00:04:55 -!- nisstyre has quit (Quit: WeeChat 0.4.3). 00:23:41 -!- Phantom_Hoover has quit (Ping timeout: 264 seconds). 00:32:05 -!- edwardk has joined. 00:36:49 -!- Phantom_Hoover has joined. 00:48:52 -!- tromp has quit (Remote host closed the connection). 00:49:26 -!- tromp has joined. 00:53:51 -!- pikhq has quit (Ping timeout: 252 seconds). 00:54:17 -!- tromp has quit (Ping timeout: 264 seconds). 00:55:51 -!- pikhq has joined. 00:58:47 -!- Phantom_Hoover has quit (Ping timeout: 240 seconds). 01:11:23 -!- drdanmaku has quit (Quit: Connection closed for inactivity). 01:11:42 -!- Bicyclidine has quit (Ping timeout: 265 seconds). 01:31:43 -!- Sorella has quit (Quit: It is tiem!). 01:35:11 -!- nisstyre has joined. 01:44:10 -!- ter2 has joined. 01:57:52 -!- edwardk has quit (Quit: Computer has gone to sleep.). 02:06:49 -!- Sprocklem has joined. 02:43:46 Chrome seems to disbelieve that I could ever want a new tab when I ctrl-click a link 02:43:56 It assumes CNN's trying to do an evil pop-up and blocks it 02:58:12 -!- edwardk has joined. 03:05:35 -!- FreeFull_ has joined. 03:07:35 -!- FreeFull has quit (Ping timeout: 252 seconds). 03:11:23 -!- oerjan has joined. 03:35:54 Chrome seems to disbelieve that I could ever want a new tab when I ctrl-click a link <-- can you do it with the right click menu instead? that's what i have to do whenever following a link out of stupid wikia. 03:36:11 Or middle-click? 03:36:56 Chrome seems to have sponaneously changed it's behavior, even ctrl-mouse pad press is working now 03:37:01 I'm sure the old behavior will come back 03:37:25 oh btw i'm not using chrome so this is just a wild suggestion. 03:39:02 wikia, of course, _does_ make an evil popup then. which i could tolerate if not for the fact that it simultaneously forgets i want a new tab. 03:39:21 that popup is p. evil 03:39:58 that popup is wikia 03:49:56 -!- edwardk has quit (Remote host closed the connection). 03:50:25 -!- edwardk has joined. 04:02:48 -!- drdanmaku has joined. 04:05:05 -!- ^v has quit (Read error: Connection reset by peer). 04:05:32 -!- ^v has joined. 04:08:09 -!- password2 has joined. 04:23:41 -!- ter2 has quit (Ping timeout: 252 seconds). 04:26:23 http://www-sop.inria.fr/marelle/Laurent.Thery/T2048.v 04:26:27 2048 in coq 04:27:21 not to be confused with cock in 2048 04:27:27 to which I could also provide a link 04:37:07 -!- curiousguy13 has quit (Ping timeout: 265 seconds). 04:50:07 http://chneukirchen.org/trivium/ looks pretty good 05:02:50 [wiki] [[Special:Log/newusers]] create * John Misciagno * New user account 05:04:04 -!- tswett has quit (Ping timeout: 252 seconds). 05:05:18 -!- adu has joined. 05:13:06 Why are there no "1" tiles in 2048, anyway? 05:16:43 they're considered too odd hth 05:17:55 -!- nisstyre has quit (Quit: WeeChat 0.4.3). 05:20:02 -!- MoALTz has joined. 05:35:34 tdh 05:36:15 yw 05:46:23 -!- adu has quit (Quit: adu). 06:00:04 -!- ^v has quit (Read error: Connection reset by peer). 06:00:31 -!- ^v has joined. 06:15:45 Lemma thm2 : forall seed, exists ms, g2048 seed ms 06:16:07 Oh, the previous one is also unproven. 06:16:46 -!- MoALTz has quit (Quit: bbl). 06:16:49 Jafet: move sequence? 06:23:18 -!- variable has changed nick to trout. 