00:06:48 -!- Sprocklem has joined.
00:17:45 -!- yorick has quit (Read error: Connection reset by peer).
00:21:29 -!- Faris has joined.
00:38:09 -!- Sgeo has joined.
00:42:40 -!- Koen_ has joined.
00:43:29 -!- Koen_ has quit (Read error: Connection reset by peer).
00:43:45 -!- Koen_ has joined.
00:50:04 -!- conehead has quit (Quit: Computer has gone to sleep.).
01:01:10 -!- Phantom_Hoover has quit (Remote host closed the connection).
01:06:54 <Sgeo> > "After analyzing SCP-1461 and SCP-1461-01, it seems that SPC-1461is just made out of silver and nothing else abnormal is found but is about ████ years old and was retrieved in ██/██/████ and it’s still unknown how it dematerializes or let the subject find the object it’s looking for, as for SCP-1461-01 it made out of the object it needs to be and was discovered during testing"
01:06:55 <lambdabot> "After analyzing SCP-1461 and SCP-1461-01, it seems that SPC-1461is just ma...
01:07:05 <Sgeo> "This sentence is really really bad. It's literally at least 4 different ideas shoved into one sentence. It's the human centipede of sentences."
01:09:08 -!- `^_^v has quit (Read error: Connection reset by peer).
01:09:13 <oerjan> are you confusing #esoteric with reddit
01:09:31 -!- `^_^v has joined.
01:14:03 <const> oerjan: is there a difference?
01:16:12 <Sgeo> Read an utterly terrifying SCP tale
01:16:27 <Sgeo> Well, not terrifying
01:16:36 <Sgeo> But it's the thing most likely to give me nightmares tonight
01:16:46 <Sgeo> http://www.scp-wiki.net/pila
01:31:24 -!- kallisti has joined.
01:41:39 -!- nooodl has quit (Quit: Ik ga weg).
01:44:30 -!- tertu has joined.
01:50:00 -!- kallisti has quit (Read error: Connection reset by peer).
01:59:21 -!- dessos has joined.
02:05:16 -!- FreeFull has quit (Ping timeout: 264 seconds).
02:06:51 -!- tertu has quit (Ping timeout: 272 seconds).
02:07:23 -!- FreeFull has joined.
02:08:00 -!- tertu has joined.
02:09:09 -!- shikhin_ has quit (Ping timeout: 248 seconds).
02:09:11 -!- shikhin__ has joined.
02:15:27 <oerjan> > ([1],"a") <> undefined
02:16:36 <Bike> > ([1],"a") <> ([2],"a")
02:17:25 <Bike> i misunderstood.
02:19:17 <Bike> oerjan: were you expecting, like, (1:undefined, 'a':undefined)?
02:19:39 <Bike> > [1] <> undefined
02:20:03 <oerjan> <> is a synonym for ++ for lists
02:20:27 <Bike> it's monoidal add right
02:20:34 <oerjan> i was mostly just wondering if the Monoid instance for tuples unwrapped them lazily.
02:21:15 <oerjan> because there was a stackoverflow answer suggesting it, and i got to think "what about infinite lists?"
02:34:43 <Bike> http://phpmanualmasterpieces.tumblr.com/post/65544023819/ i'm starting to think php could be linked to actual, major economic damage
02:38:42 <Bike> "PHP provides a SecureVar() function which is used to mark variables names as being secure variables. These secure variables can only be set directly in a PHP script, or they can come from a POST method form."
02:40:50 <kmc> «The password defaults to your unix username and falls back on “php” if it can’t figure it out.»
02:40:58 <kmc> «It mentions it’s “a good idea” to change the password but doesn’t actually mention how. »
02:42:43 <kmc> «You do not need to use this leading 0 in PHP since the functions that expect octal parameters are will simplyassume that the parameter is octal.»
02:42:54 <Bike> good use of espurr there
02:43:11 <kmc> meaning that they take decimal 755 and convert it to 0o755 = 493
02:47:30 <ais523> kmc: actually I think it's more likely a special case in the parser
02:47:52 <ais523> that will parse numbers as octal if they're literally the argument to chmod
02:48:08 <Bike> i thought of that too but i stopped thinking about it because i didn't want to think about it.
02:49:49 <kmc> the article makes it sound like what I said but I wouldn't put either kind of horribleness past them
02:49:56 <ais523> Bike: it's less insane than back-converting the base
02:50:36 <Bike> you think special casing function call syntax based on the function name is less insane than something?
02:50:51 <shachaf> http://php.net/manual/en/function.chmod.php says decimal values just don't work
02:51:26 <Bike> shachaf: this is php 2
02:52:43 <shachaf> well anyway the conversion thing sounds ambiguous
02:53:11 <ais523> Bike: yes, I think it's less insane than the alternative suggestion
02:56:54 <kmc> yeah the great thing about the conversion is that if you do write an octal literal, it'll be wrong
02:57:22 <shachaf> oh now i've actually read that page
02:58:44 <shachaf> i thought you meant it supported both but maybe it just supported decimal arguments
02:59:03 <shachaf> maybe it didn't support octal literals
02:59:33 <Bike> tht's the impression i got.
03:01:42 <ais523> oh, something I discovered recently: on Debian, ddate is in the same package as mkfs
03:01:51 <ais523> thus making it part of the "essential" functionality set that can never be removed
03:01:54 <ais523> without breaking the system
03:05:00 <shachaf> do you mean date or is that a different thing
03:05:18 <Bike> ddate, discordian date.
03:11:10 <ais523> yeah, it's the Discordian version of "date"
03:18:49 -!- ais523 has quit.
03:43:27 -!- nisstyre has joined.
03:46:34 <kmc> istr some controversy arising from this
03:47:02 <shachaf> Oh, now I see it on my Debian computer.
03:55:41 <pikhq> It's actually part of util-linux.
04:03:50 <kmc> https://twitter.com/holman/status/395674486973808640
04:05:39 <kmc> i don't have it on Ubuntu :(
04:05:48 <shachaf> since when do you use ubuntu
04:05:54 <shachaf> or is that a mozilla thing
04:06:04 <kmc> since it was one of the easy options for my micro EC2 instance
04:06:14 <kmc> i use Debian on my work and personal laptop and basically every other machine
04:06:39 <Fiora> cygwin doesn't have ddate :<
04:06:47 <Fiora> or, well, not in the default install I guess.
04:06:48 <kmc> Today is Pungenday, the 11th day of The Aftermath in the YOLD 3179
04:07:35 * Fiora should reaalllyyy run the updater on that note, and does so
04:08:28 -!- Sprocklem has quit (Ping timeout: 245 seconds).
04:08:35 -!- shikhin__ has quit (Ping timeout: 240 seconds).
04:08:57 <kmc> why are you winning cygs, anyway
04:09:21 <kmc> i mean 'why are you using cygwin'
04:09:28 <kmc> sorry if that sounds more judgmental than I meant it to be
04:09:39 <Fiora> well it's kinda nice to have like, a unixy shell on windows?
04:09:50 <Fiora> and like, a development environment
04:10:10 <Fiora> I could use visual studio or something but I'm bad at it
04:11:37 <HackEgo> Today is Prickle-Prickle, the 12th day of The Aftermath in the YOLD 3179
04:14:16 -!- Faris has quit (Ping timeout: 264 seconds).
04:14:55 <HackEgo> Today is Prickle-Prickle, the 12th day of The Aftermath in the YOLD 3179
04:15:13 <HackEgo> Thu Oct 31 04:15:12 PST 2013
04:16:15 <oerjan> i think it interpreted it as portuguese instead of pacific
04:16:30 <kmc> yeah I guess I meant "why are you using Windows"
04:16:36 <kmc> if I had to use Windows then I would install Cygwin for sure
04:16:40 <kmc> but now I sound even more judgemental!
04:16:41 <HackEgo> Thu Oct 31 11:16:41 UTC 2013
04:16:49 <kmc> ttants: UTC-7, UTF-7
04:17:16 <kmc> `run TZ=Balls date
04:17:18 <HackEgo> Thu Oct 31 04:17:17 Balls 2013
04:17:59 <HackEgo> Today is Prickle-Prickle, the 12th day of The Aftermath in the YOLD 3179
04:18:03 <kmc> `run TZ="'; DROP TABLE timezones --" date
04:18:32 <kmc> `run TZ=America/San_Francisco date
04:18:34 <HackEgo> Thu Oct 31 04:18:34 America 2013
04:19:08 <HackEgo> Today is Pungenday, the 11th day of The Aftermath in the YOLD 3179
04:19:40 <oerjan> something something switching + and -
04:19:50 <HackEgo> Wed Oct 30 21:19:49 UTC 2013
04:20:25 <Bike> `run TZ=UTC+31 ddate
04:20:26 <HackEgo> Today is Pungenday, the 11th day of The Aftermath in the YOLD 3179
04:20:41 <Bike> `run TZ=UTC+631 ddate
04:20:43 <HackEgo> Today is Pungenday, the 11th day of The Aftermath in the YOLD 3179
04:21:02 <oerjan> are you trying to collapse spacetime
04:21:14 <Bike> spin me right round
04:21:26 -!- shikhin__ has joined.
04:21:50 -!- Uguubee111118 has quit (Read error: Connection reset by peer).
04:21:57 <kmc> tomorrow: "spin me right round" featuring Bike
04:24:10 -!- Uguubee111118 has joined.
04:26:31 <Fiora> I guess bike wheels do spin around a lot?
04:28:26 -!- Faris has joined.
04:30:28 <Bike> i don't mean to brag, but...
04:36:54 -!- shikhin_ has joined.
04:39:02 -!- shikhin__ has quit (Ping timeout: 240 seconds).
05:01:31 <Sgeo> I need a stiff drink
05:04:34 <oerjan> Sgeo: those are called solids
05:05:01 <Sgeo> Will solids numb the pain of other people's stupidity?
05:05:24 <oerjan> some will. i think kmc knows more about that than me.
05:05:58 * oerjan hopes this is about an internet debate
05:06:01 <shachaf> `run quote solid | head -n1
05:06:03 <HackEgo> 245) <treederwright> enjoy being locked in your matrix of solidity
05:06:13 <Bike> never been to stupidity
05:07:18 <shachaf> is kmc really as much of a drugz expert as we make him out to be
05:08:01 <Bike> well he said he took acid the other day. that's like 10000% more expertise than me.
