00:01:38 22:44 AnMaster_ipv6: but bzr don't do the confusing thing that git does where you switch between branches 00:01:42 how .. confusing? 00:03:37 back 00:03:57 ALSO 00:03:58 http://www.ikea.com/gb/en/catalog/products/S09806763 00:03:59 best desk ever 00:04:01 have I said that? 00:04:08 specifically 00:04:12 white w/ T-leg 00:04:17 bring js 00:04:20 ehird, ? 00:04:28 doesn't look so good 00:04:34 AnMaster_ipv6: white w/ t-leg. 00:04:39 and what do you mean bring js 00:04:42 choose those and the awesome unfolds 00:04:49 and because you need JS to select those 00:04:52 so the images change 00:04:58 yes right 00:05:04 but how is that awesome 00:05:08 I fail to see how 00:05:11 how is it not 00:05:17 it's everything you need in a desk 00:05:21 as minimal as possible 00:05:22 ehird, drawers? 00:05:28 I needs lots of those 00:05:30 who needs drawers on a desk? 00:05:34 me 00:05:37 That's no desk. 00:05:48 ... 00:05:52 Use standalone drawers. 00:05:57 A desk is for putting things on. 00:06:04 use the drawers luke 00:06:10 would have fitted " That's no desk." better 00:06:53 these drawers you speak of 00:07:01 what would they be drawing 00:07:02 ehird: Coincidentally, we have those desks exclusively (80x60+120x60 combinerated for me to a single two-meter-desk with six legs, a single 120x60 for the wife); they are white, although with the A-leg. I don't really remember the justifications for leg-selecting, but I don't have a problem with that either. 00:07:24 See, the almighty fizzie uses the Ultimate Desk. 00:07:39 fizzie: Oh btw 00:07:41 Length: 120 cm 00:07:41 Width: 80 cm 00:07:42 But with the wrong legs! 00:07:45 what the fuck does that mean 00:07:52 what is the difference between length and width 00:08:00 The other is longer. :p 00:08:08 Yes but which applies and what 00:08:28 Well, both. They sell the desk parts in sizes 80x60, 120x60 and 120x80. 00:08:40 For that particular thing in the link, it's probably the 120x80 version. 00:08:50 fizzie: 00:08:52 you need to apply the symmetry theorem from table theory 00:08:53 Assembled size 00:08:53 Length: 120 cm 00:08:55 Width: 80 cm 00:08:57 Min. height: 60 cm 00:08:59 Max. height: 90 cm 00:09:01 Thickness: 2 cm 00:09:03 That looks very much like one measurement set to me. 00:09:06 Of one object. 00:09:27 Well, yes. It's the 120x80 table. So? 00:09:29 you need to apply the symmetry theorem from table theory <--- :D 00:09:37 fizzie: 120x80, what two dimensions are these? 00:09:49 I mean, height, width, thickness are all specified, so wtf is length? 00:09:53 ehird, look at pic, what is reasonable 00:10:01 maybe it between left and right 00:10:08 AnMaster_ipv6: "Width: 80cm" 00:10:09 The two dimensions of the actual table plate are "width" and "length". 00:10:20 fizzie, exactly 00:10:20 fizzie: so it's 120cm deep? 00:10:30 http://www.ikea.com/PIAimages/21139_PE106138_S4.jpg ← then this is the longest fuckin' laptop ever 00:10:34 ehird, no that is the 80 xm... 00:10:39 cm* 00:10:41 AnMaster_ipv6: what? that's "Width" 00:10:47 "Width: 80 cm" 00:10:50 since when does width mean depth 00:11:10 Width and length are the two dimensions of the table plate. It's completely up to you how you situate it. 00:11:18 Oh. 00:11:21 exactly 00:11:23 Well that's the silly. 00:11:30 ehird, are you intentionally dense? 00:11:42 The image actually looks more like it's from the 120x60 variant. 00:11:47 AnMaster_ipv6: 'width' and 'length' seem very the equal to me. 00:11:58 ehird, it was crystal clear to me 00:12:11 fizzie: so they make an 80x60 one? that'll be what I want, then 00:12:18 This dip thing is only 80cm wide 00:12:30 * AnMaster_ipv6 has a desk from the 1930s 00:12:32 very solid 00:12:36 and heavy 00:12:38 but works well 00:12:40 and looks nice 00:12:53 pix 00:13:07 oklopol, maybe tomorrow, if I get around to cleaning up the desk 00:13:15 AnMaster_ipv6: Yes, well, I have special requirements (TM). Because I am short, for instance, I need it low down enough that I can see the monitor but still have the keyboard/mous at a comfortable level 00:13:16 e 00:13:18 ehird: The "More GALANT Desk System Products" -> "Table tops with frames" has the whole set. At least here the "a complete table" combinations actually listed there had a price exactly equal to sum of the parts. 00:13:24 And it can only be 80cm wide due to space contraints. 00:13:28 And I don't like frills. 00:13:29 I wish there was more space for the legs 00:13:30 though 00:13:38 ehird: how short are you 00:13:41 In conclusion: http://www.ikea.com/PIAimages/17840_PE102272_S4.jpg FTW. 00:13:45 oklopol: Um. Short. Very. 00:13:47 especially when using rudder pedals 00:13:51 Like, I don't even know. 00:13:58 Lemme check. 00:14:04 i will let you. 00:14:05 ehird, I'm rather tall. 188.4 cm iirc 00:14:11 or maybe it was .3 00:14:14 at the end 00:14:17 AnMaster_ipv6: so 6 feet 00:14:17 I opted to hook two of http://www.ikea.com/gb/en/catalog/products/50035115 's to the underside of this two-metres-long table, too. 00:14:17 something like that anyway 00:14:22 you're as tall as my dad, AnMaster_ipv6. 00:14:23 :D 00:14:29 ehird, I'm using metric 00:14:31 (he is my one and only standard for tallity.) 00:14:40 AnMaster_ipv6: yes, well, I can't do metric for peoplesizes. 00:14:51 Misread "I can't do metric for pedophiles". 00:14:53 ehird, I can't do feet for peoplesizes 00:15:00 AnMaster_ipv6: That's why I converted it for you 00:15:04 i'm like 180, but i'm kinda crouchy, and look more like 170 00:15:09 fizzie, the dirt is in the eye of the beholder. 00:15:20 ehird, so how tall are you? In metric 00:15:30 AnMaster_ipv6: I was chceking before you interrupted 00:15:31 → 00:15:38 k 00:16:08 ehird: The 120x60 table can be seen in http://zem.fi/~fis/kotikuvei/olohuone.jpg 00:16:14 ehird: alternatively, how tall are you in feet and inches. :-P 00:16:22 * kerlo prepares Google for a metric conversion 00:16:54 129.54 centimeters. 00:17:07 Unfortunately, that number is probably wrong, seeing as how ehird hasn't actually told us how tall he is yet. 00:17:15 (There are other pictures there of this place in the same directory, although the file names are all in Finnish.) 00:17:27 111.76 centimeters. 00:17:36 91.44 centimeters. 00:17:38 fizzie: err, you're a cat? 00:17:53 what the fuck, never cared to mention that? 00:17:57 71.12 centimeters. 00:17:59 oklopol: Yes, well, don't be a specieist. 00:18:30 oklopol: See how I've been locked out on the balcony? That's the sort of treatment I get around here. 00:18:37 being a lizard, i can't say i approve of cats. 00:18:38 Also, fizzie, very clever how you managed to take a picture of yourself from so far away. 00:18:56 kerlo: Telekinetics. 00:19:02 especially being locked on the balcony 00:19:51 ehird, well? 00:20:07 ← I am 146cm = 4 feet 9 inches. And 32kg. 00:20:26 fizzie, you used telekinetics to open the door then? 00:20:35 omg you're smaller than my gf 00:20:41 146 00:20:45 that's like ... wow 00:20:48 lol 00:21:05 I'm around 82 kg btw 00:21:07 That's 144.78 centimeters, or 4 feet and 9.48031496 inches. 00:21:10 AnMaster_ipv6: No, it only works for controlling the camera. 00:21:10 The first words of a few internet-friends I met in August '08 was "you're smaller than I expected". 00:21:16 fizzie, ouch 00:21:23 fizzie, oddly specific too 00:21:24 http://zem.fi/~fis/kotikuvei/keittion_ikkuna.jpg ← finland is so pretty 00:22:00 ehird, you could see that in Sweden too 00:22:01 The name should be fi:"keittiön ikkuna" =~ en:"window in the kitchen", but I didn't dare to use ö in a file name. 00:22:02 just FYI 00:22:10 AnMaster_ipv6: Your language is ugly though. 00:22:12 The snow's pretty much gone now. 00:22:26 ehird, yeah but it doesn't look like a joke instead.. 00:22:47 A traditional example of an ugly-sounding Finnish phrase is: "Älä rääkkää sitä kääkkää." 00:22:48 (no offence meant) 00:22:51 http://zem.fi/~fis/kotikuvei/olohuone.jpg ← that desk is the pretty but the shape of this room dictates 80cm width 00:22:55 fizzie, meaning? 00:22:59 fizzie: no that's hot 00:23:00 älä rääkkää räkäkääkkää 00:23:10 "do not torture that old goat" 00:23:14 sayss google 00:23:21 käkkäränkkä vänkkää väärää määrää 00:23:26 oklopol, translate 00:23:26 Yes, that's rather close. I'm not sure how to translate "kääkkä". 00:23:32 anyway sweden is like the srs 00:23:36 and finland is the jokes 00:23:45 *känkkäränkkä 00:23:49 ehird, oh yes, we are rather timid I guess 00:23:56 oklopol, translate 00:23:57 ... 00:24:19 "älä rääkkää räkäkääkkää" would be "do not torture the old snot goat" 00:24:28 The counterpart (i.e. the beautiful Finnish phrase) is "alavilla mailla hallan vaara". 00:24:31 snot... 00:24:36 oklopol, what does that mean 00:24:48 or did you mean "snow" 00:24:50 or something 00:24:58 AnMaster_ipv6: Snot is what comes out of your nose. 00:25:00 phlegm 00:25:02 AnMaster_ipv6: snorget 00:25:04 fizzie, translate 00:25:20 olsner, jaha 00:25:35 nasal sperm 00:25:41 .. 00:25:48 .. 00:25:53 ... 00:25:55 .... 00:26:00 ... 00:26:02 ... 00:26:04 .. 00:26:06 AnMaster_ipv6: "risk of frost on low-lying areas" is one translation in the interwebs. 00:26:07 WHAT SERIES IS THIS 00:26:14 fizzie, :D 00:26:26 oklopol, tell me when you figured it out 00:26:27 NASAL SPERM. 00:26:52 alivallimailan hillan viiru 00:27:01 fizzie, "risk för frost på låglänta ytor" I think in Swedish 00:27:08 AnMaster_ipv6: figured what out 00:27:09 thought not sure about spelling of låglänta 00:27:10 Kanske. 00:27:17 oklopol, WHAT SERIES IS THIS 00:27:18 that 00:27:28 ohh. 00:27:34 låglänt? is that even a word? 00:27:49 olsner, dialektalt kanske? 00:27:54 I have heard it anyway 00:28:02 but maybe dialect yeah 00:28:16 Wiktionary "lowlands" just has translations "Dutch: laagland n" and a Serbian one. 00:28:41 I have heard låglänt or maybe låglent (but don't think so) 00:28:47 never seen it written 00:28:58 Results 1 - 10 of about 8,070 for låglänt. (0.12 seconds) 00:28:59 well 00:29:04 not unknown anyway 00:29:08 olsner, ^ 00:29:10 oklopol: google translate yields "Do not rääkkää räkäkääkkää" 00:29:18 AnMaster_ipv6: ok :) 00:29:25 olsner: i translated it 00:29:40 oklopol: you claim to have 00:29:42 :) 00:29:47 "[PDF] Bebyggelse från äldre bronsålder i låglänt terräng - ww.arkeologiuv.se/publikationer/rapporter/vast/2008/rv2008_10.pdf 00:29:48 well 00:29:52 rääkkää is a pretty basic finnish word, and räkäkääkkää should be simple to decompose, google translate sucks. 00:29:53 stuff like that 00:29:56 sounds plausible 00:30:05 olsner, I'd say it exists 00:30:09 www* 00:30:10 btw 00:30:19 and add missing ending quote 00:30:43 need to go to sleep soon methinks 00:30:49 same meh 00:31:18 "Sanakirja.org - Käännökset haulle låglänt (ruotsi-suomi). Käännöspeli. Lähdekieli: Ruotsi." <-- what 00:31:19 olsner: Google Translate does know the infinitive form of "rääkkää", which is "rääkätä". 00:31:26 found when googling for låglänt 00:31:32 maybe fizzie can translate that 00:31:59 it looks like some dictionary 00:32:07 Uh... "Sanakirja.org - Translations for the search låglänt (ruotsi-suomi). Translation game. Source language: Swedish." 00:32:13 "Käännös. Adjektiivit. 1. alava. Sana kuuluu seuraaviin luokkiin: ..." too 00:32:16 from same 00:32:20 fizzie, heh 00:32:23 Er, "ruotsi-suomi" being "Swedish-Finnish", forgot that part. 00:32:28 ah 00:32:42 "Translation. Adjectives. 1. low-lying. The word belongs to the following classes: ..." 00:32:52 * ehird unarchives 4gb 00:32:53 Well, I'm not sure if low-lying is the correct term. 00:32:59 fizzie, "sanakirja" means anything? 00:33:09 AnMaster_ipv6: "Dictionary". 00:33:12 ah 00:33:16 AnMaster_ipv6: I just didn't want to mess the DNS name. 00:33:28 fizzie, well there was "låglänt - Wikisanakirja" too 00:33:29 heh 00:44:18 Hm. 00:44:44 Perl golf challenge: Take two filenames, XOR the file contents together char-by-char. 00:45:00 <> only uses the first arg I think :-( 00:45:16 <> will read both files, but I don't think you can tell where the files change. 00:45:34 ah 00:45:37 There probably was some magic current-file variable, though. 00:47:19 -!- psygnisfive_ has quit (Read error: 110 (Connection timed out)). 00:48:30 Ah, you can use non-argument "eof" to determine when the file has closed. "perldoc -f eof" has an example. Not sure how golf-friendly that is. 00:48:49 I guess it would be reasonably short to change some sort of mode with that. 00:49:00 at this point why not just slurp argvs 00:49:26 anti-golf: which is the most verbose programming language (in general, it can of course differ between tasks) 00:49:35 I'd say SQL is high up at least 00:49:40 intercal or cobol 00:49:46 ehird, yes probably 00:50:08 ehird, intercal would be better for certain bitwise operations though :D 00:56:16 -!- Judofyr has quit (Remote closed the connection). 01:02:11 glio glio bolf glo flomog 01:02:21 -!- oklopol has changed nick to gliopol. 01:02:24 o 01:02:24 o 01:02:24 o 01:02:24 o 01:02:30 a butt 01:02:36 WHAT WHAT 01:02:47 two butts 01:03:04 WHOOOOOSE BUTTS 01:03:14 My butts 01:03:22 I have two butts. 01:03:29 well how come you have both 01:04:06 Nighty -> 01:04:34 -!- FireFly has quit ("Later"). 01:13:29 -!- pikhq has joined. 01:18:04 http://technically.us/spde/Fold ← this is neat. 01:36:17 Walmart: land of the absurdly odd bundles of Magic cards. 01:36:33 They sell packaged bundles which consist of a pack and some random cards... 01:36:47 I got cards from freaking Unlimited. 01:47:27 Octothorpal. 01:55:38 Magic: that game where one of the main ways to improve is by paying a certain company. 01:56:01 It's very unique in that respect. Completely unlike any other trading card game or MMORPG. 02:23:17 -_-' 02:50:43 -!- neldoreth has quit ("Lost terminal"). 03:38:54 -!- GreaseMonkey has joined. 04:34:31 -!- psygnisfive has joined. 05:35:59 -!- pikhq has quit (Read error: 60 (Operation timed out)). 05:50:24 -!- cherez has joined. 05:50:30 -!- cherez has left (?). 05:51:12 -!- pikhq has joined. 06:00:02 -!- GregorR has quit (Remote closed the connection). 06:00:47 -!- GregorR has joined. 07:26:29 -!- Sgeo has quit ("Leaving"). 07:51:11 -!- MigoMipo has joined. 07:59:59 -!- clog has quit (ended). 08:00:00 -!- clog has joined. 08:00:21 -!- bsmntbombdood has quit (Read error: 110 (Connection timed out)). 08:16:42 -!- kar8nga has joined. 08:46:49 -!- psygnisf_ has joined. 08:47:55 -!- psygnisfive has quit (Read error: 104 (Connection reset by peer)). 08:48:58 -!- kar8nga has quit (Remote closed the connection). 08:49:57 -!- tombom has joined. 08:53:55 -!- psygnisfive has joined. 08:55:28 -!- psygnisf_ has quit (Success). 08:58:09 -!- psygnisfive has quit (Read error: 104 (Connection reset by peer)). 09:03:51 -!- psygnisf_ has joined. 09:21:56 -!- BeholdMyGlory has joined. 09:35:37 -!- FireFly has joined. 09:54:30 -!- MigoMipo has quit (Read error: 104 (Connection reset by peer)). 10:16:09 -!- asiekierka has joined. 10:16:11 hi 10:16:19 I'm trying to make a sorta-TV-channel again 10:16:20 :/ 10:17:43 I even made a custom testcard 10:22:59 Can you write a working 'cat' in DOBELA? I thought about it a bit and failed 10:26:09 I could try 10:26:12 i mean 10:26:13 I can try 10:26:22 Why? Did you finish it? 10:26:30 (the interpreter) 10:27:06 It's at the point that it can run 'Hello!' (after I fixed the one on the wiki), now I'd like a more sophisticated test program :-P 10:28:22 I think we have to clarify the outputting a bit more 10:28:37 If there is a COMPLETE BYTE (8 bits) to output, it outputs it, and outputs all complete bytes 10:28:51 If there are any bits left (less than 8), it ignores them 10:28:57 Meh, that'd make it too easy :-P 10:29:07 Otherwise, I don't know 10:29:18 I think the more basic issue is that I don't see a way of doing 'do X every 8 iterations' 10:29:27 neither do I 10:29:42 Except if we do a loop that takes 8 iterations 10:29:46 which toggles the emission setting 10:29:49 but that's too hard :P 10:29:57 One thing that I thought of that might make it possible (and probably a lot of other things easier) is if we had a duplication instruction 10:30:04 asiekierka: I tried that and failed 10:30:11 Since the dot dies on hitting the generator 10:30:17 And also when hitting the output 10:30:23 With duplication it could be done 10:30:33 (I think, at least.) 10:30:44 Since you could then fork a dot to hit both the ^ and the : 10:30:44 As in, when hit east or west, the same dot is sent north and south? 10:30:55 Yeah 10:31:05 Yes, and while we're at it, when hit north/south it's sent east/west 10:31:09 yep 10:31:13 It would be a + sign 10:31:15 if it's not taken 10:31:20 It isn't 10:31:20 and no it's not 10:31:42 :) 10:31:47 Also, I wonder what to do with asievision 10:31:51 ustream.tv/channels/asietv 10:32:05 Maybe DOBELA's now a bit closer to 99% TC :-P 10:32:21 Well, now it's 99.25% TC I think 10:32:22 :) 10:32:44 So can you write cat for me now? ;-) 10:32:50 ...maybe 10:33:21 I have a bunch of work to do today so I probably won't have time to look at it... no worries if it takes you a week :-P 10:37:39 Well, we neet to hit two ^'s :) 10:37:41 need* 10:37:53 Why? ^ can output AND change the data of all generators 10:38:04 You need more than one generator? 10:38:10 i'm not sure 10:38:13 I'm still thinking 10:39:05 It should switch it in 2 cycles, but that's still not fast enough 10:39:13 wait 10:39:14 wait 10:39:15 no 10:39:17 it IS enough 10:39:26 cuz a generator outputs every other cycle 10:39:36 Wait, no, it'll be 3 cycles AFAIK 10:39:44 Remember that : starts on the second cycle and _ on the first 10:39:48 Yep 10:40:01 So it requires 16 cycles for _ to output 10:40:02 I was thinking that it might be easiest to start with 10:40:03 __ 10:40:03 $$ 10:40:15 So you get a bit of input every cycle 10:40:21 so 8 cycles 10:40:54 And to make it 2 cycles, :+ doesn't need a space 10:40:55 And then you need to hit a ^ at any multiple of 8 10:41:08 So cycle 2: A dot gets output 10:41:14 3: Duplicated, hitting a wall 10:41:15 4: DUH! 10:41:30 I think if it hits a generator or something directly in the same cycle as being output, both count 10:41:38 So Cycle 2: A dot gets output AND duplicated 10:41:48 cycle 3: Hits a wall and toggles the generator off 10:41:51 Cycle 4: WIN! 10:42:13 So what's the code? 10:42:18 Not ready... yet 10:42:22 In the meantine implement the + 10:42:34 Like said, I don't have time today :-P 10:42:40 -!- GreaseMonkey has quit ("X-Chat -> http://xchat.org <- At least when I quit I don't look like a lamer"). 10:42:45 Oh 10:42:55 You don't need to do it now if you don't feel like it 10:44:00 Well, now it's 99.25% TC I think <-- does such a measure even make sense? 10:44:11 AnMaster: Does DOBELA even make sense? 10:44:19 meh true 10:44:21 :P 10:44:37 so this interpreter, where can it be found Deewiant 10:44:47 It can not be found 10:44:51 oh ok 10:45:00 It doesn't implement much of anything yet 10:45:05 Deewiant, 10:45:07 ;P 10:45:13 If you really want it, I can put the binary up 10:45:16 meh 10:45:23 Well, I think I did it 10:45:24 It can only really run 'Hello world' though 10:45:25 but remove the numbers 10:45:26 not really that interested 10:45:59 maybe when it reached 99.62% TC I'll want a look 10:46:09 lol 10:46:13 http://rafb.net/p/zrKO1Q68.html - this is my idea 10:46:29 I actually tried to think of a BF interpreter in DOBELA but then soon realized that I can't even implement cat :-P 10:46:37 I think this one should work 10:46:43 And it does not need to be a multiple of 8 10:46:46 Why? 10:46:55 You know, if there are 9 bits stored, 8 are output and 1 is left 10:47:01 19 bits - 16 are output and 3 are left 10:47:07 So duh, why do you need a multiple of 8? 