06:24:24 applybot: fun merge where "merge (x#y#xs) = (if x = y then (x+y) # merge xs else x # merge (y#xs))" | "merge [x] = [x]" | "merge [] = []" 06:24:25 constants \ merge :: "'a list => 'a list" \ Found termination order: "length <*mlex*> {}" 06:24:57 applybot: value "merge [2,0,2,2,4] :: nat list" 06:24:58 "[2, 0, 4, 4]" \ ∷ "nat list" 06:25:05 applybot: value "merge [2,2,2,4] :: nat list" 06:25:06 "[4, 2, 4]" \ ∷ "nat list" 06:27:01 applybot: definition "left = map (%row. let row' = merge (filter (op ~= 0) row) in row' @ replicate (4 - length row') 0)" 06:27:02 constants \ left :: "'a list list => 'a list list" 06:27:21 applybot: value "merge [2,0,2,4] :: nat list" 06:27:22 "[2, 0, 2, 4]" \ ∷ "nat list" 06:28:10 applybot: value "left [[2,0,2,4]] :: nat list" 06:28:11 *** Type unification failed: Clash of types "_ list" and "nat" 06:28:15 applybot: value "left [[2,0,2,4]] :: nat list list" 06:28:16 "[[4, 4, 0, 0]]" \ ∷ "nat list list" 06:29:32 applybot: definition "right = map rev o left o map rev"; definition "up = transpose o left o transpose"; definition "down = transpose o right o transpose" 06:29:34 constants \ right :: "'a list list => 'a list list" \ constants \ up :: "'a list list => 'a list list" \ constants \ down :: "'a list list => 'a list list" 06:29:54 what's all this then 06:35:14 -!- ^v has quit (Quit: http://i.imgur.com/DrFFzea.png). 06:39:20 applybot: fun gen where "gen rows ((row,col,n)#ns) = (if rows ! row ! col = 0 then rows[row := (rows ! row)[col := n]] else gen rows ns)" 06:39:21 constants \ gen :: "'a list list => (nat * nat * 'a) list => 'a list list" \ Found termination order: "(%p. list_size (%p. size (fst (snd p))) (snd p)) <*mlex*> {}" 06:39:53 Hmm, that doesn't work 06:40:00 applybot: undo 06:40:01 At: definition "right = map rev o left o map rev";definition "up = transpose o left o transpose";definition "down = transpose o right o transpose" 06:40:09 that looks like some sort of ml 06:40:18 applybot: fun gen where "gen rows ((row,col,n)#ns) = (if rows ! row ! col = 0 then (rows[row := (rows ! row)[col := n]], ns) else gen rows ns)" 06:40:19 constants \ gen :: "'a list list => (nat * nat * 'a) list => 'a list list * (nat * nat * 'a) list" \ Found termination order: "(%p. list_size (%p. size (fst (snd p))) (snd p)) <*mlex*> {}" 06:43:21 isabeml 06:46:56 applybot: definition "win rows0 ns acts = list_ex (list_ex (op = 2048)) (fst (fold (%(rows, ns) mov. gen (act rows) ns) acts (rows0, ns)))" 06:46:57 *** Extra variables on rhs: "act" 06:47:04 applybot: definition "win rows0 ns acts = list_ex (list_ex (op = 2048)) (fst (fold (%(rows, ns) act. gen (act rows) ns) acts (rows0, ns)))" 06:47:04 *** Type unification failed: Clash of types "_ * _" and "_ => _" 06:48:47 -!- S1 has joined. 06:49:37 applybot: definition "win rows0 ns acts = list_ex (list_ex (op = 2048)) (fst (fold (%act (rows, ns). gen (act rows) ns) acts (rows0, ns)))" 06:49:38 constants \ win :: "'a list list => (nat * nat * 'a) list => ('a list list => 'a list list) list => bool" 06:51:29 applybot: value "win ([[1024,0,0,0],[0,2,4,8],[1024,0,0,0],[0,0,0,0]] :: nat list list) [(2,3,4)] [up]" 06:51:29 "True" \ :: "bool" 06:52:49 applybot: value "gen (up ([[1024,0,0,0],[0,2,4,8],[1024,0,0,0],[0,0,0,0]] :: nat list list)) [(2,3,4)]" 06:52:50 "([[2048, 2, 4, 8], [0, 0, 0, 0], [0, 0, 0, 4], [0, 0, 0, 0]], [])" \ :: "nat list list * (nat * nat * nat) list" 06:53:05 ah, how useful. 06:53:23 What's with the 4 06:53:39 The game places it at (2,3). 06:54:16 applybot: declare [[show_sorts]] term gen 06:54:16 *** Unrecognized command 06:54:26 applybot: raw:declare [[show_sorts]]; term gen 06:54:27 "gen" \ :: "'a::zero list list => (nat * nat * 'a::zero) list => 'a::zero list list * (nat * nat * 'a::zero) list" 06:54:56 Yep, I think this lets you play 2048 on an arbitrary group 06:55:01 oh, it's the next number, ok 06:55:07 applybot: raw:linear_undo 06:55:07 (No output.) 07:01:33 -!- MindlessDrone has joined. 