, you're using clisp
21:18:38 <AnMaster> ehird, yeah sbcl had issues for me.
21:18:59 <AnMaster> ehird, specifically, it doesn't work if you turn off overcommitting memory.
21:19:01 <ehird> after setup-server failing it works
21:19:10 <ehird> it's its whole memory model
21:19:12 <ehird> so don't turn that off
21:19:22 <AnMaster> ehird, um. I consider the program that needs it buggy
21:19:29 <ehird> no, it's by design
21:19:33 <ehird> it reserves a pool of memory
21:19:58 <ehird> what's wrong with overcommits
21:20:38 <AnMaster> ehird, the fact that the OOM killer can kill any process. Not just the offending one (which isn't trivial to figure out either)
21:20:48 <ehird> Disable the OOM killer, then.
21:21:13 <AnMaster> ehird, yes, that is what I did, when you disable it, you disable memory overcommit
21:21:44 <AnMaster> ehird, how would you handle out of memory when overcommit is in use, without an OOM killer
21:22:08 <ehird> AnMaster: you can refuse to allocate
21:22:12 <ehird> if you run out of real memory
21:22:22 <AnMaster> ehird, um, you already did when you overcommited.
21:22:32 <ehird> you ran out of virtual memory
21:23:14 <AnMaster> ehird, well I have 1.5 GB RAM. So It would refuse to allocate that 8 GB block in either case.
21:23:36 <AnMaster> ehird, then what exactly do you mean
21:23:40 <ehird> you have 1.5GB of PHYSICAL ram
21:23:44 <ehird> you only refuse to allocate when you run out of i
21:23:57 <ehird> AnMaster: the 8GB would fail if and only if you were out of physical memory
21:24:13 <AnMaster> ehird, that is what overcommit is. But it doesn't solve the issue.
21:24:24 <AnMaster> Since you could allocate those 8 GB and then start filling it!
21:24:25 <ehird> what issue? it's overcommit without OOM killer
21:24:47 <AnMaster> what would happen when it loaded data in more than 1.5 GB + 1.5 GB of swap
21:26:01 <AnMaster> ehird, But what if the app filled all but one page of that. And then syslog came along and wanted two extra pages (it also pre-allocated that).
21:26:23 <AnMaster> ehird, sure this is such a good idea?
21:26:25 <ehird> AnMaster: you're out of memory.
21:26:36 <ehird> you can't expect everything to go swimmingly when you're out of memory
21:26:41 <AnMaster> ehird, that is why you disable memory overcommit
21:26:49 <AnMaster> so apps can handle it gracefully.
21:27:02 <AnMaster> or at least log an error and exit
21:27:06 <ehird> evidently you're not listening to me, so I won't bother
21:27:18 <AnMaster> ehird, evidently you're not listening to me either.
21:27:31 <AnMaster> ehird, and, OOM killer is exactly what you suggested above
21:27:39 <ehird> ooh awesome, you have to use "wxplot2d" instead of "plot2d".
21:27:55 -!- tombom_ has joined.
21:28:04 <AnMaster> ehird, it SIGSEGVs (or SIGKILLs forgot which) the process that tries to use more than the RAM around.
21:28:12 <AnMaster> ehird, so you just reinvented the OOM killer
21:28:37 -!- coppro has quit (Read error: 110 (Connection timed out)).
21:28:48 <AnMaster> ehird, got a suggestion NOT involving the OOM killer.
21:28:49 <ais523> actually, the OOM killer uses a different algorithm to "the first process to request memory"
21:29:04 <AnMaster> ais523, right, it used to use that algo before though iirc.
21:29:23 <AnMaster> iirc nowdays it tries to find memory offender and avoid init
21:30:08 <ais523> going for today, anyway; bye everyone!
21:30:08 -!- BeholdMyGlory has quit (Remote closed the connection).
21:30:10 -!- ais523 has quit (Remote closed the connection).
21:30:26 -!- BeholdMyGlory has joined.
21:30:34 -!- BeholdMyGlory has quit (Remote closed the connection).
21:30:55 -!- BeholdMyGlory has joined.
21:39:32 <fizzie> Current OOM killer chooses the "worst" process, worst being defined as the one which gets most points from the badness() function; that's mostly based on the total memory size of the process, plus half the vmsize of its children (to kill memory-hungry forking servers), and tweaks (points /= sqrt(cpu_time), points /= sqrt(sqrt(run_time)), *2 for niced processes because they're unimportant, /4 for SYS_ADMIN/SYS_RESOURCE/SYS_RAWIO processes, ...)
21:41:10 <fizzie> Deewiant: There's some sort of justification for most things, but for the time-based ones... "CPU time is in tens of seconds and run time is in thousands of seconds. There is no particular reason for this other than that it turned out to work very well in practice."
21:42:01 <Deewiant> But in general, that. sqrt one and sqrt-sqrt the other, *2 for something, /4 for something else...
21:42:13 <fizzie> The "half of children" is there so that it kills servers with memory-hungry children, but (since it's only half) it kills a particular child if it's only that child that's eating most of the memory.
21:42:17 <Deewiant> The justification is, of course, that it works well.
21:42:26 <Deewiant> But I still don't like that kind of arbitrary stuff. :-/
21:42:39 <fizzie> Yes, the actual amount of the adjustment is pretty arbitrary.
21:43:17 -!- KingOfKarlsruhe has joined.
21:43:28 <fizzie> (It's actually first /4 for cap_sys_admin/cap_sys_resource processes, and then a further /4 for cap_sys_rawio, so I guess some processes could get /16 out of that.)
21:43:51 <fizzie> Too bad top can't show the badness values; I'd actually like to know what it would kill first here.
21:44:51 -!- tombom has quit (Read error: 110 (Connection timed out)).
