2013-04-17T00:31:03 *** amstan has joined #aichallenge 2013-04-17T00:31:03 *** ChanServ sets mode: +o amstan 2013-04-17T00:32:46 *** amstan has quit IRC (Client Quit) 2013-04-17T00:33:00 *** amstan_ has quit IRC (Quit: Konversation terminated!) 2013-04-17T01:10:56 *** UncleVasya has joined #aichallenge 2013-04-17T01:27:24 *** mceier has quit IRC (Quit: leaving) 2013-04-17T02:30:19 UncleVasya: you asked a few days ago how I found out about nodewar - aerique told me about it here 2013-04-17T02:39:33 smiley1983: ok 2013-04-17T03:06:19 *** Wren80[work] has joined #aichallenge 2013-04-17T03:17:47 *** mceier has joined #aichallenge 2013-04-17T03:41:50 *** epicmonkey has joined #aichallenge 2013-04-17T03:49:11 *** loglog has quit IRC (Ping timeout: 252 seconds) 2013-04-17T04:07:14 *** thestinger has joined #aichallenge 2013-04-17T04:09:06 *** epicmonkey has quit IRC (Ping timeout: 272 seconds) 2013-04-17T04:10:34 *** loglog has joined #aichallenge 2013-04-17T04:16:12 *** KokaKiwi has quit IRC (Ping timeout: 252 seconds) 2013-04-17T04:21:56 *** KokaKiwi has joined #aichallenge 2013-04-17T04:46:44 *** janzert1 has joined #aichallenge 2013-04-17T04:47:07 *** janzert has quit IRC (Ping timeout: 252 seconds) 2013-04-17T05:07:57 *** RagingDave has joined #aichallenge 2013-04-17T05:13:08 *** epicmonkey has joined #aichallenge 2013-04-17T06:05:50 *** UncleVasya has quit IRC (Read error: Connection reset by peer) 2013-04-17T06:25:57 *** UncleVasya has joined #aichallenge 2013-04-17T06:34:40 *** sigh has joined #aichallenge 2013-04-17T06:39:59 *** ramn[work] has quit IRC (Ping timeout: 258 seconds) 2013-04-17T07:28:05 *** RagingDave_ has joined #aichallenge 2013-04-17T07:31:43 *** RagingDave has quit IRC (Ping timeout: 264 seconds) 2013-04-17T07:35:33 *** RagingDave_ is now known as RagingDave 2013-04-17T07:45:19 *** ramn[work] has joined #aichallenge 2013-04-17T09:00:06 *** cyphase has quit IRC (Ping timeout: 264 seconds) 2013-04-17T11:00:20 *** mceier has quit IRC (Quit: leaving) 2013-04-17T11:33:44 *** mcstar has joined #aichallenge 2013-04-17T11:34:03 i guess the question is now, how did aerique find about it 2013-04-17T11:34:30 i think im missing an 'out' somewhere 2013-04-17T11:35:26 probably: 'find out about it' 2013-04-17T11:43:09 \o 2013-04-17T11:44:32 o/ 2013-04-17T11:46:19 *** rusti_ has joined #aichallenge 2013-04-17T11:47:46 *** thestinger1 has joined #aichallenge 2013-04-17T11:48:19 *** amstan has joined #aichallenge 2013-04-17T11:48:19 *** ChanServ sets mode: +o amstan 2013-04-17T11:48:24 *** rusti_ has quit IRC (Remote host closed the connection) 2013-04-17T11:48:40 *** rusti_ has joined #aichallenge 2013-04-17T11:49:10 rusti_: "hi mcstar" 2013-04-17T11:49:12 "hi mcstar" 2013-04-17T11:49:34 mcstar: how is your 3d stuff going? :) 2013-04-17T11:50:02 *** thestinger has quit IRC (Ping timeout: 245 seconds) 2013-04-17T11:50:09 *** thestinger1 is now known as thestinger 2013-04-17T12:15:09 thestinger: sry, i was afk, and going to be for a little while more 2013-04-17T12:40:17 *** mceier has joined #aichallenge 2013-04-17T12:41:24 *** Wren80[work] has quit IRC (Quit: Wren80[work]) 2013-04-17T12:50:39 thestinger: i think im back 2013-04-17T12:51:26 rusti_: for 3.times { println("yay") } 2013-04-17T12:51:27 yay 2013-04-17T12:51:27 yay 2013-04-17T12:51:27 yay 2013-04-17T12:51:28 ;p 2013-04-17T12:52:13 rusti_: retry (unsafeIOToStm launchICBMs) 2013-04-17T12:52:13 