00:14:30 -!- hppavilion[1] has quit (Ping timeout: 255 seconds). 00:16:56 -!- hppavilion[1] has joined. 00:31:07 -!- Phantom_Hoover has quit (Read error: Connection reset by peer). 00:56:30 How difficult would it be to port xterm to a new kind of operating system if you omit some features (Unicode font support, TrueType font support, Tektronix mode, and everything having to do with widget sets)? 01:09:12 hezzo38. aren't you just porting the concept of a POSIX shell then? 01:09:50 related question: are there anything else besides Tektronix and other historical weirdities? 01:10:28 boily: xterm is mostly a VT100 family emulator, zzo38 wasn't suggesting removing that 01:10:38 and it doesn't actually have a shell included, it relies on other programs for that 01:12:02 Bitmap fonts and VT420 and Sixel graphics and so on would still be included but it doesn't necessarily use a UNIX shell or a windowing system, is what I meant 01:12:43 zzo38: I suspect the hardest parts will be a) rendering, and b) replacing the pseudoterminal logic with something that isn't UNIX-specific 01:12:45 sorry, clusterconfusion on my part there. of course the terminal isn't the shell. the map isn't the territory... 01:13:19 why must rendering be hard? 01:13:30 boily: xterm assumes it's rendering to X, for obvious reasons 01:13:40 zzo38's new kind of operating system probably doesn't have an X impl 01:15:06 oh. 01:20:00 <\oren\> hi 01:21:07 \helloren\. 01:21:24 I wonder if porting the whole X shebang would be hard... 01:21:58 <\oren\> unicode font suppoirt is one of the easier things to have I think, beacuse on Windows for example you have to do work to not have it 01:22:33 <\oren\> most of the modern windows apis use unicode only 01:24:17 <\oren\> similar modern OS's have true typ and opentype font support for free 01:25:09 <\oren\> probably the hard parts are the stuff where xterm exposes its X window reference 01:29:11 Clearly you should have to omit the part where xterm exposes its X window reference 01:30:16 I also do not mean using an existing OS, but rather a new kind for specific uses 01:33:50 -!- boily has quit (Quit: SEPULCHRAL CHICKEN). 01:43:37 In an attempt to be a simpler and faster system that uses up less power on portable computers; core software may include a Forth environment, a terminal emulator, and a SSH client (with extensions for one-time-pad and Plan9 forwarding); there can also be some optional packages to add alarm-clock, CD audio, infrared communications, photography, fax, etc (some may require specific hardware to be present) 01:48:47 -!- hppavilion[1] has quit (Ping timeout: 256 seconds). 01:52:38 -!- hppavilion[1] has joined. 01:55:31 [wiki] [[MCEP]] N http://esolangs.org/w/index.php?oldid=44406 * Hppavilion1 * (+213) Created Page 02:03:57 -!- hppavilion[1] has quit (Ping timeout: 252 seconds). 02:22:52 The intention is that you can access the computer even when you travel for vacation or business and/or to other country or to hotel or whatever; most computation and storage would then be done remotely by accessing your computer at home or some other server, but you can still do some operations locally too, including to program your own if needed. 02:31:54 -!- hppavilion[1] has joined. 02:40:15 -!- hppavilion[1] has quit (Ping timeout: 252 seconds). 03:05:56 -!- MDude has changed nick to MDream. 03:15:30 -!- trout has quit (Quit: 1 found in /dev/zero). 03:18:07 [wiki] [[User:Zzo38]] http://esolangs.org/w/index.php?diff=44407&oldid=40109 * Zzo38 * (+28) 03:20:48 -!- mauris has quit (Ping timeout: 255 seconds). 03:44:00 Is there a keyboard shortcut for "close all other tabs" in Firefox? 03:44:36 Many things should have keyboard commands available but don't; I want to have as many things using keyboard command as will be possible to do so. 03:44:57 -!- adu has joined. 