21:44:51 -!- tombom_ has changed nick to tombom.
21:45:55 <fizzie> (Okay, with 783/3951MB memory use it's not likely to kill anything, but still.)
21:49:16 -!- BeholdMyGlory has quit (Remote closed the connection).
21:49:30 -!- BeholdMyGlory has joined.
21:50:57 -!- BeholdMyGlory has quit (Remote closed the connection).
21:51:23 -!- BeholdMyGlory has joined.
21:58:13 <AnMaster> fizzie, you can see it in /proc/pid/oom_* iirc
21:58:26 <AnMaster> to make it less likely to kill some specific process
21:59:11 -!- M0ny has quit ("Read error: 182 (Connection reset by beer)").
22:01:34 <fizzie> Oh, right, there's the oom_score. I did notice the oom_adj tweak though.
22:03:17 <Sgeo> I AM GOING TO BRUTALLY MURDER PAYPAL
22:03:32 -!- kar8nga has quit (Remote closed the connection).
22:03:33 <fizzie> Heh.This is a bit... suspicious. The process most likely to be killed is "vi oom_kill.c". Protecting itself, is it?
22:06:40 -!- BeholdMyGlory has quit (Remote closed the connection).
22:06:49 <oerjan> whatever you do, don't try to shut it off
22:07:01 -!- BeholdMyGlory has joined.
22:08:10 <Sgeo> Any help here?
22:08:22 <Sgeo> Oh, I forgot to say what I need help with
22:08:23 <Sgeo> Does anyone know how to force paypal to let me pay with the money that I have in there already, without connecting it to RL financials?
22:09:47 <fizzie> "for pid in [0-9]*; do echo `cat $pid/oom_score`: `cat $pid/cmdline | tr '\0' ' '`; done | sort -n | tail" seems to be a useful oom-score viewer.
22:10:20 <Deewiant> I was trying to do that and wondering why my output was all b0rken
22:10:57 <Deewiant> 1153943: C:\windows\system32\services.exe
22:11:21 <fizzie> Currently it seems to want to kill a un-offending "vi" that has some random notes in it, and "urxvtd" (which would kill all my terminal windows) is second. Third is PostgreSQL.
22:11:36 <fizzie> Would've though Firefox to be a candidate, too, but I guess it's been running for quite a while.
22:12:02 <Deewiant> Then more Wine-stuff, then init.
22:12:09 -!- Judofyr has joined.
22:13:26 <fizzie> 1153943 is quite a score; my winner just gets 19574.
22:14:03 <Deewiant> With 'tail' my lowest is 45405
22:14:29 <fizzie> I have values 9807-19574 in the... "top-10", or "bottom-10", or anything.
22:14:39 <fizzie> Init gets 15644 here, #4.
22:14:49 <Deewiant> Openbox, the current #2, is 593089
22:14:58 <fizzie> Your numbers are bigger.
22:15:18 <fizzie> Wouldn't think you have *that* much more memory (or bigger processes), though; 4G here.
22:15:38 <Deewiant> I have a bunch of zeroes with empty command lines, /sbin/udevd is the lowest 'real' process at 0
22:16:02 <Deewiant> I've got 8G of memory, init's virtual size is 3784 (K, I guess)
22:16:20 <fizzie> Those empty ones are the kernel threads, I guess.
22:16:25 <Deewiant> Maybe your uptime is just so huge?
22:16:25 -!- BeholdMyGlory has quit (Remote closed the connection).
22:16:25 <pikhq> What, and you don't have gettys?
22:16:43 -!- BeholdMyGlory has joined.
22:17:13 -!- BeholdMyGlory has quit (Remote closed the connection).
22:17:13 <Deewiant> mount.ntfs-3g has 17, being the second-lowest process with a cmdline
22:17:14 <fizzie> If you replace the $pid/cmdline cat with "ps hp $pid" you get their names too, but that's not exactly interesting since the scores are 0.
22:17:28 -!- BeholdMyGlory has joined.
22:17:46 <fizzie> My lowest-scoring ones are at score 14: five gettys, klogd and acpid.
22:18:21 -!- oerjan has quit ("Good night").
22:18:41 <Deewiant> Well, mine is less than 13 hours
22:19:31 <pikhq> 1153026: c:\windows\system32\services.exe
22:19:37 <pikhq> Definitely no surprises there.
22:21:06 <Deewiant> That took off, for some reason
22:22:07 <fizzie> Heh, I opened another "vi oom_kill.c", and it jumped to #1, with a 39198 score. It's definitely some sort of self-protecting thing.
22:22:19 <Deewiant> Wonder what that 'supervising' is about
22:22:51 <pikhq> Why is argv[0] SCREEN?
22:23:35 <pikhq> 2324: /usr/bin/../lib32/../bin/wineserver
22:23:53 <fizzie> Deewiant: "-:0" seems to be wdm, actually.
22:23:56 <Deewiant> fizzie: My OS is much less self-protective; it only gave 13841 to a vim oom_kill.c
22:24:05 <pikhq> bin/../lib32/../bin/
22:24:19 <Deewiant> Meanwhile, openbox is trying to show off
22:24:55 <fizzie> Well, a subprocess of wdm, anyway. It's a sibling of the actual X server, the one that's the immediate parent of the actual X session.
22:25:31 <Deewiant> /bin/login -- -> -zsh -> /bin/sh /usr/bin/startx -> xinit
22:25:52 <Deewiant> Which would explain it, I guess.
22:26:05 <fizzie> Funny, the ones I have score=0 (udevd and sshd) also have a -17 in the oom_adj file. I guess someone's protecting them.