foo.rs:5:30: 5:41 error: expected `,` but found `launchICBMs` 2013-04-17T12:52:13 foo.rs:5 retry (unsafeIOToStm launchICBMs) 2013-04-17T12:52:13 ^~~~~~~~~~~ 2013-04-17T12:52:26 ^ doesnt work 2013-04-17T12:52:40 rusti_: unsafe { libc::abort() } 2013-04-17T12:52:42 more than 5 lines of output, pastebinned: http://sprunge.us/aSeR 2013-04-17T12:52:54 dammit stupid unreachable check 2013-04-17T12:52:56 luckily it's dumb 2013-04-17T12:52:58 rusti_: let aichallenge = best_challenge in launch aichallenge now ;; 2013-04-17T12:52:59 foo.rs:5:42: 5:44 error: expected `;` but found `in` 2013-04-17T12:52:59 foo.rs:5 let aichallenge = best_challenge in launch aichallenge now ;; 2013-04-17T12:52:59 ^~ 2013-04-17T12:53:21 rusti_: fn die() { unsafe { libc::abort() } }; die() 2013-04-17T12:53:22 thestinger: lol 2013-04-17T12:53:22 -bash: line 1: 16 Aborted ./foo 2013-04-17T12:53:42 thestinger: did you write the IRC code too? 2013-04-17T12:54:00 mcstar: yes it's python irc bot <- serial port -> qemu <-> container with rust 2013-04-17T12:54:06 mcstar: i didn't write the irc part. 2013-04-17T12:54:10 ok 2013-04-17T12:54:15 i thought you did it in rust 2013-04-17T12:54:21 or at least suspected... 2013-04-17T12:54:23 it was a 5 min hack :P 2013-04-17T12:54:39 why qemu? 2013-04-17T12:54:50 I don't trust my ability to secure a container 2013-04-17T12:55:07 current kernels provide sandboxing afaik 2013-04-17T12:55:17 mcstar: yes the container is sandboxed 2013-04-17T12:55:19 is that the 'container'? 2013-04-17T12:55:23 alright 2013-04-17T12:55:35 rusti_: core::run::run_program(~"ls", [~"/dev/"]) 2013-04-17T12:55:36 console core fd full null ptmx pts random shm stderr stdin stdout tty urandom zero 2013-04-17T12:55:36 0 2013-04-17T12:55:46 rusti_: core::run::run_program(~"ps", [~"ax"]); 2013-04-17T12:55:48 more than 5 lines of output, pastebinned: http://sprunge.us/UcRe 2013-04-17T12:56:01 thestinger: i didnt know rust had a repl 2013-04-17T12:56:13 mcstar: it does, but the bot isn't actually using the REPL/JIT 2013-04-17T12:56:30 just putting it in a file and compiling it is faster than LLVM's silly JIT 2013-04-17T12:56:37 because it's small 2013-04-17T12:56:43 oh 2013-04-17T12:56:52 that is quite bad 2013-04-17T12:57:03 well it's partly rusti's fault for being slow 2013-04-17T12:57:10 it hits some really, really terrible slow path in LLVM 2013-04-17T12:57:14 well, i guess working in the repl for longer times does amortize this cost :) 2013-04-17T12:57:45 mcstar: 1 + 1 in the repl takes a noticeable 200ms or something 2013-04-17T12:58:07 and the repl has a startup cost too, right? 2013-04-17T12:58:09 it's doing something weird that makes LLVM waste a lot of time 2013-04-17T12:58:14 mcstar: not really 2013-04-17T12:58:16 julia's repl is very slow to start 2013-04-17T12:58:21 I think you pay it whenever you enter something 2013-04-17T12:58:24 but that might be their fault? 2013-04-17T12:58:29 mcstar: probably their fault 2013-04-17T12:58:32 cling is fast 2013-04-17T12:59:02 rust's REPL is slow because it hasn't gotten any love beyond being created yet :P 2013-04-17T12:59:17 thats still love, alright 2013-04-17T12:59:25 LLVM's JIT is broken with segmented stacks so it might be happier with segmented stacks off 2013-04-17T12:59:30 thestinger: i finished my 3d work 2013-04-17T12:59:49 i have an implementation of marching tetrahedrons, and naive surface nets 2013-04-17T13:00:10 what are surface nets? 2013-04-17T13:00:11 i can save unstructured triangles/quads with data at the vertices in VTK 2013-04-17T13:00:28 thestinger: well, a naive kind of efficient triangulation 2013-04-17T13:01:05 it produces quads, and even if you double the number(since there are 2 triangles in a quad) it still outperforms marching tetrahedrons/cubes by much 2013-04-17T13:01:10 so it's just an efficient way of representing some polygonal surface? 