10:47:08 asiekierka: I think it's better that partial bits are output 10:47:20 Ok 10:47:21 asiekierka: And as for 'why' — because that's what the specs currently say :-P 10:47:57 asiekierka: I think it's better that partial bits are output <--- ‽ 10:48:04 TURKEY BOMB‽ 10:48:14 as in a third of a bit iirc 10:48:24 s/partial/leftover/ 10:48:27 http://rafb.net/p/Zo1qa570.html ok, I did it I think 10:48:29 ah 10:48:30 Or s/bits/bytes/ whatever 10:48:34 Now it should output every 8 cycles 10:48:42 Well, 10:48:45 I'm not sure but nah 10:48:54 Should work IMO 10:49:00 afaik you can only output complete bytes on most systems 10:49:10 Well, it fills the byte with 0's if it's partial 10:49:15 :) 10:49:26 or read complete bytes 10:49:28 asiekierka: By the way, a problem with collision: What happens after ..# 10:49:44 What way are the dots going 10:49:48 Both east 10:50:09 Theoretically they should create a wall 10:50:15 Cuz of the top-left scanning 10:50:21 Not according to the current rules 10:50:24 wait 10:50:25 wait 10:50:31 Well, ok 10:50:34 asiekierka: Since '.. ' should be ' ..' next cycle 10:50:44 As far as I remember, it does a collision if it can't be resolved 10:50:52 But what happens there is that first the dot on the left moves so you have '.#' where there are two dots on the . 10:51:01 So first cycle: 10:51:02 Then the next dot turns but doesn't move 10:51:40 So a wall gets created when the next dot finds out that it's standing on top of another dot? 10:51:52 According to my rules, the first dot should stay in place, but the second hits the wall 10:52:19 :P 10:52:22 With top-left scanning the first dot should move first 10:52:28 How would it know to stay in place 10:52:36 By detecting the collision 10:52:53 That'd make moving a single dot O(n) where n is the size of the program 10:52:58 oh 10:53:00 Since it potentially has to scan n dots 10:53:02 ..........................# 10:53:13 Yay, O(n^2) to move those dots 10:53:27 And a pain in the ass to implement :-P 10:53:34 So then, I think i'll go lazy and just make the dots create a wall 10:53:34 :P 10:53:42 Or there's another way 10:53:44 maybe simpler 10:53:51 And I guess this would be the same regardless of whether it's ,.# ,,# .,# ..# 10:53:52 The first dot moves and the second bounces from the wall 10:53:58 Deewiant, only O(n^2) ? 10:53:59 meh 10:54:08 it isn't worth it if it isn't O(n!!) 10:54:16 asiekierka: Bounces? 10:54:19 as in 10:54:24 moves according to the wall rules 10:54:24 duh 10:54:37 So you'd change how walls work? 10:54:41 no 10:54:43 Or only in that kind of case? 10:54:57 seems this lang is mostly special cases 10:54:57 Well, i don't change anything 10:55:07 Because the second dot moves according to the wall rules 10:55:11 And the first dot moves 10:55:14 No 10:55:21 Because the wall rules are currently that the dot turns but doesn't move 10:55:30 ...Huh? 10:55:42 Oh 10:55:59 Wait 10:56:10 I change the rules to "turn and then move one space" 10:56:22 Okay, then what about 10:56:24 . 10:56:25 .. 10:56:29 Where the dot to the north is moving south 10:56:34 asiekierka, btw why haven't you implemented this lang yourself? 10:56:34 And the others are going east 10:56:39 I'm going to eat lunch -> 10:56:49 Well, first the north dot moves south 10:56:52 then it moves south again 10:57:02 then the middle dot moves east 10:57:11 then everything's normal 10:57:19 Tada! I waste 3 cycles on a collision! 11:01:26 -!- asiekierka has set topic: topic appoppic http://tunes.org/~nef/logs/esoteric/?C=N;O=D aproppotic *popaboom*. 11:03:43 DOBELA looks interesting, though i'm not yet clear how complex program can be achieved 11:06:33 -!- kar8nga has joined. 11:06:59 asiekierka: is there working cat program? 11:10:26 asiekierka: So then collision detection is O(n) again :-/ 11:13:52 Deewiant: Nah, then just make every pair of these create a wall 11:14:01 or whatever opposite-direction moving is handled 11:14:09 So then 11:14:10 . 11:14:11 .. 11:14:13 Becomes 11:14:16 11:14:18 #. 11:14:20 Yep 11:14:26 That's easier :-) 11:14:40 You know, DOBELA is all special exceptions OR extremely stupid ideas 11:14:53 asiekierka: But do you still think walls mean turn+move instead of just turn? 11:15:27 # 11:15:28 #.# 11:15:29 # 11:15:37 That's an infinite loop in the interpreter then :-P 11:16:28 -!- k has joined. 11:16:43 -!- kar8nga has quit (Nick collision from services.). 11:16:46 -!- k has changed nick to kar8nga. 11:17:16 Deewiant: Then not, though I'd need to rework my engine then 11:17:20 I think 11:17:31 asiekierka: I think it's best to make it mean just turn 11:17:34 yep 11:17:36 So, uh 11:17:49 I will need to rework my cat then 11:18:00 Yes, I think so :-P 11:18:12 Just make the second-duplicator loop double size 11:18:15 er, double length 11:18:29 I think the one that goes down from the first + fails 11:18:35 Yep 11:18:38 And it can't be fixed, sadly 11:18:47 Yep: you'll need another generator 11:18:53 Or something 11:18:58 That wouldn't help either 11:19:02 Well, it would 11:19:03 actually 11:19:11 :) 11:19:18 :^ 11:19:26 Below the upper : 11:19:42 Then the upper : needs just one + and some timing. 11:20:37 I think I fixed it 11:22:33 Wait, I need to think whether hitting ^ to restart generators on cycle 16 does make them work on cycle 16 11:22:55 If not, I will need to remove the E near the rightmost ^ 11:23:03 (1-E are hex digits for counting the cycles 11:23:12 http://rafb.net/p/Obwqw853.html 11:23:29 asiekierka: If they are more southeast, then yes 11:23:33 Right? 11:24:39 asiekierka: And btw, the dot that goes west from that + won't work like that: you have it turning first counterclockwise and then clockwise twice, but it should always turn counterclockwise 11:30:08 DOBELA Crescent 11:30:30 -!- ais523 has joined. 11:31:01 Deewiant, asiekierka: seriously, DOBELA does seem to have too many obscure rules. 11:31:09 AnMaster_ipv6: it's an esolang 11:31:11 AnMaster_ipv6: It's his spec, not mine. 11:31:12 so I think the interpreter should be named that... 11:31:21 AnMaster_ipv6: also, interesting nick 11:31:23 (for the benefit of ais): DOBELA Crescent 11:36:51 -!- KingOfKarlsruhe has joined. 11:38:15 AnMaster_ipv6: I don't get the reference, FWIW. 11:38:38 btw, so far the only gcc option that actually seems to break cfunge is -ffast-math (which make DATE fail in mycology), not are make it faster, but no other ones in the optimising section of the manual seem to actually produce a broken program. I'm currently checking which exact option that -ffast-math enables it is that breaks cfunge... 11:38:55 Deewiant, Mornington Crescent... 11:38:59 obscure rules 11:39:19 Found it on Wikipedia, never heard of it before. 11:39:25 what 11:39:32 it has been discussed several times in here 11:39:36 how could you miss it 11:39:44 well, Deewiant isn't British, that's just about a small excuse for not having heard of Mornington Crescent 11:40:24 ais523, I never heard of it before it was mentioned in this channel, but it has been mentioned so often in here... 11:41:10 mornington crescent's one of the best-known parts of ISIHAC 11:41:11 AnMaster_ipv6: I find no mention in my logs since February. 11:41:21 but there are others, such as the bit where they sing one song to the tune of another 11:41:24 That doesn't count as 'often' then. :-P 11:41:38 Deewiant, check 2008 11:41:46 AnMaster_ipv6: Don't have them handy. 11:41:47 ais523, ISIHAC? 11:41:59 "I'm Sorry, I Haven't A Clue" 11:42:04 the name of a British radio program 11:42:05 nice name 11:42:15 fis@eris:~/irclogs/freenode/#esoteric$ grep -i mornington 200[3-8]* | wc -l 11:42:15 9 11:42:22 I'm not sure either that this qualifies as "often". 11:42:32 fizzie: Try crescent 11:43:48 Those 9 are at: 2006-06-15 18:20:08; 2008-06-15 00:40:26; 2008-06-18 00:00:07..00:02:23; 2008-07-03 18:54:28; 2008-12-15 19:21:32..19:21:46. 11:44:09 So it's been mentioned at about five occasions. 11:44:57 The set of "crescent" matches is equal, except with one addition: [2008-06-18 00:00:48] < tusho> #nomicton-crescent if anyone wants to try it 11:46:09 ok odd 11:46:29 -ffast-math breaks cfunge under gcc 4.1 but not 4.3 it seems 11:46:35 or maybe something else is going on 11:47:32 hm 11:47:54 wow 11:47:58 dobela discussion 11:48:01 awesome 11:48:11 -!- gliopol has changed nick to oklopol. 11:48:28 oklopol: interesting nick you just changed from 11:48:31 is there a story behind it 11:48:49 it's always either social blabber, your incomprehensible unix talk, or befunge 11:49:05 err 11:49:07 yes 11:49:11 i like glio 11:49:15 it's a nice word. 11:49:23 and 11:49:30 glio glio bolf glo flomog 11:49:30 * oklopol is now known as gliopol 11:49:33 For some values of "story". 11:49:36 it doesn't suffi...xify well 11:50:09 * oklopol takes a deeper look at dobela 11:50:22 uh uh 11:51:08 Well, duh, I will fix it later 11:51:17 I want to make a better esolang though 11:51:34 oklopol: When you gaze deep into dobela, the dobela gazes into you. 11:51:49 oklopol: You must face the dobela alone. 11:52:27 Well, DOBELA was my first esolang 11:52:30 And is 11:52:31 :P 11:52:38 I must make a better esolang one day 11:52:53 Or just use Befunge 11:52:59 don't fix it too much, from the discussion above it seems it is currently the interesting kind of esolang, since no one knows it's whether it's tc 11:53:01 well 11:53:04 aha 11:53:20 I mean 11:53:24 I will fix my CATs 11:53:26 -mfpmath=sse,387 -ffast-math breaks on gcc 4.1.2 11:53:32 i mean for sure, usually you can tell by a glance 11:53:56 fizzie: i plan to, unless someone wants to share my monitor with me 11:55:04 and "do" is pronounced like the "do" in "don't" :D 11:55:15 'the "do" in "don't"' 11:55:20 * oklopol finds funny. 11:55:20 oklopol, what about slashes? 11:55:28 that is a really interesting one 11:55:41 * asiekierka notices hidden funny 11:55:42 "DOBELA -- putting the "do" back in "don't"." 11:55:46 (probably not intended,and probably not funny in any way) 11:55:50 * asiekierka laughs. Seriously. 11:56:16 AnMaster_ipv6: was slashes discussed last night? 11:56:21 -!- pikhq has quit (Remote closed the connection). 11:56:24 oklopol, no 11:57:00 -!- pikhq has joined. 11:58:07 interesting 11:58:16 GCC has an anti-inline option 11:58:21 -frtl-abstract-sequences 11:58:33 I wonder what would my new esolang be like 11:58:38 it tries to find common code and abstract it out in a shared function 11:58:55 AnMaster_ipv6: well, i agree that's one of the gems; but i'm more interested in the fact something happened here :D 11:59:11 oklopol, oh ok 11:59:17 oklopol, befunge is nice though 11:59:54 I need a language that will make some videos for me 11:59:57 that's what I need 12:00:05 but I'm afraid an interpreter will take a while to do 12:00:07 i'm not saying it's not 12:00:54 ".= turns into a =,", does that mean it moves :DD 12:01:37 hm 12:01:52 I should try to make a complete list of all gcc command line options 12:02:12 oooh interesting 12:02:23 there is a *doucmented* switch gcc --help=undocumented 12:02:24 wow 12:02:37 meh 12:02:47 some of those are documented in man page, how boring 12:02:48 -!- ais523 has quit (Read error: 104 (Connection reset by peer)). 12:03:40 dobela is insane :D 12:04:38 oklopol: I know 12:04:56 -!- ais523 has joined. 12:09:28 GCC 4.3.2 --help -v supports some 770+ unique options 12:09:39 s/supports/list/ 12:10:02 -!- Mony has joined. 12:10:19 a bit hard to count since some are listed like: -Wnormalized= Warn about non-normalised Unicode strings 12:10:24 while some are listed expanded 12:10:30 -!- Mony has changed nick to Guest12281. 12:10:52 --param l1-cache-size The size of L1 cache --param l1-cache-line-size The size of L1 cache line 12:10:54 and so on 12:10:58 (on different lines) 12:11:55 wait 12:12:05 didn't Deewiant say cat was impossible to implement 12:12:06 ? 12:12:23 i must have misunderstood something because there's a cat on the page 12:12:26 oklopol, I think they were trying to add something to support it 12:13:22 hmmmmm right actually that's not a cat 12:13:28 just collects stuff on the queue 12:13:34 but 12:13:47 there's that rules that the dots can stop each other from moving 12:13:52 for a cycle 12:13:54 -!- Guest12281 has changed nick to M0ny. 12:14:01 so maybe you could have two generators 12:14:34 and as long as stuff came out the stdin, they'd always block the dots for one cycle, and the two generators' dots would collide 12:14:36 but 12:14:43 when eof was reached 12:14:54 they wouldn't collide, and a generator would be set on 12:15:10 and dots would start coming out, emptying the queue into stdout 12:15:32 well dunno, i should start reading my readings now 12:18:28 oklopol, hm this requires that there isn't any delay in the input data right? 12:18:34 like a user typing slowly 12:19:07 i assume that delays the whole program 12:19:13 oklopol, oh ok 12:19:22 otherwise it's trivially impossible to know when there's no input left 12:19:24 umm 12:19:25 actually 12:19:39 couldn't you just output the contents of the queue all the time 12:19:44 hmm. 12:20:01 well, maybe i'm fundamentally misunderstanding something 12:20:03 dunno 12:20:08 no idea either 12:20:30 oklopol: You can't output every bit, you have to output every byte 12:20:51 I think it is wrong for esolangs to use blocking IO 12:20:54 almost all do 12:21:07 or maybe even all 12:21:13 how many esolangs have async IO? 12:21:19 anyone know a single such? 12:21:25 Deewiant: ohh 12:21:38 but, couldn't you still use the idea of waiting for eof? 12:22:55 You can try, I couldn't figure out a way of getting it to work 12:23:09 -!- pikhq has quit (Remote closed the connection). 12:23:12 just send a literal control-D when you reach the end of the file 12:23:17 hmm 12:23:19 oh 12:23:23 actually my idea doesn't work 12:23:35 there are both 1- and 2-dots 12:23:44 so clearly you cannot rely on that delaying behavior 12:23:54 because if one delays, the other is destroyed 12:24:12 and because there's no duplication, it's impossible to get any use out of 12:24:31 err 12:24:36 0- and 1-dots maybe 12:24:39 but still 12:24:47 ais523, know any esolang with async IO? 12:25:02 not offhand 12:25:10 possibly also uncomfortably parallel 12:25:12 is there a Funge fingerprint for it 12:25:21 ais523, not that I know of 12:25:29 CLC-INTERCAL can simulate a select(), but you need to run multiple programs at once and use network connections 12:25:40 ais523, I guess you could do it in Feather though? 12:25:57 Feather doesn't really define its I/O environment at all 12:26:01 oh ok 12:26:06 ais523, it is so hard to know... 12:26:09 :P 12:26:15 it just uses abstract input/output streams, the implementation decides what they're connected to, if anything 12:27:03 ais523, with implementation defined extensions for sockets and such possibly? 12:28:05 that would just be another input/output stream 12:29:21 -!- pikhq has joined. 12:33:18 I'm thinking about an esolang represented as lines 12:33:23 on a 2D space 12:33:27 wait 12:33:35 well, the lines move to the right 12:33:49 but they can go up and down and right, never back left 12:33:54 like audio, sorta 12:34:44 So basically, 2 variables: Angle and Length :P 12:34:54 or Target Y and Length 12:35:06 -!- kar8nga has quit (Remote closed the connection). 12:35:36 It would be cool if it had a mechanical interpreter 12:35:42 well, electromechanical 12:35:53 or electronic, duhh 12:36:13 asiekierka: there are angle-based graphical languages like Wierd, but i don't know any angle-and-length-based one 12:36:47 well, it is angle and length of the line going at that angle 12:36:51 and it's an array of these 12:37:22 Well, this language can be interpreted only with a ruler and a protractor 12:37:30 or a combined ruler/protractor 12:37:49 -!- pikhq has quit (Remote closed the connection). 12:38:13 and written with a pen and a ruler and a protractor 12:38:35 There's no I/O though 12:38:41 except if you interpret the results yourself 12:39:24 -5 to 5 degrees of angle: NOP for x milimetres 12:39:27 where x is the length 12:40:01 5 to 30: Increment variable (3+(x*5)), where x is the length 12:40:17 -5 to -30 is the same, but decrement 12:40:31 And there can only be 5 variables 12:40:34 0-4 12:40:50 so the length can be 3, 8, 13, 18, or 23 12:41:16 But I can't believe I'm making an esolang for which you only need a ruler and a protractor 12:41:59 er, 5 to 30 should be: Increment variable x, and the calculation for drawing the line is (3+x*5) 12:43:16 -!- pikhq has joined. 12:45:17 i made my first program that makes no sense 12:45:34 First it nop's for 10 cycles, then it increments variables 0 and 1 and decrements 2 12:45:34 :P 12:45:47 asiekierka: Add instructions to change "speed", the space being transformed from underlying space to reading space using Lorentz tranformations. :-) 12:46:09 Ilari: I'm not a maths geek so I don't know what you mean 12:46:20 But i still have 300 degrees left to use 12:48:08 I will show you how the program looks like 12:48:23 -!- pikhq has quit (Remote closed the connection). 12:48:43 Lemme make a photo though 12:48:48 as i'm too lazy to pull out my scanner 12:49:45 asiekierka: When one's speed reaches the speed of light, what he/she sees is distorted according to his/her direction. 12:50:41 I don't quite get it so nah 12:50:57 :p 12:51:34 http://en.wikipedia.org/wiki/File:Lorentz_transform_of_world_line.gif it is very excellent image explaining such behavior 12:51:49 I'm looking at it 12:52:13 It doesn't help much, but I get the idea 12:52:33 There's a much better example actually 12:52:52 There's a software that lets you pilot a little spaceship at relativistic speeds 12:52:59 It's quite nifty 12:53:03 link plz? 12:53:13 Lemme see 12:53:27 for example, http://www.adamauton.com/warp/ ? 12:53:45 there are plenty of programs there, search for "relativity simulation" for example 12:56:40 http://www.starstrider.com/ 12:57:38 I see the apparent distortion 12:57:52 hm there are lots of whois clients 12:57:56 but what about whois servers 12:58:14 but i still don't quite get it, but nah, lemme upload how a typical handmade Anglent program looks like 12:58:23 (Anglent is a temporary name, maybe it'll stay) 12:59:38 http://img26.imageshack.us/img26/6489/dsc01722q.jpg - here you go 13:00:05 the last NOP line may NOT be working but should 13:00:17 The interp should handle errors in draw 13:00:22 because the interp should be human 13:01:22 So, how do you like it 13:02:54 Hello? 13:03:00 hi 13:03:00 It's how the esolang looks like 13:03:24 sorry, I was having technical problems trying to load it 13:03:28 still am, in fact 13:03:33 I'll let you know when I can actually see the image 13:03:36 Okay 13:03:42 I can upload it to my server though 13:03:44 ah, ok 13:03:45 and not Images Hack Us 13:03:48 I can see it now 13:03:53 and? 13:03:56 and interesting 13:04:00 really? O_O 13:04:02 is it just a weird way to represent a program 13:04:07 or are you going to add restrictions 13:04:21 such as preventing the line crossing itself, or forcing it to come back to its starting point, or something? 13:04:26 It can't go left. 13:04:27 you could have multithreading by branching and rejoining the lines 13:04:34 maybe Befunge-style loops, too 13:04:39 It can't go left, as I said. 13:04:51 although those would involve going left 13:05:03 The "Audioform/Basic" version can't branch, is just a long, variously curved line 13:05:34 And I think BF could work this way 13:05:42 how do you do loops? 13:06:08 You can't do them in Audioform/Basic. Why? Cuz Audioform/Basic is the version capable of possibly being stored in a WAV file 13:06:31 Audioform/Pro allows to jump back by x milimetres, where x is the length of the line 13:06:46 But i'm going to work on Audioform/Pro first 13:07:00 So 30-45 degrees is: Jump back by x milimetres 13:07:03 so 13:07:14 anyone knows any whois server software? 