07:05:24 applybot: lemma "EX len. ALL ns. (length ns > len & list_all (%n. n = 2 | n = 4) ns) --> EX moves. win (replicate 4 (replicate 4 (0 :: nat))) ns moves" 07:05:24 *** Inner syntax error at "EX moves . win ( replicate 4 ( replicate 4 ( 0 :: nat ) ) ) ns moves" 07:05:37 applybot: lemma "EX len. ALL ns. (length ns > len & list_all (%n. n = 2 | n = 4) ns) --> (EX moves. win (replicate 4 (replicate 4 (0 :: nat))) ns moves)" -- oh come on 07:05:38 *** Type unification failed: No type arity prod :: numeral 07:06:34 applybot: lemma "EX len. ALL ns. (length ns > len & list_all (%(i,j,n). 0 <= i & i < 4 & 0 <= j & j < 4 & (n = 2 | n = 4)) ns) --> (EX moves. win (replicate 4 (replicate 4 (0 :: nat))) ns moves)" 07:06:36 Proving: 1. EX len. ALL ns. len < length ns & list_all (%(i, j, n). 0 <= i & i < 4 & 0 <= j & j < 4 & (n = 2 | n = 4)) ns --> (EX moves. win (replicate 4 (replicate 4 0)) ns moves) 07:08:46 Hmm, there's a bug in the colouring code 07:09:51 applybot: apply clarsimp 07:09:53 Proving: 1. EX len. ALL ns. len < length ns & list_all (%(i, ab). i < 4 & (case ab of (j, n) => j < 4 & (n = 2 | n = 4))) ns --> (EX moves. win (replicate 4 (replicate 4 0)) ns moves) 07:11:30 applybot: apply (subgoal_tac "ALL ns. EX moves. win (replicate 4 (replicate 4 0)) ns moves", blast) 07:11:32 Proving: 1. ALL ns. EX moves. win (replicate 4 (replicate 4 0)) ns moves 07:12:26 Ooh, colors. 07:12:29 applybot: apply (erule allI, rule_tac x = "%_. (replicate 4 (replicate 4 2048))" in exI) 07:12:29 *** Failed to apply proof method 07:12:30 Jafet: What is that language? 07:12:40 https://isabelle.in.tum.de 07:13:01 thx 07:15:09 applybot: apply (rule allI, rule_tac x = "[%_. replicate 4 (replicate 4 2048)]" in exI) 07:15:11 Proving: 1. !!ns. win (replicate 4 (replicate 4 0)) ns [%_. replicate 4 (replicate 4 2048)] 07:16:35 Wait, that won't work. 07:33:00 -!- nortti has changed nick to [Nn]ortti. 07:33:13 -!- [Nn]ortti has changed nick to nortti. 07:36:22 -!- slereah_ has joined. 07:39:31 -!- FreeFull_ has quit. 07:57:23 Chess Clock Jenga http://youtu.be/YN_F9bNuF0I 08:04:34 Disney funds 3D printing research https://youtu.be/qc-tGbMN9Ms 08:05:14 ion: awesome 08:21:52 as for freefall, i am not sure if those are the right people to bring to a debate on whether to exterminate all robots. 08:21:54 jafet: neat 08:22:26 I was hoping I'd get a less boring "spring season" street-outside-the-office-window time-lapse because today is, well, today, but all the colorful people seem to be in other parts of the campus. 08:25:35 What is so today about today? 08:25:44 http://en.wikipedia.org/wiki/Walpurgis_Night#Finland 08:26:38 Jafet: Nothing, duh 08:27:44 -!- Patashu has joined. 08:36:31 -!- augur has quit (Quit: Leaving...). 08:48:16 https://blog.mozilla.org/security/2014/04/24/10000-security-bug-bounty-for-certificate-verification 08:48:49 -!- Patashu_ has joined. 08:48:49 -!- Patashu has quit (Disconnected by services). 08:54:48 nice 08:55:49 that iranian student in the comments needs to work on his english. 08:56:03 Or to get a job 08:56:51 no jobs in iran. sanctions, you know. 08:57:40 perhaps mozilla can't even legally pay him if they wanted to... 08:57:46 Hmm, the same sanctions might prevent the transfer of a bug bounty 08:58:08 That must be why iranian hackers just go ahead and use their exploits 08:58:31 $10,000 sounds like a lot until you think about how much good software developers make 08:58:37 it all becomes clear now. 08:58:57 (in america, anyway) 08:59:02 You can be a good software developer not in california 08:59:09 kmc: so you are saying you earn more in a day? 08:59:32 oerjan: Are you just brazenly assuming kmc is a good software developer? 08:59:58 Come on, let's not get all offended or anything. 09:00:02 fizzie: i have somehow associated "filthy rich" with him. 09:00:17 Besides, this bug bounty seems to have more to do with bad software developers 09:00:47 oerjan: As a Norwegian, you're supposed to be filthy rich too, don't you all have heaps and heaps of oil money? 