03:45:14 just drag the tab out of the window 03:45:18 close other window 03:46:29 OK, that does work, but I want to know how to do it with one or two keystrokes rather than using mouse 03:46:45 (or use that sequence of kb commands, eg open in new window, alt-tab, alt-f4 03:46:52 not sure if theres a open in new window 03:47:48 ALT+TAB and ALT+F4 are for Windows (although there are similar keycodes for the system I am using) 03:48:19 apple tab, cmd w, or cmd f4 03:48:56 No they aren't, although I do know what they are on my system 03:49:08 *shrug* 03:49:20 * blurelIse is obviously not on a mac 03:49:37 Neither am I 03:49:42 (or linux) 03:51:36 you could always program said sequence into a key command 03:52:52 I don't really know how, although I would think I could put something into userChrome.js to implement such a thing, although still I don't quite know how 03:54:08 i couldnt explain not knowing all the details 03:54:24 but its probably something one could possibly googlecate themselves on with some ease 03:55:30 I know I have figured out some other things by looking in the DOM Inspector and Mozilla documentation and source-codes, but some things I did not find even with Google or other search enginges 03:55:37 'g custom keyboard commands firefox 03:55:57 -!- glowcoil has quit (Quit: Connection closed for inactivity). 03:56:14 program semantics, query languages, distributed systems, etc... https://www.youtube.com/watch?v=R2Aa4PivG0g 03:59:04 For example I have figured out how to change the way of location bar working; now it will display percent-encodings instead of displaying Unicode characters, and if I type / in the location bar it will access the root page of the current site (it now always treats the URL entered as relative, unless it is a complete absolute URL) 04:03:24 [wiki] [[Special:Log/newusers]] create * Mbomb007 * New user account 04:05:12 -!- bender| has joined. 04:06:53 -!- hppavilion[1] has joined. 04:09:00 [wiki] [[Self-modifying Brainfuck]] http://esolangs.org/w/index.php?diff=44408&oldid=38632 * Mbomb007 * (+68) /* External resources */ 04:10:29 [wiki] [[MCEP]] M http://esolangs.org/w/index.php?diff=44409&oldid=44406 * Hppavilion1 * (-14) Not sure what I did, or even if I did anything. 04:18:40 [wiki] [[User:Mbomb007]] N http://esolangs.org/w/index.php?oldid=44410 * Mbomb007 * (+179) Me 04:21:01 [wiki] [[Talk:Esoteric processor]] http://esolangs.org/w/index.php?diff=44411&oldid=35169 * Hppavilion1 * (+215) /* Ideas */ new section 04:21:13 -!- MoALTz_ has joined. 04:21:18 [wiki] [[User:Mbomb007]] http://esolangs.org/w/index.php?diff=44412&oldid=44410 * Mbomb007 * (+16) fix link 04:21:19 [wiki] [[Talk:Esoteric processor]] M http://esolangs.org/w/index.php?diff=44413&oldid=44411 * Hppavilion1 * (+98) Signed 04:22:25 [wiki] [[Esoteric Processor/GPU]] N http://esolangs.org/w/index.php?oldid=44414 * Hppavilion1 * (+94) Created page 04:22:51 Hi, MoALTz_! 04:23:01 [wiki] [[User:Mbomb007]] M http://esolangs.org/w/index.php?diff=44415&oldid=44412 * Mbomb007 * (-11) fix link 04:24:10 -!- MoALTz has quit (Ping timeout: 246 seconds). 04:24:51 ais523: did someone say "tunnel over DNS"? What would the purpose of that be? 04:25:01 Note that I'm not aware of any relevant context. 04:25:07 hi hppavilion[1] 04:25:15 tswett: when you have to tunnel over /something/ 04:25:21 whoever put that {{stub}} on the Esolang article about Feather is a genius, btw <-- why thank you. 04:25:22 Hi adu 04:25:28 DNS is one of the least likely internet-based services to be blocked 04:25:54 and it felt like a pretty eso thing to do 04:26:52 I want to design an Esoteric Computer (CPU and possibly GPU) 04:26:53 * tswett nods. 04:26:58 It's actually *shockingly* practical -- a decent number of captive portal wifi hotspots will end up still allowing DNS traffic through. 04:27:01 Now I have to go study Procesors... 04:27:05 *Processors 04:27:10 Huh. 04:27:42 hppavilion[1]: I recommend MMIX 04:27:46 pikhq: presumably, to avoid accidentally leaving a cached IP pointing at the portal for the website that the user was trying to visit? 04:28:03 I think so. 04:28:04 (because an nxdomain return, or just dropping the request, would mean that the browser didn't try to load the captive portal page) 04:29:01 Depends on the captive portal, though. 04:29:27 hppavilion[1]: I think Conway's Game of Life would be a great thing to make an esoteric computer out of. 04:29:37 Oooh 04:29:54 I assume you're familiar with http://code.kryo.se/iodine/ ? 04:29:55 Cellular Automata are definitely an idea 04:30:38 What should I call my Esoteric Computer? 04:31:50 @ask b_jonas can you send one to Ben Franklin please/ <-- let me guess, you want him to switch negative and positive charge? 