13:07:28 45-55 is: Jump back by x milimetres if variable 0 is larger than 0 13:07:35 Oh, and variables can have values from 0 to 9 13:07:52 Why? Cuz I want to make an interpreter that doesn't need any electronics 13:07:59 and can be hand-operated 13:08:25 Oh, and -30 to -40 is: Jump back by x milimetres if variable 1 is larger than 0 13:09:23 AnMaster_ipv6: I don't offhand, I assume someone does but they might not be in this channel 13:09:28 true 13:09:41 ais523, but even wikipedia's article on whois doesn't mention any 13:09:48 so I'm kind of lost here 13:09:53 And I can use -90 to 90 13:10:00 And i'm using -40 to 55 13:10:35 -40 to -50 is: Swap variable x with variable 0. The length is (3+x*5) 13:10:43 -50 to -60 is the same, but with variable 1. 13:10:57 So -60 to 55, which leaves me 30 on the minus side and 35 on the plus side 13:11:07 meh 13:11:08 Any ideas? 13:11:19 AnMaster: Write your own 13:11:24 what? 13:11:33 I'm not talking about what you said 13:11:34 whois server 13:11:37 ah 13:11:38 right 13:11:51 :) 13:11:56 asiekierka, well I found some I think, seems the RIPE one is open or shared source (not sure yet) 13:12:10 Ok, so, ais523: Any ideas what to do with the leftover command space? 13:12:32 I think I may do an input command, as in "Set whatever variable you want in variable 0" 13:12:50 55 to 65 degrees: Input to var. 0 13:13:28 ftp://ftp.ripe.net/ripe/dbase/software/whoisserver-nightly.tgz <-- hm *loading* 13:13:29 And well, you can't go left so you must do meaningless commands to be on the right track 13:13:36 as in 13:13:38 for example 13:13:49 a bunch of -4 degree "nop 3" or something 13:14:25 Or -25 "dec 0" and +10 "inc 0", that gives you -15 in 6 milimetres 13:14:55 Or -28 and +7, which gives you -21 degrees 13:14:58 in 6 milimetres 13:16:13 Remember, going over -90 or 90 degrees is an error in a computer interpreter 13:16:18 But a human interpreter needn't care 13:16:20 :) 13:16:38 Yes, I think SixXS uses the RIPE server sources too for whois.sixxs.org. 13:16:44 wow RIPE uses mysql it seems... 13:17:17 and there is an embedded imap server software it seems 13:17:52 with code around for AMIGA and DOS (in the directory for the imap part) 13:18:05 tops-20 too 13:18:05 wth 13:19:38 what a mess 13:19:42 oh well 13:20:35 Well, I'm working on a rough template of a Anglent interpreter board 13:21:22 which you need to combine with a drawing of an Anglent program 13:21:30 and then you can interpret 13:21:47 An Anglent board currently contains the state of v0 to v4 13:21:47 err 13:22:33 asiekierka: why not v6? 13:22:40 btw anyone made a non-discrete automaton yet? DOBELA is after all discrete, using ticks, and cells 13:22:49 AnMaster_ipv6: Gravity, but it's uncomputable 13:22:52 something can't go to an arbitrary position 13:23:06 v0 -> v4 are Variable 0 to Variable 4 13:23:07 -!- Judofyr has joined. 13:23:08 that's why 13:23:16 ais523, yes I know that one is uncomputable. But is it possible to make a computable automaton like that? 13:23:17 And you CAN have more than 5 variables 13:23:19 asiekierka: clearly I should make my jokes more obvious 13:23:22 but that makes too long lines 13:23:28 ais523: ipv6, i get it 13:23:29 ais523, because I don't know *why* Gravity is uncomputable 13:23:30 :P 13:23:56 Oh, and for multithreading programs you need more Anglent boards 13:23:57 AnMaster_ipv6: because you need infinite precision in solving differential equations to solve it 13:24:47 ais523, ah right. What I meant would be something like that Photon idea I had. where program was controlled by photons hitting mirrors that moved to cause sensitive switches to change and such 13:24:48 in 3D 13:24:55 I never got around to details 13:25:12 And 65 to 80 degrees is: Exchange between process x and process 0, where the line length is (2+(x*4)) mm 13:25:17 ais523, it would be ray-tracing I guess. 13:25:20 or something 13:25:32 So i have 80-90 left and -90 to -60 13:25:42 I mean, tracing the photons emitted to see what they hit 13:25:50 80 to 90 degrees is: exit program 13:25:51 possibly classical physics only 13:25:56 or maybe not 13:26:03 -!- Judofyr has quit (Remote closed the connection). 13:26:32 ais523, it might be uncomputable too 13:26:37 probably in fact 13:26:55 Forking is represented by drawing a straight line down from the process you want to fork from. 13:27:16 asiekierka, you are going to implement it yourself this time? 13:27:42 AnMaster: If by "implement" you mean "the human interpreter board", I already did that 13:27:53 Because it is intended to be interpreted by a human 13:28:08 cop-out 13:28:53 So, yep, but an interpreter is centrainly possible, but human is better for such things 13:29:46 -!- jix has joined. 13:30:32 Of course, it would be a pain to interpret if it splits in 25 processes or something 13:30:48 ANGLENT: (process amount) users required! 13:31:18 Oh, and if a process ends, the board can be reused 13:31:30 Wait, no it can't, sorta 13:31:39 Why? Because it may still copy data from that process 13:31:56 So we need -60 to -70: Remove process x, where the line length is (5+(x*5)) mm 13:31:56 asiekierka, example program please. Because I have no idea what you are trying to describe 13:32:07 it's confusing 13:32:07 I already gave a link, but it's simple 13:32:13 AnMaster: I know 13:32:51 I'm making a program showing more features 13:33:07 asiekierka, is it TC? 13:33:15 AnMaster: I don't know 13:33:18 how do you do a loop? 13:33:25 -!- Asztal_ has joined. 13:33:29 There is a "jump back if variable 0 > 0" 13:33:37 and you can swap different variables with variables 0 and 1 13:33:50 infinite memory? 13:34:15 Well, you CAN extend memory by using more processes 13:34:27 infinite number of processes then? 13:34:30 Yep 13:34:35 or are variables bignum? 13:34:44 Variables are digits from 0 to... guess. 13:34:47 256? 512? 65535? 13:34:51 meh 13:34:51 No. 0 to 9. 13:34:54 why that limit 13:35:03 Because it's meant to be interpreted by human 13:35:04 :) 13:35:32 And I don't want to make the boards too large 13:35:35 well and? Humans calculated pi to several hundred digits long before computers were invented 13:35:45 Well, this is the Basic ANGLENT board 13:35:48 You can make your own 13:35:48 digits of precision* 13:35:48 :D 13:36:00 That's the good thing, it only defines the commands 13:36:01 it doesn 13:36:12 it doesn't define anything else, the number of processes, number of digits 13:37:47 and that's the point of Anglent 13:37:55 the other point is that it can't go left 13:38:06 as in 13:38:09 the line can't go backwards 13:38:31 -!- olsner has quit ("Leaving"). 13:38:51 asiekierka, you said you could jump though? 13:39:13 jump back 13:39:17 but that's not drawing a line backwards 13:39:21 true 13:39:23 it's moving the IP backwards, sorta 13:39:52 By "the line can't go backwards" I meant "the line can't be drawn backwards, it must go right, even if by a minimal amount" 13:40:00 asiekierka, you need to write a reference implementation that interprets svg images 13:40:06 The exception is forking, cuz you draw a straight line down for that 13:40:09 write it yourself that is 13:40:31 AnMaster: Why can't I just show you an image of a program and how it works when interpreted 13:40:42 asiekierka, less fun 13:40:53 asiekierka, so you can make a fork bomb? 13:40:54 AnMaster: Or make a video where I interpret a program myself? 13:41:02 by jumping back over a fork 13:41:07 and forking multiple times 13:41:20 what about making each program recursively fork new ones 13:41:30 so you get exponential growth 13:41:38 Well, you need to remove a process for it to not count 13:41:42 So theoretically you can 13:41:46 asiekierka, ? 13:41:50 By making a fork and jumping back so you can make the fork again 13:41:58 without removing the previous process's data 13:42:02 but how do you do exponential growing forks 13:42:16 Do you need that? 13:42:36 Well, you can't clone the program you're doing 13:42:36 asiekierka, can the forked process jump back to the previous program to make a fork too 13:42:43 AnMaster: No 13:42:50 The forked process is it's own set of commands 13:42:58 AnMaster_ipv6: Incidentally, I just solved that "matrix too short" issue I had. 13:42:58 but it may exchange data between any process 13:43:00 so you can't make an exponential growing fork bomb 13:43:04 fizzie, oh? 13:43:05 Probably no 13:43:13 But you still can make a fork bomb 13:43:25 asiekierka, how do you make an if 13:43:54 Well, I need to add a "jump forward" command for that 13:43:58 I hope you can skip code depending on if a variable is set or not 13:44:16 asiekierka, do all processes run at the same rate 13:44:31 and how is that rate defined 13:44:36 -70 to -80: Jump forward x milimetres if v0 > 0, length is (4+(x*5)) 13:44:44 So now you can skip code 13:44:47 k 13:44:51 And to make a fork bomb 13:44:55 you use the "jump back" command 13:45:08 AnMaster_ipv6: My original sound source was a .mp3 file recorded from one of those interweb radios; the script used sox to turn that to a raw audio file for the Octave analysis. I was correctly looking at the raw audio file for the amount of samples, but accidentally reading the from the unconverted MP3. So one of the blocks was shorter because it hit the end of the MP3 file, and the blocks after that were full-length since fseek-past-end-of-file was actually 13:45:08 silently failing in Octave, and they started reading from the beginning of the file. 13:45:12 yes sure, but I can't see how you could make each process be a sub-fork-bomb 13:45:20 is gravity tc? 13:45:28 Well, you can 13:45:31 Sorta 13:45:32 fizzie, ah! 13:45:33 I will show how 13:46:01 (Of course reading the .mp3 file as a raw audio file also meant the results were, shall we say, suboptimal. Whelp, that's about twenty hours of CPU time wasted.) 13:46:10 AnMaster: Well, I think you can make each fork process be an infinite loop 13:46:39 asiekierka, hm? 13:46:51 Well 13:46:55 With the "jump back" command 13:46:59 You can make infinite loops 13:47:09 And you can make an infinite fork-creating loop in the main process 13:47:17 asiekierka, yep but that is not what I asked about 13:47:28 fizzie: how much was wasted in human money? 13:47:31 You can't make every process a sub-fork-bomb 13:47:38 Except if a program could fork-clone itself 13:47:50 oklopol: Depends on the amount of $$$s you give for my sanity, I guess. 13:47:58 asiekierka, can't it draw a straight line up to a previous program? 13:48:20 AnMaster: Well, I did think of that but it may get too messy for a human 13:48:27 hm 13:48:28 I'm thinking of drawing a straight line up to CLONE a process 13:48:34 fizzie: maybe 200€ 13:48:42 nah 100 13:48:43 bbl food 13:48:44 as in, like forking, but it does whatever the process forked from is doing 13:48:47 only on it's own space 13:48:58 oklopol: Well, some vaguely definable part of that was lost. Maybe around one euro, then. 13:49:18 -!- FireyFly has joined. 13:49:19 -!- FireFly has quit (Read error: 104 (Connection reset by peer)). 13:49:40 fizzie: well i assume you value your sanity more than me 13:49:51 it's a bit subjective 13:50:14 -!- FireyFly has changed nick to FireFly. 13:51:19 Yes. Well, a more objective number would be the amount of electrons wasted, but I don't care enough to start approximating that. 13:54:06 Uploading the fork bomb... 13:58:08 well, it nops 5, inc's v0, clones itself, nops 10 and jumps back just to the moment of cloning 13:58:26 http://img441.imageshack.us/img441/9668/dsc01723m.jpg 13:58:39 back 13:58:55 I did the forkbomb 13:59:24 Well, it will work on a COMPUTER interpreter 13:59:34 cuz HUMAN interpreters will probably realize it's a forkbomb 13:59:51 asiekierka, and execute PLEASE GIVE UP instead? 14:00:09 No, they'll just stop execution 14:00:24 ... 14:00:26 duh 14:00:42 asiekierka, that was what I said basically... 14:00:59 Well, humans can detect a fork bomb 14:01:39 asiekierka, you completely failed to detect the INTERCAL reference right? 14:01:47 I don't even know INTERCAL well! 14:01:54 so, yes 14:01:55 i failed 14:01:56 I don't claim to know it well 14:02:21 oh 14:02:26 "PLEASE GIVE UP" is the reference 14:02:45 yep 14:02:49 AnMaster_ipv6: "execute PLEASE" is redundant 14:02:58 you should just say "and PLEASE GIVE UP instead" 14:03:13 INTERCAL's syntax was designed to be embedded into natural-lanuage English sentences, and you even fail to use that feature? 14:03:14 mhm 14:04:07 ais523, yeah I said I wasn't good at INTERCAL 14:04:18 well get better then 14:05:13 yes AnMaster_ipv6 how dare you not be good at it 14:05:20 oklopol, meh 14:05:28 there is an infinite amount of crap in my keyboard 14:05:35 when i turn it upside down 14:05:40 oklopol, don't clean it! 14:05:42 wait 14:05:47 if it is infinite 14:05:52 it would 1) fill your room 14:06:03 2) figure out a way to use it as memory then make an UTM 14:06:03 an infinite stream of all kindsa stuff snows down 14:06:28 oklopol, don't waste the possibility to make an UTM! 14:08:30 ais523, interesting, gcc has an option to try to make use of delayed branch slots... 14:08:32 -fdelayed-branch 14:08:32 If supported for the target machine, attempt to reorder instructions to exploit instruction slots available after delayed branch instructions. 14:08:59 I wish x86 had that 14:09:04 Is that turned on by default? One would think it would be. 14:09:15 fizzie, yes it is 14:09:20 with -O and higher 14:15:18 fizzie, there are lots of interesting flags to gcc, did you see what I said above there seems to be roughly over 770 of them 14:15:21 in gcc 4.3.2 14:15:45 (a bit hard to measure, the --help -v output is not that simple to parse) 14:15:45 Did not notice that, no. 14:16:03 I removed assembler and linker flags info 14:16:34 also: GCC has an anti-inline option 14:16:34 -frtl-abstract-sequences 14:16:34 it tries to find common code and abstract it out in a shared function 14:16:53 there is a *doucmented* switch gcc --help=undocumented 14:16:59 some of those are documented in man page, how boring 14:17:16 fizzie, around that is info about the 770 flags in logs 14:17:22 should be easy to grep for 14:20:24 Yes, well, I did catch the gist of it: there are some 770+ options. 14:22:31 AnMaster_ipv6: the documentation for --help=undocumented is obviously lying 14:22:40 yes 14:22:41 now you can decide whether it counts as a documented switch or not 14:23:09 heh 14:23:37 ais523, It seems to mean there is no --help text stored for the switch 14:23:53 maybe the listing is auto generated 14:23:54 So, well, I gotta collect all Anglent commands 14:23:55 well, yes, that's a good definition of undocumented 14:24:04 and it is auto-generated, gcc has command-line option definition files 14:24:15 its whole command-line option infrastructure is /really/ over the top 14:24:28 but I suppose with 770 options, it needs a pretty heavyweight infrastructure 14:24:31 ais523, Why doesn't it use GNU getopt_long? After all it is a GNU project so why... 14:24:53 AnMaster_ipv6: probably it does internally, but are you going to pass 770 options to it by hand? 14:25:00 also, the options depend on things like compile-time flags 14:25:00 ais523, no way! 14:25:07 hm true 14:25:24 so it has a heavy infrastructure for trying to make sure the options end up in the binary 14:25:28 and also in the documentation 14:26:38 ais523, yeah openmp is a compile time option for gcc itself. So the existence of -openmp (or whatever) at compile time(1) depends options passed at compile time(2) 14:26:42 (1) of program 14:26:45 (2) of gcc 14:26:55 :D 14:27:18 afraid I didn't manage to mess up that more 14:27:20 in gcc-bf, all the compile-time options are actually link-time, passed with -Wl 14:27:41 ais523, even that -m open I suggested to select EOF value? 14:27:47 EOF behaviour* 14:28:22 AnMaster_ipv6: I haven't implemented that yet, but it's a library issue not a main-program issue 14:28:52 heh 14:29:18 ais523, does it work fully nowdays or? 14:29:35 I haven't worked on it for a while 14:29:38 I mean can you get it to generate a runnable program 14:29:44 ok 14:29:45 I can run some very simple programs 14:29:53 ais523, so it generates bf now? 14:29:57 the current state is that some features, like multiplication, are missing 14:30:00 and most of the others are buggy 14:30:10 but it does generate bf, sometimes the bf even works 14:30:27 ais523, the patch to gcc I could host, the full gcc would be a bit too large. 14:30:43 well, let me create something to host first 14:30:50 and not right now 14:31:00 I haven't slept for 5 or 6 nights in a row 14:31:04 ais523, ouch! 14:31:07 ais523, why not 14:31:08 although I have been sleeping over the day to make up 14:31:08 :( 14:31:34 and because I've been getting my University project report finished, it's worth a massive number of marks and the deadline's tomorrow 14:31:54 ais523, ah 14:32:00 ais523, finished yet then? 14:32:21 pretty much 14:32:25 good :) 14:32:26 I came in today to look up a reference 14:35:25 oh 14:36:32 wow, some spam I can actually read 14:36:33 Heh, we have this pair-wise university project that is now approximately two years late of the deadline, because the deadline was a "soft" one. It's "almost done", hoping to finish it tomorrow. 14:36:34 "We apologize for contacting you at this time of the day on 08:16 AM , and we hope that we haven't interrupted you in anyway, but we wanted to make sure that you received the message that we sent you last week. We have checks ready to send you for offering us your honest opinion on various online surveys that only take a few minutes to complete." 14:36:48 it's even spelt correctly and almost has correct grammar 14:36:54 something's gone wrong with the world 14:37:48 (I set my email client to never show HTML email; if there's no plain text it just shows nothing. This has the nice side effect of meaning most spam doesn't show up.) 14:38:56 ais523, indeed 14:39:13 ais523, does anyone actually enable HTML email support? 14:39:27 AnMaster_ipv6: you'd be surprised 14:39:31 huh 14:39:35 pretty much all internet users have it on by default 14:39:37 and never changed it 14:39:49 they should change the default in clients then 14:40:08 AnMaster_ipv6: but then you couldn't send people emails with smileys and pictures in 14:40:19 * ais523 wonders if it's possible to send an HTML form by email 14:40:23 ais523, emails? attach file? 14:40:40 I have nothing against MIME to attach files 14:40:49 pics, patches and such 14:40:59 AnMaster_ipv6: you don't get the point, the point isn't to send information 14:41:03 the point is that the email /looks pretty/ 14:41:08 oh 14:41:10 hm 14:41:18 ais523, then don't use email duh 14:41:57 ais523, and you can send smilies: 14:42:05 :D :) and so on 14:42:10 that is what smiles originally are 14:42:12 pure text 14:42:30 you could make your email client convert them in pure text messages 14:42:43 * AnMaster_ipv6 has seen a bug tracker that converted smilies everywhere 14:42:54 made C++ backtraces look horrible 14:42:55 You can't have that pretty pink background in your message. 14:43:00 and yes a C++ project used it 14:43:02 AnMaster_ipv6: emails sent in Outlook by default not only use HTML, but with some awful Microsoft out-of-bound information 14:43:16 also, replies come up in blue 14:43:23 personally I think it looks hideous 14:43:31 out-of-bound or out-of-band 14:43:33 or both? 14:43:40 I meant out-of-band 14:43:44 MSN messenger converts (or at least used to) some pretty common character combination to a smiley, I think it was ":s". 14:43:47 I haven't slept for almost 24 hours, please forgive me 14:44:01 fizzie: in what way is :s common? 14:44:12 fizzie, add notice at top saying: For best viewing experience set your terminal emulator to use pink background. 