09:01:01 I assume it works so that you just get money for free. 09:01:30 heh 09:01:40 sadly, i am not filthy rich 09:01:42 fizzie: surprisingly, i cannot get access to my 1 million NOK part of it at all. 09:02:03 (yep, the oil fund is that much) 09:02:21 kmc: shocking. 09:02:31 kmc takes baths 09:02:37 ah. 09:02:53 in something expensive, i assume. 09:02:53 I just assume kmc with mushrooms. 09:02:59 Associate, not assume. 09:03:23 if you have running water in your house, you are pretty rich by global standardz 09:04:02 Oh no, it's the Perspective Man 09:04:22 Does whatever a perspective can? 09:04:23 come to think of it, i saw something about san francisco it workers suing companies for collusion to keep down wages, so perhaps you are actually dirt poor. (don't remember if it was kmc who linked it.) 09:04:49 possibly i saw it on reddit. 09:04:51 You're either filthy rich or dirt poor, only the middle class keeps clean. 09:05:12 -!- mhi^ has joined. 09:07:12 I am making too many slides. :/ 09:07:46 http://www.symmetrymagazine.org/article/march-2014/physics-by-hand "To encourage discussion and engagement, a physics forum has banned PowerPoint slides in favor of low-tech whiteboards." 09:08:45 I don't think there will even be a whiteboard there. 09:08:52 -!- curiousguy13 has joined. 09:08:54 Or at least I wouldn't take one for granted. 09:09:08 what are you presenting? 09:09:30 A boring paper. 09:10:01 For 15 minutes, plus 5 for "questions from the audience and some extra time to introduce the next presenter". 09:11:23 -!- drdanmaku has quit (Quit: Connection closed for inactivity). 09:12:07 I don't think there will be a lot of "discussion and engagement". 09:12:58 just auction off the extra slides hth 09:13:41 for advertising? 09:14:23 just tell them they'll be worth a lot when you become famous. 09:15:42 -!- S1 has quit (Quit: http://www.kiwiirc.com/ - A hand crafted IRC client). 09:17:01 I'm not sure which ones are the extra ones, to be honest. 09:24:45 tricky. 09:37:08 -!- conehead has quit (Quit: Computer has gone to sleep). 09:37:17 -!- oerjan has quit (Quit: leaving). 10:19:48 -!- boily has joined. 10:28:55 -!- Phantom_Hoover has joined. 10:30:20 -!- yorick has joined. 10:44:07 Hmm, the zsh builtin time doesn't work on zsh functions. 10:56:41 -!- curiousguy13 has quit (Read error: Connection reset by peer). 10:58:59 -!- curiousguy13 has joined. 10:59:36 -!- mhi^ has quit (Quit: Lost terminal). 11:00:38 -!- boily has quit (Quit: HUMID CHICKEN). 11:14:41 -!- password2 has quit (Ping timeout: 264 seconds). 11:19:15 -!- nooodl has joined. 11:22:02 -!- curiousguy13 has quit (Ping timeout: 255 seconds). 11:24:23 -!- curiousguy13 has joined. 11:24:25 Help I can't use tmux 11:43:04 Taneb: use screen instead 11:50:47 screen doesn't support four-byte utf-8 sequences 11:51:15 -!- S1 has joined. 11:53:31 `unidecode Щ 11:53:32 [U+0429 CYRILLIC CAPITAL LETTER SHCHA] 11:54:18 -!- nucular has joined. 11:54:18 -!- nucular has quit (Changing host). 11:54:18 -!- nucular has joined. 11:59:01 `run python -c 'pass; pass; pass' 11:59:02 No output. 11:59:11 Why 11:59:41 Who did make HackEgo? 12:04:26 S1, Gregor 12:04:54 when? 12:06:40 A long time ago, in a galaxy 12:06:54 ok 12:07:58 Hmm, the old hg log is gone. 12:08:11 This means we'll... never know? 12:08:25 we could just ask, could'nt we? 12:09:31 You're ruining the moment. 12:09:50 You're right. We'll probably never know... 12:10:30 HackEgo is ancient technology 12:11:23 I made a thing help 12:11:24 https://github.com/Taneb/webqueues/blob/master/Main.hs 12:11:27 And now I have to go 12:11:38 Am I showing off? Do I want code reviews? 