04:31:50 Consider it noted. 04:32:00 (No one better make an hppavilion[1] joke) 04:32:40 HiggledPiggledy perhaps? 04:32:53 *Higgledy Piggledy? 04:32:58 Dammit, how do you spell that? 04:33:12 hppavilion[1]: or you could base it on the work of Alexey Radul and Gerald Jay Sussman 04:33:14 <\oren\> higLdEpigLdE 04:33:23 Nah 04:33:26 Jigoldipígoldi. 04:33:29 I'll just go with higgledy piggledy 04:33:50 HiggledyPiggledyPowerAutomatonVirtualInstructibleLayoutInsensitiveOperatingNetwork[1] 04:33:59 You should call it Ambronse Alphose, of course. 04:36:05 [wiki] [[Higgledy Piggledy Processor]] N http://esolangs.org/w/index.php?oldid=44416 * Hppavilion1 * (+373) Created Page (WIP) 04:48:18 [wiki] [[User:Zzo38/Programming languages with unusual features]] N http://esolangs.org/w/index.php?oldid=44417 * Zzo38 * (+4821) Created page with "Here I list various programming languages and VMs and computers and so on with some kind of unusual features (and stuff I found interesting); if you disagree you might change ..." 05:19:03 -!- bb010g has joined. 05:27:29 hppavilion[1]: Emoculus? really? 05:27:39 I'm not writing that page 05:28:53 adu: Are you angry at the pun? 05:29:03 Because it can be changed. 05:29:07 it sounds like homonculus 05:29:16 NO clue what that is 05:29:33 The name can be changed; it was just the best I could come up with 05:30:21 [wiki] [[User:Hppavilion1/Hedwig Notta]] M http://esolangs.org/w/index.php?diff=44418&oldid=44383 * Hppavilion1 * (+12) There, adu 05:30:28 http://medical-dictionary.thefreedictionary.com/Homonculus 05:30:43 hppavilion[1]: ahh that's better 05:30:53 Dear god 05:31:02 You just retroactively ruined my childhood 05:31:35 hppavilion[1]: how so? 05:31:54 it just means "little man" 05:31:54 [wiki] [[User:Hppavilion1/Hedwig Notta]] M http://esolangs.org/w/index.php?diff=44419&oldid=44418 * Hppavilion1 * (+41) Compartmentalization! 05:32:09 adu: I saw the word "sperm". I assumed things. 05:32:17 * hppavilion[1] is not immature 05:32:54 I think the PC term is "little person" 05:35:29 http://www.thefreedictionary.com/homunculus 05:35:35 also spelled with a u 05:35:46 wow you can learn so much from a dictionary 05:46:29 [wiki] [[User:Zzo38/Programming languages with unusual features]] http://esolangs.org/w/index.php?diff=44420&oldid=44417 * Zzo38 * (+4192) 05:47:12 [wiki] [[User:Zzo38/Programming languages with unusual features]] http://esolangs.org/w/index.php?diff=44421&oldid=44420 * Zzo38 * (+15) 05:47:13 Vorpal: Re. nested k I just think that there's no reason to special-case it since the standard doesn't say you should, so I implement k to the letter and observe the resulting "interesting" nested k :-P 05:49:32 `` ls /var/tmp 05:49:33 ls: cannot access /var/tmp: No such file or directory 05:49:39 `` ls /tmp 05:49:40 No output. 05:49:49 `` ls /var 05:49:50 ls: cannot access /var: No such file or directory 05:49:57 `` ls /usr 05:49:58 bin \ games \ include \ lib \ local \ sbin \ share \ src 05:50:08 `` ls /usr/share 05:50:09 aclocal \ adduser \ alsa \ application-registry \ applications \ apps \ aptitude \ apt-xapian-index \ avahi \ base-files \ base-passwd \ binfmts \ bison \ bug \ build-essential \ ca-certificates \ ca-certificates-java \ calendar \ cmake-2.8 \ common-licenses \ dbus-1 \ debconf \ debhelper \ debianutils \ dict \ doc \ doc-base \ dpkg \ emacs \ file 05:50:27 `` echo hi >/tmp/test 05:50:27 No output. 05:50:32 `ls /tmp 05:50:33 No output. 05:52:34 ais523.................. <-- it was inevitable, really. 05:52:58 `` hg log --removed wisdom/culprit | grep summary: 05:52:59 summary: revert accbc9c5c7ec \ summary: echo wisdom/* | shuf | head -n 10 | xargs rm \ summary: le/rn culprit/`culprits` is a program that lists the lists the nicks responsible for a wisdom entry. Usage: `culprits wisdom/ENTRY \ summary: le/rn culprit/`culprits` is a program that lists the lists th 05:53:17 i know 05:54:57 my attempts to delete random entries have apparently ended up putting me in the version history of all of them 05:55:57 doesthiswork1: I've got better at mentioning Feather without actually /thinking/ about it 05:56:02 I have way too many other things to think about atm 05:56:18 Hppavilion1 put one thing in the userspace because it is not esolang and you are free to move to the main namespace if it belong there, but there is also something in my userspace I do not know if it is esolang or not (although someone on this channel told me to put it there, I think); originally it said don't move to main namespace but now it says it can be moved if it is agreed to belong in main namespace. 