14:44:14 I'd say MSN is probably acting according to the wishes of the majority of its users 14:44:16 Problem solved :D 14:44:29 AnMaster_ipv6: err... how many people view email on a terminal emulator nowadays/ 14:44:55 ais523, well ok, but usually GUI clients doesn't have features like that 14:45:12 features like what? 14:45:13 thus making text based ones more flexible (as usual) 14:45:21 they do have HTML email support, always in my experience 14:45:25 I know of no exceptions 14:45:29 ais523, ... for text only stuff 14:45:43 + you can get pink for the inbox listing too in pine I guess 14:45:50 if you set your terminal's colour scheme that wya 14:45:51 way* 14:45:56 * ais523 is suddenly annoyed that ehird isn't here to watch 14:46:10 probably possible in konsole at least 14:46:31 both konsole and gnome-terminal could handle it easily 14:46:32 ais523, well I'm glad he isn't because he wouldn't just watch, he would take active part 14:46:37 you are still missing the point of HTML email, anyway 14:47:02 if your alternative is to insist that everyone who reads your email uses a terminal emulator just so they can override the background of your email to pink 14:47:06 then you are missing the point 14:48:15 ais523: Finnish as all those suffixes instead of prepositions, and a lot of them start with 's'; OTOH, for abbreviations a common style is to use : before the suffix. So "in C++" would be translated as "C++:ssa", and MSN would convert that to "C++[silly smiley]sa". 14:48:41 you are still missing the point of HTML email, anyway 14:48:42 well yeah 14:48:49 maybe because there isn't any? 14:48:53 fizzie: ah, OK 14:49:00 I assume that most MSN Messenger users aren't Finnish 14:49:03 I would get very annoyed if I got a pink email 14:49:09 just based on statistics 14:49:10 or a pink mail at all 14:49:13 probably most are American 14:49:44 Well, maybe not "a lot of them" start with s, but both the "-ssa" and "-sta" suffixes (which correspond to "in" and "from" prepositions, roughly) do, and those are rather common. 14:50:46 fizzie, can't you turn that off in your msn client? 14:51:56 AnMaster_ipv6: I have trouble just trying to close MSN when someone else has left it running 14:52:06 ais523, hm 14:52:09 it is a very hard program to use correctly, the user interface is seriously unintuitive 14:52:12 I suspect the smileys are configurable. I do MSN with Bitlbee in an IRC client, anyway. 14:52:18 ais523, Ctrl-Alt-Del and kill the process? 14:52:25 AnMaster_ipv6: in a cybercafe? 14:52:28 oh ok 14:52:38 worrying enough that the session didn't end when the person before me ran out of time 14:52:53 rue 14:52:56 true* 14:53:31 more worrying still that control-alt-delete actually worked there and you could kill the thing that logged you off with it 14:53:38 but I didn't do that as it probably would have been theft 14:53:48 hm, irritating when music have very quiet and very loud parts after each other. you have to constantly tune the volume to either be able to hear the music or not destroy your ears 14:53:48 :( 14:54:46 AnMaster_ipv6: apply some compression to it? 14:55:56 ais523, hm how 14:56:21 ais523, I'm listening using a radio. Not a computer 14:56:22 AnMaster_ipv6: compression as in the musical effect 14:56:35 and probably most radios don't have it available 14:56:40 hm 14:57:45 So I have 12 commands in my esolang 14:57:51 And I minimized it to just use -35 to 35 14:57:54 One of the higher-end home theater -style systems could probably do it; at least it certainly could if they'd bothered to program that part in, since it's all done with DSP anyway. 14:57:55 which gives me a lot more space 14:58:02 ais523, hm wouldn't it also destroy the feeling of the music kind of 14:58:24 for something like Per Gynt (full opera) at least... 14:58:47 Tuning the volume around mid-piece is probably not any better than processing the signal otherwise, for the "intent" of the composer. 14:59:02 fizzie, true :/ 14:59:07 hmm, is there any XML-based esolang? i'm designing one but i'd like to see how other XML-based one does. 14:59:15 lifthrasiir: XSLT. :p 14:59:33 fizzie: that's true, but what others? 15:00:24 fizzie, they should limit the dynamic range, some sort of regulation is needed 15:02:50 Ok, -50 to 50 15:03:48 Well, unless you have an excessively fancy radio, you'll probably have to run it through a computer or something if you want to do that at your end. 15:05:54 fizzie, well this isn't such a fancy radio 15:06:56 -!- ais523_ has joined. 15:08:43 i can't stand graphical smileys, it's not that they're big and yellow, that's okay, it's just it looks really weird when the face is ...upside-up 15:08:56 lol 15:09:10 A simple infinite loop in Anglent looks like a curve 15:09:27 as in, Anglent 0.9 15:10:08 well okay i like ^__^ 15:10:12 oklopol: ☺ 15:10:15 maybe it's the yellow thing after all. 15:10:21 fizzie: so weird. 15:10:31 maybe it's the fact it's in a circle 15:10:54 can you do _ plus umlaut for me? 15:10:58 You only like people who do not have a physical head, just the face flying around. 15:11:10 ⚍ is pretty close. :p 15:11:19 well. that looks fine. 15:11:38 if only i had any idea how to do it 15:11:51 how do people do unicode anyway? 15:12:18 oklopol: I just pick from the "gucharmap" application. That one was U+268D DIGRAM FOR LESSER YIN. 15:12:19 do you have like a 2d chart and an approximate knowledge of what's where 15:12:27 oklopol: I use the character map application on here 15:12:35 and approximate knowledge to find the character quickly 15:12:39 gucharmap and "view / by unicode block", that's usually enough. 15:12:46 some I have memorised, like é and → 15:12:57 → 15:13:01 Em-dash with combining diaeresis: —̈ 15:13:02 it's on my keyboard 15:13:37 é too, just ' plus e (' only shows the correct dottie when used with the e) 15:14:00 fizzie: omg that's cute :> 15:14:12 except the name 15:14:16 There's a dead key in the default fi layout which adds ´ and ` to letters. 15:14:20 What about Basilisk's Eye 15:14:34 diaeresis sounds like a plural of diarrhea 15:14:41 -!- ais523 has quit (Connection timed out). 15:14:43 13:45 ais523 is suddenly annoyed that ehird isn't here to watch 15:14:57 AnMaster_ipv6 suffers from "I don't like this feature, why isn't everyone an übergeek like me"-itis 15:15:06 The mutant: ⋮) 15:15:20 -!- ais523_ has changed nick to ais523. 15:15:29 the third eye is not on the face here 15:15:50 that's a colon + a third dot almost on the line above 15:15:58 It's probably not been optimized for smiley usage; it was just the vertical ellipsis in mathematical operators. 15:16:09 well. noodle time i thinks 15:16:15 ah' 15:16:17 *ah 15:16:28 is it bad that I like xslt sort of 15:16:30 is there a 15:16:32 . 15:16:33 . 15:16:35 . 15:16:41 oklopol: is ⁖) better? 15:16:49 yes! 15:16:53 fizzie: Eyes are not so fat! 15:16:56 or big 15:16:59 And yes, there's both ⋰ and ⋱ 15:19:06 that makes no sense either as a spambot or as a please send me the codes 15:19:21 what's going on there? 15:19:53 I am think it a Chingrish, sir 15:20:25 *ahem* Yep, centrainly Chingrish or a Chinese spambot 15:21:08 fizzie: Eyes are not so fat! or big <-- ... never heard that story 15:21:09 :P 15:21:43 I'm going to work on many "graph" languages which can be interpreted by human 15:22:26 well, not many 15:22:32 But Anglent is a good start 15:27:09 -!- Judofyr has joined. 15:31:31 -!- Asztal_ has quit (Read error: 110 (Connection timed out)). 15:35:39 -!- Asztal_ has joined. 15:38:33 I made AngF too, BF which looks like Anglent code 15:40:59 And I drew Cat in AngF 15:41:46 http://lifthrasiir.jottit.com/xumul well, i'm designing this stupid language now. 15:42:06 lifthrasiir: You can't get more stupid than AngF 15:42:12 Which is basically Graph BF 15:42:17 wow. 15:42:25 But Anglent is quite more original 15:42:37 with forking and human-interpreting 15:42:43 WELL, AngF can also be human-interpreted 15:42:44 but why? 15:43:35 asiekierka: why not? 15:44:11 IRP is still going to be the best human-interpreted esolang around 15:45:10 I know 15:45:23 But Anglent can be BOTH human-interpreted AND computer-interpreted 15:45:37 And possibly electromechanical-interpreted 15:47:01 anything that can be computer-interpreted can be human-interpreted 15:47:18 Well, yep 15:47:42 Except Malbolge, because every human will give an error OR destroy his brain before finishing "Cat" 15:47:59 And my esolang is specifically crafted to be human-interpreted 15:48:09 well what do you mean by that? 15:48:18 that when done on paper you don't need to erase much? 15:48:26 You don't need to erase AT ALL 15:48:30 with the right ANGLENT board 15:48:33 which is very easy to make 15:49:03 As in, you move a black strip of paper attached to the board via cuts, there are 5 of these, each for a variable 15:49:07 Input is done by moving it 15:49:13 and output is done by writing it on a sheet of paper 15:49:22 You can do it with BF though, too 15:49:25 :P 15:49:28 Well, no 15:49:29 you can't 15:49:33 cuz it has TOO MUCH STUFF 15:49:35 in the pointer 15:49:37 as in 15:49:39 in the array 15:49:42 http://lifthrasiir.jottit.com/xumul well, i'm designing this stupid language now. <-- how do you compute in it? 15:49:57 AnMaster_ipv6: i'm writing "Command semantics" section now 15:50:00 ah 15:50:41 Ok, how can you have a 13.923568-bit-long integer 15:51:12 yesssss, I perfected linux font rendering 15:51:15 asiekierka, why do you want human interpretation 15:51:19 and I thought it couldn't be done! 15:51:25 ehird, details 15:51:33 asiekierka: 1114111 + 1 becomes 0. So that's mod-1114112 integer with strange description. 15:51:49 AnMaster_ipv6: well, it probably won't suit your tastes: it optimizes for typographical clarity over thinness, and also for high-DPI lcds 15:51:51 but: 15:52:02 I'd still like to know 15:52:31 Enable subpixel rendering. Disable autohinting. Use "Slight" hinting. Make a fontconfig block that conditions on font size >16 (uh, I'm not sure if it's points or pixels... Over 16 pixels, anyway) to set hinting to full 15:52:51 == no blue/red edges on large fonts, but text fonts keep their full body 15:53:09 AnMaster_ipv6: Well, uh, what's the point of Wheel then? 15:53:09 Disable autohinting. <-- I mentioned that before already... 15:53:12 or what was it called 15:53:14 Whirl, I think 15:53:16 It cannot contain any entities. (%foo; etc.) 15:53:19 lifthrasiir: &foo; 15:53:21 asiekierka, ? 15:53:44 ehird, hm interesting idea 15:53:44 Whirl, the one where you have 2 wheels you rotate 15:54:00 ehird: it will only permit one of < > & " { ઼, since DOCTYPE is not permitted 15:54:02 esolangs.org/wiki/Whirl 15:54:06 asiekierka, never heard of it before 15:54:08 lifthrasiir: its' not %foo 15:54:10 it's &foo 15:54:19 ehird: then that's error. 15:54:28 AnMaster_ipv6: yeah -- with full hinting on everything, fonts all look pretty much the same and letters are very thin and undefined (esp. with subpixel) 15:54:28 it cannot be in initial source code. 15:54:33 but with slight larger fonts get artifacts 15:54:41 It uses 2 rings and it can also be easily human-interpreted 15:54:43 ehird, hm right 15:55:00 asiekierka, also I didn't say there was anything wrong with human interpretation. But rather I was wondering why you seemed so fixated on them.. 15:55:06 lifthrasiir: He's trying to tell you that your spec says "any entities (%foo; etc.)" which is not right. 15:55:20 oops. 15:55:20 Since %foo; is not an entity. 15:55:29 fizzie: I appoint you my translator. 15:55:31 ehird, details about fontconfig config hm? 15:55:44 maybe i misread the XML spec.... hmm, 15:55:45 AnMaster_ipv6: Sure, sec 15:56:20 what's wrong with zsnes: 15:56:23 nasm -w-orphan-labels -D__UNIXSDL__ -f elf -DELF -D__OPENGL__ -O99999999 -D__RELEASE__ -o video/makev16b.o video/makev16b.asm 15:56:24 hm 15:56:27 -O99999999 15:56:29 crazy 15:56:36 AnMaster: Cuz it's fun to make something that can be interpreted by human without pain 15:56:37 also it doesn't make much sense for an assembler 15:56:50 asiekierka, it is fun to write an interpreter too 15:56:51 -!- Judofyr has quit (Remote closed the connection). 15:56:57 or compiler 15:57:07 -!- mib_ueqe1k has joined. 15:57:19 AnMaster_ipv6: Hi from linux (so I can copypaste) 15:57:22 asiekierka, a bf-to-C compiler isn't too hard for example 15:57:24 mib_ueqe1k, heh 15:57:26 ehird: hmm, what is %foo; called then? i got confused while reading XML spec. 15:57:35 lifthrasiir: %foo; is just plaintext in xml. 15:57:38 Do you mean a dtd? 15:57:43 %foo means something in a dtd. 15:57:45 AnMaster: Well, I just don't want to make an interpreter 15:57:47 But I could 15:57:52 And I may later, even 15:57:52 mib_ueqe1k: ah, got it. thank you. 15:58:01 And I may even write a printing app too 15:58:22 asiekierka, AnMaster doesn't highlight me atm, only this nick does. So what is wrong with tab completion? 15:58:30 AnMaster_ipv6: Okay, here's how to set it up: Open $desktop_environment's font configuration panel, enable subpixel rendering, set hinting to Slight. Then, in ~/.fonts.conf or whathaveyou: 15:58:30 16 hintslight 15:58:35 AnMaster_ipv6: Oh, didn't know about that 15:58:43 mib_ueqe1k, hm... 15:58:44 Then, log out and log back in again. 15:58:45 AnMaster_ipv6: Can you not set up highlighting so that AnMaster highlights as well? 15:58:53 Deewiant, I could 15:58:55 but why? 15:58:59 Why not? 15:59:12 AnMaster_ipv6: Note of course that any new font rendering settings take getting used to 15:59:12 AnMaster_ipv6: It sort of makes sense for nasm, as there the n in "-On" (when n>1) is the limit for the number of optimization passes to do (when making branch-offset-type instructions use less bytes), so if you want to specify "do it as many times as necessary", you need a ridiculous number. 15:59:14 Deewiant, too much work. 15:59:22 and now: why? 15:59:27 I'm wondering what other "draw" Esolang will I make 15:59:32 also yes it would be a few key presses away 15:59:36 I'm thinking of circles 15:59:39 about one line of elisp iirc 15:59:39 AnMaster_ipv6: Would you like a screenshot? 15:59:42 -!- ais523 has quit (Read error: 104 (Connection reset by peer)). 15:59:46 mib_ueqe1k, sure 16:00:08 mib_ueqe1k, since my screen is low dpi I couldn't use the same anyway 16:00:10 AnMaster_ipv6: Helps people that don't tab-complete to get to you, or people who just mention your nick in the middle of a sentence (which often implies lack of tab completion) 16:00:17 mib_ueqe1k, but that paste above was messy with no newlines 16:00:21 why not a pastebin? 16:00:22 :) 16:00:28 AnMaster_ipv6: laze 16:00:31 ok 16:00:34 valid reasin 16:00:42 AnMaster_ipv6: also, what dpi's your screen? 16:00:45 Mine's 96dpi 16:01:01 I can (and do) tab complete in the middle Deewiant. What sort of client is missing that? 16:01:14 Any ideas on what easy-to-human-interpret esolangs are there? 16:01:14 mib_ueqe1k, you said it was higher iirc? 16:01:17 except Anglent and Whirl 16:01:17 AnMaster_ipv6: I didn't say that clients don't support it, just that people don't always do it 16:01:30 mib_ueqe1k, and how do I calculate DPI now again.. 16:01:41 AnMaster_ipv6: IME people are much more likely to tab complete when addressing than anywhere else 16:01:56 AnMaster_ipv6: "xdpyinfo | grep resolution" 16:02:08 Assuming your X knows the correct DPI. 16:02:26 fizzie, I don't know if it does 16:02:27 $ xdpyinfo | grep resolution 16:02:27 resolution: 86x86 dots per inch 16:02:39 AnMaster_ipv6: how big's your screen? 16:02:44 A measuring tape can also help. 16:02:45 and how big's your resolution 16:03:18 mib_ueqe1k, res: 1400x1050 16:03:22 * AnMaster_ipv6 gets a ruler 16:03:25 Wow, how do you cope? 16:03:30 xdpyinfo reports "94x94 dots per inch" here; I'm not sure how well it handles this "two monitors, not exactly identical DPI" case. 16:03:37 I find this 1680x1050 monitor smothering. 16:03:44 Or some other word. 16:03:50 -!- ais523 has joined. 16:03:55 screen on diagonal is more than my 50 cm ruler anyway 16:04:06 which is the longest ruler I have 16:04:06 ais523: Hi. I solved linux font rendering, by the way. 16:04:17 I'm going to be rich. And famous. Rich and famous. 16:04:20 You want either X or Y, so that you can divide it with that number of pixels. 16:04:21 AnMaster_ipv6: no tape measure? 16:04:29 fizzie: you want diagonal 16:04:32 ~30.5 cm high ~40.8 cm wide 16:04:46 +/- a few mm 16:04:51 a bit hard to measure 16:05:14 mib_ueqe1k, no measure tape around here anyway that I can find 16:05:16 87.44262295081967213116 dpi in the Y side, 87.15686274509803921570 in X. 16:05:17 AnMaster_ipv6: your monitor box has something on it about dpi, probably 16:05:25 fizzie: ok, so his X is right 16:05:37 that's quite a low dpi 16:05:37 Close enough, anyway. 16:05:39 mib_ueqe1k: what's your solution? 16:05:49 mib_ueqe1k, monitor box? You think I store the original box it came in? 16:05:50 also, this keyboard at the university on the computer next to me is crazy 16:06:05 ais523: subpixel + no autohinting + a fontconfig that conditions: pixelsize<=16 = slight hinting, >16 = full hinting 16:06:11 I'm uploading a screenshot as-we-speaketh to show andreou . 16:06:12 Er. 16:06:14 AnMaster_ipv6. 16:06:23 Hm 16:06:29 "screen #0: dimensions: 2944x1280 pixels (795x345 millimeters)". That's rather far from correct, since the screen is not even actually rectangular. 16:06:30 It is optimized for high-dpi LCDs, like I said. 16:06:33 mib_ueqe1k, anyway what would you recommend for this res then? 16:06:35 :P 16:06:42 it has three modifier keys in the bottom-left: one marked Ctrl, one marked with the Windows logo, Alt, and Option (all on the same key), and one marked with Alt, the Apple logo, and that weird loopy thing that you see on Apple keyboards sometimes 16:06:52 yes, that means two keys are marked alt 16:07:05 AnMaster_ipv6: Ehm, anything's going to look ugly on that IMO, but my settings might work ok if you could get over your allergy of subpixel rendering 16:07:20 mib_ueqe1k, I can't I tried before 16:07:21 ais523: Windows logo on an Apple keyboard? 16:07:21 What? 16:07:26 -!- oerjan has joined. 16:07:27 and I'm happy with what I have atm 16:07:29 Also, that's the place of interest sign, aka the command-key. 16:07:29 mib_ueqe1k: it isn't an apple keyboard I don't think 16:07:37 I don't think it's a windows keyboard either 16:07:44 ais523: It's probably trying to be some sort of "works for both windows and OS X" keyboard, with dubious results. 16:07:53 fizzie: that's my guess too 16:08:08 AnMaster_ipv6: I'll uploaderate this screenshot anyway, so you can imagine you have a high-dpi display and imagine drooling. 16:08:20 still, even on OSX, would you have command and apple on the same key/ 16:08:25 "screen #0: dimensions: 2944x1280 pixels (795x345 millimeters)". That's rather far from correct, since the screen is not even actually rectangular. <--- ? 16:08:28 ais523: they are the same key 16:08:29 so yes 16:08:38 Anyway, i'm going to work on something else 16:08:44 Mine: " " 16:09:04 I want one with the linux logo 16:09:05 it's very confusing, anyway 16:09:14 to start with I assumed they'd multiplexed alt and start 16:09:16 actually I don't 16:09:17 so that tapping alt opened the start menu 16:09:22 AnMaster_ipv6: Tux isn't very ... logo. 16:09:23 I want one with Meta and Super 16:09:24 .. 16:09:31 mib_ueqe1k, mascot then 16:09:59 AnMaster_ipv6: I have one 1280x1024 TFT in a 90-degree angle (so it's actually 1024x1280), and another 1920x1200 TFT the right way around, so the xdpyinfo dimensions reports the bounding box, even though there's a 1920x80 pixels of unusable area at the lower-left corner. 16:10:04 AnMaster_ipv6: added "command semantics" section now. 16:10:14 fizzie, hah 16:10:18 OK, this thing just won't upload. 16:10:26 Anyone got, like, an ftp server? 16:10:28 fizzie, DPI should be separate between screens 16:10:35 Also I have no idea how it calculated the physical size and DPI since they differ. 16:10:38 mib_ueqe1k, upload where? 16:10:42 and what 16:10:44 AnMaster_ipv6: anywhere. 16:10:47 my screenshot 16:10:49 mib_ueqe1k, ompload? 16:10:53 should work 16:11:01 It probably is separate, also, but there's some sort of combined value for non-multi-screen-aware apps. 16:11:04 there is some box on the web page 16:11:13 AnMaster_ipv6: no s are working 16:11:25 mib_ueqe1k, what file format are you using? 