05:08:24 <shachaf> kmc: is that true i didn't hear about that
05:14:57 -!- shikhin__ has joined.
05:16:22 <kmc> you could have jello shots
05:16:30 <kmc> shachaf: yes I took acid last Saturday
05:17:30 -!- shikhin_ has quit (Ping timeout: 272 seconds).
05:17:30 <Fiora> he might not take acid very often but he definitely knows the basics
05:17:33 * Fiora hides under the pun blanket
05:18:22 -!- shikhin has joined.
05:20:24 -!- shikhin__ has quit (Ping timeout: 246 seconds).
05:34:37 -!- dessos has quit (Ping timeout: 246 seconds).
05:51:00 <Koen_> I read your acid-flavoured joke and open a page in firefox (which has esolangs.org/wiki/Special:Random as its default page) and I immediately stumble upon "ACIDIC is an esoteric programming language made by User:iconmaster. "
05:52:17 -!- asie has joined.
05:58:02 -!- ^v has quit (Quit: Ping timeout: 1337 seconds).
06:03:32 -!- dessos has joined.
06:07:51 -!- rodgort` has quit (Ping timeout: 260 seconds).
06:08:21 -!- monotone has quit (Ping timeout: 245 seconds).
06:13:13 -!- monotone has joined.
06:18:39 -!- rodgort has joined.
06:24:20 -!- Faris has quit (Ping timeout: 256 seconds).
06:26:49 -!- blueupro_ has joined.
06:27:45 -!- blueupro_ has left ("Textual IRC Client: www.textualapp.com").
06:41:23 -!- oerjan has quit (Quit: Owobble).
06:48:49 -!- epichero has joined.
06:51:01 -!- epichero has left.
06:55:17 -!- Koen_ has quit (Quit: Koen_).
07:01:00 -!- tertu has quit (Ping timeout: 240 seconds).
07:38:53 -!- asie has quit (Quit: My MacBook Pro has gone to sleep. ZZZzzz...).
08:18:37 -!- Taneb has joined.
08:22:04 -!- shikhin has quit (Ping timeout: 264 seconds).
08:26:22 -!- shikhin has joined.
08:28:34 -!- shikhin has quit (Max SendQ exceeded).
08:29:38 -!- shikhin has joined.
08:33:56 -!- shikhin_ has joined.
08:35:08 -!- shikhin has quit (Ping timeout: 245 seconds).
08:39:04 -!- shikhin_ has quit (Ping timeout: 246 seconds).
08:51:58 -!- shikhin_ has joined.
09:09:14 -!- Taneb has quit (Quit: Leaving).
09:50:38 -!- shikhin_ has quit (Ping timeout: 240 seconds).
09:59:24 -!- yorick has joined.
10:26:02 -!- KingOfKarlsruhe has joined.
10:35:16 -!- KingOfKarlsruhe has quit (Ping timeout: 264 seconds).
10:41:13 -!- Uguubee111118 has quit (Ping timeout: 246 seconds).
10:41:59 -!- Uguubee111118 has joined.
10:42:13 -!- Yonkie has quit (Ping timeout: 245 seconds).
10:44:15 -!- nooodl has joined.
10:56:08 -!- Jafet has quit (Read error: Connection reset by peer).
11:05:17 -!- nooodl has quit (Quit: Ik ga weg).
11:24:35 -!- Koen_ has joined.
11:33:43 -!- Sgeo has quit (Read error: Connection reset by peer).
11:44:34 -!- Faris has joined.
11:50:38 -!- boily has joined.
11:50:49 -!- metasepia has joined.
11:50:55 <boily> good milk tea morning!
11:51:34 -!- rodgort has quit (Remote host closed the connection).
11:53:39 -!- nooodl has joined.
11:58:44 -!- rodgort has joined.
11:58:48 -!- Uguubee111118 has quit (Quit: Uguubee111118).
12:14:23 -!- Uguubee111119 has joined.
12:19:18 -!- Uguubee111119 has quit (Quit: Uguubee111119).
12:19:39 -!- Uguubee111119 has joined.
12:26:36 -!- Faris has quit (Ping timeout: 252 seconds).
12:33:24 -!- nooodl has quit (Read error: Connection reset by peer).
12:34:21 -!- shikhin has joined.
12:46:43 -!- Faris has joined.
12:49:40 -!- Faris has quit (Client Quit).
13:07:03 -!- nooodl has joined.
13:51:06 -!- Phantom_Hoover has joined.
13:54:49 -!- mrhmouse has joined.
14:06:27 <fizzie> "The Committee makes no claims that a program written using trigraphs looks attractive." (C99 Rationale document.)
14:23:07 <nooodl> the ??!??! operator is cute though
14:24:24 <boily> I fail to understand the rationale behind C/C++'s {di,tri}graphs.
14:25:29 <Fiora> http://stackoverflow.com/a/1234596 seems to explain it?
14:26:23 <boily> “...this appears to be infeasible for some...” say wut?
14:26:58 <boily> the next answer (mentioning 3270es and suchlike) make a little bit more sense.
14:27:26 <Fiora> like when C was originally standardized not all keyboards/character sets had the necessary characters, I think
14:27:55 <boily> I know of at least one local corpo that shalln't be named that uses plenties of 3270es.
14:28:15 -!- Lymia has quit (Ping timeout: 268 seconds).
14:28:21 <boily> (a friend had to suffer through them. he's permanently scared from the experience.)
14:28:24 <Fiora> oh gosh is that like the classic green screen terminal
14:28:45 <fizzie> boily: "While [trigraphs provide] a solution for the known limitations of EBCDIC (except for the exclamation mark) and ISO/IEC 646, the result is arguably not highly readable." (C99 Rationale, still.)
14:29:19 <boily> EBCDIC has no exclamation mark???!?!!!?!!!one!!six???
14:29:41 * boily is utterly terrified, which is a very suiting emotion today
14:30:07 <fizzie> I think that depends on the EBCDIC.
14:30:29 <fizzie> There's all kinds of EBCDICs.
14:30:40 <fizzie> The table in the Wikipedia article does include a !.
14:31:55 <fizzie> Also, I've done a bit of data entry over some tn3270-for-Windows-over-X.25 thing.
14:32:17 <fizzie> (My first summer job, in fact.)
14:32:55 -!- Taneb has joined.
14:33:19 <Taneb> How would I cite an IRC log
14:34:10 <Bike> "personal communication" maybe
14:35:36 <Bike> https://owl.english.purdue.edu/owl/resource/560/10/ owl has "online interview" and "nonperiodical web document"
14:36:07 <Taneb> Also I'm writing an essay on the impact of On Computable Numbners, With An Application TO The Entscheidungsproblem
14:37:23 <Bike> 'made hilbert very sad'
14:37:53 -!- conehead has joined.
14:38:05 -!- conehead has quit (Remote host closed the connection).
14:38:21 -!- conehead has joined.
14:38:33 <Taneb> I was wondering if I could talk about ais523's prize-winning proof
14:38:50 <Bike> i don't see why not.
14:38:59 <Bike> hell you could 'interview' him.
14:40:12 <Taneb> That's what I was thinking
14:40:17 <fizzie> Bike: There's some joke about "personal communication" in that jokey-methods-of-proof list.
14:40:45 <fizzie> http://pauillac.inria.fr/~xleroy/stuff/how-to-prove-it.html "Proof by personal communication: 'Eight-dimensional colored cycle stripping is NP-complete [Karp, personal communication].'"
14:43:08 -!- Koen_ has quit (Read error: Connection reset by peer).
14:43:17 -!- Koen_ has joined.
14:43:58 <Bike> i was going to ask my PI how "unpublished data" citations worked but i didn't have the balls
14:44:32 <fizzie> (Some of the other ones on the page are kind of funny (and perhaps sad-but-true) too.)
14:45:44 <Bike> 'To see that infinite-dimensional colored cycle stripping is decidable, we reduce it to the halting problem' heh
14:48:33 -!- MindlessDrone has joined.
14:49:30 <Phantom_Hoover> meanwhile: the person who marks all my core assignments this year is, as it turns out, a total idiot
14:50:02 -!- Sprocklem has joined.
14:50:49 <Taneb> Phantom_Hoover: :/
14:51:41 <Phantom_Hoover> i spent most of an hour-long supervision trying to get her to explain why you couldn't assume that the existence of an eigenvalue implied the existence of a corresponding eigenvector
14:55:12 <Bike> what would an eigenvalue without an eigenvector mean
15:02:31 -!- Uguubee111119 has quit (Quit: Uguubee111119).
15:04:06 -!- nooodl has quit (Quit: Leaving).
15:06:29 -!- Uguubee111119 has joined.
15:07:23 -!- MindlessDrone has quit (Ping timeout: 272 seconds).
15:07:45 <fizzie> Bike: It's a \lambda such that A = \lambda. See, you just leave the vectors out.
15:08:09 <Phantom_Hoover> Bike, well she seemed to take it to mean complex eigenvalues of real matrices
15:10:34 <Bike> doesn't that just mean you get complex eigenvectors
15:11:30 <Fiora> maybe it's like the thing where a root of a polynomial "doesn't exist" (code for "complex")?
15:12:04 <Bike> yes but that's silly and complexification is a godsend >_>
15:12:28 <Phantom_Hoover> best part was that the question was asking about matrices over an arbitrary field anyway
15:12:31 <Bike> i'm no good at linear algebra but i think that it is in fact the same thing
15:12:42 <Bike> since the field should be algebraically closed bla bla
15:12:59 -!- nooodl has joined.
15:14:16 <Phantom_Hoover> oh and also she marked four of us down because of this and gave another guy who'd used the exact same reasoning full marks
15:15:50 <Taneb> Phantom_Hoover: identify the relevant authorities and file a complaint
15:16:25 -!- ^v has joined.
15:16:50 <Bike> ehat's the deal with that anyway
15:17:37 <Phantom_Hoover> they're not getting very much of the tuition fees the government keeps jacking up to 'pay for students' education'
15:18:48 <Bike> sounds like murka
15:19:05 <boily> sounds like canada.