06:01:20 (The page about Hedwig Notta is not bad as it is though, although it could be improved; unfortunately I do not know what to write.) 06:10:22 hello world 06:17:19 -!- hppavilion[1] has quit (Ping timeout: 252 seconds). 06:24:23 -!- hppavilion[1] has joined. 06:24:57 -!- ais523 has quit (Quit: going home). 06:41:12 -!- Alcest has joined. 06:56:51 zzo38: I always start with "Lorem Ipsum" when I don't know what to write on a page. It's kinda like "Once upon a time", but can go on for days if you have enough latin 06:57:15 -!- adu has quit (Quit: adu). 07:02:21 I don't know enough Latin 07:03:36 http://www.lipsum.com/ now you do 07:04:45 alternatively, theres bacon https://baconipsum.com/ 07:07:15 if thats too non-kosher, theres a great document you can probably use, entitled "Chicken", that basically repeats the word chicken repeatedly, but is nicely formatted and pnctuated 07:08:01 https://esolangs.org/wiki/Chicken 07:08:38 brilliant 07:12:40 -!- Patashu has quit (Ping timeout: 250 seconds). 07:12:41 -!- hppavilion[1] has quit (Ping timeout: 252 seconds). 07:14:43 -!- Patashu has joined. 07:41:05 -!- AnotherTest has joined. 07:51:46 @tell ais523 It found one real crash that happens on the normal binary so far. Related to stack stack starting ({ that is) 07:51:46 Consider it noted. 07:52:51 @tell ais523 "PXIF"4({XJJCdM-^? and "PXIF"4({XJJCyM-^? both trigger variants of it. And those are minimized. 07:52:51 Consider it noted. 07:59:39 -!- doesthiswork1 has quit (Quit: Leaving.). 08:06:43 @tell ais523 Oh I see, it is trying to allocate around 2^62 funge cells, and for some reason it isn't failing at malloc, but when trying to memset it all to zero 08:06:44 Consider it noted. 08:10:41 -!- MoALTz_ has changed nick to MoALTz. 08:16:56 malloc never fails. 08:19:10 -!- Patashu has quit (Ping timeout: 260 seconds). 08:20:38 -!- sc00fy has joined. 08:21:25 `cc #include \ int main(void) { for(int i=0;; ++i) { assert (malloc(1e9)), printf("%d\n", i); } } 08:21:30 : In function ‘main’: \ :2:2: error: missing terminating " character \ compilation terminated due to -Wfatal-errors. 08:22:56 `cc #include \ #include \ #include \ int main(void) { for(int i=0;; ++i) { assert (malloc(1e9)), printf("%d%c", i, 10); } } 08:22:58 a.out: :4: main: Assertion `malloc(1e9)' failed. \ Aborted 08:23:08 For some values of never. 08:23:12 `uname -a 08:23:13 Linux umlbox 3.13.0-umlbox #1 Wed Jan 29 12:56:45 UTC 2014 x86_64 GNU/Linux 08:23:28 `stat /sys/vm 08:23:30 stat: cannot stat `/sys/vm': No such file or directory 08:28:51 -!- AnotherTest has quit (Ping timeout: 256 seconds). 08:36:11 -!- gamemanj has joined. 08:44:37 -!- adu has joined. 08:53:18 -!- bb010g has quit (Quit: Connection closed for inactivity). 08:58:39 -!- sc00fy has quit (Ping timeout: 240 seconds). 09:16:25 [wiki] [[Talk:Brainfuck]] http://esolangs.org/w/index.php?diff=44422&oldid=44390 * Martin Büttner * (+275) /* Would BF still be TC with do-while loops? */ 09:27:48 -!- adu has quit (Quit: adu). 09:40:50 -!- Phantom_Hoover has joined. 09:55:29 -!- oerjan has quit (Quit: leaving). 09:59:32 Insane, a language that changes how it works when you use it, but gives you details of how it changes. 10:30:06 -!- Patashu has joined. 10:36:47 -!- Patashu has quit (Ping timeout: 244 seconds). 10:36:55 -!- Patashu has joined. 11:09:12 Wait, that question is still considered to be open? 11:34:27 -!- Sprocklem_ has joined. 11:36:51 -!- Sprocklem has quit (Ping timeout: 265 seconds). 11:47:48 -!- Phantom_Hoover has quit (Ping timeout: 264 seconds). 11:49:04 -!- AnotherTest has joined. 12:21:46 Guys, are there compilers that compile to BF? 12:22:41 There are (BFBASIC is one), but wouldn't that take out the fun of it? 12:30:27 -!- J_Arcane_ has joined. 12:32:00 -!- J_Arcane has quit (Ping timeout: 244 seconds). 12:32:07 -!- J_Arcane_ has changed nick to J_Arcane. 12:33:28 -!- Phantom_Hoover has joined. 12:33:45 -!- Phantom_Hoover has quit (Changing host). 12:33:45 -!