00:07:41 -!- sebbu2 has joined. 00:08:01 -!- sebbu2 has quit (Changing host). 00:08:01 -!- sebbu2 has joined. 00:09:24 -!- sirdancealot7 has quit (Remote host closed the connection). 00:10:58 -!- sebbu has quit (Ping timeout: 245 seconds). 00:18:47 -!- GreyKnight has quit (Ping timeout: 255 seconds). 00:22:19 -!- GreyKnight has joined. 00:35:39 -!- Nisstyre-laptop has quit (Read error: Connection reset by peer). 00:39:11 -!- GreyKnight has quit (Ping timeout: 252 seconds). 00:42:37 -!- Vorpal has joined. 00:44:07 -!- GreyKnight has joined. 00:53:12 -!- GreyKnight has quit (Ping timeout: 276 seconds). 00:53:27 That's cool: 00:53:31 "'2013' is non-prime in any base b: 2013_b = 2b³ + b + 3 = (b+1)(2b² - 2b + 3)" 00:55:07 -!- GreyKnight has joined. 00:59:36 Sgeo: that's essentially the division by 11 test when the result is exactly 0 so doesn't need to be repeated 01:00:49 hi oerjan: thanks for the notes 01:00:56 yw 01:01:41 `run quote 652 > paste 01:01:42 bash: paste: Is a directory 01:01:55 `bin/run quote 652 > paste 01:01:56 â€‹/home/hackbot/hackbot.hg/multibot_cmds/lib/limits: line 5: /hackenv/bin/run: No such file or directory \ /home/hackbot/hackbot.hg/multibot_cmds/lib/limits: line 5: exec: /hackenv/bin/run: cannot execute: No such file or directory 01:02:30 hagb4rd: actually understanding how shell commands work is also recommended :P 01:03:01 `help 01:03:01 Runs arbitrary code in GNU/Linux. Type "`", or "`run " for full shell commands. "`fetch " downloads files. Files saved to $PWD are persistent, and $PWD/bin is in $PATH. $PWD is a mercurial repository, "`revert " can be used to revert to a revision. See http://codu.org/projects/hackbot/fshg/ 01:03:03 `run echo hi 01:03:05 hi 01:03:13 oh 01:03:14 `revert 01:03:17 GreyKnight: you scoundrel 01:03:17 Done. 01:03:23 er 01:03:23 Gregor: 01:03:50 just want to know if i can use pipe to output sth on the pastesite returning a link 01:04:02 `bin/run echo "i thought he said it wouldn't work anyway" 01:04:03 i thought he said it wouldn't work anyway 01:04:16 hagb4rd: yes but > is not the correct syntax for that 01:04:22 lol 01:04:23 ok 01:04:28 you're right 01:04:33 is it just paste? 01:04:43 oerjan: i convinced him it would 01:04:45 `bin/run quote 652 | paste 01:04:49 http://codu.org/projects/hackbot/fshg/index.cgi/raw-file/tip/paste/paste.22907 01:04:52 great 01:04:54 I didn't do nothing! 01:05:23 `bin/run quote 692 | paste 01:05:27 http://codu.org/projects/hackbot/fshg/index.cgi/raw-file/tip/paste/paste.26677 01:06:25 elliott: i meant the `bin/run thing 01:06:45 hagb4rd: you don't actually need the bin/ part, that was just part of an elliott experiment 01:07:16 oerjan: yes i convinced him it'd be ok 01:07:34 `/bin/echo hi 01:07:35 hi 01:08:12 elliott: yes but you played around with a lot of env stuff, because he said something about only alphanums, which apparently isn't true 01:08:22 you can just use absolute paths 01:09:02 oerjan: ??? 01:09:06 i have no idea wtf you are talking about 01:09:12 ...which should be obvious anyway given the existence of `? 01:10:22 21:07:29: But `/hackenv/run wouldn't work. 01:10:22 21:07:33: why not 01:10:22 21:07:34: Because it's a-x 01:10:30 this is what i'm referring to 01:10:32 i'm not sure if hackego first tries to `run as builin function before looking up in /bin 01:10:38 i guess it does 01:10:51 oerjan: ...do you know what a-x means 01:10:55 hagb4rd: it does, although they just discussed whether to change that 01:11:02 oh 01:11:07 oerjan: (permissions) 01:11:13 -!- sirdancealot7 has joined. 01:12:18 elliott: i misunderstood what that was all about then, and you didn't help much by never actually testing `bin/run :P 01:12:29 which of course didn't make sense anyway. hm... 01:12:41 `run cp bin/run . 01:12:44 -!- sebbu has joined. 01:12:45 No output. 01:13:00 `./run echo test 01:13:01 test 01:13:05 -!