15:19:36 <Phantom_Hoover> student loans here are lenient enough that i find it very hard to believe raising fees actually makes any money at all
15:20:54 -!- MindlessDrone has joined.
15:27:06 -!- Taneb has quit (Quit: Page closed).
15:31:02 -!- Sprocklem has quit (Ping timeout: 265 seconds).
15:43:29 -!- Uguubee111119 has quit (Read error: Connection reset by peer).
15:51:03 -!- Uguubee111119 has joined.
15:51:42 <boily> I feel like relcomming somebody...
15:52:19 -!- oerjan has joined.
15:54:56 <oerjan> fizzie: please tell the email spammers i'm not finnish thx
15:58:48 <oerjan> <boily> (a friend had to suffer through them. he's permanently scared from the experience.) <-- that might be even worse than being permanently scarred.
16:02:01 <oerjan> i think the first page of google hits for "permanently scared" contains 2, _maybe_ 3 that are not misspellings.
16:02:18 <oerjan> of course personalization yada yada
16:03:04 <oerjan> one hit is about us politics, one is about a rat.
16:04:06 <HackEgo> somebody: Welcome to the international hub for esoteric programming language design and deployment! For more information, check out our wiki: http://esolangs.org/wiki/Main_Page. (For the other kind of esoterica, try #esoteric on irc.dal.net.)
16:19:43 -!- asie has joined.
16:22:57 <oklofok> "Proof by example: The author gives only the case n = 2 and suggests that it contains most of the ideas of the general proof."
16:23:17 <oklofok> i have whole articles that say "we only do the case d=2, the general case is similar"
16:24:17 <oerjan> and of course 3 and 4 is where everything usually breaks down
16:24:33 <nooodl> http://joeyh.name/blog/entry/license_monads/ this is good but i'll be damned if it hasn't been linked here
16:26:21 * oerjan sometimes thinks about the uncited "naturality of the pimsner-voiculescu exact sequence" we left in a paper
16:26:39 <oerjan> it's true, because i checked it (eventually)
16:27:05 <oklofok> i'm pretty sure at least 20% of my published results are wrong
16:27:12 <oklofok> that's why i never do two papers on the same topic
16:28:45 <oerjan> but we never found an actual citation, since it's not the kind of thing non-category theorists would care to prove.
16:29:31 <oerjan> but you can just keep track of the morphism in the usual proof.
16:30:24 <oerjan> well things have to be natural if you aren't doing something requiring arbitrary choices.
16:30:29 -!- asie has quit (Quit: My MacBook Pro has gone to sleep. ZZZzzz...).
16:30:32 <oklofok> btw we sent a paper to ergodic theory and dynamical systems, did i mention
16:31:04 <oklofok> the referee said we suck and that we should cite these 6 papers on orbit equivalence and other largely unrelated topics
16:31:29 <oklofok> i probably did but i'm a bit tired and cannot recall who i told this story
16:31:53 <oerjan> well you did mention a referee saying you sucked
16:32:21 <oerjan> also that some of the papers were by my advisor
16:32:35 <oklofok> that's why i talked about it
16:32:56 <oklofok> journals have really annoying referees
16:33:18 <oklofok> they never find any actual errors, but they can still nag endlessly
16:34:17 <oklofok> (but like little ones that no one cares about)
16:34:23 <oerjan> we once had a paper that was rejected for what we thought were lousy reasons, next time we submitted we explicitly asked them exclude a particular person as referee ("for priority reasons") and it went straight through.
16:35:41 <oerjan> well theoretically that's supposed to mean that that person might be competing with us over the proof we're submitting, i think
16:36:27 <oerjan> it was my collaborator who had the hunch to do this, anyway
16:37:20 <oklofok> the paper about substitutions was rejected because we didn't discuss existing methods
16:37:39 <oklofok> the thing is we haven't read _anything_ on the topic
16:38:04 <oklofok> i mean why would we right, clearly they're doing something wrong since this problem was open
16:39:08 <oklofok> (but mainly it's all some weird spectral/ergodic stuff and i don't understand it)
16:39:54 <oklofok> but yeah probably there's some little detail that we skipped that breaks the whole proof and all that literature is about solving that little thing
16:40:29 <Phantom_Hoover> so the person marking like half of my assignments this year took two weeks to mark a linear algebra assignment
16:40:39 <oklofok> yeah i heard, that's a bit silly
16:40:54 <oklofok> a bit of a coincidence, i checked 100 linear algebra exams today
16:41:08 <oklofok> (but just one out of four problems)
16:41:48 <oerjan> Phantom_Hoover: sounds like when i had a breakdown while TA'ing, did basically the same thing and they had to pass it on to someone else
16:42:57 -!- tertu has joined.
16:43:08 <oklofok> i think laziness is more common than breaking down though
16:44:13 <oklofok> so the problem i checked was something like prove that {(x,y,z) | y = 5x-3z} is a subspace of R^3
16:44:19 <oerjan> nah all the people you are calling lazy are really having breakdowns
16:44:26 <Phantom_Hoover> If it takes her this long to mark 3 calculation questions and an easy proof then I don't think I'm ever going to get any analysis marks back this term.
16:45:25 <oerjan> ah finally i got that piece stuck between my teeth
16:45:35 <oklofok> and also you had to find a basis for something like {(2a+b,a+b-c,-a-c,4a+b) | a,b,c \in R}
16:46:05 <oklofok> erm {(2a+b,a+b-c,-a-c,4a+2b) | a,b,c \in R}
16:46:41 <oklofok> most people said something like 2(2a+b) = 4a+2b, so you can drop 4a+2b. thus, 2a+b, a+b-c and -a-c are the basis.
16:49:40 <oklofok> some took the vectors (2,1,-1,4), (1,1,0,2), (0,-1,-1,0), and checked that they are linearly independent, thus a basis
16:49:52 <oklofok> "well that's a bit hard to show since it's false" you might say
16:50:07 <oklofok> you just compute the determinant of the matrix whose rows are those three vectors
16:52:14 <oklofok> at least one of them were, it has multiple different determinants
16:52:56 <oerjan> at least it is a generating set
16:54:33 -!- Guest62574 has changed nick to ralz.
16:54:36 <oklofok> yeah i basically gave two points out of 12 if you managed to take that finite generating set and say it's a basis/generating set.
16:54:36 -!- ralz has changed nick to realz.
16:54:43 -!- realz has quit (Changing host).
16:54:43 -!- realz has joined.
16:56:21 <oklofok> also note that many students on the course are philosophers and such
16:56:21 <oerjan> it's easy if you're determined enough
16:56:36 <oklofok> there's a separate linear algebra course for math people
16:56:58 * FireFly was confused for a while what the determinant of the 4x3 matrix would be
16:57:02 <oklofok> apparently it's shoppe time now
16:57:03 <oerjan> everyone sucks at different levels
16:57:11 <FireFly> I thought I was missing something first..
16:58:31 <oklofok> is there a notion of determinant where the determinant would be something like the volume of the image
16:59:17 <oklofok> kind of like there's a hausdorff measure but when it's zero, you can compute the (d-1)-dimensional measure and such
17:02:03 <oerjan> problem is, that won't be basis independent
17:02:50 <oerjan> like, if you stretch the space in a direction normal to the image
17:03:19 <oerjan> composing that with your matrix won't change it
17:18:34 <fizzie> oerjan: Contrary to what you might believe, I don't know all email scammers personally.
17:21:13 <fizzie> (Also I've got some students coming to meet me tomorrow morning, eugh.)
17:23:31 -!- metasepia has quit (Ping timeout: 245 seconds).
17:29:04 -!- Lymia has joined.
17:29:04 -!- Lymia has quit (Changing host).
17:29:04 -!- Lymia has joined.
17:37:29 -!- tertu has quit (Ping timeout: 272 seconds).
17:40:15 -!- asie has joined.
17:41:54 -!- oerjan has quit (Quit: leaving).
17:42:16 -!- Phantom_Hoover has quit (Ping timeout: 246 seconds).
17:57:45 -!- Phantom_Hoover has joined.
17:59:14 -!- ais523 has joined.
17:59:33 -!- ^v has quit (Read error: Connection reset by peer).
17:59:50 -!- ^v has joined.
17:59:56 <boily> back from lunch and multiple things and lots of rain and halloween preparations and...
18:00:57 -!- ais523 has changed nick to ais523\unfoog.
18:05:57 -!- shikhin_ has joined.
18:08:48 -!- shikhin has quit (Ping timeout: 240 seconds).
18:12:02 -!- KingOfKarlsruhe has joined.
18:25:19 -!- shikhin_ has quit (*.net *.split).
18:25:20 -!- Fiora has quit (*.net *.split).
18:25:21 -!- lambdabot has quit (*.net *.split).
18:27:57 -!- shikhin_ has joined.
18:35:54 -!- Phantom_Hoover has quit (Ping timeout: 256 seconds).
18:37:35 -!- Phantom_Hoover has joined.
18:44:47 -!- Fiora has joined.
18:50:29 -!- Sprocklem has joined.
18:51:56 <fizzie> What could make an image (with a data: URI) behave so (in Chromium) that the "Open Image in New Tab" and "Save Image As..." options of the context menu are grayed out? It didn't used to do that, before.
18:52:03 -!- epicmonkey has joined.
18:52:29 <olsner> maybe they just removed the feature to save data: urls?
18:52:54 <fizzie> But, I mean, it worked like two days ago.
18:53:10 <fizzie> Though perhaps I upgraded the browser. I guess that's possible.
18:55:10 -!- Phantom_Hoover has quit (Ping timeout: 256 seconds).
18:56:16 <fizzie> If I make it an <a href="data:..."> instead, and right-click that link, I get a "He's dead, Jim!" Chromium crash page.
18:56:38 <boily> looks like you... *puts on glasses*... up-grayed your browser...
18:57:27 <boily> upgraded with gray.
18:57:44 <fizzie> I wonder if there's some way to get an actual error message corresponding to the he's-dead thing.
18:57:53 * boily hides under his desk in shame
18:58:06 <olsner> but maybe it should be upgrayeded as in more-grayed-made
18:58:29 <fizzie> No, I mean, I get *that*.