- Phantom_Hoover has joined. 12:55:24 [wiki] [[Talk:Brainfuck]] http://esolangs.org/w/index.php?diff=44423&oldid=44422 * 213.162.68.156 * (+1756) /* Would BF still be TC with do-while loops? */ 12:56:50 a bit tedious, perhaps, but not hard 13:05:17 [wiki] [[Special:Log/newusers]] create * Int-e * New user account 13:06:52 [wiki] [[Talk:Brainfuck]] M http://esolangs.org/w/index.php?diff=44424&oldid=44423 * Int-e * (+85) /* Would BF still be TC with do-while loops? */ de-anonymize 13:07:00 -!- Phantom_Hoover has quit (Ping timeout: 264 seconds). 13:07:06 [wiki] [[Nyarlathotep]] M http://esolangs.org/w/index.php?diff=44425&oldid=44404 * CosmoConsole * (+0) /* Operation */ a minor change in operation 13:08:20 [wiki] [[User:Int-e]] N http://esolangs.org/w/index.php?oldid=44426 * Int-e * (+43) Created page with "I wasn't bored enough to say anything here." 13:22:08 Nonymising. 13:24:00 oerjan will probably have a fit about the missing -- and the broken thread... but I'm not sure what to do about the latter 13:24:16 and can't be bothered to add the former right now 13:24:37 :D 13:41:33 -!- MDream has changed nick to MDude. 13:51:37 The calling conventions of bf-targeting compilers are even more esoteric than their target language 13:54:52 [wiki] [[Special:Log/newusers]] create * Sait2000 * New user account 13:58:49 -!- boily has joined. 14:14:47 -!- bender| has quit (Quit: [000]). 14:24:13 -!- MoALTz_ has joined. 14:26:00 [wiki] [[Talk:Insomnia]] http://esolangs.org/w/index.php?diff=44427&oldid=39840 * Sait2000 * (+146) 14:26:25 -!- MoALTz has quit (Ping timeout: 256 seconds). 14:30:30 -!- JesseH has quit (Remote host closed the connection). 14:31:04 -!- doesthiswork has joined. 14:39:39 Could you train a markov chain on an image and then have it generate an image? 14:41:22 FreeFello. probably so! 14:41:40 I'm thinking maybe start with the middle pixel, then grow concentric circles around it? 14:42:31 -!- Phantom_Hoover has joined. 14:42:35 -!- Patashu has quit (Ping timeout: 250 seconds). 14:43:09 Hmm, it'd be something to try out 14:45:00 Seems it is something that has been done 14:45:05 ah? 14:45:56 http://www.mat.ucsb.edu/~g.legrady/academic/courses/08s594/prj/db/ 14:46:03 Doesn't seem to have been too successful 14:46:19 The generated images don't really look like the training data 14:47:13 enhance the algorithm! 14:47:32 -!- mauris has joined. 14:48:37 I'm thinking about exploiting anisotropy: different probabilities according to the growth vector. 14:49:04 maybe treat the colour channels separately? split tables for each of them. 14:49:12 surely a context of one pixel can't be enough... 14:49:22 that, too. 14:50:37 translate RGB to Lab, normalize data, create an Expectation-Maximisation model? use multiple pictures at the same time for training? 14:50:48 combine chickens! 14:52:45 -!- int-e has left ("CHICKEN POULET HUHN КУРИЦА"). 14:52:45 -!- int-e has joined. 15:00:57 * int-e should add a russian key layout... and re-learn russian... 15:06:24 what would be the most efficientest way to cover all of Unicode on a layout? modifier keys to reach different blocks? 15:07:02 Escape-Meta-Alt-Control-Shift-SouthIndic-Tamil? 15:09:26 hmm, efficient in what sense... 15:10:17 you can do a frequency analysis and make a huffman tree using all keys of your keyboard ;) 15:10:28 something along those lines :D 15:11:13 I remember that when I was using rxvt-unicode I could hold Ctrl-Shift and type codepoint hex codes to get any wanted character. 15:11:22 just pick 24 keys and use binary chording 15:12:00 Phellontom_Hoover. that's a possibility, but not enough "humanlike" hth 15:12:33 right, better limit yourself to 8 simultaneously pressed keys at most ;) 15:22:05 -!- int-e has set topic: The Unicode Pianist | ɛ̃ˈglɪʃ spɛˈliŋ ʀɘfɔʀm/ | https://dl.dropboxusercontent.com/u/2023808/wisdom.pdf http://codu.org/logs/_esoteric/ http://tunes.org/~nef/logs/esoteric/ | https://esolangs.org/. 15:41:47 [wiki] [[Talk:Brainfuck]] http://esolangs.org/w/index.php?diff=44428&oldid=44424 * Int-e * (+321) /* Would BF still be TC with do-while loops? */ oops. 15:44:34 hmm. "Notice will be provided in clear and conspicuous language" ... is "conspicuous" a legal term? 15:49:51 <\oren\> the most effiecient! 