2013-04-17T13:01:24 not representing, calculating one 2013-04-17T13:01:32 the representation is just triangles/quads 2013-04-17T13:01:44 ah 2013-04-17T13:01:58 but if you have a scalar field, you have to calculate which triangles (and how) belong to the isosurface 2013-04-17T13:02:21 let me upload an image 2013-04-17T13:03:34 http://i.imgur.com/BieSfSg.png 2013-04-17T13:03:46 this is a growing dendrite in a phase field model 2013-04-17T13:03:59 the box is around 380x380x1500 divisions 2013-04-17T13:04:32 and the phase field determines structures(liquid vs. solid) and the isosurface shown is the 0 valued surface 2013-04-17T13:04:43 and it is colored by the concentration field 2013-04-17T13:04:58 concentration of what? energy? 2013-04-17T13:04:59 originally there are 2x860MB of floats 2013-04-17T13:05:11 and now the file is around 200MB 2013-04-17T13:05:21 wow, looks cool 2013-04-17T13:05:23 thestinger: different materials 2013-04-17T13:05:33 it is a 2 component system 2013-04-17T13:05:37 ah 2013-04-17T13:06:30 is it something biological? 2013-04-17T13:06:36 *** mviel has quit IRC (Remote host closed the connection) 2013-04-17T13:06:50 nope 2013-04-17T13:07:39 :( 2013-04-17T13:07:54 think metallurgy 2013-04-17T13:08:19 im not sure how many biological stuff have dentritic structure 2013-04-17T13:08:39 but pattern formation is relevant for biology 2013-04-17T13:09:01 mcstar: I googled it and neurons show up before the actual term about crystals 2013-04-17T13:09:03 ;p 2013-04-17T13:09:07 yeah 2013-04-17T13:09:16 they have a similar structure 2013-04-17T13:09:21 they branch and stuff 2013-04-17T13:09:33 but it has nothing to do with neurons, ofc 2013-04-17T13:09:40 (these dendrites i mean) 2013-04-17T13:09:58 this is not my field though 2013-04-17T13:10:07 just helping out the team with this code 2013-04-17T13:10:34 I hope you'll get a job in biological informatics field and you'll start to post relative pictures of your work :3 2013-04-17T13:11:00 thats deltor 2013-04-17T13:11:46 the code is ugly... i dont want to show it 2013-04-17T13:12:20 mcstar: what language is it? 2013-04-17T13:12:22 VTK is soo incredibly overengineered though, its furtunate i only use it to save into a common file format 2013-04-17T13:12:27 c++ 2013-04-17T13:12:38 *** Accoun has quit IRC (Ping timeout: 246 seconds) 2013-04-17T13:12:40 does it actually have to "grow" that before rendering it? 2013-04-17T13:12:56 im not sure i understand 2013-04-17T13:13:09 are you just visualizing existing data, or actually calculating something? 