18:58:36 <fizzie> What I don't get is why it behaves like this.
18:59:05 <fizzie> I think I'll just upgrade it further. Can't get any worse, right?
18:59:30 * boily dons a hazmat suit and gets in a bunker
18:59:40 <boily> go ahead! nothing bad can happen!
19:02:59 -!- epicmonkey has quit (Ping timeout: 272 seconds).
19:03:27 <fizzie> It didn't get worse, but it didn't get any better either.
19:04:22 <fizzie> A non-gray "print..." selection seems to have appeared in the context menu, but other than that it's still the same, and right-clicking the link still crashes the page.
19:06:05 <fizzie> https://code.google.com/p/chromium/issues/detail?id=97108 should work fine, as far as I can tell.
19:10:31 -!- shikhin_ has changed nick to shikhin.
19:11:06 <fizzie> Also if I make the <a> element have a download="foo.png" attribute -- which "Can be used with blob: URLs and data: URLs, to make it easy for users to download content that is generated programmatically" -- Chromium still crashes when I click the link.
19:11:12 <fizzie> I guess it must be some sort of a bug.
19:11:13 -!- Lymia has quit (Ping timeout: 272 seconds).
19:15:12 -!- Koen_ has quit (Quit: Koen_).
19:15:44 <boily> weird. on my chromium “open in new tab” and “save as...” both work, but I have to double-right-click to open the contextual menu...
19:16:02 <boily> (version 30.0.1599.114)
19:16:23 <fizzie> I'm pretty sure it *worked* in this same Chromium. But I might be wrong about that. I'm sure it worked on a Google Chrome.
19:17:23 <fizzie> Also a plain "data:text/plain;charset=UTF-8,hello world" link (not <img>, of course) behaves well (doesn't crash when right-clicked, saves to disk when left-clicked), but not the thing I get from canvas.toDataURL('image/png');.
19:17:39 <fizzie> (Even though it works in img.src fine.)
19:18:11 -!- lambdabot has joined.
19:18:16 -!- KingOfKarlsruhe has quit (Quit: ChatZilla 0.9.90.1 [Firefox 24.0/2013091200]).
19:18:40 <fizzie> It seems to be some sort of a length-related thing.
19:18:47 <fizzie> Because when it's a small picture, it works fine.
19:19:10 -!- augur has quit (Remote host closed the connection).
19:19:29 -!- augur has joined.
19:19:38 <fizzie> I think I'm going to binary-search the exact limit. (If there is one.)
19:19:41 -!- augur has quit (Remote host closed the connection).
19:19:59 <boily> what version are you running?
19:20:07 -!- augur has joined.
19:24:38 -!- augur has quit (Ping timeout: 256 seconds).
19:24:43 <fizzie> 30.0.1599.101, apparently.
19:25:36 <fizzie> Also I can manage a .png of up to 1572663 bytes, but the next step larger from there goes and crashes.
19:25:49 <fizzie> > 1572663*4/3 -- for base64
19:26:12 <fizzie> I'm going to go ahead and say it crashes for data: URIs longer than two megabytes.
19:26:35 -!- Taneb has joined.
19:27:04 <fizzie> The save-as and such also work for smaller images.
19:27:19 <fizzie> I guess it could be some sort of a designed-in limitation. But that's annoying.
19:27:28 <fizzie> Two megs isn't really a whole lot for an image, for one thing.
19:28:16 <fizzie> https://code.google.com/p/chromium/issues/detail?id=45395 heh.
19:29:25 <fizzie> "The issue we had was that arbitrary length URIs (including data URIs) could easily consume excessive resources in the browser process and cause it to crash. As I remember we originally capped it at 10M, but still found ways to trigger crashes until we capped it at 2M."
19:30:08 <Bike> kind of curious why you want to do this
19:30:43 <fizzie> For the obvious reason. I produce an image programmatically, and want to offer a "save it" option.
19:31:32 <Taneb> I've taken out a book on category theory from my understoo
19:31:36 <fizzie> (And Chrome doesn't support canvas.toBlob at all, apparently.)
19:31:52 <Taneb> ...university's library
19:32:05 <Taneb> I may tomorrow ask for help
19:32:27 <Fiora> are there like, javascript jpeg encoders for that?
19:33:07 <Bike> encode image as javascript, kolmogorov minimize that shit
19:33:20 <Bike> a new age of demosceners!!
19:33:24 <fizzie> It's not the kind of image I'd encode as JPEG.
19:33:48 <fizzie> canvas.toDataURL could probably produce a JPEG, anyway.
19:33:55 <fizzie> Still, two megs is two megs.
19:34:06 <fizzie> A PNG is what I'm already producing.
19:34:20 <Bike> encode as xbm, for portability
19:34:35 <Bike> man i fucking love that format
19:34:59 <boily> split the canvas in small canvas, then save the smaller images?
19:35:28 <Taneb> An ascii format like ppm?
19:35:29 <fizzie> boily: I'm sure the hypothetical user would love getting MxN small images and a popup "please assemble the final result in Gimp or something manually" message.
19:36:00 -!- Taneb has quit (Quit: movie starting).
19:36:17 <fizzie> Even if I had a pure-JS PNG encoder, I don't see how a JavaScript encoder could provide a "save as" option except via packing the thing up in a data: URI, in which case I'd run into the same problem. (Unless maybe by making a Blob and a blob object reference URL instead.)
19:37:06 <fizzie> Maybe I can just take the data: URI and base-64 decode it into blob, that's not hard.
19:37:27 <fizzie> (Also shortcuttable if someone bothers to add canvas.toBlob at some point.)
19:38:51 <olsner> maybe you could encode it as ascii instead of base-64?
19:39:17 <fizzie> I don't think I can choose what canvas.toDataURL produces.
19:39:45 <fizzie> Though I could un-base64, theoretically, I guess. Still, it's just that 4/3 factor.
19:40:09 <fizzie> Maybe I'll just not care. Two megs should be enough for everyone.
19:40:10 -!- Yonkie has joined.
19:47:45 <HackEgo> Yonkie: Welcome to the international hub for esoteric programming language design and deployment! For more information, check out our wiki: http://esolangs.org/wiki/Main_Page. (For the other kind of esoterica, try #esoteric on irc.dal.net.)
19:50:51 -!- Sprocklem has quit (Ping timeout: 272 seconds).
19:54:25 <fizzie> Welp, the blob method (via base64 decoding the canvas.toDataURI string, packing it up in a Blob, and doing URL.createObjectURL for the image) works just fine even for large images.
19:55:52 <mrhmouse> Nobody gave me a rainbow welcome..
19:56:07 <HackEgo> mrhmouse: Welcome to the international hub for esoteric programming language design and deployment! For more information, check out our wiki: http://esolangs.org/wiki/Main_Page. (For the other kind of esoterica, try #esoteric on irc.dal.net.)
19:56:20 <Fiora> not everyone gets welcomes, and some people get like, 7 welcomes
19:56:28 <Fiora> it's just, like, a distribution centered around one welcome
19:57:31 -!- Bike has quit (Ping timeout: 245 seconds).
19:57:51 <mrhmouse> I see.. I don't suppose Plugnburn is here under a different name?
19:58:06 <fizzie> Is it a... Poisson distribution?
19:58:24 -!- augur has joined.
19:58:27 <Fiora> Poisson or Laplacian maybe?
19:58:37 <kmc> `relcome ais523\unfoog
19:58:39 <HackEgo> ais523\unfoog: Welcome to the international hub for esoteric programming language design and deployment! For more information, check out our wiki: http://esolangs.org/wiki/Main_Page. (For the other kind of esoterica, try #esoteric on irc.dal.net.)
19:59:20 <fizzie> Fiora: I don't know how make a discrete Laplacian, or give someone non-integral welcomes.
19:59:44 <Fiora> maybe like, round to nearest?
20:00:02 <Fiora> maybe someone should make a fractional welcome script.
20:00:09 <HackEgo> fizzie: Welcome to the international hub for esoteric programming lang
20:00:30 <fizzie> ais523\unfoog: That's still quite discrete. (If admittedly non-integral.)
20:00:52 <olsner> `run relcome fizzie | head -c $RANDOM
20:00:55 <HackEgo> fizzie: Welcome to the international hub for esoteric programming language design and deployment! For more information, check out our wiki: http://esolangs.org/wiki/Main_Page. (For the other kind of esoterica, try #esoteric on irc.dal.net.)
20:01:07 <olsner> ... that was spectacular
20:01:37 <Fiora> `run relcome | sed `s/\(..\)\(.\)\(.*\)/\1\3/g'
20:01:38 <HackEgo> bash: -c: line 0: unexpected EOF while looking for matching ``' \ bash: -c: line 1: syntax error: unexpected end of file
20:01:45 <Fiora> `run relcome | sed 's/\(..\)\(.\)\(.*\)/\1\3/g'
20:01:47 <HackEgo> Welcome to the international hub for esoteric programming language design and deployment! For more information, check out our wiki: http://esolangs.org/wiki/Main_Page. (For the other kind of esoterica, try #esoteric on irc.dal.net.)
20:02:11 <boily> what the fungot is going on. I leave my desk for like 5 mins top, and what do I see?
20:02:12 <fungot> boily: libcurl might be more than 26 single-letter languages, that's really something a compact camera can offer that most slrs don't remember
20:02:25 <boily> fungot: huh? cameras support unicode now?
20:02:26 <fungot> boily: huh.......interesting....... doesn't everyone have an imaginary part and a little bit take some time to play with some preliminary stuff at http://chicken.humankraft.com/ fnord
20:02:27 <Fiora> `run relcome | sed 's/\(..\)\(.\)/\1/g'
20:02:30 <HackEgo> Wecoe 02o 06heineratonl 04ubfo eotri pogamin lngag dsin 13nddelomet!Fo mreinoraton cec ot 13urwii:htp:/eolng.og/ik/MinPae.(Fr 09heoterkid 13f 04soerca ty 08esteiconir.dl.et)
20:02:44 <Fiora> there. it's like, half a welcome.
20:02:52 <boily> fungot: oh, so the support is imaginary. guess I didn't have to expect much more from hardware manufacturers.