16:11:32 also crazy.. 16:11:35 png. It's a problem with the VM. 16:11:39 hm 16:11:47 how strange... 16:12:04 mib_ueqe1k, I don't have that no, but couldn't you scp it to the host? 16:12:11 from the vm 16:12:36 I'm thinking about an Esolang that uses only 2 shapes 16:12:39 i could just use this "Shared Folders" thingy-bob-ma. 16:12:54 mib_ueqe1k, heh 16:14:18 great, i have to install this proprietary shitware to get the shared folders working 16:14:21 boy oh boy I want my quad core soon 16:14:32 so I can run this natively... 16:14:44 OS type: Xorg version: -------- ------------ Unbuntu 7.04 7.2.0 Ubuntu 6.10 7.1.1 Ubuntu 6.06 7.0.0 16:14:49 messing around with virtualisation has taught me one interesting fact 16:14:50 fucking up to date support you got ther 16:14:50 e 16:14:52 UNBUNTU 16:14:58 fuuuuuuuuck you parallels corporation 16:15:00 -!- Judofyr has joined. 16:15:08 ais523: oh? 16:15:10 which is that mounting a drive simultaneously from two operating systems tends not to work 16:15:14 at least with ext3 16:15:40 that's a pity, it ought to, it's certainly an expressible operation with obvious semantics 16:15:45 just the OSes seem unable to handle it 16:15:53 (I tried with Knoppix + Ubuntu, last time) 16:16:05 well, they'd start fucking each other's files 16:16:27 grrrr, I can't figure out how to get this out of the box 16:16:59 ais523: what's rutian's ip? 16:17:16 mib_ueqe1k: I don't know 16:17:23 I always acessed it via domain name 16:17:28 *accessed 16:17:36 yes well it'll have one of them soon 16:17:49 since bogons needs you to have a reverse dns for a static ip 16:17:55 why do virtualisers care about which OS you're virtualising? 16:18:08 ais523: so they can integrate to let you move stuff out of the box 16:18:57 I like the low-tech qemu method, but unfortunately simultaneous mounts would make it so much better 16:20:04 Some of those cluster-use-designed file systems support multiple mounting, I think. 16:20:28 simple solution 16:20:32 nfs 16:20:37 (So you can use them with something like the network-block-device, instead of using a NFS server.) 16:20:43 nfs is never a solution 16:20:45 I don't think I usually associate NFS and simple together. 16:21:16 Although it's certainly a solution; most of the stuff at work and at the university works over NFS. 16:21:18 what? 16:21:23 it was simple to setup for me 16:21:46 took about 5-10 minutes of reading up on it and then about 5 more minutes to set it up 16:21:49 Have you tried to use it with systems that have very different UID numberings? 16:21:53 Wanna hear a joke guys? 16:21:54 NFS! 16:22:06 fizzie, hm no, they have similar ones in fact 16:22:13 and I used nfs4 btw 16:22:37 Screenshot.png 100% 728KB 727.9KB/s 00:00 16:22:41 and it's actually 0kb on the server 16:22:41 (that took another 5 minutes of reading up on) 16:22:44 WHAT IS THIS TRAVESTY 16:22:57 It's certainly far from simple internally, even if package-maintainers or such have made it simple. 16:22:58 mib_ueqe1k, used scp to copy it over? 16:23:07 AnMaster_ipv6: It's actually hanging 16:23:10 after printing that line 16:23:12 oh 16:23:14 and the file is 0bytes on the server 16:23:18 mib_ueqe1k, how strange 16:23:36 take a screenshot from OS X of the screenshot in the vm? 16:24:06 AnMaster_ipv6: Nice try, xzibit. 16:24:11 mib_ueqe1k, err? 16:24:13 what 16:24:20 why would it not work? 16:25:30 Speaking of network file systems, someone has managed to configure Samba on this NAS appliance box in a really strange way: 16:25:32 fis@eris:/e/video/misc$ umask 077 16:25:32 fis@eris:/e/video/misc$ echo x > foo 16:25:32 fis@eris:/e/video/misc$ umask 0 16:25:32 fis@eris:/e/video/misc$ echo x > bar 16:25:33 fis@eris:/e/video/misc$ ls -l 16:25:35 total 8 16:25:37 -rwxrwxrwx 1 root media 2 2009-03-27 07:57 bar 16:25:39 -rwx-wx-wx 1 root media 2 2009-03-27 07:57 foo 16:25:47 AnMaster_ipv6: It would work but I'd have to crop it all after and fffffffffff. 16:25:58 fizzie, that makes no sense 16:26:02 There's a "force create mode = 777" (presumably to avoid any permission-related issues) but for some reason it fails to force the 'read' bits there. 16:26:10 Yes, I agree. 16:26:54 mib_ueqe1k, hm ok 16:27:07 mib_ueqe1k, well why would ftp work any better? Not that I have any ftp 16:27:10 ftp server* 16:27:43 fizzie, too much +x always make me irritated 16:27:50 AnMaster_ipv6: I just wnt to know what the fuck is up. 16:28:00 AnMaster_ipv6: You should've seen how it worked when I mounted with "-o nounix". I can't even describe it, it was that strange. (Making food now, semi-away.) 16:28:07 mib_ueqe1k, indeed it makes no sense since your current connection works 16:28:14 fizzie: what does -o nounix do? 16:28:49 mib_ueqe1k, maybe you hit a bit pattern that cause a bug in the virtual ethernet card or something? 16:29:05 Only one thing for it. 16:29:06 uuencode. 16:29:27 ais523, can't find it in man mount.cifs at least 16:29:38 ais523: It's a quasi-documented flag for mount.cifs not to negotiate the CIFS unix extensions with the remote server. 16:29:48 mib_ueqe1k: use shar, it's like uuencode only more high-tech 16:29:48 It's not in the man page, but it's documented somewhere elsewhere. 16:30:37 ais523: uuencode is part of sharutils 16:30:52 mib_ueqe1k: yep 16:30:59 shar is a wrapper around uuencode 16:31:02 mib_ueqe1k, encode as a data url? 16:31:03 /usr/share/themes/Human/gtk-2.0/gtkrc:82: Murrine configuration option "highlight_ratio" will be deprecated in future releases. Please use "highlight_shade" instead. 16:31:05 with fun stuff like creating directories 16:31:07 and pastebin dataurl 16:31:09 That sort of thing irritates me 16:31:18 I get 3 of those errors every time something gtky happens 16:31:20 Maybe I could patch the theme 16:31:47 mib_ueqe1k, you used to like dataurls iirc 16:32:51 elliott@elliott-desktop:~/Desktop$ scp Screenshot.png.uuencode ehird@208.78.103.223: ehird@208.78.103.223's password: Screenshot.png.uuencode 100% 1003KB 1.0MB/s 00:00 16:32:56 STILL HANGS 16:32:58 Life sux 16:33:08 Okay, MEGA LAST TRY 16:33:11 I will pastebin the uuencode 16:33:14 decode it on my mac end 16:33:15 and imgur.com it 16:33:16 mib_ueqe1k: your password's "Screenshot.png.uuencode"? 16:33:20 Then I will kill myself 16:33:22 ais523: :-D 16:34:02 NOTE TO SELF: 16:34:11 Pasting 1MB of uuencode into a text field in a browser is EL DUMBO 16:34:17 hahaha 16:34:30 why didn't you just use imagebin.ca or somewhere like that? 16:34:39 ais523: I told you -- just like scp, they hang 16:34:40 and never upload 16:34:43 ah 16:34:50 so I uuencoded it --still hanged-- 16:34:54 does the same happen for small files? 16:34:59 so now I'm pastie.org'ing the uuencoded, looking it up on my mac, uudecoding it 16:35:03 and uploading it to imgur.com 16:35:07 ais523: It's only 700kb 16:35:15 yes, but I mean smaller still 16:35:20 can you send a few bytes of binary, for instance? 16:35:43 Almost certainly; I can load http pages. 16:35:50 -!- Judofyr has quit (Remote closed the connection). 16:36:43 hm 16:36:51 ehird, maybe it has problems with large packages 16:36:57 that could explain why irc works 16:37:01 No. 16:37:02 Mibbit works. 16:37:05 well yes 16:37:14 but you don't send a lot of data in each package 16:37:18 it isn't bulk transfer 16:37:55 ehird, for example I had a problem some time ago with a bug in kernel ethernet drivers that made ssh and such work fine but caused massive rsync to hang 16:38:15 This is fuckin' ridiculous. I'ma try the proprietary parallels crap 16:38:18 actually it wasn't ethernet driver, it was somewhere else in the stack 16:38:33 related to tcp window size iirc 16:38:44 ehird, linux auto tunes tcp window size 16:38:56 so do most other modern OS 16:44:25 * AnMaster_ipv6 ponders a SQL based funge space. 16:44:29 no not for cfunge 16:44:58 -!- Judofyr has joined. 16:45:09 it would be slow anyway, but have interesting properties 16:45:40 -!- mib_ueqe1k has quit ("http://www.mibbit.com ajax IRC Client"). 16:45:54 ais523, what do you think? 16:46:04 for example you could SELECT to find next non-whitespace 16:46:43 allowing slowdown.b98 to be reasonably fast without tracking bounds 16:46:47 and other crazy things 16:47:24 meh, anyone there? 16:47:24 AnMaster_ipv6: I thought you meant SELECT the operator from INTERCAL 16:47:27 it might be interesting, though 16:48:25 ais523, efunge has some crazy code to get a block from funge space somewhat like that.. 16:49:07 ais523, http://paste.lisp.org/display/78049 16:49:25 plan was to use it for o and such 16:49:49 also I have thought about ATHR (that was put on hold due to implementation issues) 16:49:56 and I have some ideas there.. 16:50:11 it would mean changing the draft quite a bit 16:50:32 to be "optimistic non-synced funge space" or something like that 16:50:56 it would make it more esoteric (becuase it would be a nightmare to program in) 16:51:07 program with* 16:51:59 basically, if ATHR wasn't loaded everything would be normal, but when you start an ATHR thread the way they interact change 16:52:16 I would tread funge space as consisting of blocks 16:52:26 somewhat like cache lines 16:52:38 and only one thread can own a specific funge space block at a time 16:55:20 * ehird tries to optimize jvm startup tim 16:55:20 e 16:55:22 it gets interesting when you change the semantics by marking a block as non-synced or something, each thread could modify it's local copy and what would happen when they were flushed back to funge space server would be interesting 16:56:10 and anyway, no changes would be visible right away 16:56:15 to other threads 16:56:30 not until the cache line was flushed back 16:56:32 err 16:56:36 funge space block 16:56:59 (but trying to access in that block would make it flush back usually, somewhat like cache lines) 16:57:08 ais523, do you think this is crazy or? 16:57:32 AnMaster_ipv6: sorry, I'm busy with something else, I haven't been paying attention 16:57:38 meh ok 16:59:36 -!- Judofyr has quit (Remote closed the connection). 17:03:08 writing XUMUL spec is getting harder, should i specify how to parse invalid XML? :S 17:05:27 lifthrasiir: no such thing 17:05:35 the XML spec does not allow agents to process invalid xml 17:05:38 if you do, you're breaking the spc 17:05:44 (and also being sane, but xml isn't) 17:06:43 ehird, what about xml validators? they parse broken xml. But maybe they aren't agents. Or maybe the spec guys didn't know because they were _secret_ agents... 17:07:09 AnMaster_ipv6: an xml validator is not an xml-processing UA 17:07:24 also, validators report the first parsing error they see then give up 17:07:31 thus, they comply with the spec 17:08:15 http://www.lyx.org/images/about/insert_menu.png ← WOW those window decorations are ugly 17:08:17 not all validators do. Some try to list more than one error 17:08:26 AnMaster_ipv6: no, that's only post-parsing 17:08:32 ehird: initial document should be valid, but in XUMUL the document is modified in-place and if we don't allow broken XML at least for a while it become uninteresting 17:08:42 lifthrasiir: then it isn't xml 17:08:44 so don't call it xml 17:08:48 ehird, ah when validating against DTD or such you mean? 17:08:52 AnMaster_ipv6: ye 17:08:52 s 17:10:02 ehird: if i call such internal document processed by the program is "similar to XML but possibly invalid", are you right? 17:10:21 lifthrasiir: what you have isn't an xml language, if it deals with invalidity 17:10:22 or "XML-like document", anyway 17:10:28 it's a string-processing-language-that-starts-off-as-xml 17:10:59 lifthrasiir: also, http://www.xsharp.org/ 17:11:02 (it's serious) 17:11:28 If you want to make it all-so-very-XML but allow self-modification, you might make the self-modification use a DOM-style API. 17:11:37 aww 17:11:40 they removed the freaky xml syntax 17:11:41 :-( 17:14:22 http://martiansoftware.com/nailgun/background.html ← ahh, this is exactly what i want 17:14:43 last release 2005 >_< 17:14:59 Didn't XML spec allow parsers to attempt to continue parsing after first error trying to discover additional errors (but the parse has still failed)? 17:15:05 Ilari: hm, maybe 17:15:11 all I know is: 17:15:12 ais523, tell me when you read that bit above though 17:15:19 they are forbidden from sending any more non-error data 17:15:20 strictly 17:15:25 no thanks to tim bray 17:15:26 AnMaster_ipv6: probably not for a while, or ever, tell me when I'm less busy 17:15:31 ok 17:15:33 okay, i agree to you ehird. i should look for another method so it can modify its own source and still remains valid. 17:15:41 lifthrasiir: base something on the DOM ap 17:15:42 i 17:15:49 except... less, you know, Java 17:16:15 ... 17:17:25 -!- ais523 has quit (Remote closed the connection). 17:21:26 -!- Judofyr has joined. 17:35:13 [[Video 17:35:13 Interface to Apple's QuickTime for using a camera, playing movie files, and creating movies.]] 17:35:15 — processing.org 17:35:22 Aw c'mon, you don't have to depend on QuickTime for that. 17:45:00 ehird, apple wouldn't be happy if they heard you say that 17:45:12 ;P 17:54:32 -!- Judofyr has quit (Remote closed the connection). 18:00:44 Wow. I managed to make DejaVu Serif look non-ugly. 18:00:48 Thought I'd see the day, never/ 18:06:47 -!- tombom has quit ("Peace and Protection 4.22.2"). 18:09:20 ehird: http://lifthrasiir.jottit.com/xumul okay, how about this? 18:09:37 * It cannot contain any comments or CDATA sections. Just use normal text for it, since they will be ignored. 18:09:38 it's not xml 18:09:44 * It doesn't use XML namespaces. foo:bar is just a normal identifier, not to mention xml:lang, xmlns:foo and so on. 18:09:45 also not xml 18:09:48 * It should be coded in ASCII-compatible encodings. This excludes, for example, UTF-16 encodings. 18:09:50 ditto 18:09:51 it is just a restricted subset. 18:09:59 right, but the point is that it's not xml 18:10:40 A cell past the first cell, at offset -1, maps to 20th-to-the-infinity character. This means it doesn't actually map to any character in the original XML document, but if it is assigned, it is treated as like in the last character in XML document. For example, suppose that current document consists of . If offset -1 is set to X, current document becomes X, where the whitespace is actually infinitely long but treated as one whitespace. Simila 18:10:43 rly if offset -2 is set to Y, current document becomes Y X. This remains same to other out-of-bound offset, e.g. offset 10 for 199-character-long document, so if offset 0 is set to Z, current document becomes Z Y X. 18:10:48 Being able to set arbitrary text allows for non-xml 18:11:21 Oops, that IS right. I probably change the rule. 18:11:43 I probably have to change it* 18:12:36 Ah, no, 18:13:42 ehird, do you know if Intel CPUs have different instruction dispatch units for SSE and x87? 18:13:53 AnMaster_ipv6: no 18:13:57 as in i don't know 18:13:59 ah 18:14:00 hm 18:14:00 well, i thought that if the original document is and the program changes it to < / f > it is correct, but i realized < and / cannot be separated. 18:14:07 what the heck. 18:14:44 ehird, becuase if they do, using long double for about half of the math you do and double for the rest, and -mfpmath=sse,387 might in effect double the throughput I think 18:14:46 a crazy idea 18:14:51 and I can't even test it on cfunge 18:15:02 because it doesn't use much floating point 18:16:12 lifthrasiir, I'd go the other way instead. Allowing full SGML, not a subset of xml 18:16:26 that way you can have short tags and other fun stuff 18:18:59 AnMaster_ipv6: good idea! but sgmllib in python standard library is removed in python 3... :p 18:19:49 lifthrasiir, why do it in python? 18:19:53 (and sgml is quite hard to parse correctly iirc) 18:21:58 oh, i found http://www.jclark.com/sp/. looks fine except that it's written in C++ :p (just kidding!) 18:22:18 meh I dislike C++ 18:23:23 well, but anyway i have to modify the existing parser or write new parser from scratch since it have to manage character offset of callers 18:23:33 it has* 18:23:57 ah 18:24:23 that's one reason for using a subset of XML 18:26:09 If you just need to know where you are, you can always use expat and the CurrentByteIndex property. It might get a bit tricky with self-modification, though. 18:27:08 Personally I'd write an XML language so that the interpretation part would also be defined in terms of the DOM, as would the tree-modification. But whatever floats your XML-boat, I guess. 18:28:20 woo 18:28:22 shared folders works 18:28:26 screenshot time vsoon 18:28:52 fizzie++ 18:28:56 xml operating on strings is stupid 18:29:00 and xml operating on xml is, uh, xslt :P 18:29:10 fizzie: i was adding all the strange things for keeping the name (XUMUL) but anyway thanks. 18:29:19 ehird, that took quite some time... 18:29:29 I was wondering what you were talking about 18:29:31 AnMaster_ipv6: Yes, well, i was the doing the other the stuff 18:29:32 had forgot it all 18:29:41 the other the stuff? 18:29:45 Yes. 18:29:45 i dislike xslt, and dislike xml operating on xml generally 18:29:50 I'm the talking like the fizzie. 18:30:05 The Other... The Stuff! Soon at a lame tagline generator near you 18:30:08 ehird: XSLT doesn't really operate on the XSLT style-sheet itself, but maybe that's such a small difference. 18:30:19 fizzie: you could make it 18:31:54 AnMaster_ipv6: http://imgur.com/FLORQ.png. Some windows look better than others. Only looks good ona n LCD. May not look good on an LCD that isn't high-dpi. May cause shock from the font-rendering-differing effect thing that everyone gets. Do not use while pregnant. 18:31:59 Er. 18:32:01 It resized it down. 18:32:06 Forget that. 18:32:26 ehird, link broken? 18:32:27 Sec. 18:32:28 or times out 18:32:31 AnMaster_ipv6: no. 18:32:35 your internet is broken. 18:32:38 Worked fine for me. 18:32:41 ehird, meh it was just very slow 18:32:45 AnMaster_ipv6: I said 18:32:48 it's resized down 18:32:51 yeah 18:32:52 so you can't see it properly 18:32:54 so i'm rehosting it 18:32:56 try ompload 18:33:00 no, try xs.to 18:33:10 never heard of that one before 18:33:13 which, while having a shitty design, isn't designed as a shock-site hoster. 18:34:16 Meh, it's broken now 18:34:21 Fine, omploader. 18:35:00 AnMaster_ipv6: http://omploader.org/vMWhnNw. Some windows look better than others. Only looks good ona n LCD. May not look good on an LCD that isn't high-dpi. May cause shock from the font-rendering-differing effect thing that everyone gets. Do not use while pregnant. 18:35:27 ehird, with pixel order? 18:35:30 which* 18:35:33 RGB. 18:35:37 Horizontal. 18:35:42 Like everyone else apart from fizzie's upside down monitor. 18:35:44 Er, sideways 18:35:46 monitor 18:36:10 I think the Nintendo DS Lite screen is BGR, but I could be remembering wrongly. 18:36:29 ehird, I found BGR to be least bad when it comes to colour bleeding on this monitor... 18:36:35 but still bad 18:36:40 AnMaster_ipv6: Have you looked at your subpixels? 18:36:46 If they're not BGR, using BGR is efeating the point. 18:36:49 Entirely. 18:36:52 Since it's not subpixel any more. 18:36:59 what? 18:37:06 >_< 18:37:09 also I can't see individual sub pixel elements 18:37:13 You have to use the pixel order of your monitor. 18:37:17 That's the whole point. 18:37:21 AnMaster_ipv6: get a magnifying glass 18:37:24 ehird, yep, I know that is the point 18:37:39 Anyway, your monitor is RGB. 18:37:42 Nobody manufactures anything else. 18:37:55 Apart from nintendo based on fizzie-rumours. 18:38:20 ehird, tried with magnifying glass, still can't see the sub pixels 18:38:37 AnMaster_ipv6: The only leakage in my screenshot is: the Avant Window Navigator body text, the Fishbowl monospaced tagline, and ever so slightly on the fishbowl body text 18:40:14 The fishbowl text is something I did notice to be a bit bluey-red-y. I'm not sure how much it would annoy. 