15:50:49 <\oren\> Try simply having compose + base-64-encoded-code-poin 15:52:48 <\oren\> or maybe have a choose-language key, which is followed by a language code 15:53:23 <\oren\> choose,D,E gives you a german keyboeard 15:54:00 <\oren\> choose,S,E gives you a swedish one 15:54:22 <\oren\> choose,J,P gives you a japanese IME 15:56:20 -!- sc00fy has joined. 16:01:21 [wiki] [[Talk:Brainfuck]] http://esolangs.org/w/index.php?diff=44429&oldid=44428 * Int-e * (+229) /* Would BF still be TC with do-while loops? */ oops again 16:02:48 -!- sc00fy has quit (Ping timeout: 244 seconds). 16:06:22 [wiki] [[Talk:Brainfuck]] M http://esolangs.org/w/index.php?diff=44430&oldid=44429 * Int-e * (+2) /* Would BF still be TC with do-while loops? */ ...details, details 16:10:10 -!- augur has quit (Quit: Leaving...). 16:16:58 int-e: I have a "phonetic Russian" keyboard layout set up. 16:17:28 But I don't speak Russian. 16:17:34 Perhaps I should. 16:19:23 i had that keyboard layout set up on my laptop 16:19:33 \oren\: oh I see what happened, I had somehow conflated "conspicuous" and "suspicious" 16:19:39 but i havent used my laptop seriously in months 16:35:09 -!- augur has joined. 16:36:22 <\oren\> quintopia: are you using it now? 16:39:44 <\oren\> in other news I still havent' found any need to install linux on my new laptop. Instead, I'm currently sshing into my linux laptop and using the brand new one as a thin client. 16:41:33 int-e: i hear you know the secrets to 84-byte haskell fizzbuzz (but not henkma's 82) 16:41:39 <\oren\> so yeah. still on 32-bit up in this bitch 16:42:37 <\oren\> 84 bytes? isn't that rather a lot 16:42:43 people are collecting the very shortest fizzbuzzes on codegolf.SE now that the competition has sort of died out! so i shared my 85B and maybe we should compare solutions? http://codegolf.stackexchange.com/a/58800/3852 16:42:58 main=mapM(putStrLn.f)[1..100];f n|d<-drop.(*4).mod n=max(show n)$d 3"Fizz"++d 5"Buzz" 16:44:11 <\oren\> oh, haskell, not some esolang 16:44:16 <\oren\> well then 16:45:31 at those code densities any language is an esolang :) 16:46:28 exceptt perl, where this is considered normal 16:47:22 <\oren\> it's interesting that the lengths of the haskell and C versions are so close 16:48:00 <\oren\> despite the two languages being basically at odds in so many ways 16:49:25 <\oren\> then there is a cluster around 54 or Perl Python PHP and Powershell 16:51:10 -!- boily has quit (Quit: TRIBULATIVE CHICKEN). 16:52:56 <\oren\> Myabe these clusters reflect some degree of similarity of the "density" of the syntax, as an emergent quality of the language (the maximum "meaning per character" expressible in the syntax) 16:53:49 mauris: I have no idea where you heard that but you heard wrong 16:54:26 http://codu.org/logs/_esoteric/2014-09-14.txt ? 16:55:54 If I had 84 that day I didn't record the solution. But I probably just miscounted and didn't bother to report the fact. 16:56:05 haha, ok 16:57:13 (in my file there's a 85 character solution) 17:06:51 -!- TieSoul has joined. 17:12:50 -!- FreeFull has quit (Quit: BBL). 17:15:38 hi 17:20:59 tix=0 tid=0 x=11 y=0: / (47) 17:20:59 Stack has 2 elements, top 15 (or less) elements: 17:20:59 -1 -9223372036854775808 17:21:04 Floating point exception 17:21:09 whaty 17:22:32 -!- Phantom_Hoover has quit (Ping timeout: 272 seconds). 17:22:39 -!- Phantom__Hoover has joined. 17:23:40 (gdb) print -9223372036854775808LL / -1LL 17:23:40 $3 = 0 17:23:42 Uh? 17:23:52 And the program crashes with SIGFPE on that 17:25:05 I don't get it 17:25:30 > (-2^63) `div` (-1) :: Int 17:25:31 *Exception: arithmetic overflow 17:25:52 Vorpal: look at the manual for sigfpe 17:26:08 it's stated there to exactly do that 17:26:26 Ah 17:26:38 myname, I don't have a man page for sigfpe? 17:27:10 so which manual do you mean? The C specification? 17:30:00 Vorpal: I did that on cfunge just the other day. 17:30:07 Vorpal: Didn't think it counted as a bug when you asked. 17:30:23 That's exactly the one thing I was trying to use <(echo ...) for. 17:30:24 fizzie, oh? Well the standard only mentions division by zero 17:30:39 But I would like to handle whatever this is 17:31:36 fizzie, so why is -2⁶³ / -1 an SIGFPE? 17:31:36 You get SIGFPE on the DE exception, and x86 throws the DE exception if the result is out of range. 