20:02:53 <fungot> boily: i'm not insulting him, etc. what have i been using display?
20:03:07 <boily> fungot: some samsung monitor, 1920×1080.
20:03:09 <FireFly> Fiora: I think you're better of with applying rainwords *after* that sed line
20:03:14 <Fiora> oh. that makes sense.
20:03:24 <fizzie> fungot: To be entirely honest, your box is actually headless.
20:03:24 <fungot> fizzie: i'm getting an " fnord jump instruction", i don't even know
20:03:26 <Fiora> `run welcome | sed 's/\(..\)\(.\)/\1/g' | rainwords
20:03:29 <HackEgo> Wecoe o heineratonl ubfo eotri pogamin lngag dsin nddelomet!Fo mreinoraton cec ot urwii:htp:/eolng.og/ik/MinPae.(Fr heoterkid f soerca ty esteiconir.dl.et)
20:03:37 <FireFly> 08esteiconir.dl.et sounds like a fishy page
20:03:37 <Fiora> yeah that works better <.<;
20:03:49 <fizzie> FireFly: Oh, I thought it was kind of better the way it was.
20:04:13 <Fiora> .et is apparently a TLD
20:04:37 * FireFly wonders what the urwii:htp protocol is
20:05:08 <Fiora> this channel is all about eotri pogamin lngag!!
20:05:16 -!- asie has quit (Quit: My MacBook Pro has gone to sleep. ZZZzzz...).
20:05:18 <fizzie> .og is sadly not a TLD.
20:05:23 <boily> are there any ethiopian domains out there? I never seen any.
20:05:23 <FireFly> I like that the parens are still matched
20:05:34 <boily> fungot: since when do you paren match?
20:05:35 <fungot> boily: fnord doesn't seem to work correctly
20:05:48 <fungot> FireFly: then i can set up something more... usual for later.
20:05:52 <boily> fungot: oh. continue not fnording, then, if there's a conflict between the two.
20:05:52 <Fiora> http://www.mfa.gov.et/ ?
20:05:53 <fungot> boily: that sounds like a huge, bloated program referred to as 2006. e3 can be referred to as ' stem cells' haven't been proven to be. i've not gotten my head around
20:06:19 <Fiora> asdfsdfd they use the <marquee> tag
20:06:42 <olsner> fun fact: marquee is the best html tag
20:06:50 <fizzie> <marquee onmouseover="this.stop()" onmouseout="this.start()"> very classy.
20:07:49 * boily whacks fizzie with the reification of an animated .gif background
20:07:50 -!- Bike has joined.
20:07:54 <fizzie> They really shoud get on with it and go with style="marquee-style: alternate; marquee-speed: fast; marquee-play-count: infinite;" or something.
20:07:59 <FireFly> olsner: I prefer blink. fun fact, "hello".blink() in most JS engines gives you "<blink>hello</blink>"
20:08:32 <Bike> what's et, estonia?
20:08:50 <Bike> ethiopia. "darn, i was close"
20:08:53 <boily> Bike: estonia is .ee.
20:09:01 <fizzie> And there's a lot of .ee domains out there.
20:09:12 <Bike> i wonder if there's an internationalized domain name thingie in ge'ez
20:10:34 <boily> i18ned domain names are fun! http://西貢租盤.香港/
20:11:05 <kmc> Eesti Nõukogude Sotsialistlik Vabariik
20:11:07 -!- KingOfKarlsruhe has joined.
20:11:22 <fizzie> http://esolangs.онлайн/ would be the hottest for the wiki.
20:11:25 <boily> (and I really do have to learn Estonian)
20:11:42 <Bike> do you really?
20:11:43 <fizzie> Or possibly esolangs.сайт. Or both!
20:11:50 <kmc> huh I had forgotten that estonian uses õ
20:12:11 <kmc> fizzie: I think you're not allowed to mix scripts, though :/
20:12:12 -!- Sprocklem has joined.
20:12:19 <Bike> i read that as o with tilde every time and i hate it
20:12:39 <Bike> son of a bitch.
20:12:49 <kmc> so we could be эсолангс.сайт
20:13:06 <kmc> do we have a welcome.ru?
20:13:28 <boily> who speaks Russian in this chännel?
20:13:28 <fizzie> kmc: ЕсоЛангс.онлайн then.
20:14:03 <fizzie> Oh, I didn't see you said that already.
20:14:13 <fizzie> (Had to pick letters from gucharmap laboriously, I don't have a system set up.)
20:14:31 <kmc> I used Google Translate's transliterate-as-i-type feature
20:14:54 <fizzie> I've done one course's worth of Russian at the university.
20:15:06 <fizzie> Not enough to make a proper welcome.ru, though.
20:15:09 <kmc> it really should be эсꙮлангс.сайт
20:15:45 <boily> esquarelangs.site?
20:15:49 <fizzie> What do you type in Google Translate to get ꙮ?
20:16:01 <fizzie> `run unidecode ꙮ # what else could it be?
20:16:03 <HackEgo> [U+A66E CYRILLIC LETTER MULTIOCULAR O]
20:17:19 <kmc> `paste bin/unidecode
20:17:21 <HackEgo> http://codu.org/projects/hackbot/fshg/index.cgi/raw-file/tip/bin/unidecode
20:17:46 <kmc> `unidecode abcdef
20:17:48 <HackEgo> [U+0061 LATIN SMALL LETTER A] [U+0062 LATIN SMALL LETTER B] [U+0063 LATIN SMALL LETTER C] [U+0064 LATIN SMALL LETTER D] [U+0065 LATIN SMALL LETTER E] [U+0066 LATIN SMALL LETTER F]
20:18:02 <kmc> I should learn Python's new .format()
20:18:06 <kmc> now that Rust has adopted similar syntax
20:18:13 <olsner> `unicode эсꙮлангс.сайт
20:18:29 <kmc> `unidecode эсꙮлангс.сайт
20:18:31 <HackEgo> [U+044D CYRILLIC SMALL LETTER E] [U+0441 CYRILLIC SMALL LETTER ES] [U+A66E CYRILLIC LETTER MULTIOCULAR O] [U+043B CYRILLIC SMALL LETTER EL] [U+0430 CYRILLIC SMALL LETTER A] [U+043D CYRILLIC SMALL LETTER EN] [U+0433 CYRILLIC SMALL LETTER GHE] [U+0441 CYRILLIC SMALL LETTER ES] [U+002E FULL STOP] [U+0441 CYRILLIC SMALL LETTER ES] [U+0430 CYRILLIC S
20:19:19 <kmc> Bike: wp even says "Not to be confused with Ő, O with double acute."
20:19:26 <kmc> are you disobeying them??
20:19:51 <olsner> "Not to be confused with the completely normal letter ö"
20:20:24 <kmc> 'Even in the Encyclopædia Britannica one can find "ostrov Khiuma", where "ostrov" means "island" in Russian and "Khiuma" is back-transliteration from Russian instead of "Hiiumaa" (Hiiumaa>Хийума(а)>Khiuma).'
20:20:49 <Bike> fizzie: is it just me or are the rules for when variable settings in a parfor affect the surrounding context... kind of obscure
20:21:39 <fizzie> Bike: I've never really bothered to got the hang of them, which perhaps says something.
20:21:59 <Bike> i'm just wondering if this running average will work.
20:22:05 <Bike> oh well, i'll just try it, that's totally sensible
20:22:26 <Bike> oh, mlint hits them. i uh, think.
20:22:26 <fizzie> Bike: If I want something out from a parfor loop (that doesn't go in a file), I just make outputs = cell(n, 1); and put outputs{ix} = stuff; inside.
20:22:38 <fizzie> That seems to work sensibly.
20:25:19 <fizzie> olsner: Estonian also has an ö that makes, well, an ö (/ø/). The õ is the /ɤ/ instead.
20:25:52 <fizzie> Bike: Did I already mention what I hate about parfor dispatching?
20:25:58 <kmc> Bike: oh is this MATLAB again?
20:26:08 <Bike> fizzie: don't think so
20:26:59 <fizzie> Bike: It decides on a uniform distribution of loop indices to workers a priori, assuming every iteration takes constant time, even though it still keeps in touch with all the workers dynamically (for moving data back and forth) and could therefore just as easily give them loop indices as they become free.
20:27:37 <fizzie> Currently if you do parfor f = 1:nfiles; processafile(files{f}); end and the files are, say, audio files with different durations, some of the workers will finish earlier and then just lay idle for hours.
20:29:11 -!- metasepia has joined.
20:29:11 <fizzie> (Especially if there's some pathological pattern instead of random variation in the durations, since it always allocates every N'th index to a particular worker, unless I misremember.)
20:29:23 <olsner> wikipedia's sound player is doing something weird, sounds like it's starting twice when I click play
20:29:42 <olsner> sometimes more than twice, with a varying delay
20:30:06 <boily> olsner: maybe they applied a Steve Reich remix to all their sound tidbits?
20:30:44 <olsner> maybe, how does that sound?
20:31:13 <boily> olsner: http://youtu.be/FcFyl8amoEE
20:35:46 <olsner> they should've hidden the clappers in the audience
20:40:07 <fizzie> Ohhhh. Well, *that* explains it. I had some OCRemix Mega Man thing -- http://mm25.ocremix.org/ -- playing in another tab.
20:40:15 <fizzie> Made that YouTube link quite different.
20:40:25 <Bike> http://www.mathworks.com/support/solutions/en/data/1-188VX/index.html sigh
20:40:29 <fizzie> I was wondering why it was already playing things when they were obviously just setting up.
20:41:30 <fizzie> "Since UNIX uses asynchronous signals, it is possible to break out of any operation and into the signal handler at any time." Well, now, that's not quite true.
20:41:47 <Bike> indeed, it contradicts itself later.
20:41:48 <fizzie> At least I've been in situations where ^C in MATLAB (in Linux) did nothing in particular.
20:43:34 <mrhmouse> Perhaps MATLAB is just ignoring the signals?
20:43:59 -!- Lymia has joined.
20:45:50 -!- tertu has joined.