18:40:17 ehird, I see colour bleeding everywhere in that 18:40:54 AnMaster_ipv6: Either your eyes are so good that you could be making a lot of money just using them now, or you're sitting an inch away from your screen, or your monitor is one of the 3 non-RGB subpixel orders in the world, or it's a CRT. 18:40:58 Pick one. 18:41:03 It's rather horrible in the sideways-up monitor, though. 18:41:05 actually as far as I can make out, sub pixels are BRG here, or possible RBG 18:41:08 hard to say 18:41:13 or any other in fact 18:41:14 RBG? 18:41:16 Um, no. 18:41:18 That doesn't exist. 18:41:22 Now does BRG... 18:41:24 *Nor 18:41:28 ehird, indeed. I can't see any sub pixels 18:41:35 AnMaster_ipv6: Are you sure you don't have a CRT. 18:41:57 ehird, it's way flatter yes. 18:41:58 fizzie: IMO, the avant one is much worse than fishbowl 18:42:01 AL2017 18:42:03 is the model 18:42:03 and I browsed for a while and it wasn't too botherating 18:42:16 fizzie: If you want to make it use the full hinting to fixerate that some more, you can ++ the font size above 16px 18:42:20 Acer AL2017 18:42:41 You're just a hallucinating kind of person then 18:42:56 ehird, anyway I can't make out sub pixle 18:43:00 pixels* 18:43:02 The subpixel order should be rather easy to see if you do (in Gimp in 800%-zoom or something) a file which has a single-pixel vertical line of alternating blue and red pixels, and then look at it at native size, with the magnifying glass or otherwise. 18:43:06 but I can see the colour bleeding 18:43:18 AnMaster_ipv6: If you don't have subpixels of course you see bleeding. 18:43:22 At least here it's very easy to see which pixels (red or blue) are on the left side and which on the right. 18:43:27 But really I don't think so. 18:43:35 It's probably your monitor's DPI being awfully low. 18:43:42 ehird, yep 18:43:46 it is low 18:43:50 Here's a nickel; go buy a decent monitor. 18:44:07 ehird, nickel is slang for money? 18:44:14 ... 18:44:18 * ehird explodes with laughter 18:44:23 ? 18:44:45 http://en.wikipedia.org/wiki/Nickel_(United_States_coin) 18:44:54 Do you know _anything_ about non-Swedish things? 18:45:08 ehird, thought US used dollar and cents 18:45:12 HAHAHAHAHA 18:46:53 * lifthrasiir seriously thinks about ehird exploding 18:46:57 ehird, still it says this nickel is 5¢ 18:47:12 lifthrasiir: I explode literally quite often 18:47:15 there's a little me inside of me 18:47:24 and it grows to be the size of me when I explode leaving only the little m 18:47:24 e 18:47:24 -!- ais523 has joined. 18:47:29 ehird, not my fault they use funny names when they mean "5 cent coin" 18:47:33 . 18:47:33 you may see this as a metaphor for man babies 18:47:46 I assume you have that TFT plugged in with a DVI cable, though? How much would the analog-blur-effect affect that sort of stuff, anyway? 18:47:52 fizzie, VGA 18:47:57 no DVI supported 18:47:59 i misunderstood ehird too, for a sec, thought nickel was slang for "tip" 18:48:07 AnMaster_ipv6: geez 18:48:12 then i realized maybe you just meant nickel 18:48:17 oklopol: yeah but you're detuned from reality in a funny way, I mean it's oklopolific 18:48:19 ehird, not supported on video card indeed. 18:48:26 oklopol: so do i. 18:48:44 lifthrasiir: you still do? 18:48:46 isn't nickel the material they're mostly made of? 18:49:02 okay, so did i. 18:49:30 fizzie, also I have tested this monitor with both dvi and vga cables before (on a different computer. Saw no difference 18:49:39 "75 % Copper, 25 % Nickel" says the linked page. 18:49:54 Well, except the nickel-free nickels of ww2. 18:50:01 that's funny. 18:50:41 bbl 19:07:01 AFAIK, one of the most signaficant sources of analog blurring is the video card. 19:10:00 -!- asiekierka has quit (Read error: 110 (Connection timed out)). 19:10:42 Also, isn't there both Analog DVI and Digital DVI (some systems might support both)? 19:12:57 back 19:13:21 ais523, hi 19:13:50 hi 19:14:14 ais523, so what did you think about the idea for fungespace coherency? 19:14:25 (for ATHR) 19:14:37 AnMaster_ipv6: very busy atm, I'll talk to you later 19:14:40 ok 19:17:12 -!- ais523 has quit ("http://www.mibbit.com ajax IRC Client"). 19:26:32 -!- Dewi has quit (Read error: 110 (Connection timed out)). 19:37:13 Deewiant, btw efunge should be run on R12B-5 to be mycology compatible, I believe (from changelog reading) it will break in 0-byte handling under R13A (that is an alpha version), and for now I'll just say R13 isn't supported until I have time to test R13 and fix any bugs there. 19:38:23 And how do I know what I have 19:38:44 Erlang (BEAM) emulator version 5.6.5 [source] [64-bit] [async-threads:0] [hipe] [kernel-poll:false] 19:38:53 5.6 means R12 19:39:25 So you just have to know what the major version numbers mean? :-P 19:39:28 hm 19:39:47 wait I'm wrong 19:40:09 Deewiant, you know because the tarball you downloaded was called otp_src_R12B-5.tar.gz or such 19:40:19 No I don't because I didn't download a tarball 19:40:29 I have programs which do these things for me 19:40:46 Deewiant, ok I think it is the version of the application kernel that matters (erlang application) 19:41:00 $ ls -d /usr/lib/erlang/lib/kernel* 19:41:00 /usr/lib/erlang/lib/kernel-2.12.5 19:41:08 that is R12B-5 19:41:27 -!- neldoreth has joined. 19:42:20 Deewiant, anyway in R13 all console IO will be UTF-8, previously it was all ISO-latin-1. 19:42:57 Deewiant, and there is no way around it. Erlang is semi-closed-world style. 19:43:23 Sure there is, you can decode/encode UTF 19:44:02 Deewiant, no, because R13's console IO code will throw exceptions on non-UTF it seems 19:44:10 anyway I can read the file still 19:44:14 depends on how I do it it seems 19:44:38 Yees, but if you want to output non-UTF you just encode it into UTF first 19:44:41 And likewise for input 19:44:58 yes 19:46:03 Deewiant, anyway erlang's console IO will output most non-printable as ^@ or whatever when running in interactive mode. This is a feature (again can't be turned off). 19:46:13 oh and interactive mode is one valid way to start efunge in 19:49:51 -!- oerjan has quit ("Globus"). 19:50:01 AnMaster_ipv6: and you whine about closed world systems 19:50:23 at least fully closed systems let you do things flexibly from inside 19:50:40 ehird, true. I never claimed this was good. 19:50:54 ehird, but now that I started I have to bite the bullet :P 19:50:59 (I believe that is the idiom?) 19:52:29 "ASCII English text, with CRLF, CR, LF line terminators" 19:52:32 huh 19:53:06 Wut 19:53:47 oh I managed to mess up 19:53:56 I thought the file was in LF and wanted to convert to CR 19:53:59 but it was CRLF 19:54:21 so recode added something strange 19:54:43 that explains it 19:55:41 I'm amused by the fact that efunge of all interpreters doesn't support t 19:56:26 Deewiant, well it doesn't currently indeed 19:56:31 Deewiant, plan was to add ATHR 19:56:40 anyway that hit some issues and was put on hold 19:56:42 $ ./efunge ~/src/cfunge/trunk/tests/refc-invalid-deref.b98 19:56:42 No REFCCFER 19:56:44 interesting 19:56:48 * AnMaster_ipv6 fixes his test code 19:57:05 anyway I just pushed a fix in efunge for form feed handling 19:58:14 Deewiant, anyway the threading in erlang is async and message passing 19:58:21 Yes, I am well aware of this 19:58:32 Deewiant, I feel supporting t would be wrong in it 19:58:37 o will be supported at some point 19:58:39 t won't 19:58:42 How 'wrong' :-P 19:59:01 How is it more wrong than supporting any other feature of Befunge-98 19:59:27 Deewiant, because it doesn't allow me to take advantage of the message passing concurrency in erlang 19:59:30 Deewiant: quick, make pervasive use of t in mycology 19:59:38 ehird, it is optional 19:59:43 in the spec 19:59:52 AnMaster_ipv6: And my point was, neither does the rest of Befunge :-P 20:00:41 Deewiant, ATHR will be better. I guess I can have a lock step compat mode that sends a token around to support t 20:00:50 it sounds like a fun way to implement t 20:01:07 and remember, efunge isn't about speed, as long as I don't have to wait too long I'm happy with efunge 20:01:21 efunge seems decently fast to me 20:01:32 Takes a while to shutdown but I guess that can't be helped 20:01:46 Deewiant, indeed it can't, it is because erlang need to stop internal processes 20:01:58 well you could bypass it, but then I couldn't send back exit code either afaik 20:02:04 (I think I tested) 20:02:49 hm efunge trunk locks up on jumpwrap.b98 for some reason, will have to debug that 20:03:02 AnMaster_ipv6: efunge is way faster than pyfunge FWIW 20:03:07 hah 20:03:30 Deewiant, well it is decent if you use ERL_COMPILER_OPTIONS='[inline,native,{hipe,[o3]}]' make clean all 20:03:34 Over twice as fast including the shutdown time 20:03:35 AnMaster_ipv6: I did 20:03:49 Deewiant, it is horribly slow without using native/hipe 20:04:08 * Deewiant checks what AnMaster_ipv6 thinks is 'horribly slow' 20:04:19 heh, 20:04:20 Deewiant, I don't like waiting 30 seconds for mycology 20:04:26 or 20 or so 20:04:35 I mean, 5-15 is ok for efunge 20:04:42 but I got other stuff to do 20:04:51 I see almost no speed difference, did I muck something up 20:04:59 I'm in supervisor-tree FWIW 20:05:18 oh ok, supervisor-tree is known buggier, a feature branch for ATHR work 20:05:58 Deewiant, if there is spurious output at start/end (or in middle) in that branch I will just point you to trunk for now 20:06:21 Deewiant, and you need to do make clean between 20:06:32 and make sure the env variable is no longer set 20:06:39 I did 20:06:47 Speed difference is around 0.1s 20:06:50 ah 20:06:53 Deewiant, erl 20:06:58 what does the first line say 20:06:59 Dominated by the 1+s it takes to shutdown 20:07:05 Erlang (BEAM) emulator version 5.6.5 [source] [64-bit] [smp:4] [async-threads:0] [hipe] [kernel-poll:false] 20:07:09 hm ok 20:07:11 -!- kar8nga has joined. 20:07:12 it does have hipe 20:07:14 so not that then 20:08:03 Hmm, pyfunge doesn't have a way of disabling fingerprints so it's not fairly benchmarkable 20:08:13 Deewiant, here it is 2 seconds vs. 7 seconds (trunk) and 3 seconds vs. 14 seconds in the supervisor branch 20:08:15 But then efunge doesn't either 20:08:20 Deewiant: use -f ''. 20:08:23 AnMaster_ipv6: Both around 1.5s here in supervisor 20:08:25 (maybe) 20:08:29 -!- KingOfKarlsruhe has quit (Remote closed the connection). 20:08:30 Deewiant, yes it has, called "edit source" 20:08:35 lifthrasiir: I tried that, it crashes :-P 20:08:42 lifthrasiir: ImportError: No module named fp_ 20:08:43 Deewiant, well supervisor branch can make use of smp 20:08:43 oops, :S 20:08:50 potentially 20:08:56 you probably need to pass erlang some option 20:08:58 lifthrasiir: I also tried just specifying 'standard' but it didn't seem to change anything 20:09:00 hmm i didn't check for null argument... well. 20:09:04 and the message passing overhead may make it slower 20:09:08 -!- jix has quit ("Computer has gone to sleep"). 20:09:08 AnMaster_ipv6: 'Edit source' isn't an option 20:09:32 Deewiant: it will disable other fingerprints (e.g. STRN) except for standard one. you mean it wasn't? 20:09:36 Deewiant, well just comment out in the listing of fingerprints in src/efunge_fingerindex.erl 20:09:37 -!- tombom has joined. 20:09:41 so only 20:09:43 lookup(_Fingerprint) -> notfound. 20:09:45 is left 20:09:53 of the lines starting in "lookup" 20:09:59 lifthrasiir: Still runs HRTI, TOYS etc 20:10:10 lifthrasiir: Oh right, by standard you mean the Cat's Eye ones 20:10:14 So it did work 20:10:17 yes 20:10:20 There's just no way of disabling them all 20:10:33 * lifthrasiir thinks of misnamed fingerprint module 20:10:48 lifthrasiir: I would've expected a per-fingerprint list and not a category list :-) 20:10:55 Deewiant, in efunge branch overhead is rather large, thus I won't add it, adding a option for tracing led to a 0.5 second slowdown or so when I tested 20:11:02 0.5-1 second 20:11:26 AnMaster_ipv6: You don't need to branch at every instruction, it's just a check when you do ( or ) 20:11:32 so for tracing, you should uncomment the relevant line and recompile 20:11:35 :D 20:11:40 Deewiant: i should have done so, but didn't do yet :p 20:11:53 lifthrasiir: No worries :-) 20:12:09 It's not like being easily benchmarkable is what Funge interpreters should aspire to... 20:13:11 oh interesting 20:13:12 hm 20:13:35 hm 20:13:50 D in FIXP behaviour questionm 20:13:52 question* 20:14:00 (D = random number) 20:14:07 what if the parameter is 0 20:14:16 efunge currently reflects and cfunge pushes 0 20:15:10 Deewiant, are both correct? 20:15:18 it seems undef 20:15:40 AnMaster_ipv6: Why would it reflect 20:15:50 hm ok 20:16:17 Anyway, I guess the interpreters in current speed order are: cfunge, RC/Funge-98, CCBI, efunge, pyfunge, Language::Befunge 20:17:03 random:uniform(0) throws an exception 20:17:16 Although I'm a bit hesitant to include RC/Funge-98 in that list since it still uses stuff like gets() :-P 20:17:23 It's full of buffer overflow bugs 20:17:29 Deewiant: i pushed the version contains --disable-fprint option, and it takes 1.7s for mycology with psyco enabled 20:17:55 ..and 2.7s without psyco enabled 20:17:59 Deewiant, anyway I recommend efunge trunk atm for stability, the cool new stuff in the supervisor branch is so far only under the hood and it is buggy 20:18:12 lifthrasiir: Around 1.7s for me too 20:19:01 Deewiant, I fixed two bugs the last few minutes that have not been ported to the supervisor branch yet 20:19:59 AnMaster_ipv6: Their speed is about the same for me, 0.6s of CPU time and 1.6 in total 20:20:43 right 20:20:50 Deewiant, still it is less buggy atm. 20:23:50 Deewiant, btw for make in efunge -j have absolutely no effect 20:24:04 because it actually just runs erl -make 20:24:22 (not for make clean though) 20:28:11 Deewiant, about speed there above btw... Language::Befunge is slow? 20:28:13 I never tried it 20:30:16 -!- ais523 has joined. 20:31:05 hi ais523 20:31:11 hi 20:31:16 and this time, I'm not busy 20:31:26 AnMaster_ipv6: By your standards, definitely. :-P 20:31:31 although I've been awake for 28 hours straight, so may not be coherent 20:31:39 ais523, heh, well I was talking about funge space coherency protocol before 20:31:42 hah 20:33:24 it might solve the issue I had with implementing ATHR, along with defining that bounds may not be exact (even for wrapping purposes, when other threads writes cause bounds to extend) if ATHR is loaded and a sync instruction have not been executed 20:33:35 it will always see it's own writes in a coherent way 20:33:44 how could bounds be wrong for wrapping purposes? 20:33:48 but between processes some stuff updates lazily 20:33:53 ais523, remember ATHR? 20:33:56 AnMaster_ipv6: It takes 12s on Mycology here 20:33:59 oh, if one thread writes to another's Lahey-line 20:34:07 beyond where it would have bounced 20:34:07 So yes, it's by far the slowest 20:34:07 ais523, or anywhere else 20:34:10 and yes I remember ATHR 20:34:15 Deewiant: which implementation? 20:34:26 Deewiant, oh my that is bad 20:34:29 ais523: Language::Befunge, the Perl one. 20:34:40 Deewiant, anyway your computer is way faster 20:34:44 ah, I've been meaning to mess with that 20:34:47 you got much better speed even for efunge 20:34:48 AnMaster_ipv6: Yes, I know. :-P 20:34:53 is it on cpan, or does it need to be downloaded from elsewhere? 20:34:59 It's on CPAN. 20:35:12 * ais523 starts CPAN invocations 20:35:34 $ sudo chown -R ais523:ais523 ~/.cpan 20:35:52 $ cpan -i Language::Befunge 20:35:53 What, you normally have it root:root? :-P 20:36:03 can't you install it per-user? 20:36:08 Deewiant: no, bits keep on ending up root:root by accident anyway 20:36:27 the chown is a bit of black magic for using cpan I developed, it seems to work 20:36:42 I never install software outside home without using my distro's package manager.. 20:36:55 ~/.cpan is not very outside home. 20:36:57 oh, I don't install outside home or /usr/local without package manager 20:37:01 but this is inside home 20:37:23 I don't use /usr/local, well on freebsd the package manager uses that 20:37:26 besides, cpan is a package manager, just a confusing and buggy one 20:37:40 but on linux my /usr/local is just some empty directories 20:37:46 on Debian-based systems, /usr is for the package manager, except /usr/local is guaranteed package-manager free 20:38:05 so it's a good place to test things like my C-INTERCAL packaging 20:38:25 I have some systems-administrationary custom scripts in /usr/local/sbin/ it seems. 20:38:33 ais523, well *bsd is: / and /usr are for base system /usr/local for ports collection (the package manager) 20:38:49 /usr/home contains your home directories 20:38:55 (/home is a symlink) 20:39:11 fizzie, /root/bin for me 20:39:26 $ perldoc Language::Befunge 20:39:28 seems to be working 20:39:37 I had to switch to root and back though as it was trying to install man pages 20:39:44 then I needed another chown to clear up 20:40:36 R13 will have wxErlang hm 20:40:40 interesting 20:40:50 better than the old Tk based thingy at least 20:43:13 "Can't locate UNIVERSAL/require.pm in @INC" 20:43:19 I'm having CPAN problems, as usual 20:44:41 why use cpan then 20:45:16 AnMaster_ipv6: you have to 20:45:20 the alternatives are far worse 20:45:24 hm 20:45:27 i.e., either no perl libraries, or chasing 5 billion dependencies 20:45:31 and yes, there are always that many dependencies 20:45:36 perlists are promiscuous 20:45:42 ehird: even with cpan, you have no perl libraries and are chasing 5 billion dependencies 20:45:55 ais523: except you can get cpan to chase them for you 20:45:56 hm 20:45:57 if you're lucky 20:46:18 usually I just install the relevant part using my distro's package manager 20:46:23 ehird: it usually gets about half 20:46:26 it seems to have lot of perl packages 20:46:26 which is an improvement, I suppose 20:46:31 perl-archive-tar 20:46:32 and what not 20:46:38 grr, one thing that irritates me about jvm conventions: the package convention is com.foo.app, so you _need_ a domain 20:46:40 lots of them certainly 20:46:57 brb → 20:47:00 ehird, you could use org.eso-std 20:47:02 ;P 20:47:06 or com.example. 20:47:07 ← 20:47:09 AnMaster_ipv6: I don't own that domain. 20:47:13 except you can't any more 20:47:14 :P 20:47:22 lifthrasiir: You must use a domain you own. 20:47:23 → 20:48:18 what if java package name can contain URN or so? urn.uuid_6e8bc430_9c3a_11d9_9669_0800200c9a66? 20:48:30 * lifthrasiir gives up thinking 20:48:50 lifthrasiir, they aren't urls to begin with 20:48:54 just reverse domain names 20:48:56 so no 20:49:15 AnMaster_ipv6: of course, but i said "what if" :p 20:50:41 It's not a "need" since it is just a convention. 20:51:14 the convention makes the need to comply with. 20:51:28 .. 20:51:29 (which can be ignored sometimes) 20:51:38 the convention makes the need to comply with. <-- wut? 20:52:05 i mean that widespread convention becomes de facto standard by itself 20:52:12 ah 20:52:21 even though it doesn't need to be so 20:53:13 fizzie, what about relative question rates. That is relative % ending in question marks to all lines said by that person. Combine ehird and tusho btw 20:53:39 Admittedly the convention is documented in the Java Language Specification, so it's a rather strong convention. 20:53:47 "You form a unique package name by first having (or belonging to an organization that has) an Internet domain name, such as sun.com." 20:54:44 yay, Language::Befunge is working 20:54:53 I only had to chase up 7 or 8 of the dependencies by hand 20:55:02 I should probably write some sort of framework for log-analysis. But I guess I can quickly hack that relative-question-percentage for you, if it's that interesting. 20:55:03 ais523, sounds like a crappy package manager 20:55:12 cpan isn't really a package manager, although it acts like one 20:55:18 it doesn't have an uninstall feature, for one 20:55:19 ais523: Is your cpan broken or something? 