22:26:33 <Deewiant> Indeed, udevd has -17 here as well
22:26:34 <fizzie> "export SSHD_OOM_ADJUST=-17" in Debian's /etc/init.d/ssh script.
22:26:43 <Deewiant> sshd has no such protection and is at 405
22:27:04 <Deewiant> dhcpcd and crond are much lower, at 64 and 45 respectively
22:27:47 -!- KingOfKarlsruhe has quit (Remote closed the connection).
22:27:48 <fizzie> I guess they're trying to avoid the "tra-la-la, I'm remotely administering my system that's 500 miles away in a locked room... hmm, I seem to be running a bit low on memory, let's see... hey, why did it get disconnected?" case.
22:28:13 <Deewiant> That makes sense, but I can't really conceive of a case where sshd would be the one to be killed
22:28:17 <fizzie> Though Murphy's law says the frustrated OOM killer is just going to kill the shell.
22:28:52 <Deewiant> The 'sort -n' on the output of that has a higher value than firefox
22:29:24 <fizzie> I've wondered about SCREEN uppercasing its name earlier, too. Not sure why it does that.
22:30:37 -!- tombom has quit ("Peace and Protection 4.22.2").
22:30:49 <fizzie> if (!strncmp("screen", ap, 6)) { strncpy(ap, "SCREEN", 6); /* name this process "SCREEN-BACKEND" */
22:30:56 <fizzie> That's not really a justification there.
22:31:05 <fizzie> It doesn't even name it SCREEN-BACKEND.
22:31:17 <Deewiant> strncpy(ap, "JOE", 3); /* name this process JACK */
22:32:13 <fizzie> That test is done "while (ap >= av0)" and after that there's ap--;
22:32:20 <AnMaster> <fizzie> Heh.This is a bit... suspicious. The process most likely to be killed is "vi oom_kill.c". Protecting itself, is it?
22:32:37 <ehird> 22:09 bsmntbombdood: hehe
22:32:37 <ehird> 22:09 bsmntbombdood: tmpfs is awesome
22:35:06 <AnMaster> wow, would be enough to build open office ON A RAM DISK right?
22:35:53 <AnMaster> bsmntbombdood, would be a tight fit if you had source tree on ram disk too though
22:36:45 <ehird> AnMaster: not crazy
22:36:50 <ehird> 12GB is common for i7
22:37:11 <ehird> Deewiant: i7 is the uncommon thing.
22:37:50 <AnMaster> ehird, "crazy being able to build OpenOffice on a ram disk"
22:38:13 <Deewiant> ehird: I'd say 6 Go is more common
22:38:41 <ehird> but it's more like 6GB: 55% 12GB: 45%
22:38:55 <Deewiant> As Core i7 increases in popularity that ratio will shift
22:38:57 <ehird> since i7 is, well, excessive.
22:39:06 <fizzie> 87 % of statistics are all made up on the spot.
22:39:08 <AnMaster> <fizzie> Giga-octets. <-- really?
22:39:18 <ehird> Deewiant: i7 IS for mega-super-gamer-moneywasters...
22:39:42 <Deewiant> ehird: For now, since it's the new thing.
22:39:55 <Deewiant> ehird: Having more than one CPU core was like that a couple of years back.
22:39:59 <ehird> Deewiant: They're planning on a separate model for regular people.
22:41:14 <fizzie> I'm not quite sure where they're getting the numbers from.
22:41:39 <ehird> http://en.wikipedia.org/wiki/Intel_Core_i5
22:41:44 <ehird> Core i5 (codenamed Lynnfield)[1] is an upcoming family of Intel desktop x86-64 microprocessors. It is scheduled to be released in the third quarter of 2009[2] using the Intel Nehalem microarchitecture and is a mainstream variant to the Intel Core i7 family
22:41:57 <ehird> so, I predict that 6GB only being slightly more popular than 12GB for i7 will continue
22:42:05 <ehird> also typing without one finger is hard
22:43:25 <fizzie> You were in a gruesome industrial accident and lost a finger? Oh, not that? Aww. :/
22:44:05 <ehird> bsmntbombdood: Because everyone is excessive like you :)
22:46:10 <fizzie> Well, "mainstream" *is* an anagram for "mate rams in".
22:48:09 -!- olsner has quit ("Leaving").
22:49:27 <ehird> bsmntbombdood: there seems to be a tendency to overupgrade when you have obsolete hw :)
22:50:05 <AnMaster> bsmntbombdood, since you have so much ram, would you mind compiling something for me, which is too large for me to test. GCC OOMs here.
22:50:59 <AnMaster> bsmntbombdood, current LostKing from my BF-to-(C|other backends planned) compiler
22:51:05 <AnMaster> bsmntbombdood, it is auto generated.
22:51:08 <ehird> AnMaster: it'd take hours
22:51:11 <fizzie> Oh, I hadn't run into "slabtop" earlier. It seems to be for when you want to see how the kernel's wasting memory.
22:51:29 <AnMaster> ehird, I'd be happy with -O0 or -O1
22:51:46 <ehird> enable overcommit :P
22:52:47 <fizzie> I could compile it on that "64GB of memory" shell server, but I guess that would be abusing it somewhat. (And I wouldn't be too surprised if they'd actually added some ulimits there.)
22:53:03 <AnMaster> fizzie, http://rage.kuonet.org/~anmaster/LostKng.b.c.gz compile with: gunzip LostKng.b.c.gz; gcc -std=c99 LostKng.b.c
22:53:36 <fizzie> Well, the shell server is in test use right now, and they did tell us to report any problems. :p
22:53:37 <AnMaster> anyway I'm just interested in if it produces a correct program atm.
22:54:05 -!- Judofyr has quit (Remote closed the connection).
22:54:05 <pikhq> AnMaster: ... You still gzip things?