20:50:16 <mroman> can't you just ignore the signals
20:50:24 <mroman> and if that does not work
20:50:31 <mroman> you can setjmp out of a sigsegv
20:50:47 <mroman> possibly sigkill as well
20:51:08 <kmc> is that like jumping out of a crashing helicopter seconds before it explodes
20:51:16 <kmc> also you can't trap SIGKILL
20:51:37 <kmc> not really
20:51:48 <Fiora> missing ctrl-c might happen if like, there's a signal handler, but the signal handler can get stuck, right?
20:52:01 <Fiora> like maybe ctrl-c calls a handler that says "start terminating the current operation cleanly"
20:52:10 <Fiora> but if the termination gets stuck then ctrl-c just doesn't do anything
20:52:39 <Bike> "you can setjmp out of a sigsegv" well, thanks for reminding me that it's not just matlab that's terrifying, i guess
20:52:48 <mroman> you can ignore sigterm
20:52:58 <kmc> ^C is SIGINT, I thought
20:52:58 <Fiora> yeah, you can ignore everything except... sigkill and sigstop?
20:53:30 <Fiora> oh... huh, SIGINT and SIGTERM are different things
20:53:30 <mroman> Bike: are you the francophone guy
20:53:37 <mroman> or was that someone else with a B
20:53:44 <Bike> you're thinking of boily, probably
20:54:04 <kmc> Fiora: yeah... there's also SIGQUIT
20:54:15 <kmc> 'The SIGQUIT signal is similar to SIGINT, except that it's controlled by a different key—the QUIT character, usually C-\—and produces a core dump when it terminates the process, just like a program error signal. You can think of this as a program error condition “detected” by the user. '
20:54:34 <Fiora> does that like, ask the process to core dump, or does the shell do that?
20:54:52 <kmc> I don't know
20:54:59 <kmc> might just be the default disposition involves coredumping
20:54:59 <Fiora> that sounds kinda convenient actually
20:55:05 <kmc> except coredumps are usually disabled globally
20:55:06 <olsner> the jvm traps quit and dumps stack traces for all threads
20:55:12 <kmc> but maybe that's a ulimit you can raise
20:55:26 <Fiora> I've remembered having to futz with ulimits a lot to try to get a core dump of a running process that's infinite looping or something >_<
20:55:40 <fizzie> Fiora: Missing ^C can also easily happen when multiple processes are involved.
20:55:57 <kmc> I loved that Linux coredump crontab.d vulnerability....
20:56:04 <Fiora> oh huh. SIGBUS is when the memory address doesn't exist
20:56:14 <Fiora> instead of when it's just protected
20:56:18 <kmc> basically you could make a setuid root process dump core in /etc/cron.d/
20:56:31 <fizzie> I think my cases have been the kind of things where you have a for i = 1:bignumber; dosomething(i); end where dosomething involves a system() call or something, and you can't get ^c in fast enough to get one passed over to MATLAB instead of the executed process.
20:56:39 <kmc> and the cron daemon will gladly ignore megabytes of binary garbage preceeding and following a legitimate crontab entry
20:56:44 <fizzie> (Except with an out-of-band kill.)
20:56:45 <kmc> \rainbow{POSTEL'S LAW}
20:57:04 <kmc> Fiora: ah, interesting
20:57:20 <kmc> Fiora: I almost never get SIGBUS on x86 Linux... it's common on OS X though, for things that would be segfaults on Linux
20:57:27 <fizzie> SIGBUS is whatever SIGBUS wants to be. (Sometimes it's bad alignment.)
20:57:28 <kmc> i think it's a pretty arbitrary distinction
20:57:39 <Fiora> kmc: oh gohs, I looked up that vulnerability
20:57:40 <kmc> similarly SIGFPE is used mainly for /integer/ arithmetic errors
20:57:43 <myname> gentleman, i present you: http://esolangs.org/wiki/Wolfgang
20:57:49 <kmc> Fiora: that was the vuln that motivated the creation of Ksplice
20:57:59 <Fiora> kmc: yeah, that makes sooo little sense to me. SIGFPE is.... integer divide by zero???
20:58:02 <Bike> wait, which gentleman
20:58:12 <kmc> a SIPB server that Jeff Arnold admined got pwned between the announcement of the vuln and the update window
20:58:13 <Fiora> though I think maybe sigfpe is overloaded for um. floating point exceptions? I'm not sure, I've never actually turned them on
20:58:16 <fizzie> I think systems on architectures with actual alignment restrictions generally do SIGBUS on bad alignment.
20:58:18 <boily> mroman: bonjour! j'étais momentanément absent, mais je suis de retour.
20:58:24 <Bike> myname: nice little thing though
20:58:57 <kmc> Fiora: yeah, divide by zero, or INT_MIN / -1
20:59:06 <kmc> (the latter is a corner case that people almost never think about)
20:59:06 <Fiora> oh gosh, right, that latter one is extra evil
21:00:53 <fizzie> You get SIGFPE (on x86) also if you manage to overflow a division, which isn't possible (except for the /-1 case) when the operands are equally wide as the result, but the underlying division has a double-wide numerator.
21:01:03 <Bike> Does the variance of two summed datasets equal the sum of the variances of the sets?
21:01:40 <Bike> looks like yes as long as they're not correlated...?
21:01:42 <kmc> Var(X + Y) = Var(X) + Var(Y) + 2 Cov(X, Y)
21:02:22 <Bike> hm, these probably aren't actually uncorrelated, guess i shouldn't risk it
21:02:46 <fizzie> `run echo 'int d(int x) { asm("div %0" : : "r" (x), "a" (0), "d" (-1)); } int main(void) { return d(1); }' | gcc -x c - -o /tmp/x; /tmp/x; rm /tmp/x
21:02:49 <kmc> https://xiph.org/~xiphmont/demo/daala/demo1.shtml (through demo4.shtml) shows off some really cool video codec work
21:02:49 <HackEgo> bash: line 1: 293 Floating point exception/tmp/x
21:02:55 <kmc> (from my friends at Mozilla Research)
21:03:06 <kmc> I don't know much about video codecs but it's interestingly different from what I do know
21:03:14 <Bike> "cov(x), if X is a vector, returns the variance" thanks
21:03:21 <fizzie> I ran across Daala the other day from that article about Cisco doing that thing w.r.t. H.264.
21:03:32 <kmc> overlapping, variable-size blocks, intra-frame prediction in frequency domain
21:03:47 <Fiora> fizzie: huh, so like. "idiv" can overflow if your denominator is 1 and your numerator is 64-bit?
21:03:51 <Fiora> and that'll sigfpe?
21:04:26 <kmc> the approach for avoiding block artifacts is to apply a reversible filter before encoding which makes the video extra blocky
21:04:39 <kmc> somehow this works and is a more principled way to approach the ad-hoc deblocking that all codecs do
21:05:20 <fizzie> Fiora: "idiv" and "div" both, yes. And of course for all operand sizes. It's just that you can't (generally) get a well-behaving compiler to perform a 64x32-to-32 (or 32x16-to-16, or 16x8-to-8) division, since if one operand is 64 bits, it's going to go with 64x64-to-64. (Implemented using the 128x64-to-64 one, or with software.)
21:05:22 <kmc> also it does chroma-from-luma prediction
21:05:38 -!- MindlessDrone has quit (Quit: MindlessDrone).
21:06:38 <kmc> (again in the frequency domain)
21:06:44 -!- ^v has quit (Quit: Ping timeout: 1337 seconds).
21:08:15 <myname> so, have to check if it's turing complete
21:08:46 <myname> http://esolangs.org/wiki/Wolfgang
21:10:28 <mroman> boily: j'ais commencé d'apprendre le francais de nouveau ;)
21:13:49 <Fiora> fizzie: gosh. that is so freaky. I guess that explains why compmilers don't like to use it
21:13:53 <boily> mroman: c'est bien!
21:14:04 <mroman> si tu es d'accord je demanderai toi de temps en temps si j'ais besoin d'aide
21:15:28 <fizzie> Fiora: It's a common (FSVO) cause of questions from people who are puzzled why their "div x" is misbehaving (because they've not properly zeroed or sign-extended-to ah/dx/edx/rdx).
21:16:11 <Fiora> yeah, I figured that would be an issue but I had no idea it would misbehave quite that badly
21:16:53 <Fiora> are there like, security vulnerabilities because of SIGFPEs?
21:17:24 <mroman> shouldn't it just terminate?
21:17:33 <Fiora> I guess you could like, DoS a server?
21:17:55 <mroman> if you can trick a server into divbyzero
21:18:05 <fizzie> Passing in values that'd cause it to do that INT_MIN / -1 thing sounds feasible.
21:18:13 <kmc> reminds me of http://www.exploringbinary.com/php-hangs-on-numeric-value-2-2250738585072011e-308/
21:18:19 <mroman> sounds feasible to crash a server yes
21:18:24 <kmc> which created a lot of remote DoS vulns
21:18:26 <Fiora> oh gosh. that thing.
21:18:37 <kmc> i can't think immediately of how SIGFPE could have a worse consequence than DoS, but there probably is a way
21:18:40 <Fiora> x87's endless wonderfulness~
21:18:59 <kmc> one of the io.sts levels is a contrived program where the SIGFPE handler gives you a shell and you have to come up with a way to hit it without dividing by zero
21:19:57 <kmc> ais523\unfoog: main() takes two numbers on the command line, and divides them, but only if the second is nonzero
21:20:12 <kmc> and it's a setuid program (that's the point of exploiting it) so you can't just change that behavior
21:20:43 -!- asie has joined.
21:21:03 -!- Sprocklem has quit (Ping timeout: 272 seconds).
21:21:15 <ais523\unfoog> if there's any other solution, it probably involves a parser abuse
21:21:25 <boily> mroman: pas de problème. je suis toujours ouvert!
21:23:23 <kmc> ais523\unfoog: you mean exploiting atoi()?
21:23:49 <ais523\unfoog> kmc: hmm, well IIRC atoi doesn't handle error that well…
21:24:28 <kmc> there were some awesome JS engine security holes arising from NaN boxing and string -> float conversions which would let the attacker specify the unused NaN bit
21:26:16 <Fiora> kmc: :o any good links on that?