2013-04-17T13:13:11 this is a frame in a long running simulation, later times 2013-04-17T13:13:32 thestinger: well, we want to only save this data for each frame, not the whole 3D array 2013-04-17T13:13:52 because this is what we would do anyway, when postprocessing it 2013-04-17T13:14:16 so, this isosurface code is not yet integrated into the simulation but it will be 2013-04-17T13:14:30 possibly i will have to write one for the GPU too 2013-04-17T13:14:53 to spare pciexpress bandwith... and transfer only the isosurface to the host 2013-04-17T13:18:29 google doesnt love me anymore 2013-04-17T13:18:46 yesterday, for a certain term it showed my image as the first result 2013-04-17T13:18:57 *** Accoun has joined #aichallenge 2013-04-17T13:21:27 mcstar: don't you aware of some computer projects in myrmecology? 2013-04-17T13:22:39 aren't you 2013-04-17T13:22:45 no, im not 2013-04-17T13:22:52 i dont even familiar with that term 2013-04-17T13:22:55 pff 2013-04-17T13:22:58 ok 2013-04-17T13:22:58 im not even* 2013-04-17T13:23:01 ;p 2013-04-17T13:23:11 thestinger: i changed my mind midway... 2013-04-17T13:23:41 mcstar: the part of your brain that thinks in english arguing with the part that doesn't? :) 2013-04-17T13:24:10 i wanted to write: i dont even know that term 2013-04-17T13:24:19 but changed to im not even familiar... 2013-04-17T13:24:54 UncleVasya: was that some kind of joke, that we all programmed 'ants' ? 2013-04-17T13:28:22 i should have made a screenshot yesteday 2013-04-17T13:28:39 *** UncleVasya has quit IRC (Ping timeout: 256 seconds) 2013-04-17T13:29:02 irc >:) 2013-04-17T13:29:09 google showed as the first result one of my plots and then subsequently several others whe searching for 'bcc interface free energy anisotropy' 2013-04-17T13:30:59 *** UncleVasya has joined #aichallenge 2013-04-17T13:31:54 UncleVasya: so, would you explain why did you ask about myrmecology? 2013-04-17T13:32:52 mcstar: i thought there is a good chance you know something interesting :) 2013-04-17T13:33:50 btw, this is the group homepage http://www.szfki.hu/~grana/crystal.html 2013-04-17T13:33:53 *** amstan has quit IRC (Ping timeout: 245 seconds) 2013-04-17T13:35:50 Phase field <--- sounds like technology in space game for me 2013-04-17T13:36:52 yeaaah, it is just on order parameter field, and a field is nothing like a force field... so bad terminology as far as scifi goes 2013-04-17T13:37:59 UncleVasya: what was this challenge aerique posted? i totally forgot its name, it start with an N 2013-04-17T13:38:05 are you participating? 2013-04-17T13:38:30 i cant even guess how much 10 bitcoins worth in real money 2013-04-17T13:38:50 mcstar: according to their website its about 2k$ 2013-04-17T13:39:00 o.O 2013-04-17T13:39:05 mcstar: nodewar.com 2013-04-17T13:39:08 ah 2013-04-17T13:39:17 no, I'm not participiating. 2013-04-17T13:40:04 But I've registered and did watch replays. 2013-04-17T13:40:07 " we reserve the right to pay $2,000USD instead." 2013-04-17T13:40:21 didnt they mean 20 dollars? 2013-04-17T13:40:27 2k is a lot 2013-04-17T13:40:28 best open source bot is now at 5th place or something 2013-04-17T13:40:46 and this is paid in the alpha stage 2013-04-17T13:40:57 i mean, this seems too much for me 2013-04-17T13:41:19 so maybe 200$ 2013-04-17T13:41:23 20 2013-04-17T13:41:26 XD 2013-04-17T13:41:36 UncleVasya: so, whats holding you back? 2013-04-17T13:42:23 UncleVasya: the winning strategy for you, is to pick Rust, and write a bot while learning it 2013-04-17T13:42:37 mcstar: i'm bad in physics, geometry, etc. 2013-04-17T13:43:40 and also I have some programming stuff that is more interesting and still I'm too lazy to do even this stuff. 