22:54:07 <AnMaster> bsmntbombdood, If you do get a working binary: tell me how large it is
22:54:32 <ehird> bsmntbombdood: do -mtune=i7
22:54:35 <AnMaster> bsmntbombdood, yeah, but not Intel specifc. -march=athlon64 here :P
22:55:04 <fizzie> bsmntbombdood: It's a generic shell server for all the university's... 10k or so students, I guess.
22:55:12 <AnMaster> actually, -march=athlon64-sse3 for me, if gcc is recent enough
22:55:39 <pikhq> -march=athlon64-sse3 instead of -march=athlon64 -msse3? Whoo.
22:55:48 <bsmntbombdood> nice -n 19 gcc LostKng.b.c -std=c99 -O2 -march=athlon64-sse3
22:55:50 <AnMaster> pikhq, scheduling differences iirc
22:55:54 <ehird> http://www.youtube.com/watch?v=JXnvVlNOcoM Pentium 3!
22:55:59 <ehird> bsmntbombdood: err
22:56:05 <ehird> how about neutral amd64
22:56:07 <fizzie> rogue ~ 32 % gcc -o LostKng LostKng.b.c -O0
22:56:07 <fizzie> rogue ~ 33 % ls -l LostKng
22:56:07 <fizzie> -rwxr-xr-x 1 htkallas users 1698778 May 18 00:55 LostKng
22:56:11 <ehird> so you can actually run it?
22:56:15 <ehird> fizzie: wow, that was fast
22:56:36 <AnMaster> -march=generic -mtune=athlon64-sse3
22:56:38 <ehird> bsmntbombdood: brainfuck adventure game
22:56:42 <fizzie> rogue ~ 34 % ./LostKng
22:56:42 <fizzie> (C) Jon Ripley 2004, 2005
22:56:42 <fizzie> Brainfuck Edition v0.11
22:56:46 <fizzie> It seems to mostly work.
22:56:49 <ehird> translated from BF-BASIC
22:57:01 <fizzie> AnMaster: Well, I mean, as far as I've tested.
22:57:08 <fizzie> AnMaster: I just pasted you the "ls -l" up there.
22:57:22 <ehird> bsmntbombdood: 64GB of ram
22:57:26 <ehird> so I assume several Xeons
22:57:30 <fizzie> bsmntbombdood: "2 x E5450 @ 3.00GHz"
22:57:31 <AnMaster> fizzie, if it is something I can run, could you put it somewhere I can download it
22:57:51 <ehird> are thos edual core?
22:57:55 <ehird> they're xeon harpertowns
22:58:13 <ehird> AnMaster: I was asking about the system.
22:58:21 <fizzie> AnMaster: I didn't give any flags, so I guess it should be quite generic.
22:58:32 <AnMaster> fizzie, can ou put it somewhere for download then!
22:58:36 * pikhq goes to recompile his entire system
22:58:44 <fizzie> AnMaster: http://users.tkk.fi/~htkallas/LostKng should work.
22:58:51 <pikhq> AnMaster: Changed CFLAGS.
22:58:56 <ehird> that's what they do
22:59:06 <AnMaster> pikhq, um, and? You can recompile it as you go
22:59:08 <fizzie> AnMaster: Just wget it.
22:59:30 <fizzie> It didn't quite guess the content-type, since there was no extension.
22:59:33 <AnMaster> Deewiant, yeah, I just told what wget said
22:59:37 <fizzie> Maybe I should've gzipped it.
22:59:45 <fizzie> Runs on my Athlon X2, anyway.
22:59:49 <pikhq> Eh, I've changed a few other things than just CFLAGS.
22:59:52 <AnMaster> better guess for application/octet-stream
22:59:58 <pikhq> CFLAGS is just the straw that broke the camel's back.
23:00:11 <Deewiant> Why is octet-stream application/ anyway
23:00:24 <ehird> bsmntbombdood: just ^C it and "-O3 -march=i7 -mtune=i7" if you're in for the long haul
23:00:35 <AnMaster> application, image, text, audio, video?
23:00:36 <ehird> bsmntbombdood: then you'll get a result you can actually use... and that is super-fast
23:00:50 <fizzie> IANA has defined application, audio, example, image, message, model, multipart, text and video.
23:00:54 <ehird> bsmntbombdood: bah :)
23:01:05 <AnMaster> ehird, and I'd like to be able to run it
23:01:14 <ehird> AnMaster: fizzie just compiled it for you.
23:01:16 <fizzie> VRML, for example, is model/vrml.
23:01:26 <pikhq> ehird: -march implies -mtune.
23:01:40 <ehird> AnMaster: he didn't do -O2
23:01:48 <Deewiant> ehird: mtune is "do generic stuff that is faster on this arch", march is "do generic and this-arch-specific stuff that is faster on this arch"
23:02:09 <bsmntbombdood> fizzie: time -O2, i want to see how much faster than xeon is
23:02:26 <ehird> bsmntbombdood: its not current-gen xen
23:02:39 <fizzie> VRML, for example, is model/vrml.
23:02:44 <ehird> Annoying thing with OS X's invert screen feature: it doesn't fix up subpixel rendering.
23:02:50 <ehird> fizzie: VRML, for example, is model/vrml.
23:02:51 <fizzie> Waiting for the O2 now.
23:02:55 <ehird> I repeat myself. I repeat myself.
23:02:58 <AnMaster> fizzie, what about example and message?
23:03:04 <ehird> bsmntbombdood: you got 8M of l3
23:03:22 <fizzie> AnMaster: "The 'example' media type is used for examples. Any subtype following the media type syntax may be used in those examples. No subtype can be registered with IANA."
23:03:41 <AnMaster> ehird, and GCC is single threaded.