- sebbu has quit (Changing host). 01:13:05 -!- sebbu has joined. 01:13:36 ais523: Speaking of those tape heatmaps, I told the hierarchical-cluster plot -- http://zem.fi/egostats/plot_cluster.png -- that used scores for distances to do the same thing except with Euclidean distance between normalized average (across duels, but not tape length/polarity) tape heatmaps, and it spit out http://zem.fi/egostats/plot_heatclust.png -- don't know if those are "more ... 01:13:42 ... appropriate" strategywise or anything. 01:13:43 oh hm 01:13:50 i was still misunderstanding 01:14:56 -!- GreyKnight has quit (Ping timeout: 255 seconds). 01:15:13 * oerjan _thinks_ he's got it now, but better keep his mouth shut :P 01:15:58 -!- sebbu2 has quit (Ping timeout: 245 seconds). 01:18:38 -!- nooga has quit (Ping timeout: 255 seconds). 01:18:46 -!- GreyKnight has joined. 01:19:14 oerjan: if it's any consolation I understand less of what you say 01:19:23 yay! 01:21:06 elliott: i somehow managed to misinterpret the conversation to get into my head that ` commands couldn't contain '/' , which somehow seemed to fit the rest of your experimentation 01:21:18 (back earlier today) 01:21:23 *tonight 01:23:44 -!- GreyKnight has quit (Ping timeout: 252 seconds). 01:25:00 mm 01:25:52 -!- GreyKnight has joined. 01:26:24 -!- GreyKnight has quit (Client Quit). 01:36:57 -!- Frooxius has quit (Ping timeout: 240 seconds). 01:40:13 -!- carado has quit (Ping timeout: 246 seconds). 01:40:42 -!- Nisstyre has joined. 01:41:28 Hmm 01:41:40 I am not amused by DrRacket saving files in an opaque way 01:41:40 https://github.com/mflatt/scratchy/blob/master/scratchy/examples/fish.rkt 01:48:00 `fetch https://www.googleapis.com/customsearch/v1?key=AIzaSyBfdt4VN9qT_Y8LAl100TBRCWAOiSTaqSY&cx=009459769908964694922:yb5esesqxmg&q=robocop&searchType=image&fileType=jpg&imgSize=xlarge&alt=json&safe=off 01:48:03 2013-01-02 01:48:02 URL:https://www.googleapis.com/customsearch/v1?key=AIzaSyBfdt4VN9qT_Y8LAl100TBRCWAOiSTaqSY&cx=009459769908964694922:yb5esesqxmg&q=robocop&searchType=image&fileType=jpg&imgSize=xlarge&alt=json&safe=off [10585] -> "v1?key=AIzaSyBfdt4VN9qT_Y8LAl100TBRCWAOiSTaqSY&cx=009459769908964694922:yb5esesqxmg&q=robocop&searchType=image&fileTy 01:48:32 `fetch https://www.googleapis.com/customsearch/v1?key=AIzaSyBfdt4VN9qT_Y8LAl100TBRCWAOiSTaqSY&cx=009459769908964694922:yb5esesqxmg&q=robocop&searchType=image&fileType=jpg&imgSize=xlarge&alt=json&safe=off | paste 01:48:32 https://www.googleapis.com/customsearch/v1?key=AIzaSyBfdt4VN9qT_Y8LAl100TBRCWAOiSTaqSY&cx=009459769908964694922:yb5esesqxmg&q=robocop&searchType=image&fileType=jpg&imgSize=xlarge&alt=json&safe=off%20%7C%20paste: \ 2013-01-02 01:48:32 ERROR 400: Bad Request. 01:49:02 `ls bin 01:49:04 â€‹? \ @ \ ï¼·ï¼¥ï¼¬ï¼£ï¼¯ï¼ï¼¥ \ addquote \ allquotes \ anonlog \ calc \ define \ delquote \ etymology \ forget \ fortune \ frink \ google \ hatesgeo \ interp \ joustreport \ jousturl \ json \ karma \ karma- \ karma+ \ learn \ log \ logurl \ lua \ luac \ luarocks \ luarocks-admin \ macro \ maketext \ marco \ No \ pastaquote \ paste \ pastefortune 01:49:26 `run ls v* 01:49:28 v1?key=AIzaSyBfdt4VN9qT_Y8LAl100TBRCWAOiSTaqSY&cx=009459769908964694922:yb5esesqxmg&q=robocop&searchType=image&fileType=jpg&imgSize=xlarge&alt=json&safe=off 01:50:01 `run mv v* robocop 01:50:05 No output. 01:50:13 `file robocop 01:50:17 robocop: ASCII English text, with very long lines 01:50:36 fizzie: looking now 01:50:56 err, hmm 01:51:00 it's better in some places 01:51:02 but worse in others 01:51:11 ^url robocop 01:51:19 like, my turtles are correctly grouped close to each other 01:51:20 `url robocop 01:51:25 http://codu.org/projects/hackbot/fshg/index.cgi/raw-file/tip/robocop 01:51:32 but space_hotel is in there too and it works entirely differently 01:51:33 what the hell 01:52:00 i assume that's what they call json 01:53:18 hagb4rd: in any case you can use `url instead of paste when the file is already in HackEgo; paste doesn't do anything other than making a copy and giving the result of url on that 01:53:43 so the first command worked? 