17:31:43 DE? 17:31:52 Oh is that a x86 flag 17:31:54 One of the x86 hardware exceptions. 17:32:00 hm 17:32:07 And it's a SIGFPE because INT_MIN / -1 would have the value (INT_MAX+1), which doesn't fit. 17:32:10 fizzie, well the standard says nothing about this issue 17:32:16 And presumably it is x86 specific 17:32:35 What does the IEEE say? 17:32:41 Nothing, because it's an integer operation. 17:32:46 It's signed overflow, which is undefined 17:32:49 Oh of course 17:33:00 fizzie, so what does the funge standard say? 17:33:07 And what does ccbi do 17:33:07 Probably nothing. 17:33:21 I'm not sure what you get on other architectures, but the problem itself is common to everything that's twos'-complement. 17:33:25 -!- aretecode has joined. 17:33:34 fizzie, so what do you think should happen? 17:34:10 I think that's really a matter of opinion. You could argue that it's quite similar to division-by-zero, and could yield zero. 17:34:12 The standard doesn't mention the issue, CCBI doesn't special-case it 17:34:22 fizzie, also how did you test this btw? I'd like to know if you managed to beat the fuzzer to the program length 17:34:26 The other "typical" answer is to return INT_MIN again. 17:34:42 $ echo '2:*:*:*:*:*:2/*:.01-/' > tmp.bef; cfunge tmp.bef 17:34:43 Deewiant, well I'd like my program to not crash :P 17:34:54 > 2^32 * 2^32 :: Int 17:34:54 fizzie, or do: "PXIF"4(R:S/3"0 17:34:55 0 17:34:57 The one :. is extra, too, left for debugging. 17:35:34 2:*:*:*:*:*:2/*01-/ is still a bit longer, but at least it's fingerprint-free. 17:35:34 fizzie, I haven't bothered figuring out how that triggers it, have fun if you want to 17:35:48 fizzie, I think it multiplies the fingerprint itself or something 17:36:19 Vorpal: Oh wow, I didn't realize that actually causes a SIGFPE on x86. 17:36:23 I don't even know what the FIXP representation is. 17:36:36 I always thought it just yields INT_MIN again. 17:36:40 It's a classic denial-of-service technique. 17:36:56 Deewiant, right, so now you will patch ccbi? 17:37:09 I'd rather not 17:37:15 If somebody actually cares, then yes 17:37:22 fizzie, is it just LONG_MIN / -1 that can cause this? 17:37:28 fizzie, or some other combination too? 17:37:34 I'd rather not catch SIGFPE 17:37:35 Vorpal: INT_MIN works just fine too. 17:37:45 Deewiant, no, I have 64-bit cells :P 17:37:51 I for one would prefer if INT_MIN/-1 silently gave INT_MIN, but the division by zero trap in that case is hard-coded in the signed division operation of x86, so you can't just change this cheaply without wrapping every division operation. 17:38:22 Vorpal: It's the only combination when your input and output operands are the same size. 17:38:23 Vorpal: (LONG_MIN isn't necessarily 64-bit) in that case, yes, only that case (and division by zero which you know and handle). 17:38:30 b_jonas, well you could catch SIGFPE and mess with the registers, but urgh 17:38:32 You can certainly wrap the division, but it would be crazy to expect for the C language to do that for you. 17:38:43 Vorpal: yes, you can certainly do that. 17:38:46 fizzie, hm? 17:38:57 Ah 17:38:58 I see 17:39:12 The "SIGFPE if result is too big" is also the bane of everyone who assumes the x86 div/idiv instructions are 64x64 to 64 when it's actually 128x64 to 64. 17:39:12 but you can't avoid the trap unless you make sure in advance to give the right arguments 17:39:15 Deewiant, I think I have a FUNGE_CELL_MIN define or something *somewhere* 17:40:00 b_jonas, that is the more portable solution I will go for probably 17:40:05 I don't want to mess with SIGFPE 17:40:17 (And analogously for 64x32-to-32, 32x16-to-16 and 16x8-to-8.) 17:40:27 There's been a number of people on ##asm who don't clear (or sign-extend, as appropriate) the high half. 17:41:32 also, some cpu do crazy stuff if you divide by 2^x with the lower 64 bit being 0 17:41:34 "ARIN has reached depletion of the general IPv4 free pool today, 24 September 2015." 17:41:39 fizzie: FIXP provides 4 decimal digits of precision. 