23:03:48 <ehird> AnMaster: 256K L2 per core.
23:03:51 <ehird> And I know, goddamn!
23:03:56 <ehird> Stop fucking saying that!
23:03:58 <ehird> I never disputed it...
23:03:59 <AnMaster> ehird, yep, but it will only use one
23:04:10 <fizzie> There's message/http (RFC2616), message/disposition-notification (RFC3798) and so on. http://www.iana.org/assignments/media-types/ is the definitive source.
23:04:18 <fizzie> bsmntbombdood: Well, it's not actually *my* xeon.
23:04:22 <AnMaster> bsmntbombdood, upload it somewhere?
23:05:41 <ehird> hey bsmntbombdood, upload it
23:06:13 <fizzie> The -O2 is being slow... I'm not sure I'll bother running that for very much longer, given that there are other users for the system. Though using just one core out of eight is maybe not making them so unhappy.
23:06:24 <ehird> does clang multithread
23:06:37 <fizzie> A progress bar would be nice, though. It's taken at least four minutes now.
23:06:44 <fizzie> bsmntbombdood: How long did your compilation take, anyway?
23:06:46 <AnMaster> ehird, not afaik, You can't really multi thread on a single file well.
23:06:56 <ehird> AnMaster: sure you can
23:07:01 <ehird> procedures, for instance
23:07:31 <AnMaster> ehird, good compilers nowdays consider the whole file at once, to be able to do interprocedural optimisation
23:07:46 <ehird> AnMaster: yes, but once they get down to compiling each procedure...
23:07:50 <AnMaster> ehird, and some even considers the whole program at once. Like icc's -ipo mode
23:08:00 <AnMaster> ehird, the code generation phase you mean?
23:08:11 <AnMaster> normally that is a tiny part of it
23:08:13 <fizzie> I'll have to just state "-O2 takes longer than 5 minutes" and be done with it; it's not really my box, can't abuse it like that.
23:08:19 <ehird> AnMaster: orly? -O0 can be slow.
23:08:32 <Deewiant> With LDC, -O0 is slower than -O1
23:08:34 <AnMaster> ehird, for C or C++? Parsing both takes a lot of the time
23:08:42 <ehird> AnMaster: C is easy to parse...
23:08:48 <Deewiant> ehird: Less code generated speeds up the LLVM backend
23:08:50 <ehird> fizzie: i'd hate having access, but not reign, to/on a powerful server
23:10:33 <fizzie> Which gcc, btw? "4.3.3-5ubuntu4" at that server.
23:11:10 <ehird> fizzie: that thing's running ubuntu?
23:11:19 <ehird> that's sort of worrying
23:11:38 <ehird> i mean, it's an excellent desktop OS, but...
23:11:46 <fizzie> ehird: I guess they want to just learn one thing well; all the Linux workstations run Ubuntu too.
23:12:34 <ehird> bsmntbombdood: er, lm-sensors thing?
23:12:42 <Sgeo> It's a miracle! ehird's in Sine!
23:12:50 <ehird> Get over yourself :P
23:13:02 <ehird> Sgeo: kyevan dragged me in yesterday
23:14:21 <ehird> AnMaster: nothing.
23:14:27 <ehird> (you asked that yesterday...)
23:14:58 <fizzie> The motherboard hardware-monitoring chipsets are all over the place, but I think "modprobe coretemp" + sensors should work rather more reliably. (Well, assuming they've updated the coretemp driver for i7 models, and it still supports the same things. I have no idea whether that's true.)
23:14:59 <ehird> bsmntbombdood: -march=i7
23:15:06 <AnMaster> bsmntbombdood, -fwhole-program -combine -O3 -march=athlon64-msse3 -fgcse-sm -fgcse-las -fgcse-after-reload -fsee -fipa-pta -fipa-cp -ftree-loop-linear -ftree-loop-im -ftree-loop-ivcanon -fivopts
23:15:08 <ehird> bsmntbombdood: speed
23:15:19 <AnMaster> ehird, that is suboptimal for me though
23:15:27 <fizzie> bsmntbombdood: Well, the system couldn't have been *much* faster than yours, since it hadn't finished in 5 minutes.
23:15:32 <ehird> AnMaster: he already compiled it for you.
23:15:34 <bsmntbombdood> FATAL: Error inserting coretemp (/lib/modules/2.6.26-2-amd64/kernel/drivers/hwmon/coretemp.ko): No such device
23:15:47 <fizzie> 2.6.26 is not very new.
23:16:01 <ehird> core != i7? What a meaningless equation.
23:16:13 <fizzie> i7 is of the Core family, though.
23:16:20 <AnMaster> then that core module is too old
23:16:24 <fizzie> Given that even Atom does the coretemp stuff, I'd guess i7 does too.
23:16:35 <fizzie> bsmntbombdood: What's your /proc/cpuinfo "model" number?
23:16:39 <ehird> AnMaster: he's running debian testing
23:17:29 <fizzie> bsmntbombdood: Yes; that works with a newer coretemp module. So you'll get it sooner or later.
23:17:51 <ehird> bsmntbombdood: arch might have the newer module :-P
23:18:02 * AnMaster has it on that arch computer already
23:18:17 <ehird> http://www.archlinux.org/packages/core/x86_64/gcc/ # 4.4!
23:18:23 <pikhq> AnMaster: Gentoo is not bleeding edge with its compilers.
23:18:27 <ehird> Clearly, it's 0.1 better.
23:18:33 <ehird> AnMaster: I was looking for it as you typed.
23:18:57 * pikhq just switched to 4.3 a few months ago
23:19:15 <fizzie> For the record; compiled LostKng with -O2 also at home with this Athlon X2. It took 9 minutes, 18 seconds. Used up very little memory for most of the time, then jumped to some 2.3 gigabytes in the end.