12:11:39 `run echo $BASH_VERSION; perl -ewarn\$]; ruby -v 12:11:39 4.2.37(1)-release \ 5.014002 at -e line 1. \ bash: ruby: command not found 12:11:42 YOU WILL NEVER KNOW 12:12:17 Yes. I will never know o.o 12:13:37 `run python --version; python3 --version 12:13:37 Python 2.7.3 \ bash: python3: command not found 12:14:59 Taneb: you seem rather RESTless 12:16:10 Nevermind, I misread 12:17:22 `run (shopt -s nullglob; IFS=:; for x in $PATH; do echo -n "$x"/ruby*; done); 12:17:23 No output. 12:17:30 `run (shopt -s nullglob; IFS=:; for x in $PATH; do echo -n "$x"/python*; done); 12:17:31 /usr/bin/python /usr/bin/python2 /usr/bin/python2.6 /usr/bin/python2.7 12:17:36 `run (shopt -s nullglob; IFS=:; for x in $PATH; do echo -n "$x"/perl*; done); 12:17:37 /hackenv/bin/perl-e/usr/bin/perl /usr/bin/perl5.14.2 /usr/bin/perlbug /usr/bin/perldoc /usr/bin/perlivp /usr/bin/perlthanks 12:18:07 Nevernevermind, it is restless 12:18:59 `file bin/perl-e/usr/bin/perl 12:18:59 bin/perl-e/usr/bin/perl: ERROR: cannot open `bin/perl-e/usr/bin/perl' (Not a directory) 12:19:38 `file bin/perl-e 12:19:39 bin/perl-e: Bourne-Again shell script, ASCII text executable 12:19:43 `cat bin/perl-e 12:19:43 #!/bin/bash \ perl -e "$@" 12:36:34 `perl-e 1while(1x++$_)=~/^(11+)\1+$/||print"$_\n" 12:37:04 1 \ 2 \ 3 \ 5 \ 7 \ 11 \ 13 \ 17 \ 19 \ 23 \ 29 \ 31 \ 37 \ 41 \ 43 \ 47 \ 53 \ 59 \ 61 \ 67 \ 71 \ 73 \ 79 \ 83 \ 89 \ 97 \ 101 \ 103 \ 107 \ 109 \ 113 \ 127 \ 131 \ 137 \ 139 \ 149 \ 151 \ 157 \ 163 \ 167 \ 173 \ 179 \ 181 \ 191 \ 193 \ 197 \ 199 \ 211 \ 223 \ 227 \ 229 \ 233 \ 239 \ 241 \ 251 \ 257 \ 263 \ 269 \ 271 \ 277 \ 281 \ 283 \ 293 \ 307 12:42:22 For a moment I thought 1while was some strange new keyword 12:43:26 Also, why is 1 printed? 12:49:29 `perl-e 1while(1x++$_)=~/^1|(11+)\1+$/||print"$_\n" 12:50:00 No output. 12:50:02 -!- ais523_ has joined. 12:50:15 Jafet: 1 cannot be factored into two factors >= 2. 12:51:20 -!- FreeFull has joined. 12:54:09 `perl $==1,(1x$_)=~/(^)(1|11\1)*(?{$=++})^/,print$=,$/for 0..20 12:54:09 Can't open perl script "$==1,(1x$_)=~/(^)(1|11\1)*(?{$=++})^/,print$=,$/for 0..20": No such file or directory 12:54:14 `perl-e $==1,(1x$_)=~/(^)(1|11\1)*(?{$=++})^/,print$=,$/for 0..20 12:54:15 2 \ 3 \ 5 \ 8 \ 13 \ 21 \ 34 \ 55 \ 89 \ 144 \ 233 \ 377 \ 610 \ 987 \ 1597 \ 2584 \ 4181 \ 6765 \ 10946 \ 17711 \ 28657 12:54:32 `perl-e $==0,(1x$_)=~/^(|()1(?1)(?1)\2)$(?{$=++})^/,print$=,$/for 0..13 12:54:36 1 \ 1 \ 2 \ 5 \ 14 \ 42 \ 132 \ 429 \ 1430 \ 4862 \ 16796 \ 58786 \ 208012 \ 742900 12:54:49 -!- Sgeo has quit (Read error: Connection reset by peer). 12:57:55 b_jonas: What sequence is that? 12:58:42 Are you trying to do fibbonacci? 13:00:09 aaah! 13:01:08 `perl-e $a=1;1while($c=$a+$b,$a=$b,print$b=$c,"\n") 13:01:08 1 \ 1 \ 2 \ 3 \ 5 \ 8 \ 13 \ 21 \ 34 \ 55 \ 89 \ 144 \ 233 \ 377 \ 610 \ 987 \ 1597 \ 2584 \ 4181 \ 6765 \ 10946 \ 17711 \ 28657 \ 46368 \ 75025 \ 121393 \ 196418 \ 317811 \ 514229 \ 832040 \ 1346269 \ 2178309 \ 3524578 \ 5702887 \ 9227465 \ 14930352 \ 24157817 \ 39088169 \ 63245986 \ 102334155 \ 165580141 \ 267914296 \ 433494437 \ 701408733 \ 1134 13:01:09 F_n = number of ways to decompose the string 1^n into 1 or 11. Right. 13:01:32 which is approximately what (1x$_)=~/(^)(1|11\1)*(?{$=++})^/ does. 13:02:21 I see 13:02:55 There probably is some formula for it 13:03:11 Is there a simple way to make C display some very simple graphics on linux? 13:03:21 I'm talking like a few pixels wide 13:03:22 Greyscale 13:03:34 writing to framebuffer? 13:03:54 Depends on what kind of graphics hardware you have 13:04:02 A regular computer kind 13:04:12 Framebuffer then 13:04:18 output xpm? 13:04:40 Let's try 13:04:52 I'm assuming plain linux, without even GNU tools or such 13:05:02 Just the kernel and you 13:05:15 I'm trying to check how well my neural network is working 13:05:19 `perl-e $==1,(1x$_)=~/^11(1|11)*(?