20:55:20 fizzie, yes very interesting 20:55:40 fizzie, since I talk so much I believe the numbers mentioned yesterday are not really relevant 20:55:44 Deewiant: yes, it tries to write outside my home dir and errors, then on the next run decides either the last one worked so it won't try again, or the last one didn't work so it won't try again 20:56:23 it doesn't have an uninstall feature, for one <-- crappy. I won't use it then certainly 20:56:24 Sounds like you have it configured somehow incorrectly 20:56:25 hmm... which fingerprint is REFC? 20:56:33 One of Cat's Eye's 20:56:38 ais523, the memory leakage one 20:56:41 AnMaster_ipv6: it makes it a bit hard to run Perl programs as a result 20:57:07 ais523, no: gentoo portage have lots of perl packages as gentoo packages that are properly install/uninstallable 20:57:16 the recommended way is to write gentoo packages for any missing 20:57:21 makes sense 20:57:24 there is in fact an automated tool 20:57:32 Ubuntu package the bits of cpan they depend on, but I think they don't do it automatedly 20:57:37 ais523, called g-cpan before, but replaced with something else now 20:57:48 talking about Perlists about the gentoo cpan thing, apparently it doesn't work in every case, or something 20:57:51 forgot details since I never use it 20:57:56 but then, neither does cpan 20:58:11 ah no it is still g-cpan 20:58:19 http://www.gentoo.org/proj/en/perl/g-cpan.xml 20:58:19 cpan's worked just fine for me on Linux 20:58:23 Windows is another story, of course 20:58:35 Deewiant, if it can't uninstall I'm not using it 20:58:41 Your loss 20:58:46 I want to have a easy to maintain system 20:58:52 Deewiant: does Mycology test deep y as a pick instruction? 20:58:56 ais523: Yes 20:59:12 I can't find the bit in the output where it reports on it 20:59:19 ah, found it now 20:59:29 thus every file outside /home /etc /var /usr/portage /root and a few more places has to belong to a package in the distro package manager IMO 20:59:34 that is a rule I use 20:59:55 it's nice to have a second /usr tree for testing things that shouldn't be interfered with by the manager 21:00:04 ais523, why would I need that? 21:00:19 AnMaster_ipv6: what if you're writing a package manager yourself? 21:00:34 or just an install/uninstall script designed to work in a /usr tree as well as in the home dir 21:01:02 anyway, my jqbef98 seems to pass latest mycology 21:01:04 ais523: chroot into $HOME/foo which contains usr? 21:01:25 ais523, well the latter makes no sense, it is up to package manager to handle it, it should be trivial to write a package, if it isn't then the package manager is too hard to use. 21:01:49 ais523: Hmm, where'd you get jqbef98, I just grabbed the example from the perldoc and used that 21:01:50 for gentoo I usually need a 5-10 lines long file, and 5 minutes of work. 21:01:53 hmm... no wait, jqbef98 fails mycouser I think 21:01:57 for arch about the same 21:02:01 Deewiant: from cpan, it installed as part of the package 21:02:02 but more like 15 lines there iirc 21:02:10 ais523: Didn't for me. Oh well. 21:03:04 AnMaster_ipv6: Here is top-20 of question-%s for everyone who's spoken more than 1000 lines during 2006-2008. (Without the filtering there's quite an amount of 100%s.) → http://zem.fi/~fis/q.txt 21:03:42 Heh, EgoBot. 21:03:48 see 21:03:51 Huh? 21:03:54 it's not so bad for me at all 21:04:07 This time it was "contains a ? anywhere". 21:04:12 fizzie, where did ehird end up 21:04:34 AnMaster_ipv6: Number 44, with 3.2 %. 21:04:38 meh 21:04:53 AnMaster_ipv6: He talks even more than you, after all. 21:04:57 true 21:05:15 "With Jergen's, who needs the sun?" 21:05:21 Sometimes commercials are immeasurably stupid. 21:05:48 AnMaster_ipv6: you're cheating, you're asking questions without a question mark at the end 21:05:49 Sometimes? 21:06:02 ais523, yes I started that recently 21:06:08 err 21:06:13 actually, I often typo ? as /, so I'm cheating just as much 21:06:13 AnMaster_ipv6: That wasn't a question. 21:06:33 s///;s/yes/yes 21:06:34 there 21:06:34 ;P 21:06:56 /^s/s/$/\// 21:07:10 err... that has too many forward slashes 21:07:15 shouldn't some of them be escaped? 21:07:19 ais523, no 21:07:24 ais523, that is quite different... 21:07:29 it is matching line first 21:07:34 to the one starting with ^s 21:07:42 oh, I see 21:07:47 /^blah/s|foo|bar 21:07:52 that's sed not Perl 21:07:54 and you must use forward slashes there 21:07:57 ais523, well duh 21:07:59 I always use sed 21:08:01 never perl 21:08:07 and in Thutu, it would be /^s/$/=\// 21:08:08 so that's like... expected 21:08:37 ais523: how does the power of thutu compare to that of sed? 21:08:39 ;ais523, why did you think I didn't use sed;337**,@ 21:08:52 SimonRC: they're both turing-complete 21:09:27 thutu needs to be wimpmoded to really be useful, although it's a relatively nice language, I should detarpit it and remove some of the insanities 21:09:31 I mean expressive power 21:09:32 ais523, still, why perl(3 3 7 * * p) 21:09:36 and I may end up with a language that's worth using 21:09:41 ah, ok 21:09:43 why did you even assume I would use perl... 21:09:45 ais523, ! 21:09:53 AnMaster_ipv6: Perl sets the standard for regular expressoins 21:10:01 anyway, I'm going home now, bye everyone 21:10:04 -!- ais523 has quit (Remote closed the connection). 21:10:07 ... 21:10:21 I didn't use pcre, I used s syntax, which means it is sed when it comes to me... 21:10:37 "s syntax" is present in Perl. 21:10:38 ditto 21:10:48 Deewiant, yes it is, but when I do it it is sed 21:21:53 once perl added (?...) extension syntax to regexp, and perl 5.10 adds (*...) "verb". i wonder perl 5 will ever add (+...). 21:22:34 lifthrasiir, um? 21:22:47 isn't (?) for ( or something 21:23:01 I forgot syntax of lookaround 21:23:16 (?=...), (?!...), (?<=...), (? ah 21:23:28 err 21:23:30 that seems wrong 21:23:32 and even there is unrelated (?>...) 21:23:35 ?= ? 21:23:52 I remember 4 variants in PCRE 21:24:05 (negative|positive) look(behind|ahead) 21:24:36 < is for lookbehind, = is for positive and ! is for negative. 21:24:45 ah 21:24:53 what is the > one then 21:24:57 and (?>...) is for atomic match 21:25:06 atomic match... 21:25:15 i.e. do not backtrack into this parenthese 21:25:16 s* 21:25:16 as in, no concurrency issues 21:25:19 ah 21:25:58 once Larry Wall criticizes this as a bad huffman coding... :p 21:26:27 what 21:27:12 well, perl 6 has a new regexp syntax which shortens such constructs which is used quite often but too long to type 21:27:47 but i still think perl 6 showed new dimension of esoteric language... heh 21:28:22 * lifthrasiir away, finally 21:30:54 20:56 AnMaster_ipv6: it doesn't have an uninstall feature, for one <-- crappy. I won't use it then certainly 21:30:59 CPANPLUS has uninstall 21:31:07 20:55 fizzie: I should probably write some sort of framework for log-analysis. But I guess I can quickly hack that relative-question-percentage for you, if it's that interesting. 21:31:11 Log analysis framework? Count me in. 21:31:24 Gah, ais is gone. 21:31:29 I was hoping for some micro-optimization help. 21:31:40 Wouldn't AnMaster be the one to ask? :-P 21:31:45 also, 28% of sgeo is quetsions 21:31:46 I knew it 21:31:54 I predicted a large % all along 21:31:55 since like last year 21:32:11 Deewiant: no, ais523's microoptimizations give a better speed boost and are less obnoxiously _posix_fuseless 21:32:15 IME 21:32:16 Heh 21:32:24 CPANPLUS has uninstall <-- interesting 21:32:42 ehird: Can cpanplus uninstall stuff cpan has installed? 21:32:48 yes 21:32:53 i think 21:33:13 http://blogs.discovermagazine.com/80beats/2009/04/03/sleep-may-prepare-you-for-tomorrow-by-dissolving-todays-neural-connections/ ← Humans need concurrent gc 21:33:39 ehird: What are you micro-optimizing? 21:33:41 ehird, I may still be able to help microoptimising 21:33:59 Deewiant: I haven't written it yet, but it'll only be ~100 lines -- a tripcode (http://en.wikipedia.org/wiki/Tripcode) cracker 21:34:03 well, bruteforcer 21:34:16 there's a few but I have some microoptimization ideas & I want to take advantage of multi-core systems 21:34:42 ("For example, 2channel translates <, >, and " to <, >, and "." ← I love perverse edge-case programming errors becoming defacto standards) 21:35:33 2channel < wat 21:35:47 Slereah_: Wat is greater than 2channel? 21:35:54 Oh, it's a discussion about tripcode 21:36:08 It seemed weird to see 2ch in that window 21:37:17 Really, most of the time, you don't even need a tripcode cracker 21:37:26 Most people use insecure tripcodes 21:37:45 There's a list of all letter-based tripcodes 21:37:46 Yes, well, it's a fun thing to optimize. 21:37:57 So you just have to google it 21:38:27 Not to mention the very common tripcodes :o 21:38:38 I wonder what the character set of crypt() output is 21:38:49 I don't think it's all printable ascii chars, there's no space for instance 21:39:13 ehird, simple, use the SSE8 PWCJIS to convert to JIS to begin with 21:39:15 ;P 21:39:27 It's probably something like ASCII from 33 to 126 21:39:39 AnMaster_ipv6: SHIFT jis. And the article then goes on to say that 4chan doesn't do the shift-jis conversion, so it's probably not worth bothering. 21:39:43 oh 21:39:51 hm ok 21:40:01 Hooray for de facto standards. 21:40:07 ehird, anyway SSE is so feature bloated in recent versions it wouldn't surprise me.. 21:40:10 If you want to learn about tripcodes, you can go to /soc/ 21:40:19 Slereah_: I used to frequent there ages ago 21:40:26 Okay, so there's 48398230717929318249 unique tripcode outputs. I think. 21:40:48 Now to figure out what -illion that is. 21:41:16 Um, alotillion. 21:41:28 48.398.230.717.929.318.249 -> 48 quintillions 21:41:31 48 398 230 717 929 318 249 so... yeah 21:41:45 Assuming you're US, of course. 21:41:52 Yeah 21:41:58 Or modern UK, or whatever crap. 21:42:03 It's 48 trillions otherwise 21:42:05 Trillions. 21:42:07 Yes. 21:42:07 the us system is retarded 21:42:21 oklopol: it's the world wide system now 21:42:27 -!- Asztal_ has quit (Connection timed out). 21:42:28 quintillions. lovely 21:42:28 Only in English. 21:42:32 Is not 21:42:34 Yeah 21:42:35 Well, and other languages. 21:42:38 But not all. 21:42:43 well duh 21:42:44 well Swedish doesn't use the US system there 21:42:51 ehird: so is english, and it's retarded too. 21:43:13 According to Wikipedia, besides English-speaking countries only Brazil and Wales use the short scale 21:43:21 Hmm, I wonder what the best way to utilize multiple cores is 21:43:29 maybe start thread 1 at 0, thread 2 at (half of tripspace) 21:43:29 etc 21:43:34 And 9 others use the short scale + milliard 21:43:36 ehird: todo what? 21:43:37 what do you call 1,000,000? 21:43:40 And greece uses myriad 21:43:41 AnMaster_ipv6: million 21:43:42 AnMaster_ipv6: Million. 21:43:44 hm 21:43:46 what do US call it? 21:43:50 Deewiant: your facts are nicer than ehird'sa 21:43:51 *ehird's 21:43:52 million too 21:43:56 SimonRC: Crack http://en.wikipedia.org/wiki/Tripcode s. Learnt to read :P 21:43:56 ah yes it breaks above 21:43:57 *Learn 21:44:04 what is the next step above million 21:44:08 billion 21:44:10 miljard in Swedish 21:44:12 oklopol: Thanks, I guess. 21:44:13 milliard :o 21:44:20 Slereah_: Nobody says milliard. 21:44:22 : 21:44:22 :P 21:44:24 miljardi 21:44:28 Million, milliard, billion, billiard, trillion, trilliard, etc. 21:44:31 10^9 = billion 21:44:32 ehird: I actually do. 21:44:34 Deewiant, agreed 21:44:39 billiard? 21:44:40 Deewiant: You don't count as a person :| 21:44:40 Deewiant: where are you from? 21:44:44 that's used? 21:44:45 SimonRC: Finland. 21:44:50 oklopol: Sure, it goes all the way up. 21:45:01 vigintilliard 21:45:24 SimonRC: keyspace is 72057594037927936 actually 21:45:25 sez wp 21:45:35 oklopol, well vigintillion already sounds silly so meh... 21:45:35 which is a fucktonion 21:45:40 Vigintilliard would be 10^123 21:46:52 unde- and duode- are used to get the ones before it, but i'm a bit fuzzy on the details, not having any idea where those come from. 21:47:12 and is nonillion 10^6^9? 21:48:35 In the long scale, yes 21:49:12 -!- Judofyr has joined. 21:49:16 well yes the sensible scale, naturally 21:49:20 btw 21:49:27 it was a question, where are the names from 21:49:28 latin? 21:50:01 Obviously, yes 21:50:22 why is it obvious? 21:50:58 YOU KNOW COULD'VE BEEN GREEK... OR JAPANESE 21:51:01 "Take the second and third characters of the string obtained by appending H.. to the end of the input. " 21:51:05 I love how meaningless this is. 21:51:07 It's a clusterfuck 21:51:22 oklopol: Except that it isn't. :-P 21:52:04 E.g. decillion = 10^6^10 where the latter 10 is from Latin decem, 10. 21:52:21 well i don't know latin, so a bit hard to know; i guess i have to admit it was the only possible choice tho 21:52:29 ehird, the H.. is for handling very short strings I guess 21:52:32 that could explain it 21:52:35 AnMaster_ipv6: yes, but why 'H..' 21:52:37 it's hilarious 21:52:45 i guessed from oktaavi -> nooni -> desimi 21:52:47 ehird, agreed. 21:52:50 it is still stupid 21:52:53 WACKY JAPS 21:53:05 (octave, octave+1 -interval, octave+2 -interval) 21:53:24 also could've guessed from simply deci i guess 21:53:28 cuz it's like 1+ 21:53:29 *10 21:53:32 BUT I DIDN'T 21:53:48 ehird, maybe tripecode then 21:54:38 ... 21:54:44 -!- M0ny has quit ("PEW PEW"). 21:55:08 ehird, prod 21:55:21 what 21:55:26 oklopol: I'd say the most obvious is that it's quadrillion and not tetrillion or something. 21:55:31 ehird, it should have been called tripecode 21:55:34 why 21:55:41 http://wordnetweb.princeton.edu/perl/webwn?s=tripe 21:55:43 second sense 21:55:45 .. 21:55:59 It shouldn't have been called that because that's not funny. 21:56:00 It is called $Bec0e&,ec&e%w(B :3 21:56:00 Tripe ration 21:56:03 Next. 21:56:04 Deewiant: you mean it's obvious it is the former, or that based on that, latin is obvious? 21:56:17 i don't see why quadrillion is more obvious than tetrillion 21:56:20 oklopol: Based on that, Latin is obvious. 21:56:21 Deewiant: I tried to figure some sort of oerjan-class pun related to tripecode and the % character, but couldn't. 21:56:24 right. 21:56:33 not to me, i don't know latin. 21:56:37 fizzie: :-) 21:56:58 where's tetri- from? 21:57:37 oklopol: Greek tetra-, tetr-: the number four as a prefix 21:57:48 not japanese? 21:57:48 AnMaster_ipv6: Given that it's in a file, I couldn't really justify the "head -n 20" in q.txt, so http://zem.fi/~fis/q.txt now has the complete list of everyone who's spoken >1000 lines in those logs. 21:58:07 oklopol: No, four in Japanese is shi or yon. 21:58:18 fizzie: YOU FORGOT TO MERGE ZUFF AND EHIRD 21:58:20 WAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAH 21:58:24 fizzie, for bsmnt you should count "..." as a qustion mark 21:58:32 AnMaster_ipv6: you too. 21:58:33 '..' 21:58:34 '..' 21:58:36 etc 21:58:37 fizzie, no need to do that 21:58:41 it is fine as it is 21:58:54 it shows ehird's split personality 21:58:58 or something 21:59:24 It's a bit crappy in other ways too. There's also oklopol/oklofok, curiously next to each other; very consistent question-percentage there. 21:59:34 heh 21:59:34 Same goes for Pikhq/pikhq. 21:59:50 clearly we need the Loggitude framework 22:00:12 yay, i'm #23 22:00:16 I had somewhere a log-parsing script that combinated nicks based on hostmasks, nick similarity wrt. edit-distance, and some other random heuristics, but I seem to have lost it. 22:00:22 ehird, no... the log ness framework you mean 22:00:28 So fungot /does/ ask questions. 22:00:28 Deewiant: 18. some hard-boiled things can be cracked." he unfolded the paper as he spoke, ' is the fnord" ( making little marks on the ground as she spoke, and rested his head against her shoulder. " what size will you be?" 22:00:49 Hah, that's a question. 22:00:57 lament: you should put that on a t-shirt 22:01:19 $ loggitude -e'map { my @lines = search(-from => $_); my @questions = filter { /?$/ } @lines; something to output percentage here } @users' 22:01:23 You get the idea. 22:01:33 ehird, loggness 22:01:34 ... 22:01:36 no 22:01:38 loggitude 22:01:40 no! 22:01:58 why would loggitude be better than loggness 22:02:06 loggness is a really bad pun after all 22:02:11 so is loggitude 22:02:12 while loggitude sounds enterprisy 22:02:14 logs, attitude 22:02:17 loggness is 8.3 22:02:19 loggitude. 22:02:20 So it gets my vote 22:02:25 logitud 22:02:26 Deewiant, :D 22:02:30 logitude even 22:02:39 anyway I'm writing it so fu 22:02:40 :P 22:02:49 ehird, well if you are actually writing it 22:03:10 sure, I just need to get my log-scraper done 22:03:21 then anyone can analyze #esoteric logs with two invocations: 22:03:33 $ perl download-esoteric-logs.pl 22:03:39 $ perl loggitude.pl -e'...' 22:03:54 (download-esoteric-logs.pl will update your logs if you already have some, btw, it only downloads what you need) 22:03:59 that would be $ logness-fetch esoteric 22:04:01 or something 22:04:05 (also rename to YYYY-MM-DD and changes times to UTC) 22:04:16 or even logness-sync 22:04:17 (I have the renaming part done and some of the downloading) 22:04:31 loggness* 22:05:15 or for Deewiant loggness fetch esoteric 22:05:52 Separate scripts for separate tasks. 22:07:00 Yay, I have gotten recomputed 34/39ths of the Octave stuff I messed up last night. 22:07:29 fizzie how come all you ever do is awesome computing processing stuff 22:07:35 i'm a kid and I don't have that much fun :< 22:07:56 ehird, git has all in one 22:08:05 yet you like git 22:08:06 AnMaster_ipv6: It didn't used to. 22:08:08 Used to be git-foo. 22:08:13 Anyway, all git operations are linked 22:08:22 They all manipulate the same datastore in much the same ways 22:08:31 But downloading logs vs processing logs are entirely separable tasks 22:08:52 AnMaster_ipv6: It didn't used to. <-- used or use 22:08:59 s/used/use/ 22:09:03 It's the reporting bias; I do not much advertise the times when I clean up the sewage input pipe in the bathroom, or other pleasant things like that. 22:09:03 And if I required strict adherence to my philosophies I'd never use anything. 22:09:10 ah 22:09:14 fizzie: do it, so I can feel better about myself 22:09:41 Will try to remember to. 22:10:16 Fuck the JVM. Fuck Parrot. I'm compiling all my VM-langs to Infocom's Z-machine in future: http://en.wikipedia.org/wiki/Z-machine 22:10:23 -!- kar8nga has quit (Read error: 145 (Connection timed out)). 22:10:38 -!- neldoreth has quit (Read error: 104 (Connection reset by peer)). 22:10:38 :-D 22:10:44 -!- neldoreth has joined. 22:10:52 ZBefunge was probably the first Befunge interp I used, actually. 22:11:01 wait 22:11:08 zbefunge compiles befunge to z machine code? 22:11:09 It's quite crap 22:11:10 ok, that's hot 22:11:20 No, it's just an interpreter written to work in a Z machine. 22:11:23 No, it's a befunge interpreter written in Z-Machine code 22:11:29 oh 22:11:31 still 22:11:31 Also Befunge-93 only and so on. 22:12:10 Seems that there's a ZedFunge which does Funge-98. 22:12:29 Deewiant: New mycology contender :P 22:12:42 when I heard zbefunge I first thought "zsh" 22:12:45 not "z-machine" 22:12:52 That's the crap one IIRC 22:13:03 I may misremember. 22:13:18 At least ZBefunge's debugger had a nice text-based view of where the IP was going. 22:13:40 That's the crap one IIRC <-- thought that was either RC/Funge or FBBI 22:13:55 RC/Funge is crap? 22:13:57 There are other crap ones. 22:14:02 I know you dislike the code, but it works, doesn't it? 22:14:02 It's somewhat crap. 22:14:05 It passes Mycology? 22:14:12 It uses static buffers and gets() 22:14:13 I think comparing RC/Funge to FBBI is giving MKRY a great disservice 22:14:16 It'd be trivial to make it segfault. 22:14:45 It's good as long as you don't feed it too much data :-P 22:14:52 -!- tombom has quit (Connection timed out). 