23:19:18 <AnMaster> and I didn't recompile everything!
23:19:57 <pikhq> Actually, recompiling for a new version of the compiler is highly recommended; the ABI sometimes changes.
23:20:06 <ehird> % gcc -march=deathstation9000
23:20:07 <ehird> i686-apple-darwin9-gcc-4.0.1: no input files
23:20:10 <Sgeo> http://www.techeblog.com/index.php/tech-gadget/college-students-use-glass-bottles-to-play-tetris-theme
23:20:14 <AnMaster> fizzie, interesting. And I only have 1.5 GB
23:20:20 <ehird> pikhq: er, what about binary blobs?
23:20:29 <ehird> AnMaster: lern2share
23:20:32 <AnMaster> pikhq, yeah I read the GCC changelog to figure out when it does
23:20:50 <ehird> (ehird: lern2words)
23:21:06 <AnMaster> ehird, yes I have swap too. another 1.5 GB. But using it == slow
23:21:14 <ehird> he said it just jumped at the en
23:21:21 <ehird> so it'd only be a few seconds, I assume
23:21:26 <pikhq> ehird: Weird shit happens.
23:21:41 <ehird> i rely on binary blobby sort of things
23:21:42 <AnMaster> ehird, minutes. I have 512 MB ram *free* + 244 MB swap in use
23:21:55 <ehird> fizzie: how long did it use >2
23:21:58 <AnMaster> total used free shared buffers cached
23:22:18 <fizzie> ehird: I wasn't actually looking so closely. Not over a minute.
23:22:48 <ehird> AnMaster: -/+ buffers/cache sez 748 free
23:22:51 <ehird> swap says 276 free
23:23:01 <ehird> AnMaster: you DO know that the kernel reserves a ton of memory for itself, right?
23:23:09 <ehird> unaligned irc client
23:23:11 <AnMaster> ehird, where does it end up with 1 GB
23:23:31 <ehird> AnMaster: "fail2"? lern2meme
23:24:41 <AnMaster> in your i = Use i as referring to math
23:25:05 <fizzie> Beauty is in the eye(5) 5x5 identity matrix of the beholder.
23:25:08 <ehird> actually you just came off as a drug-addled incoherent
23:25:16 <ehird> bsmntbombdood: srsly?
23:25:22 <ehird> your machine is smoking. gimme an ssh connection
23:25:35 <AnMaster> no no. Give me an ssh connection!
23:26:03 <ehird> wait, same binary size?
23:26:09 <ehird> that must mean that -O3 did not much
23:26:12 <ehird> or a very big coincidence
23:26:15 <AnMaster> http://en.wikipedia.org/wiki/Superoptimization
23:26:35 <ehird> AnMaster: superoptimization is totally impractical
23:26:53 <AnMaster> ehird, it is still interesting.
23:27:51 <bsmntbombdood> didn't someone predict this compilation would take hours?
23:28:11 <ehird> but i7s are too fast ;-)
23:28:44 <ehird> bsmntbombdood: do "-O3 -march=i7 -msse4.1"
23:28:46 <AnMaster> actually I got that OOM message after just a few minutes on my own Sempron
23:28:49 <ehird> bsmntbombdood: it'll be ridiculously fast
23:28:54 <ehird> bsmntbombdood: or try icc
23:28:59 <AnMaster> GCC is RAM bount, not memory bound
23:29:09 <ehird> bsmntbombdood: do "-O3 -march=i7 -msse4"
23:29:17 <ehird> it should take about the same time
23:29:21 <ehird> AnMaster: I said that ages ago
23:29:32 <ehird> bsmntbombdood: you should do it anyway, it'll just take 8 minutes
23:29:33 <AnMaster> ehird, you seem to be forgetting it though
23:29:38 <AnMaster> bsmntbombdood, do it on GCC 4.4
23:29:41 <ehird> and it'll be a ridiculously fast computation for ridiculously fast hardware
23:29:49 <AnMaster> with the graphite loop optimiser
23:32:02 <bsmntbombdood> i once tried to compile the output of gperf on a large dictionary
23:32:31 -!- coppro has joined.
23:32:40 <ehird> eaten any feces lately?
23:32:41 <AnMaster> bsmntbombdood, you mean profiled feedback?
23:33:09 <bsmntbombdood> that was just from handling a large array of strings though
23:33:16 <AnMaster> bsmntbombdood, oh the perfect hash one
23:33:36 <AnMaster> bsmntbombdood, gperf works best on small sets. For large ones better use other hash libs.
23:35:00 <AnMaster> bsmntbombdood, it worked well for me on large sets
23:35:26 <AnMaster> bsmntbombdood, http://cmph.sourceforge.net/
23:35:57 <AnMaster> bsmntbombdood, yes. testing that it works
23:36:14 <AnMaster> bsmntbombdood, next step is to improve the BF-to-C translation
23:36:34 <AnMaster> and, esotope-bfc is a way better compiler
23:36:51 <AnMaster> than in-between (the one that generated that output you compiled)
23:37:35 <AnMaster> I do most the the basic stuff it does though. What I don't yet (since I haven't worked out a good way to do it reliably and that works on nested loops): loop-to-polynom
23:37:53 <ehird> bsmntbombdood: compile mandelbrot.b with esotope-bfc, hthen
23:38:03 <ehird> bsmntbombdood: compile mandelbrot.b with esotope-bfc, then "gcc -O3 -march=i7 -msse4"
23:38:11 <ehird> bsmntbombdood: it'll be jaw-droppingly ... fast.