2013-04-17T13:44:14 like ocaml card game from september and Game of Life bot for hackerrank 2013-04-17T13:44:26 i think you have time management issues 2013-04-17T13:44:41 listen to david allen: getting things done recordings 2013-04-17T13:44:50 nah, everything is ok :) 2013-04-17T13:45:06 doitfuckingnow.com :D 2013-04-17T13:45:10 you know best, boss 2013-04-17T13:45:19 let me check that out 2013-04-17T13:45:34 rofl 2013-04-17T13:45:51 it does exist, i though you made it up 2013-04-17T13:45:55 thought 2013-04-17T13:46:05 Steve Nojobs, CEO of iPoor 2013-04-17T13:46:58 it's not like I can't handle time well. I can't force myself to do things. I start and soon I switch to movie/game. Even for interesting things. 2013-04-17T13:47:29 so the problem is not to start but to keep going. 2013-04-17T13:47:42 i know the situation, i was like that once 2013-04-17T13:48:01 well, I'm making progress. do some work for android project :) 2013-04-17T13:48:21 the car thing? 2013-04-17T13:48:47 UncleVasya: whats your email address? 2013-04-17T13:50:02 yep. yesaterday Denis found one more of our program analogs. Before it was 'don't mind, we can do better' and now it's 'damn, this app looks good' 2013-04-17T13:50:06 nvm, i just wanted to submit it on that 'blog' 2013-04-17T13:50:57 UncleVasya: you could make a similar app, but for flying car drivers 2013-04-17T13:51:11 couple years, and they will be needing it 2013-04-17T13:51:16 maybe decades 2013-04-17T13:51:30 anyway, you guys would be the first to provide such an app 2013-04-17T13:52:24 in fact we already support flying cars 2013-04-17T13:52:27 i have to make a note that marco hasnt been around for quite a long time 2013-04-17T13:52:37 i guess he got productive 2013-04-17T13:52:52 UncleVasya: great! 2013-04-17T13:53:00 * UncleVasya wants to google him 2013-04-17T13:53:19 just remember, when you market the thing, you stress the support for flying cars :) 2013-04-17T13:54:04 he has a commit at the end of february 2013-04-17T13:54:34 *** sigh has quit IRC (Remote host closed the connection) 2013-04-17T13:56:10 thestinger: how familiar are you with conflict detection/resolution in revisioning systems? 2013-04-17T13:56:24 as in git? 2013-04-17T13:56:34 as in any distributed system 2013-04-17T13:56:49 i think git is quite primitive in this sense 2013-04-17T13:57:04 like an eventual consistency db? 2013-04-17T13:57:18 they just timestamp stuff and throw out the conflicts with newer timestamps :P 2013-04-17T13:57:26 what i was interested in recent days, is that how a textual conflict detection can cause semantic changes/bugs 2013-04-17T13:58:08 these things merge changes if lines do not conflict 2013-04-17T13:58:22 i think thats a poor criterion 2013-04-17T13:58:38 *** Accoun has quit IRC (Remote host closed the connection) 2013-04-17T13:58:53 i was actually surprised by this, and went googling, if im the stupid one, or this problem does exist 2013-04-17T13:59:31 darcs has something called 'cherry picking', ill check it out 2013-04-17T13:59:49 mcstar: so does git 2013-04-17T14:00:14 mcstar: git can apply only parts of a patch 2013-04-17T14:00:51 mcstar: a commit is just generating a patch of the old HEAD to the index, so 'git commit -p' is the same thing 2013-04-17T14:01:03 but git doesnt have any semantic meaning of the code, so, it cant safely apply something at one point, and be sure it doesnt affect some kind of global consistency 2013-04-17T14:01:30 yep, you can stage/unstage/reset selected lines instead of whole file 2013-04-17T14:01:55 *** Accoun has joined #aichallenge 2013-04-17T14:02:21 mcstar: anyway, that's what unit tests are for! :P 2013-04-17T14:02:28 i heard that before 2013-04-17T14:02:35 and it is a sin to say that 2013-04-17T14:02:35 nothing should be merged to master in a serious project before being tested 2013-04-17T14:02:41 (in my eyes at least) 2013-04-17T14:02:43 mcstar: why ;p 2013-04-17T14:02:53 if it's not tested it's going to break anyway 2013-04-17T14:03:16 im not a software engineer and i dont believe in unit tests 2013-04-17T14:03:21 that's why rust's REPL is broken, it's one of the few things without regression tests :) 2013-04-17T14:03:39 if i were to write a serious program, i would choose a language that you can only validate with unit tests 2013-04-17T14:03:47 i wouldnt* 2013-04-17T14:03:56 natural language sucks... 2013-04-17T14:05:08 mcstar: are you talking about case when DeveloperA edit partA of file and DevB edits partB of file. DevB delets something in PartB and partA can't work without it. But for git these changes isn't a conflict because edits done in different parts of file. right? 2013-04-17T14:05:19 it feels to me like software is growing out of control, and the development cant be properly isolated and divided among developers 2013-04-17T14:05:46 mcstar: not *just* unit tests 2013-04-17T14:06:01 UncleVasya: afaik that can happen in the same file also, but yeah, between files it totally goes undetected 2013-04-17T14:06:35 mcstar: strong typing and memory safety, but *also* unit tests ;P 2013-04-17T14:06:48 mcstar: if you have a 200k line project developed collaboratively, stuff is going to break 2013-04-17T14:06:56 some sweeping change across the codebase to migrate to a new API will do it 2013-04-17T14:07:22 mcstar: in git you must pull from server latest version of code before you can push your changes. 2013-04-17T14:07:23 thestinger: imagine, that the revision control system had a typechecker 2013-04-17T14:07:41 mcstar: rust's revision control does ;P, only a CI bot merges changes 2013-04-17T14:07:44 it wouldnt run unit tests(maybe it would idk) but it could enforce some consistentcy 2013-04-17T14:07:47 tency* 2013-04-17T14:07:51 it forces you to check your code works with this version. 2013-04-17T14:08:09 so if someone broke something you depend on you'll notice. 2013-04-17T14:09:55 person A pulls master, B pushes change, A edits file meanwhile, tried to push, cant, has to pull, pulls, when it does that, does he get notified of changes that B introduced? 2013-04-17T14:10:44 mcstar: yes it will make you deal with the merge conflict 2013-04-17T14:10:53 unless you aren't a guy who doesn't ensures his code works before submitting 2013-04-17T14:10:55 mcstar: if there wasn't a merge conflict, it would let you push. 2013-04-17T14:10:58 but there isnt necessarily a conflict! 2013-04-17T14:11:03 *unless you are 2013-04-17T14:11:08 mcstar: which is why you don't push to master 2013-04-17T14:11:28 no one should push to master, the CI software should be doing that 2013-04-17T14:11:28 mcstar: yes, there can be no conflict in terms of git. 2013-04-17T14:12:06 but to ensure there isn't a conflict in semantics you run your code right before you use push button. 