22:15:22 a world wide network of toaster 22:15:23 s 22:16:24 Deewiant, seems ehird is a RCS fanboy for unknown reason... 22:16:35 I'm not a fanboy. 22:16:37 Er, no. 22:16:48 The code sucks. I agree. 22:17:05 I just don't like how you continually say that RC/Funge sucks and MKRY can't implement befunge properly etc etc 22:17:14 Why not invite him in and say that? 22:17:42 well he should come on irc himself 22:17:55 For what? 22:17:56 to be insulted? 22:18:14 if he is interested he would be in here 22:18:20 in befunge in general I mean 22:18:24 Deewiant, add some tests to mycology that make rc/funge segfault? 22:18:31 22:18 AnMaster_ipv6: if he is interested he would be in here 22:18:31 22:18 AnMaster_ipv6: in befunge in general I mean 22:18:32 bullshit 22:18:44 he'd be in here if he's interested in talking about the topics we talk about in here with the people who are in here at the times he would come in here 22:18:57 AnMaster_ipv6: He might just increase the buffer sizes ;-P 22:19:06 Not worth it anyway 22:19:10 Deewiant, yeah would be his style 22:19:19 ugh static buggers 22:19:54 (typo) 22:20:24 oklopol, I thought it was intentional oko.. 22:20:35 it probably was. 22:20:48 well mostly it was a typo. 22:20:54 I don't see a typo 22:21:01 static buffers are against my face. 22:21:08 Deewiant: well there was one 22:21:15 Oh, you meant to say buffers 22:21:20 something like that maybe 22:21:22 The sentence is perfectly parseable as-is 22:21:31 ohh 22:21:37 buggers, like from bug 22:21:38 hah 22:21:42 that's actually pretty good 22:21:53 Or just, you know, http://dictionary.reference.com/browse/bugger?jss=0 22:22:01 i thought like, you know, buggers 22:22:01 "an annoying or troublesome thing" 22:22:06 yes 22:22:17 but with that meaning i'm not sure it made that much sense 22:22:27 "Damn those static things" 22:22:28 I'LL BUGGER YOU IF YOU KNOW WHAT I MEAN 22:22:30 LOL OLOL LOL 22:22:31 OWTTE. 22:22:36 But the hedgehog can never be buggered at all. 22:22:43 Deewiant: yeah well okay sure 22:22:54 i guess i have to agree my fingers outsmarted me again. 22:23:41 Deewiant, OWTTE? 22:23:54 http://www.justfuckinggoogleit.com/?q=owtte 22:24:52 Deewiant: what gfx card did you say you had again 22:24:59 ehird: Radeon HD4870. 22:25:06 yar. 22:25:23 The lack of comment makes me wonder if AnMaster_ipv6 got the Pratchett reference. Although I guess it's not very commentable. 22:25:24 I'm trying to get the best passively cooled one I can from the limited choice because I can't be fucked to mess around with coolers and shit 22:26:12 I have in these machines a passively cooled GeForce 7600-something and a 8600-something, but I am in no way recommending these, and it's very much not bleeding-edge high-performance 3D-powerfest. 22:26:14 I don't know ~anything about passively cooled ones 22:26:39 Yes 22:26:44 I was just figuring out what yours was 22:26:46 So that I can exclude it 22:26:50 From my lookings 22:27:20 Looks like it'll be 512 MB ATI Radeon 4550 PCI-Express x16 GDDR3 22:27:33 Any gamers in here are welcome to take a few second break to laugh at me heartily for saying that 22:27:43 "Such a wimpy thing! Couldn't even play Wolfenstein 3D!" 22:27:56 ehird: That's worse than my previous graphics card 22:28:07 Deewiant: :-D What was your previous one? 22:28:12 A Radeon X800XL 22:28:23 and I assume your opinion of it is "crap" 22:28:34 Well, for certain values of "crap" 22:28:58 I managed to play through Fallout 3 with that one 22:29:11 Deewiant: The most graphically intensive game I've ever played is http://en.wikipedia.org/wiki/Tom_Clancy's_Splinter_Cell_(video_game) 22:29:20 So I don't think it matters which card I pick, really. 22:29:31 That didn't run well on the card before my X800XL 22:29:38 Which was... 22:29:44 GeForce 3 Ti200, that's right 22:29:56 Deewiant: But the card before your X800XL is worse than a 512mb Radeon 4550, right? 22:30:05 My current card: 22:30:08 ATI Radeon X1600: 22:30:08 Chipset Model:ATY,RadeonX1600 22:30:08 ehird: Yes, it is. 22:30:09 Type:Display 22:30:12 Er, was. 22:30:12 Bus:PCIe 22:30:13 Whatever. 22:30:14 PCIe Lane Width:x16 22:30:16 VRAM (Total):128 MB 22:30:23 So, my X1600 is better than the 4550? Heh. 22:30:34 ehird: About the same, I think. 22:30:39 Ah. 22:30:50 I'd get a non-passively-cooled one if I wasn't worried about it whirring all the time. 22:30:59 I guess it's just the high-end cards that do that, but I don't really like risking things. 22:31:02 http://www.jathardware.com/2/video.html (warning: Finnish) sorts GPUs by power, approximately 22:31:13 what? 22:31:16 * AnMaster_ipv6 was afk 22:31:32 I'll start saying 'what?' every time I come to the computer too 22:31:41 Deewiant: Well, whatever I get has to be less powerful than yours since you say it whirrs even when idle 22:32:01 ehird: It has to be anyway, since I have the next-most powerful card AMD currently sells. 22:32:07 The most powerful being the X2 version. 22:32:07 heh 22:32:11 Deewiant, what to The lack of comment makes me wonder if AnMaster_ipv6 got the Pratchett reference. Although I guess it's not very commentable. 22:32:12 ... 22:32:16 So... {512MB,1GB} 4850, 512mb {4830,or 4550} 22:32:44 The small selection I have to pick from is because apparently they assume everyone buying their top performance product is a gamer. 22:32:44 I think my passively-cooled one is a 8600 GT, which is there between GeForce 7800 GS and Radeon HD 2900 GT. 22:33:07 AnMaster_ipv6: It was a hedgehog song reference, before your owtte query. 22:33:12 I _could_ get the lowest I can and replace it with one I buy myself, but fuck, I don't want to do more than switch the power supply :-P 22:34:59 Phew, the relief is palpable: the scripts I've written for that AI tournament university course thing I've been talking about every now and then seem to be working. 22:35:04 http://www.silentpcreview.com/ ← Huray, someone else is as obsessive as me!! 22:35:12 "* Anechoic Test Chamber" 22:35:14 Er, make that more obsessive. 22:35:39 No, just more scientific. :-P 22:35:47 Oh, I thought they meant like 22:35:52 They actually use them in one of those 22:35:55 Which would be lol 22:36:00 :-D 22:36:15 There are other reviewers too that are very silently-oriented, but I've relied a lot on silentpcreview too. 22:36:29 fizzie: what tournament? tell everything you've told sofar. 22:36:40 I just love the sound of an idle room, really. 22:37:15 -!- Slereah has joined. 22:37:26 I guess the silent pc guys watercool everything 22:37:38 oklopol: Just read http://wiki.tkk.fi/display/T934400/Home if you can't wait; have to go away for 15 minutes or so. 22:38:04 i'm not in a hurry, have tons to read and exam tomorrow 22:38:11 (wait actually 10 pages left) 22:38:22 fizzie: see? your life is unbelievably fun. 22:38:27 this is solid proof. 22:38:54 fizzie: Gah, why Java 22:39:09 Well, the answer is 'because the professor said so' 22:39:10 but i will probably do one more quick reread, i confused a detail about lzj and lzmw last time 22:39:11 But still. 22:39:19 One day we'll invent completely silent systems that make no noise ever. 22:39:25 Even when you use up 100% resources. 22:39:30 * 22:39:31 *lie 22:40:10 I guess what I wanted to ask was 'is there any particular reason they specify Java' 22:40:51 our profs say something like "it's used a lot in companies", usually 22:40:57 i can't say i see the relevance of that 22:41:16 or "university policy" 22:41:24 oklopol: they're trying to teach you to go work at boringbigcorp. 22:41:53 oklopol: That's a reason to teach it, not to force its use in a course about concepts not tied to any language 22:42:07 Deewiant: hmm true. 22:42:28 my AI courses allowed any language 22:42:41 I think I might know why I hate loud computers. 22:42:55 I think I probably have tinnitus, so silence = faint whining 22:43:02 So quiet noise + faint whining = annoying 22:43:41 Deewiant: i think the actual reason is those profs who aren't actually programmers at heart don't really know anything but java, since it's the official lang; hard to accept other languages 22:44:05 Those profs aren't going to actually read the students code, are they? 22:44:10 +' 22:44:39 STOP COUNTERING MY ARGUMENTS 22:44:53 :-P 22:45:13 i don't really know, maybe you can guess next 22:45:38 The TAs might read the code but I don't see any good reason for that, in turn 22:45:55 Deewiant: Actually it's "because students said so". 22:46:05 Deewiant: It was still Scheme two years ago. 22:46:07 reason for what, *ta's* reading code, or *reading code*? 22:46:15 fizzie: students dislike scheme? 22:46:16 fizzie: Why force any particular language? 22:46:17 How shit. 22:46:21 ehird: Of course, they're idiots on average. 22:46:27 >:( 22:46:30 Deewiant: And just about everyone complained about Scheme, since it's no longer used for teaching. 22:46:43 ehird: But it also helps that it's not taught. 22:46:58 Bah. 22:46:59 -!- Slereah_ has quit (Read error: 110 (Connection timed out)). 22:47:00 Deewiant: Because the point is to provide a system for the students so that they can focus on the presumably-AI-related-parts of the assignment instead of generic code-writing. 22:47:04 Give them a fucking copy of the little schemer. 22:47:32 fizzie: So, in fact, any language would be okay as long as it can hook into what you provide? 22:48:31 Deewiant: Well, yes, if it can compile to JVM bytecode so that it works in the tournament system (which needs a sandboxy thing and so on), it's okay. We had someone use Scala-or-whatsitcalled last year. 22:48:46 Scala is nice. 22:48:50 Alright, then that page is just misleading. 22:48:59 It's like Haskell + Ruby + Java. 22:49:01 I might do the course next year and learn Clojure while I'm at it. 22:49:06 Ew 22:49:33 ehird: i doubt most students are at a level where they can learn a new language by themselves as they go 22:49:34 Deewiant: Assuming it still exists next year. Markku (the professor responsible for it) is retiring, I think. At least he said something like that. 22:49:52 ehird: Scala seems somewhat lame to me, just a mishmash of some stuff but nothing really cool 22:49:58 Maybe I haven't looked into it enough? 22:50:03 fizzie: Markku who? 22:50:03 Deewiant: It looked like that to me until today 22:50:07 I dug into it 22:50:08 Deewiant: Markku Syrjänen. 22:50:10 It's really quite nice 22:50:19 It is a mishmash, but it's a smooth mishmash 22:50:28 There's not like different segments of piled on features like D 22:50:29 wait was Deewiant at helsinki too 22:50:48 Not at Helsinki University, no. 22:50:54 In Helsinki, yes. :-P 22:51:04 err okay, where then? 22:51:09 TKK+HSE. 22:51:11 Deewiant: Its syntax is very lenient, very DSL-y, scripting code is just concise and simple like you'd expect, java libraries don't feel out of place, and its functional features are pretty much on par with haskell's, sans lazy evaluation 22:51:12 are you a bartender 22:51:20 Deewiant: Also -- it has really good concurrency, Actors and the like 22:51:20 Nope. :-P 22:51:35 Deewiant: is that some kinda university 22:51:43 oklopol: No, that's two. 22:52:01 Deewiant: Anyway, the page is "misleading" because non-Java languages aren't officially supported by the course personnel (i.e. me), so I don't go out of my way to advertise that possibility, just to keep things simple. 22:52:02 okay, i don't want to know the gory details 22:52:06 -!- Judofyr has quit (Read error: 60 (Operation timed out)). 22:52:17 Deewiant: The only thing you can't do too well is scripts that run for pretty much 0 time and are called really often (of which there aren't many *at all*), since the JVM startup time penalty kicks in 22:52:23 but, err, for some reason i thought you lived in tampere, which is kinda weird, i usually remember this stuff 22:52:26 fizzie: Meh, I'd just add a disclaimer. 22:52:39 oklopol: Do you? 22:52:52 no turku 22:53:11 ehird: Yes, I find the JVM somewhat repellent for that reason. 22:53:22 -!- Judofyr has joined. 22:53:31 Deewiant: How many scripts do you write that (1) finish instantly and (2) are called really often? 22:53:41 I can't really think of any apart from things that you should do e.g. as shell aliases 22:53:44 Deewiant: are you like a second year student? 22:53:56 Deewiant: Even a little bit of processing will offset the jvm cost 22:53:59 oklopol: Yep, for some values of 'like' 22:54:17 ehird: Sure, but it's still noticeable compared to straight-to-executable languages. 22:54:37 I mean, most things I write are the type that process for less than 1 second. 22:54:44 Deewiant: i guess the value i was going for was null. 22:54:59 Deewiant: The JVM startup time is just like 0.1s 22:55:02 Deewiant: Well, you're such a pedant. Personally I can't say I see a reason to say "Java (or any other language that can compile Java VM bytecode, but this option is not officially supported by course personnel)" when I can simply say "Java" and have the one people who wants to do anything "special" ask me in IRC about it. 22:55:19 Deewiant: I mean, even a script that processes for 0.1s almost offsets the JVM startup cost 22:55:31 fizzie: I just believe in not hiding options. :-P 22:55:50 ehird: Yes, it runs in 0.2s then, so only 50% of the time is startup. 22:56:13 Deewiant: Yes, but when it comes down to it, a 0.2s script does _not_ feel 50% more responsive than a 0.1s script 22:56:16 Humans aren't that simple 22:56:29 AnMaster_ipv6: It was a hedgehog song reference, before your owtte query. <-- didn't see the line. 22:56:30 It does if you run it 5 times. 22:56:34 that's why 22:56:41 and yes I was afk again 22:57:05 Deewiant: So the majority of programs you write are scripts that take 0.1s or less to run and are called many times in quick succession? 22:57:19 Deewiant: I really think that's quite a minority; and most of those "programs" could probably be done as shell functions. 22:57:25 I agree JVM startup time could be better 22:57:30 but I don't think it's that much of an issue. 22:57:42 ehird: The majority of programs I've written are either quick scripts or esolang interpreters, which tend to be tested in that fashion. 22:57:56 Deewiant, same here 22:58:09 quick scripts: Usually take more than 0.1s to execute. 0.5-1s, I'd say, on average. Which offset the JVM startup time just fine. 22:58:15 I hardly ever wrote long running programs apart from some irc bots 22:58:17 Besides that I've written libraries. 22:58:23 Also, since the JVM itself is blazing fast, post-startup, what would take 1s in Perl could take 0.7s in the JVM 22:58:30 And the remaining stuff I've written is in the minority. 22:58:38 Esolang interpreters: Your test cases are too trivial. 22:58:40 ehird: And 0.3s in D or whatever. 22:58:45 ehird: Mycology? :-P 22:58:55 Deewiant: http://kano.net/javabench/ 22:59:01 (↑ kidding around) 22:59:05 Deewiant, yeah, mycology is a quick script 22:59:18 around 0.080 seconds when not redirecting output 22:59:22 Deewiant: There's no way you can get Mycology to go in less than 0.5s unless you're microoptimizing C like AnMaster_ipv6. 22:59:22 so yes a quick script 22:59:28 :P 22:59:34 ehird, jitfunge could do it too 22:59:38 ehird: Depends on the machine: times make no difference in isolation. 22:59:47 CCBI with fingerprints on and output to terminal runs in 0.3s here. 22:59:48 and yes what Deewiant said 22:59:54 Deewiant: 0.4s vs 0.3s 23:00:01 Hmm, why did I say 'difference' instead of 'sense' 23:00:02 I just tested 23:00:03 sleep 0.4 23:00:04 sleep 0.3 23:00:08 I could barely tell the difference 23:00:11 Deewiant, ccbi takes about 1 second here for mycology when redirecting stdout to /dev/null 23:00:17 iirc 23:00:20 does mycology have loops? 23:00:23 I'd say 0.4 is like 3% less responsive feeling than 0.3 23:00:23 oklopol, yes 23:00:28 lots of them even 23:00:30 And I doubt you could tell unless you really, really wanted to 23:00:32 ehird: Doing it repeatedly makes the 0.3 seem noticeably faster. 23:00:35 i mean 23:00:41 But that's just my brain exaggerating the difference. 23:00:42 what in the world could it have that takes long? 23:00:46 But then, my brain is the one who cares. :-P 23:01:02 oklopol: The Perl interpreter runs it in 12 seconds. 23:01:19 Deewiant: Yes, well, for a megafast befunge interpreter that can run mycology in 0.3 that you wish to run it on again and again without any delay... don't use the JVM. 23:01:19 what's the bottleneck? 23:01:27 oklopol: Everything, it's just turtle-slow. :-P 23:01:35 oh, right 23:01:39 But I would imagine after running mycology the next step would be either to fix the source to progress further, or to stop running mycology/ 23:01:42 *mycology. 23:02:07 oklopol, for cfunge the bottleneck is sting pushing of environment in y when Deewiant wants a delay in the test of HRTI 23:02:23 * AnMaster_ipv6 turns of HRTI to compare 23:04:27 sting pushing? 23:04:33 string* 23:04:42 err right' 23:04:44 *-' 23:04:48 0.080s when outputting to stdout with HRTI.. 23:04:57 assumed that was some kinda befunge term 23:05:03 0.051s with HRTI, to /dev/null 23:05:31 hrti? 23:05:54 0.65 without HRTI to stdout 23:06:08 0.40 without HRTI to /dev/null 23:06:19 still the env is large 23:06:25 so reducing that reduces it even moer 23:06:27 more* 23:09:25 err 23:09:34 0.65 without HRTI to stdout <-- meant 0.065 23:09:43 and 0.040 for the /dev/null one 23:10:07 and clean environment it is 0.035 or so 23:10:10 to /dev/null 23:10:24 oklopol, HRTI = High Resolution Timer Interface iirc 23:10:37 provides a microsecond timer to funge 23:10:52 of course I actually get 1-2 ns resolution really 23:10:54 but meh 23:11:01 Deewiant, irritating isn't it 23:11:08 that you can't show this 23:11:16 Just make a fingerprint that does, if you care 23:11:23 Deewiant, meh... 23:12:10 Deewiant, it would be properly designed then with implementation defined units, so you used some instruction to get "ticks per second" from interpreter 23:14:23 oooooooooooooooooooooooooooooooooo 23:14:25 ooooooooooooooooooooooooooooooooooooooo 23:14:27 oooooooooooooooooooooooooooooooooooooooo 23:14:54 School in 9 hours; I shall sleep now. 23:15:07 exam in 7.5 hours 23:15:13 i shall watch south park now 23:15:14 :-P 23:15:24 i will ace it anyway :| 23:15:36 I just enjoy sleeping 23:15:39 More than South Park, anyway. 23:15:48 me too, when i'm tired. 23:16:01 I'm happy to sleep even when not really tired. 23:16:14 But I am a bit tired now, so I shall sleep. 23:16:16 -> 23:16:21 have fun. 23:16:57 -!- Judofyr has quit (Remote closed the connection). 23:17:08 the actual reason i don't want to go to sleep now is that i need to wake up soon 23:18:10 oh fuck rent. 23:18:11 -> 23:21:31 shit. 23:21:35 so my account was locked 23:21:41 and i go to the bank, get it fixed 23:21:47 now, one day late, i start paying rent 23:21:50 and it's still locked 23:21:55 \o/ 23:22:55 what the fuck 23:22:59 well. this is gonna cost. 23:27:23 hey oklo 23:27:31 -!- neldoreth has quit ("leaving"). 23:28:09 -!- neldoreth has joined. 23:28:23 hey psyggie 23:28:29 can you fix my problem 23:29:47 yes 23:30:08 first we must insight the proletariate to rise up 23:30:32 then, once the revolution has taken all land and property from the bourgeoisie, you wont need to pay rent. 23:30:37 problem fixed! 23:32:29 will that get it payed tonight? 23:32:36 * SimonRC goes 23:32:58 you may kid but I keep seeing communists campaigning ni the town center 23:33:00 * SimonRC goes 23:33:35 simonrc: im only partially kidding. 23:33:48 since i AM an insurrectionary anarchist. 23:34:44 that word did not be contained in mine head lexicon. 23:35:08 practical-anarchists are so tdeious 23:35:33 i don't know "tdeious" either 23:35:37 god i suck at this language 23:35:41 tedious. 23:35:42 :P 23:35:55 -!- jix has joined. 23:37:31 quite. 23:53:58 okay now maybe south park 23:54:13 i wonder how i managed to spend half an hour not starting the ep