23:38:19 <ehird> bsmntbombdood: AnMaster can :-P
23:38:28 <ehird> AnMaster: link to esotope-bfc
23:39:07 <ehird> bsmntbombdood: i will supply this code on the condition that you use "gcc -O3 -march=i7 -msse4" :-P
23:39:24 <pikhq> ehird: -march=i7 implies -msse4.
23:39:38 <ehird> bsmntbombdood: "gcc -O3 -march=i7" then.
23:39:41 <GregorR-L> http://google.com/search?q=esotope+bfc you mean :P
23:39:47 <pikhq> Now we're talking.
23:40:16 <GregorR-L> !bf_txtgen http://google.com/search?q=
23:40:26 <EgoBot> Supported commands: addinterp bf_txtgen daemon daemons delinterp fyb help info kill mush userinterps 1l 2l adjust asm axo bch bct befunge befunge98 bf bf16 bf32 bf8 bfbignum boolfuck c chiqrsx9p choo cintercal clcintercal cxx dimensifuck echo forth glass glypho hello kipple lambda lazyk linguine malbolge ook pbrain perl qbf rail rhotor rot13 sadol sceql sh show slashes test trigger udage01 underload unlambda whirl yodawg
23:40:39 <AnMaster> GregorR, I suspect quoting fail
23:40:43 <GregorR-L> Oh, hah, it's downloading that page :P
23:40:47 <EgoBot> 33 ++++++++++[>++++++>+>><<<<-]>+.>. [99]
23:41:00 <ehird> bsmntbombdood: http://pastie.org/481059.txt?key=nlsidb0ldiwfqybsv76tcg; "time gcc -O3 -march=i7 mandelbrot.b.c -o mandelbrot"
23:41:01 <AnMaster> GregorR, so how do you force it as text
23:41:03 <ehird> don't forget the time!
23:41:13 <GregorR-L> !bf_txtgen htp://google.com/search?q=
23:41:17 <EgoBot> 201 +++++++++++++++[>+++++++>+++++++>++++>+<<<<-]>-.++++++++++++.----.>>--.-----------..<--.++++++++..--------.<----.>--.>-.<--.<+++.--.>>+.<<++++++.>++.----.<-.>++.+++++.>++++++++++++++++.<<-.>>--.>-----. [730]
23:41:19 <ehird> bsmntbombdood: then "time ./mandelbrot > /dev/null" 5 times
23:41:35 <GregorR-L> !bf +++++++++++++++[>+++++++>+++++++>++++>+<<<<-]>-.++++++++++++..----.>>--.-----------..<--.++++++++..--------.<----.>--.>-.<--.<+++.--.>>+.<<++++++.>++.----.<-.>++.+++++.>++++++++++++++++.<<-.>>--.
23:41:37 <AnMaster> <ehird> don't forget the time!
23:41:44 <AnMaster> ehird, that isn't so much different from mine
23:41:54 <ehird> AnMaster: it's hard to optimize such a program
23:41:54 <GregorR-L> !bf +++++++++++++++[>+++++++>+++++++>++++>+<<<<-]>-.++++++++++++..----.>>--.-----------..<--.++++++++..--------.<----.>--.>-.<--.<+++.--.>>+.<<++++++.>++.----.<-.>++.+++++.>++++++++++++++++.<<-.>>--.>-----.
23:42:02 <ehird> but esotope seems to have dnoe a good job
23:42:17 <ehird> while (p[7] != 0) p -= 9;
23:42:18 <EgoBot> 142844 +++++[>+++++++++++++++++++++++++>++++++++++++++>+++++++++++++++++>++++++++++++++<<<<-]>>----------.>>++++++++++++++++++++++++++++++++++.<<<---------.>+++++++++++++++++++++++++++++++++++++++++++++++++.>+++++++++++++++++++++++.<<------------------------------------------------------.--.>-----.<+++++++++++++++++++++++++++++++++++++++++.----.>----.>----------------------------------------------.>--------------------------------------------.<+++++++
23:42:26 <bsmntbombdood> mandelbrot.c:1: error: bad value (i7) for -march= switch
23:42:33 <AnMaster> <ehird> while (p[7] != 0) p -= 9;
23:42:34 <ehird> bsmntbombdood: upgrade gcc
23:42:43 <ehird> AnMaster: i looked at it and my brain broke for a second
23:42:47 <ehird> for (loopcnt8 = p[43]; loopcnt8 > 0; --loopcnt8) {
23:42:56 <ehird> bsmntbombdood: srsly, upgrade gcc just for this
23:43:04 <ehird> we need i7/sse SHPEED goodness.
23:43:11 <AnMaster> ehird, yes, I did that in "before", and I have the needed metadata already to do it.
23:43:15 <GregorR-L> !addinterp google bf +++++++++++++++[>+++++++>+++++++>++++>+<<<<-]>-.++++++++++++..----.>>--.-----------..<--.++++++++..--------.<----.>--.>-.<--.<+++.--.>>+.<<++++++.>++.----.<-.>++.+++++.>++++++++++++++++.<<-.>>--.,[.,]
23:43:15 <EgoBot> Interpreter google installed.
23:43:26 <ehird> !google let me google that for you
23:43:27 <EgoBot> http://google.com/search?q=let me google that for you
23:43:31 <AnMaster> ehird, but I haven't got around to writing that pass yet.