21:26:56 <kmc> http://wingolog.org/archives/2011/05/18/value-representation-in-javascript-implementations talks about NaN boxing
21:27:29 <kmc> http://stackoverflow.com/questions/9435338/what-is-the-aim-of-js-canonicalize-nan-in-spidermonkey-engine
21:27:37 <kmc> i don't have a link about the vuln though
21:28:06 <Fiora> wow, fixnums. it's like lisp
21:28:55 <Fiora> "Anyway, the other way of looking at NaN-boxing is that there are 264 - 248 values in a pointer that aren't being used, and we might as well stuff something in them, and hey, the whole space of valid double precision numbers fits!"
21:29:00 <Fiora> wow. that is. phenomenally evil
21:31:40 -!- tertu has quit (Read error: Connection reset by peer).
21:31:52 <mroman> what's that about non-used pointer?
21:32:22 <Fiora> it's the thing where like, most systems only have like 48 out of 64 address bits available?
21:32:32 <kmc> address sizes : 36 bits physical, 48 bits virtual
21:32:59 <mroman> they aren't actually 64bit systems/
21:33:20 <kmc> mroman: it would require a lot more space for page tables to support a full 64-bit space
21:33:29 <kmc> the tree would be deeper
21:33:36 <kmc> that's one concern anyway
21:33:39 <kmc> also just more wires on the chip
21:33:48 <mroman> I don't know how paging works in longmode
21:33:55 <kmc> mroman: I don't know about in general, I know about AMD64 specifically
21:34:09 <kmc> on architectures with soft-filled TLBs the pagetable thing wouldn't be an issue
21:34:14 <Fiora> I think they're "64-bit" in the sense that like, the native register size is 64-bit? just it saves hardware to not support 2^64 possible addresses or something
21:34:32 <kmc> AMD64 has this notion of "canonical address" where the 16 unused address bits have to match the top one
21:34:38 <kmc> when you dereference a pointer
21:34:44 <kmc> this is to prevent people from abusing those bits for tagging tricks
21:34:51 <kmc> so that they won't break forwards-compat
21:35:35 <kmc> http://blog.xen.org/index.php/2012/06/13/the-intel-sysret-privilege-escalation/ an interesting vuln arising from that feature and from subtle differences between Intel's and AMD's implementation of x86_64
21:36:10 <kmc> mroman: of course even if the hardware supports full 64-bit virtual addresses, you might decide to leave some of the space unmapped to use it for tag bits
21:36:19 <kmc> but then (as on AMD64) you would need to mask those out before deref'ing
21:36:45 <Fiora> wait, intel actually /lets/ you dereference those invalid addresses? O_O
21:37:01 <Fiora> or oh, you mean amd64 as in 64-bit
21:37:04 <Fiora> not amd's implementation
21:37:27 <kmc> Intel and AMD chips differ regarding what happens when executing SYSRET if RCX is a non-canonical address
21:38:01 <kmc> specifically, regarding whether the exception is thrown in the old or the new privilege mode
21:38:19 <kmc> "So if the guest can induce the hypervisor to have a non-canonical RIP to return to, it can set the RSP to any value it wants to in hypervisor memory, and get the hardware (in delivering a #GP) to overwrite it. This can in turn be leveraged into a full exploit (the details of which are left as an exercise to the reader)."
21:39:22 <Fiora> see, lambdabot agrees
21:39:31 <kmc> Fiora: yeah, this wasn't just a vuln in Xen either; it hit a lot of OSes at various times
21:40:37 -!- boily has quit (Quit: halloween!).
21:40:39 -!- metasepia has quit (Remote host closed the connection).
21:49:07 -!- Koen_ has joined.
21:52:42 <Bike> wow, parfor almost makes everything more sensible, since suddenly matlab has to be able to do lexical analysis.
21:53:47 -!- augur has quit (Remote host closed the connection).
21:54:23 -!- augur has joined.
21:59:09 -!- augur has quit (Ping timeout: 272 seconds).
22:03:16 <kmc> you just used "MATLAB" and "sensible" in the same sentence
22:03:39 <Bike> i also used "almost" and "more"
22:04:03 <Bike> if it's any consolation it only rejects the hard to analyze code at runtime instead of telling me in lint.
22:04:16 <Bike> `thanks MATLAB
22:05:19 -!- Frooxius has quit (Read error: Connection reset by peer).
22:05:52 -!- Frooxius has joined.
22:15:15 <fizzie> The MATL*a*b* colorspace is perceptually uniform but also does numerical computing.
22:16:25 -!- mrhmouse has quit (Quit: Leaving.).
22:17:00 -!- oerjan has joined.
22:17:56 <kmc> hey fizzie do you like video codecs
22:18:03 <kmc> then you should read the links i sent about Daala
22:19:13 <shachaf> maybe i should read The Illuminatus! Trilogy
22:19:23 <kmc> you definitely should
22:19:27 <kmc> no maybe about it
22:21:01 <fizzie> kmc: [23:04:08] <fizzie> I ran across Daala the other day from that article about Cisco doing that thing w.r.t. H.264.
22:21:08 <fizzie> kmc: Already doned it!
22:21:37 <kmc> oh, I missed you saying that
22:21:37 <fizzie> (Okay, I kind of glanced over some parts of it.)
22:21:38 <shachaf> but should i read it before or after i learn all of maths
22:23:04 <Bike> do you need to know math to read illuminatus, now
22:23:27 <olsner> just a fungot of math will serve, I think
22:23:27 <fungot> olsner: yes; you can just drive her crazy!! crazy!!! fnord fnord teens fnord fnord fnord
22:23:45 <fizzie> I got a copy of (the Finnish translation of) The Illuminatus! Trilogy as a birthday present from my mother, I think. (No idea about translation quality.)
22:24:02 <Bike> i'd be curious to see how they translated the band names.
22:24:22 <fizzie> I could theoretically take a look, it's right here in the bookshelf.
22:24:30 <fizzie> If you know approximately whereabouts I could find them.
22:24:42 <shachaf> my father tried to make me read a copy of _Sinuhe egyptiläinen_ translated into hebrew
22:25:06 <Bike> geez, i don't know. it's like a three page long list and i think it's in the last part of the trilogy.
22:25:08 <shachaf> i think that was the one, at least?
22:25:14 <fizzie> I've read Sinuhe, but on the other hand I'm Finnish.
22:25:23 <fizzie> It's kind of required reading.
22:28:55 <shachaf> kmc: should i register that
22:29:03 <shachaf> i think maybe i could register that but maybe not
22:29:15 <shachaf> do i need a finnish postal address to register a .fi domain
22:29:44 <fizzie> Bike: I found it, and they've not translated the names.
22:30:56 <fizzie> You kind of need a Finnish postal address.
22:31:04 <fizzie> It's not exactly a postal address, but something reasonably similar.
22:31:44 <fizzie> "You will be granted a fi-domain name if you
22:31:44 <fizzie> have a Finnish personal ID number; and
22:31:44 <fizzie> have a permanent address in a Finnish municipality.
22:31:44 <fizzie> Also, fi-domain names are granted to
22:31:46 <fizzie> companies and private entrepreneurs;
22:31:49 <fizzie> Finnish public bodies and state enterprises;
22:31:51 <fizzie> independent public corporations and public associations; and
22:31:54 <fizzie> diplomatic missions of a foreign state
22:31:56 <fizzie> registered with the trade register or the registers of associations or foundations in Finland."
22:32:10 <shachaf> what, not to regular people?
22:32:39 <fizzie> Those are two separate categories.
22:32:52 <shachaf> well, i meant under the regular people category
22:33:14 <fizzie> Under the regular people category, you need to fill those first three conditions.
22:33:24 <fizzie> Which are slightly stricter than just having a postal address.
22:33:25 <shachaf> oh, wait, i didn't even see those four lines
22:34:04 <shachaf> So I don't have a permanent address in a Finnish municipality.
22:34:28 <fizzie> Do you have a Finnish personal ID number?
22:34:47 <fizzie> (I did not know that.)
22:34:49 <ion> shachaf: What is it?
22:35:00 <fizzie> You're not supposed to reveal those!
22:35:19 <shachaf> Are you also not supposed to reveal your birth date?
22:35:22 <ion> I’m sad about that. They’re aking to usernames, not passwords.
22:35:36 <Bike> is this like ssns
22:35:44 <ion> Entities keep using them as passwords.
22:35:48 <ion> bike: s/like // yes
22:35:48 <fizzie> shachaf: I think that's more of a matter of preference. But the latter half is kind of "secret".
22:36:19 <shachaf> fizzie: I thought it was just assigned in order to people born on that day.
22:36:32 <shachaf> Anyway it's only three digits.
22:36:45 <fizzie> It is. It's still "secret".
22:37:05 <shachaf> Anyway, it's not all that fancy.
22:37:18 <shachaf> You and ion also have one.
22:38:11 <fizzie> It's not *terribly* secret, but it's a kind of the default opinion that you're not supposed to publicize it, and entities do keep doing the password thing.
22:38:21 <ion> Not only do entities keep using it as a password, you also have to give it everywhere.
22:38:31 <shachaf> Like when ion asks you on IRC.
22:38:33 <fizzie> E.g. the health provider for the university employees asks for it as a login key.
22:41:16 -!- asie has quit (Quit: My MacBook Pro has gone to sleep. ZZZzzz...).
22:41:45 <fizzie> There's also a second personal ID number ("SATU", "sähköinen asiointitunnus", approximately "electronic services identifier") that people have, but I doubt many people know theirs. (I don't know mine. But it's what the smart ID cards are keyed to, and so on.)
22:42:02 <fizzie> That one's just random numbers, AIUI.
22:43:18 <fizzie> Or possibly it's an incrementing counter instead.
22:43:39 <ion> I don’t know mine either.
22:43:55 <fizzie> Incidentally, I don't think I've seen any of the new this-millennium "SSN"s yet.
22:44:08 <fizzie> Even though the oldest people having those are, what, 13 years already?
22:44:08 <shachaf> I don't know mine either. Do I have one?
22:44:30 <ion> fizzie: Depends on time dilation.
22:44:49 <shachaf> fizzie: Those use 'A', right?
22:44:49 <fizzie> shachaf: It might be allocated on-demand.