{$=++})^/,print$=,$/for 0..10 13:05:19 1 \ 1 \ 2 \ 3 \ 5 \ 8 \ 11 \ 13 \ 15 \ 17 \ 19 13:05:27 And I have a big one in my skullbox 13:05:35 So lookin' will probably easier 13:05:38 int-e: Not sure that looks right 13:05:39 strange :) 13:05:43 Let me calculate by hand 13:05:45 FreeFull: it doesn't. 13:06:02 5 8 13 21... 13:06:32 So that's the point of the \1 13:06:43 `perl-e $==1,(1x$_)=~/(^)11(1|11\1)*(?{$=++})^/,print$=,$/for 0..10 13:06:44 1 \ 1 \ 2 \ 3 \ 5 \ 8 \ 13 \ 21 \ 34 \ 55 \ 89 13:06:50 regex fibonacci? 13:07:24 exploiting the backtracking implementation in perl 13:07:45 oh, not exactly, it's generating a number in unary, and then regexing against it to try to produce the fibonacci sequence in the number of matches 13:08:20 I'm calculating it by hand and I'm getting fibbonacci so far.. 13:09:05 Are you sure this series isn't just the fibbonacci series? 13:09:15 FreeFull: 1 \ 1 \ 2 \ 3 \ 5 \ 8 \ 11 \ 13 \ 15 \ 17 \ 19 wasn't 13:09:49 int-e: When I calculate it by hand I get 1 2 3 5 8 13 13:10:19 FreeFull: ... 13:10:25 yes, that's what you should get 13:12:58 So it is just fibbonacci 13:19:17 -!- nooodl has quit (Quit: Leaving). 13:26:46 FreeFull: fibonacci and catalan numbers 13:27:05 FreeFull: I showed two perl scripts, one computes fibonacci, one computes catalan 13:27:51 here: http://www.perlmonks.com/?node_id=796712 13:29:01 it links to my explanation: http://www.perlmonks.com/?node_id=796712 13:29:45 > scanl(+)0[1,1..] 13:29:47 [0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,2... 13:30:01 > scanl(+)1[0,0..] 13:30:03 [1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1... 13:30:07 > "http://www.perlmonks.com/?node_id=796712" `compare` "http://www.perlmonks.com/?node_id=796712" 13:30:09 EQ 13:30:43 duh 13:30:44 sorry 13:31:06 find where it links yourself then 13:31:41 I forgot how to fibbonacci with scanl 13:32:17 FreeFull: does that involve the case C specifically forbids, where the input overlaps the output? 13:33:15 > let f = 1:scanl1(+)f in f 13:33:17 [1,1,2,4,8,16,32,64,128,256,512,1024,2048,4096,8192,16384,32768,65536,131072... 13:33:38 b_jonas: Does what? 13:34:03 oh, you said scanl, not scanf 13:34:04 sorry 13:35:49 I don't think you can fibbonacci with scanf, although I might be wrong 13:45:06 -!- Patashu_ has quit (Ping timeout: 252 seconds). 13:50:53 -!- Sprocklem has quit (Read error: Operation timed out). 13:51:45 oh wow is Github annoying 13:51:54 I registered for it because I have to for this new job 13:52:03 they sent me an email "ais523 + GitHub = <3" 13:52:11 also they don't require a password confirm to create an account 13:52:24 I had to change the password to itself just to get a password confirm box and thus be relatively sure it was typed correctly 13:53:17 -!- curiousguy13 has left. 13:55:19 > let f = 0:scanl(+)1f in f 13:55:20 [0,1,1,2,3,5,8,13,21,34,55,89,144,233,377,610,987,1597,2584,4181,6765,10946,... 13:57:41 Now I have 30 slides for 15 minutes. :/ 13:58:19 Wee. Gotta talk fast :P 14:01:42 Some of them are just overhead, though. (Title, references, an otherwise empty slide with a giant question mark for the questions, and 7 automatically inserted "outline with position indicator" slides to have some structure in it.) 14:02:01 So 20 with any content, and some of *those* are just pictures. Maybe it isn't so bad. 14:04:34 int-e: That's it 14:10:41 `perl-e $==0,(1x$_)=~/(^)(1|11\1)*$(?{$=++})^/,print$=,$/for 0..10 14:10:42 1 \ 1 \ 2 \ 3 \ 5 \ 8 \ 13 \ 21 \ 34 \ 55 \ 89 14:12:00 is that \1 simply there to defeat the regex optimizer? 14:12:08 yes 14:15:03 `perl-e $==0,(1x$_)=~/^(1|11)*$(?{$=++})^/,print$=,$/for 0..10 14:15:03 1 \ 1 \ 2 \ 2 \ 2 \ 2 \ 2 \ 2 \ 2 \ 2 \ 2 14:16:04 -!