2013-04-17T14:12:14 thestinger: but the CI (continuous integration ?) can miss this scenario, of there is no such test to detect this particular bug 2013-04-17T14:12:25 of->if 2013-04-17T14:12:27 mcstar: https://github.com/mozilla/rust/pull/5913 pull request -> a reviewer reviews it and comments r+ -> bot merges to a branch, builds, tests, then fast-forwards trunk 2013-04-17T14:12:40 *** wentam has quit IRC (Ping timeout: 272 seconds) 2013-04-17T14:13:01 except the buildbot is off atm ;p 2013-04-17T14:13:21 how often are there breakages in a large project with CI? 2013-04-17T14:13:35 i cant believe that this scheme always works 2013-04-17T14:13:39 mcstar: there aren't, if you qualify a breakage as breaking tests 2013-04-17T14:13:40 or even, most of the time 2013-04-17T14:13:46 no 2013-04-17T14:13:49 if it doesn't make tests fail nothing is broken :D 2013-04-17T14:13:57 breakage, is when this goes undetected 2013-04-17T14:14:10 well if you weren't testing for it, you didn't expect it to work 2013-04-17T14:14:18 if it doesn't make tests fail rewrite tests :D 2013-04-17T14:14:43 mcstar: https://www.sqlite.org/testing.html *this* is how you develop software properly 2013-04-17T14:15:02 ;P 2013-04-17T14:15:12 ok 2013-04-17T14:15:16 for real? 2013-04-17T14:15:32 mcstar: ok here's an example 2013-04-17T14:15:41 some library the project uses changes the API 2013-04-17T14:15:43 i meant, are you serious? 2013-04-17T14:15:47 is it a good example? 2013-04-17T14:15:48 someone has to go over the entire codebase and port to the new API 2013-04-17T14:15:54 mcstar: sqlite? yes 2013-04-17T14:16:00 it's like the most sanely developed software out there 2013-04-17T14:16:16 they have an enormous amount of testing/fuzzing 2013-04-17T14:16:38 sqlite is pretty primitive isnt it? 2013-04-17T14:16:54 not in functionality, but it is a disk-based database 2013-04-17T14:17:01 or file based 2013-04-17T14:17:32 it is not even a daemon with multiple concurrent connections to it 2013-04-17T14:17:44 it has concurrent users of the file 2013-04-17T14:17:56 but basically it is a lib 2013-04-17T14:17:59 not a service 2013-04-17T14:18:03 with WAL, at least 2013-04-17T14:18:03 afaik 2013-04-17T14:18:06 yeah 2013-04-17T14:18:13 a lib and a cli utility 2013-04-17T14:19:06 they have a simple query parser/compiler/optimizer, the data structure implementation (b-tree in sqlite3, lsm-tree in sqlite4) 2013-04-17T14:19:51 ... of C code 2013-04-17T14:20:10 mcstar: ? 2013-04-17T14:20:21 By comparison, the project has 1124 times as much test code and test scripts - 91421.1 KSLOC. 2013-04-17T14:20:24 wtf 2013-04-17T14:20:31 i hope they are mostly auto generated 2013-04-17T14:20:35 mcstar: some is 2013-04-17T14:20:57 mcstar: a lot of companies have invested in the tests 2013-04-17T14:21:11 thestinger: well, it is written in C so ofc, you have to be strict and test thoroughly 2013-04-17T14:21:25 mcstar: you still have to be as strict with haskell. 2013-04-17T14:21:31 the data races they are testing for are mostly external 2013-04-17T14:21:49 i dont necessarly want haskell 2013-04-17T14:22:03 (i do, but i dont want such projects to necessarlily use haskell) 2013-04-17T14:22:06 and they are handling out-of-memory in every case 2013-04-17T14:22:14 and bubbling it up 2013-04-17T14:22:17 for embedded stuff 2013-04-17T14:22:37 Small. Fast. Reliable. 2013-04-17T14:22:39 Choose any three. 