23:43:57 <bsmntbombdood> with -march=native, running it takes about 1.2 seconds
23:43:59 <AnMaster> ehird, I have all the analysing data needed to do it all. Just need to write the optimiser pass that uses it :(
23:44:18 <EgoBot> http://google.com/search?q=
23:46:26 <EgoBot> http://google.com/search?q=esotope+bfc
23:46:33 <AnMaster> ehird, what does estope generate for the program in !google
23:46:45 <ehird> AnMaster: i'll try
23:46:51 <ehird> 23:43 bsmntbombdood: with -march=native, running it takes about 1.2 seconds
23:46:55 <AnMaster> I will be able to do that once I get the loop unrolling done
23:46:56 <ehird> the top interp takes ~13 secs
23:47:15 <ehird> PUTS("http://google.com/search?q=");
23:47:20 <AnMaster> ehird, output from my compiler running it takes about 2 seconds here. And my CPU is slow
23:47:49 <ehird> AnMaster: not perfect, though
23:47:51 <AnMaster> ehird, I get this currently: http://pastebin.com/m1dc332dd
23:47:54 <ehird> it doesn't need to assign to the tape
23:48:01 <ehird> well, i guess it has to use a variable
23:48:30 <AnMaster> ehird, I already do the const, out -> out_const. Just not the loop flatterning
23:48:32 <ehird> AnMaster: I'll try life.b now
23:48:44 <ehird> http://www.linusakesson.net/programming/brainfuck/life.bf
23:49:01 <ehird> s/bf/b/ since the person evidently doesn't know about befunge-98
23:49:14 <ehird> holy nested loops batman
23:49:45 <AnMaster> out of solidarity or something. </retcon>
23:49:45 <ehird> life and mandelbrot don't get much constantizing
23:49:54 <AnMaster> ehird, it is hard to do in them
23:49:59 <ehird> i bet someone's been testing on lostkng's stupid generated code
23:50:01 <ehird> AnMaster: yeah, still
23:50:10 <AnMaster> I have studied mandelbrot quite a lot
23:50:30 <ehird> how do I get a list of march's
23:51:04 <ehird> didn't know what to grep for
23:51:26 <AnMaster> ehird, grep for k8, that lands you in the right section (x86 specific)
23:51:27 <ehird> AnMaster: that is much shorter. over 300 lines for esotope
23:51:41 <AnMaster> ehird, pastebin your life for esotope!
23:51:56 <ehird> i missed context there for a second
23:52:21 <ehird> AnMaster: http://pastebin.ca/1425899
23:52:29 <ehird> 23:51 AnMaster: ehird, pastebin your life for esotope!
23:52:34 <ehird> i didnt know what life meant there
23:52:39 <AnMaster> examples/life.b.c lines 425-470/470 (END)
23:53:03 <ehird> due to font/browser-editor size differences
23:53:06 <ehird> and lack of indentation I guess
23:53:20 <AnMaster> ehird, indention isn't needed in auto generated code
23:53:29 <AnMaster> if I need that while reading it I do astyle on it
23:53:36 <ehird> it jsut threw me off
23:53:48 <ehird> just prescott and p4 and shit
23:53:58 <ehird> 4.0.1; it's ancient but not THAT ancient
23:54:12 <ehird> it's not older than core2.
23:54:20 <ehird> gcc-4.2 has core2, huh.
23:54:28 <ehird> % time gcc-4.2 -O3 -march=core2 mandelbrot.b.c -o mandelbrot
23:54:38 <ehird> gcc-4.2 -O3 -march=core2 mandelbrot.b.c -o mandelbrot 1.56s user 0.25s system 18% cpu 9.958 total
23:55:00 <ehird> 6.5 seconds first time
23:55:07 <ehird> bff.c does it in 13
23:55:12 <ehird> and it's an interpreter!
23:55:39 <AnMaster> ~5 seconds for -O3 here on my Sempron
23:55:55 <AnMaster> for the one generated by my compiler
23:56:03 <ehird> it's just Eddie's in the space time continuum
23:56:07 <AnMaster> bsmntbombdood, yeah, 2 GHz Sempron 3300+
23:56:25 <AnMaster> bsmntbombdood, http://pastebin.ca/1425908
23:56:35 <ehird> % time gcc-4.2 -O3 -march=core2 life.b.c -o life gcc-4.2 -O3 -march=core2 life.b.c -o life 0.21s user 0.04s system 21% cpu 1.145 total
23:56:39 <ehird> AnMaster: gimme some life.b input
23:56:47 <ehird> i forget the syntax
23:56:52 <ehird> you didn't know of it
23:57:20 <AnMaster> $ time gcc -std=gnu99 -O3 -Wall -Wextra -o life examples/life.b.c
23:57:34 <AnMaster> and I don't know how you do it
23:57:37 <ehird> "while Sweden turned down the euro in a 2003 referendum, and has circumvented the obligation to adopt the euro by not meeting the monetary and budgetary requirements."
23:57:42 <bsmntbombdood> AnMaster: your code takes 1.9 seconds, compared to 1.2 for the other compiler
23:57:47 <ehird> "Hah, take that, EU! We're gonna be poor so you can't force your currency on us."
23:57:58 <ehird> AnMaster: row/col turns it on
23:58:05 <ehird> AnMaster: then an empty line steps a simulation
23:58:22 <ehird> AnMaster: the UK just said "we're fat lazy euro-hating bastards and we'll do what the fuck we want k."
23:59:06 <ehird> AnMaster: to get a glider:
23:59:08 <EgoBot> Interpreter google deleted.
23:59:33 <GregorR-L> !addinterp google bf +++++++++++++++[>+++++++>+++++++>++++>+<<<<-]>-.++++++++++++..----.>>--.-----------..<--.++++++++..--------.<----.>--.>-.<--.<+++.--.>>+.<<++++++.>++.----.<-.>++.+++++.>++++++++++++++++.<<-.>>--.,[>[-]>[-]<<[>+>+<<-]>>>[-]++++++++[<---->-]<[[-]>+<]>-[<<[-]>+++++++[<++++++>-]<+>>[-]]<<.[-]<,]
23:59:33 <EgoBot> Interpreter google installed.
23:59:37 <EgoBot> http://google.com/search?q=hello+world