- S1 has quit (Quit: http://www.kiwiirc.com/ - A hand crafted IRC client). 14:20:33 -!- edwardk has quit (Quit: Computer has gone to sleep.). 14:23:01 `perl-e $==0,(1x$_)=~/(^)(1|11\1)+$(?{$=++})^/,print$=,$/for 0..10 14:23:02 0 \ 1 \ 2 \ 3 \ 5 \ 8 \ 13 \ 21 \ 34 \ 55 \ 89 14:23:50 which is wrong. 14:24:00 `perl-e $==0,(1x$_)=~/^1(1|11)*$(?{$=++})^/,print$=,$/for 0..10 14:24:01 0 \ 1 \ 1 \ 2 \ 3 \ 3 \ 3 \ 3 \ 3 \ 3 \ 3 14:24:19 `perl-e $==0,(1x$_)=~/(^)1(1|11\1)*$(?{$=++})^/,print$=,$/for 0..10 14:24:19 0 \ 1 \ 1 \ 2 \ 3 \ 5 \ 8 \ 13 \ 21 \ 34 \ 55 14:26:02 ack, (?1) is recursion. who in their right mind adds explicit recursion to a regex engine ... 14:26:25 int-e: because without it, you can't match brackers, or parse HTML 14:26:37 The same people who add arbitrary code execution to a regex enging 14:26:44 if you're trying to write a DSL for matching patterns in strings, recursion is something you want 14:26:52 actually, they added the arbitrary code thing to see how it was used 14:26:57 and using it for recursion was the #1 use by miles 14:27:01 so they added a separate recursion primitive 14:27:24 we used to write my $regex = qr/<(??{$regex})>/; if we wanted to match angle brackets 14:27:36 Good language design 14:28:03 nowadays /<(?0)>/ works fine 14:28:48 I wonder if there is a regex without the perl 14:29:11 shouldn't that be /<(?0)>(?0)/ 14:29:23 it depends on what you want the regex to mean 14:29:24 > let f n | n < 4 = True | otherwise = or [ f a && f (a + b) | a <- [1..n], b <- [1..n], n = 2*a + b, a > b ] in filter f [0..100] 14:29:26 :1:90: parse error on input ‘=’ 14:29:34 > let f n | n < 4 = True | otherwise = or [ f a && f (a + b) | a <- [1..n], b <- [1..n], let n = 2*a + b, a > b ] in filter f [0..100] 14:29:35 [0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,2... 14:29:41 I want to match <<><>> too, for example. 14:29:45 Jafet: PCRE is the Perl sort, but you also get mathematical regexes that use only + and *, and POSIX regexes which suck 14:29:56 > let f n | n < 4 = True | otherwise = or [ f a && f b && f (a + b) | a <- [1..n], b <- [1..n], let n = 2*a + b, a > b ] in filter f [0..100] 14:29:58 [0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,2... 14:30:01 int-e: right; that isn't what I wanted, I just wanted the simplest possble example 14:30:07 but if you want something more complex, you write a more complex regex 14:30:17 Wait, what am I typing 14:30:18 > let f n | n < 4 = True | otherwise = or [ f a && f (a + b) | a <- [1..n], b <- [1..n], n == 2*a + b, a > b ] in filter f [0..100] 14:30:22 mueval-core: Time limit exceeded 14:30:28 > let f n | n < 4 = True | otherwise = or [ f a && f (a + b) | a <- [1..n], b <- [1..n], n == 2*a + b, a > b ] in filter f [0..50] 14:30:30 [0,1,2,3,5,8,13,21,34] 14:30:49 I can't see how a regex can recognize a>b, though 14:31:25 ^(1*)1+>\1$ 14:31:37 what's the representation of a and b 14:31:43 unary? 14:31:44 you can do it with just backreferences if it's unary 14:31:57 I was talking to Jafet, yours is obviously unary 14:32:26 Hmm, it's possible to combine multiple tests using lookahead 14:33:06 anyway, last time I thought along these lines, I ended up inventing Cyclexa 14:33:24 it had things like negative matches, e.g. /^(xyz)^xyzabc$/ matches "abc" 14:33:31 because (xyz)^ matches minus an xyz 14:34:01 What is minus xyz? 14:34:27 basically, normally when you're interpreting a regex, you can think of it as a backtracking program that removes from the string 14:34:32 e.g. "x" means "remove an x" 14:34:41 thus "(x)^" means "add an x to the start of the string" 14:35:06 Ok, so minus adds. 14:36:00 Is it basically pushdown, then? 14:36:02 yeah, except it gets confusing when you do things like (abc(xyz)^def)^ 14:36:11 I can't even remember what semantics I designed for that, if anything 14:36:14 this was ages ago 14:37:10 Presumably a^b would fail to match anything 14:37:18 indeed 14:37:27 the other feature it had was control flow matching 14:37:50 you could do "(a|b)(c|d):1", which is equivalent to "ac|bd" 14:37:52 (a^)^ would... equal a? 14:37:57 Jafet: yes 14:40:20 the fun part about recursion in perl regexen is that some of the semantics is unclear if it's combined with backreferences, 14:40:49 there's now an open bug ticket that is waiting for regex guru to weigh in with their opinion on what the semantics should be, 14:41:02 then the regex engine guru will try to implement that consistently if possible. 14:41:05 they should recruit the Perl half of the anarchy golf userbase 14:41:12 the current implementation is inconsistent. 14:41:29 good idea 14:41:41 you do a backreference in Cyclexa by combining recursion with control flow matching 14:42:02 the real aim for control flow matching, though 14:42:03 I mean, I noticed this inconsistency because of that catalan numbers regex I mentioned above 14:42:21 is to do things like (a$1?)(b$2?):1(c$3?):1 14:42:36 look, now I'm not a PDA any more! 14:42:44 not sure if it's actually TC, or just some sort of super-PDA class 14:43:11 what? 14:43:39 that Cyclexa regex matches any number of 'a', followed by the same number of 'b', followed by the same number of 'c' 14:43:43 you can't do that using recursion 14:43:59 (that can match the 'a' with the 'b', but not with the 'c' as well) 14:45:02 wait 14:45:09 I thought regexes couldn't do that 14:45:13 wouldn't call that a regexp anymore, myself 14:45:40 You might be able to reduce some kind of post correspondence or tags to it 14:46:08 mroman: regexes cannot even do n times 'a' followed by n times 'b', since they are limited to regular grammars 14:46:27 Yeah. 14:46:30 Unless for a fixed n 14:46:36 PCRE, which is an extension from regexes, can do n*'a' + n*'b' 14:46:44 actual mathematical regular expressions cannot even do that 14:46:57 a{3}b{3} matches aaabbb 14:47:01 in some regexes . 14:47:14 yeah but that's just sugar for writing aaabbb literally 14:47:18 but a{n}b{n+1} kinda regexs would be much more fun 14:47:29 it's like writing (a)*3(b)*3 in BF Joust 14:47:32 Like, (a|b)^*((a1|a2|a3...)*)((b1|b2|b3...)*) simulates a post correspondence system with cards [a1,b1], [a2,b2] ... 14:48:00 Jafet: I'm unaware of that formalization 14:48:05 looks like an interesting thing to try to reduce to 14:48:10 also some if stuff 14:48:14 (with :1, defined suitably) 14:48:22 is there a term for grammars one can recognise using only '|' and '('+')' operators? 14:48:34 a{n}b{n/2}(IF n < 10) 14:48:56 (basically, mathematical regexes without klein start) 14:49:41 also the grammars that can be recognised by a bolean circuit or a bounded storage machine with bounded input or a decision tree 14:50:47 Actually, ((0|1)^*end)\1((a1|a2|a3...)*end)((b1|b2|b3...)*end):1 14:52:11 `perl-e $==0,(1x$_)=~/^(1?1?1?$|(1+)(?!\2\2)(?=(?1))1+(?=(?1))\2$)/&&print$_,$/for 0..100 14:52:20 nortti: do you not want the sequence operator 14:52:34 0 \ 1 \ 2 \ 3 \ 5 \ 8 \ 13 \ 21 \ 34 \ 55 \ 89 14:52:44 Jafet: sequence operator? 14:52:45 (parens are not normally considered operators) 14:52:52 (following Jafet's idea) 14:53:11 int-e: I merely inquired whether it was remotely possible 14:53:18 -!- Sprocklem has joined. 14:53:38 ok, found the term, "non-recursive grammars" 14:53:44 Jafet: note the (?!\2\2) which encodes the a>b condition. 14:55:21 wait, no, "star-free language" seems to be exactly what I'm looking for, but it seems like it can recognise something a decision tree cannot O_o 14:55:22 Huh. 14:56:14