2013-04-17T14:22:43 thats funny 2013-04-17T14:23:53 mcstar: and they don't make backwards incompatible changes 2013-04-17T14:24:01 sqlite3 has been backwards compatible since release 2013-04-17T14:24:14 they add _v2 APIs :P 2013-04-17T14:24:22 and sqlite4 will drop them all 2013-04-17T14:24:25 one day! 2013-04-17T14:24:27 thestinger: anyway, here comes c++ for the rescue, makes writing software a breeze http://www.vtk.org/doc/release/5.10/html/classvtkUnstructuredGrid.html 2013-04-17T14:25:23 *** Scooper has joined #aichallenge 2013-04-17T14:25:43 if you were wondering what is that you are seeing, it is object oriented programming 2013-04-17T14:30:18 *** UncleVasya has quit IRC (Read error: Connection reset by peer) 2013-04-17T14:31:26 *** UncleVasya has joined #aichallenge 2013-04-17T14:50:29 Valgrind only works reliably on Linux, which in the opinion of the SQLite developers means that Linux should be the preferred platform for all software development. 2013-04-17T14:52:22 * UncleVasya thinks that minecraft is the prefferd platform for all software development 2013-04-17T14:52:45 it's Turing complete afaik 2013-04-17T14:54:02 and people do amazing things in it like computer design. 2013-04-17T14:54:32 * UncleVasya never played minecraft himself 2013-04-17T14:54:42 game of life might be turing complete too 2013-04-17T14:54:52 neither did i 2013-04-17T14:55:05 DwarfFortress is pretty cool. Sad it's not open source. 2013-04-17T14:57:48 Static analysis has not proven to be especially helpful in finding bugs in SQLite. Static analysis has found a few bugs in SQLite, but those are the exceptions. More bugs have been introduced into SQLite while trying to get it to compile without warnings than have been found by static analysis. 2013-04-17T14:57:50 wtf 2013-04-17T14:58:21 i wonder how far they have gotten with like the first version of C 2013-04-17T14:58:27 or B 2013-04-17T14:58:36 thestinger: ^ 2013-04-17T14:59:11 ;p 2013-04-17T14:59:39 mcstar: static analysis is pretty terrible 2013-04-17T14:59:46 99% of the warnings are false positives 2013-04-17T14:59:58 strong typing and memory safety in the language itself is *so* much saner 2013-04-17T15:00:09 but i think static analysis is a broader term than they are using it there 2013-04-17T15:00:30 strong typing is a kind of static analysis 2013-04-17T15:00:33 isnt it? 2013-04-17T15:00:48 mcstar: well not really what people mean by static analysis 2013-04-17T15:00:58 by static analysis they mean heuristics 2013-04-17T15:01:03 like ghc -Wall ;p 2013-04-17T15:01:12 compiler warnings are essentially static analysis 2013-04-17T15:01:21 it could be something wrong, but it's "valid" in the language 2013-04-17T15:01:31 compiler errors are too, why wouldnt they be? 2013-04-17T15:01:59 mcstar: because that's part of just making it coherent enough to compile 2013-04-17T15:02:07 * thestinger shrugs 2013-04-17T15:02:39 it feels to me, warning cant be made errors, since they would break code 2013-04-17T15:02:47 s* 2013-04-17T15:03:01 thats the only reason they arent errors 2013-04-17T15:04:28 thestinger: anyway, my conclusion after reading the sqlite test harness, is that i need to learn tcl to be effective in testing 2013-04-17T15:04:43 mcstar: lol ;p 2013-04-17T15:04:52 mcstar: they just use tcl because it's old 2013-04-17T15:05:01 i was kidding 2013-04-17T15:05:04 if it was done today it would probably be python 2013-04-17T15:05:21 the sqlite author likes tcl though 2013-04-17T15:05:24 for whatever reason 2013-04-17T15:05:33 because he knows it? 2013-04-17T15:05:47 looks horrible though 2013-04-17T15:06:14