17:41:40 Vorpal: iirc it's a trap, not a fault, so you don't even have to mess with the registers. just return from the interrupt 17:41:44 still, that's expensive 17:41:59 fizzie, I guess I will have to patch CPLI and such too 17:42:00 (so is a division in first place, but there's a big difference) 17:42:07 if (denom != 0) { 17:42:07 stack_push(ip->stack, (ai * bi + ar * br) / denom); 17:42:07 stack_push(ip->stack, (ai * br - ar * bi) / denom); 17:42:07 } else { 17:42:08 stack_push(ip->stack, 0); 17:42:09 stack_push(ip->stack, 0); 17:42:09 Deewiant: Oh, right, I somehow skipped that line. 17:42:10 } 17:42:12 Okay that will be annoying 17:43:21 Deewiant, I'm using afl to fuzz cfunge in case you missed that yesterday 17:43:23 Just stick a static inline cell funge_div(cell nom, cell denom); somewhere and use that every time you feel like writing a /. 17:44:04 Deewiant, it found this bug, k-issues and something really really strange with broken stacks in { for massive negative arguments. Other than that it seems very good 17:44:15 fizzie, probably a good idea yes 17:45:07 Vorpal: Oh, good, let me know if you find anything else interesting. It's on my todo list for my written-in-C interpreter (which is turning out to be a bit of a fi:ikuisuusprojekti). 17:45:19 Deewiant, a what? 17:45:35 Deewiant, well it should generate a corpus that will exercise lots of code paths 17:45:41 Couldn't think of a good translation on short notice. 17:45:42 Deewiant, might be fun running ccbi on that 17:45:49 Vorpal: Can it instrument D? 17:45:51 a moment, I'm still looking it up, I'm not sure in this 17:46:05 Deewiant, no, but there is a qemu mode, (somewhat slower) 17:46:31 Deewiant, and if you can target LLVM, it might be possible, not sure. I haven't looked into the llvm stuff, but it might need clang? 17:46:48 If it needs clang it can't use D, that's for sure. 17:46:55 If it can work on raw LLVM it'd work though. 17:47:23 But I'd rather not set up the whole D development thing again... I'm hoping I won't have to touch CCBI any more. :-P 17:47:23 thanks, manual, using the symbol "#DE" for two different things. typical of your sloppy ambiguous pseudocode notation, intel. 17:48:15 Deewiant, don't like D? 17:48:33 The "D" is short for "dead language". 17:48:35 Vorpal: D 1.0 is dead. 17:49:11 Last update end of 2012. I could port to D 2.0 but I don't like it and the effort is probably comparable to rewriting in C anyway. 17:49:52 Unfortunately the rewrite turned into rewrite-and-improve so it's kind of stuck as I don't have as much time to work on it as I used to. 17:50:04 fizzie, what about -1 / -2^63? 17:50:16 Vorpal: That's just 0. 17:50:20 There is no weird underflow or something I assume 17:50:22 Right 17:50:39 Deewiant, ah 17:50:47 I'm wrong 17:50:52 the AMD manual says it's a fault 17:50:58 so you have to do something special, you can't just ignore it 17:51:03 argh 17:51:06 What is the word for the part that isn't the denominator in English? 17:51:11 that's stupid, but it's for compatibility 17:51:12 Vorpal: numerator 17:51:13 The "numerator" 17:51:14 Vorpal: numerator 17:51:15 I think 17:51:16 Thanks 17:51:27 yeah, like Númenor or Numitor 17:51:40 where is this in the intel manual though? 17:51:47 I think I've heard it called "fraction line" or something like that. 17:53:22 b_jonas: In my copy, it's in volume 3, chapter 6.15 "Exception and Interrupt Reference", subpart "Interrupt 0 - Divide Error Exception (#DE)". 17:53:28 "Exception Class: Fault." 17:53:56 "Indicates the divisor operand for a DIV or IDIV instruction is 0 or that the result cannot be represented in the number of bits specified for the destination operand." 17:54:17 fizzie, what about %? 17:54:22 fizzie, when can that fail? 17:54:30 Vorpal: that's the same instruction 17:54:36 Oh, right 17:55:07 int-e, and what should it return on -2^63 / -1 17:55:29 Rather, -2^63 % -1, and that one should clearly be 0. 17:55:32 > (-2^63) `mod` (-1) 17:55:34 0 17:55:40 > (-2^63) `rem` (-1) 17:55:42 0 17:55:43 Vorpal: we have already established what it does return 17:56:10 Ah 17:56:30 fizzie: indeed, it's there 17:56:30 fizzie, what should it return in MODU fingerprint though? 17:56:32 thanks 17:56:33 (nothing, it causes an exception) 17:56:48 Haskell actually implements special cases for INT_MIN/-1 17:57:03 which is why the mod and rem don't fail 17:57:11 fizzie, because that is all the cases of signed result, unsigned result and reminder 17:57:31 > (-2^63) `mod` (-1) :: Int 17:57:32 0 17:57:53 ;-) 17:57:53 > (-2^63) :: Int 17:57:55