22:45:08 <ion> Using A is just evil.
22:45:18 <ion> They shouldn’t use A imo
22:45:24 <fizzie> They had + for the 1800s, - for 1900s, A for 2000s and presumably it's going to continue with B, C ... for 2100s, 2200s and so on.
22:45:30 <shachaf> imo ion's o is the same as my o
22:45:44 <ion> shochof: truly
22:46:03 <shachaf> ion: there's a channel for that
22:46:13 <fizzie> "Ian" is a perfectly respectable name.
22:46:36 <kmc> fizzie: do you have a national smartcard?
22:47:04 <FireFly> I think we use - except sometimes + is used to explicitly indicate "last century"
22:47:07 <fizzie> kmc: Yes, but it's used very very rarely.
22:47:29 <ion> I wish there was a national, official entity that signed your PGP key for proof of identity and perhaps a small price.
22:47:33 <kmc> you should check with some friends and try to GCD your keys http://smartfacts.cr.yp.to/
22:47:43 <fizzie> kmc: Most people don't have one, and you can't really use it for all that much, and most people don't have card readers anyway.
22:47:59 <ion> I don’t have one.
22:48:18 <kmc> I'm told that British bank cards are smartcards, and if you want to log into your bank website they give you a challenge code and you have to take your smartcard to a shop and put in the code on a reader and get a response (???)
22:48:24 <fizzie> I had one for a while, but it's only valid for... five years, I think.
22:48:31 <kmc> or maybe this is only for sending large amounts of money through the website
22:48:47 <fizzie> All our bank cards are smartcards, but they're not generally used for e-banking.
22:49:19 <ion> We tend to use a username, a personal password and a sheet of short OTPs for banking.
22:49:19 <fizzie> AIUI, many (most?) Swedish banks generally do use theirs, with inexpensive reader thingies distributed to customers.
22:49:55 <ion> The OTPs are used to log in and to confirm transactions.
22:50:12 <ion> and to modify some e-banking settings etc.
22:50:12 <FireFly> fizzie: I believe so, at least the ones I know of do that
22:51:53 <shachaf> oh you're in ##categorytheory
22:53:02 <fizzie> There's some statistics (from a Finnish IT-and-stuff magazine, in a 2007 article) that the government had used 20 million euros for the "electronic ID card" project, and that during 2006, for the tax and social services tasks, people used one in a total of 500 transactions (0.1 % of all), and the costs divided across all e-ID authentications amounted to something like 4000 eur/transaction.
22:53:09 <fizzie> That's kind of ungood.
22:53:51 <fizzie> Anyway, "identify using your e-banking account" has gone to become the de-facto standard for "strong authentication" here.
22:54:14 <fizzie> You can change the university account password with one. And do taxes, and apply for social security benefits.
22:54:57 <fizzie> You could technically log onto some university workstations with the ID card too (they had some card readers), but I don't know anyone who has gotten that to work, and anyway I think that's gone completely bitrotten and/or disabled nowadays.
22:55:17 <oerjan> i _used_ to have a card reader that required inserting my bank card but they discontinued those, so now i have one that you only push a button. (what they _really_ want us to do now is to use a smartphone with an app but i don't have one.)
22:55:22 <FireFly> kmc: what kind of authentication do you use for online payments?
22:55:32 <kmc> lolamerica
22:55:41 <fizzie> "You may set or change the password via the VETUMA service, using either your Internetbank or your electronic ID card issued by the Finnish police. You may use the electronic ID card provided you have a card reader." (password.aalto.fi)
22:55:47 <kmc> if I log in from a different computer than usual, it also tries to email or text me a cod
22:55:59 <oerjan> and long before that i got otp cards
22:56:01 <ion> I wish my bank tried to text me a cod.
22:56:06 <kmc> om nom nom
22:56:24 <kmc> I'd worry about the SMS delivery fees on live fish might
22:56:30 <kmc> s/ might//
22:56:32 <shachaf> i heard Google Nose supports that
22:58:06 <ion> The SMS delivery fees on live fish might indeed.
22:59:10 <fizzie> AIUI, the banks take some amount of money from each identification transaction done using the banking accounts, which is kind of a sore point for non-profit organizations that'd also like to support "strong" authentication.
22:59:30 <fizzie> AIAUI, the Estonian version of the smartcard ID thing is used really pervasively over there.
22:59:56 <fizzie> Like, even for signing contracts between private individuals and such.
23:01:49 <fizzie> I guess there's also the "electronic ID in your mobile phone" which maybe perhaps has been slowly gaining some traction?
23:02:23 * oerjan looked up that sinuhe book and found out that the person translating it to swedish was linus torvalds' grandfather
23:02:43 <fizzie> I think I might have known that factoid once.
23:04:38 <kmc> `run printf 'fun fact 0 = 1\n | fact n = n * fact (n - 1)\n' > wisdom/fun\ fact
23:04:43 <kmc> `? fun fact
23:04:45 <HackEgo> fun fact 0 = 1 \ | fact n = n * fact (n - 1)
23:04:48 <oerjan> another similar fun fact is that the person who translated the first Foundation book to norwegian was Arne Treholt, a convicted spy. Well unless there are two people by that name.
23:06:13 <shachaf> is a factoid like a typed fact
23:06:30 <shachaf> http://ncatlab.org/nlab/show/horizontal+categorification
23:07:04 <shachaf> copumpkin: have you been oidified yet
23:08:49 <oerjan> hm checking that fact seems harder than i though; no article on foundation in the norwegian wikipedia.
23:09:09 <fizzie> Fun translation fact: the same person is responsible for the Finnish translations of both The Hitchhiker's Guide to the Galaxy and Bored of the Rings.
23:09:30 <fizzie> oerjan: The Second Foundation has hidden all that knowledge.
23:09:42 <oerjan> fizzie: sounds plausible.
23:09:43 -!- Bike has quit (Ping timeout: 265 seconds).
23:10:13 <oerjan> unless it was the robots.
23:11:33 -!- fungot has quit (Ping timeout: 248 seconds).
23:11:59 -!- Lymia has quit (Ping timeout: 272 seconds).
23:13:30 <oerjan> `run sed -i '1N; s/ *//' wisdom/fun\ fact
23:13:42 <HackEgo> funfact 0 = 1 \ | fact n = n * fact (n - 1)
23:14:17 <oerjan> `run sed -i '1N; s/\n *//' wisdom/fun\ fact
23:14:26 <HackEgo> fun fact 0 = 1| fact n = n * fact (n - 1)
23:14:53 <oerjan> `run sed -i 's/|/ |/' wisdom/fun\ fact
23:15:22 <oerjan> i wish there was a simpler way of joining lines together in a pipe.
23:16:02 <oerjan> actually i suppose that doesn't help with -i editing much.
23:16:35 <oerjan> especially with that missing.
23:16:46 <HackEgo> fun fact 0 = 1 | fact n = n * fact (n - 1)
23:18:01 <fizzie> 'paste' can be useful for that; but it is indeed hard in-place without sponge.
23:18:50 <fizzie> `run (echo foo; echo bar; echo baz) | /usr/bin/paste -s -d '|'
23:19:00 <fizzie> Also HackEgo's bin/paste is kind of a problem for that.
23:19:45 <fizzie> And of course people do use xargs for line-joining.
23:19:58 <shachaf> U+2424 SYMBOL FOR NEWLINE []
23:20:04 <oerjan> `run (echo foo; echo bar; echo baz) | tr '\n' '|' #does this work?
23:20:52 <fizzie> xargs does automatic whitespace normalization, which can be a win.
23:20:53 <fizzie> `run (echo 'foo '; echo ' bar'; echo ' baz ') | xargs
23:21:00 <shachaf> `run (echo foo; echo bar; echo vaz) | tr '\n' ''
23:21:41 <shachaf> `run (echo foo; echo bar; echo vaz) | tr '\n' $'\u2424'
23:22:40 <oerjan> hm right xargs might work
23:24:09 <fizzie> coreutils tr is completely multibyte-braindead, I don't think you really can win there.
23:31:41 -!- Sprocklem has joined.
23:33:53 -!- augur has joined.
23:38:52 <oerjan> <fizzie> Welp, the blob method (via base64 decoding the canvas.toDataURI string, packing it up in a Blob, and doing URL.createObjectURL for the image) works just fine even for large images. <-- fancy
23:40:27 -!- Bike has joined.
23:41:28 <shachaf> well, blobs have to be large by definition
23:41:41 <oerjan> one is an ordinary boy. the other is a human-eating shapeless mass. together they fight crime!
23:43:24 -!- tswett_ has joined.
23:43:38 <fizzie> Hey, where's funfact? I mean, fungot.
23:44:15 -!- dessos has quit (Quit: leaving).
23:44:16 <oerjan> was here when i arrived.
23:44:20 <tswett_> A machine that computes using "fits": like bits, except that every 1 fit must be preceded in memory by a 0 fit. A fyte is a sequence of thirteen fits, where the first fit is 0. Arithmetic is done using Fibonacci coding.
23:44:37 <oerjan> pinged out 33 minutes ago
23:44:43 <Bike> discussion is illegal
23:44:45 -!- fungot has joined.
23:47:42 <oerjan> tswett_: the fact that fyte doesn't take up a constant amount of space might prove inconvenient.
23:48:24 <tswett_> Memory consists of a bunch of fits. A fyte is always thirteen of those.
23:48:42 <tswett_> I suppose it would be fine for a fyte to begin with a 1, as long as the preceding fyte ends with a 0.
23:49:31 <oerjan> otherwise bignums might get tricky
23:51:03 -!- Taneb has joined.
23:51:41 <oerjan> Taneb: ah your sleeping schedule has returned to abnormal?
23:53:28 <oerjan> or possibly your irc client just turned sapient by itself.
23:54:34 -!- nooodl has quit (Read error: Connection reset by peer).
23:55:10 <shachaf> kmc: you know the weird backwards non-regular tree, data T f a = L a | B (T f (f a)) ?
23:56:54 <Taneb> oerjan, yeah, it's back to a sensible abnormal
23:58:40 -!- nooodl has joined.
23:59:02 <shachaf> copumpkin: maybe you should drop the co- for today