01:53:59 yes 01:54:13 how is it named robocop anyway? 01:54:33 i missed that point 01:55:18 hagb4rd: i just used mv 01:55:36 (or to put it this way: where do i find the result of `fetch..) 01:55:36 ok 01:55:47 `fetch gave it a long name containing everything after the last / in the url 01:56:19 and that's what HackEgo gave after the -> in the `fetch response 01:56:21 oerjan: what's the verdict on free monads 01:56:58 shachaf: free monads are cool, they feel more reified somehow 01:57:19 could we have an alias for that command taking 1 argument passed as query string? do we have any conventions on creating such aliases? 01:57:31 oerjan: what about cofree comonads 01:57:38 shachaf: no f idea 01:57:45 are the ol? 01:57:50 s/the/they/ 01:58:04 oerjan: Cofree is just like Free except it uses (,) instead of Either. 01:58:11 hagb4rd: `fetch is a builtin command which can only be used alone 01:58:19 So Cofree f a = (a, f (a, f (a, f (... 01:58:36 hagb4rd: i.e. you cannot use it from within other HackEgo commands 02:00:08 hagb4rd: there _is_ a system for contacting a whitelist of web sites from within HackEgo commands, but it is currently broken 02:00:31 is that a convention to be satisfied`? l just would like to have some aliases binding to webservices.. which imho would be a nice method extending things ya know 02:00:50 hagb4rd: we used to have some 02:01:17 but the api's changed, so many of them broke again, and then the web proxy itself broke 02:01:28 only Gregor can fix it 02:02:21 `url https://www.googleapis.com/customsearch/v1?key=AIzaSyBfdt4VN9qT_Y8LAl100TBRCWAOiSTaqSY&cx=009459769908964694922:yb5esesqxmg&q=strawberry&searchType=image&fileType=jpg&imgSize=xlarge&safe=off | fetch 02:02:22 http://codu.org/projects/hackbot/fshg/index.cgi/raw-file/tip/https://www.googleapis.com/customsearch/v1?key=AIzaSyBfdt4VN9qT_Y8LAl100TBRCWAOiSTaqSY&cx=009459769908964694922:yb5esesqxmg&q=strawberry&searchType=image&fileType=jpg&imgSize=xlarge&safe=off | fetch 02:02:44 ls v* 02:02:57 meh 02:04:57 `run ls v* 02:04:58 ls: cannot access v*: No such file or directory 02:05:14 -!- WeThePeople has joined. 02:05:17 where is the output of url going to? 02:05:55 shouldn't be the stdout? 02:06:42 hagb4rd: url doesn't fetch anything, it takes a filename already existing in HackEgo and tells you where to browse it on Gregor's website 02:06:56 also, you cannot combine fetch with anything, i said 02:08:19 hagb4rd: basically url doesn't do anything other than prepending http://codu.org/projects/hackbot/fshg/index.cgi/raw-file/tip/ to its argument 02:08:59 -!- TeruFSX has joined. 02:10:15 so what is the bullshit fetch is sending to the stdout instead of the response? 02:10:35 basically fetch, like run and help, are builtin HackEgo commands that aren't looked up inside the sandbox. and unlike the two others fetch cannot be emulated fully. 02:10:52 can i just use wget instead? 02:11:01 or is it restricted too 02:11:41 hagb4rd: timestamp the://UrlYouGaveIt [size] -> theFileNameWhereItPutIt 02:12:06 wget would be restricted, and currently broken because the web proxy from inside the sandbox is broken 02:12:28 `run which wget 02:12:30 â€‹/usr/bin/wget 02:12:49 `wget http://www.google.com 02:12:50 â€‹--2013-01-02 02:12:50-- http://www.google.com/ \ Connecting to 127.0.0.1:3128... failed: Connection refused. 02:13:25 iirc google.com is in the whitelist, so should work if the proxy does 02:13:39 @quote Jafet 02:13:39 Jafet says: Help, I'm trapped in this channel I tried to escape but ghc said occurs check 02:13:55 it would be SOOOO NICE to use REST services :( 02:14:31 hagb4rd: you'll just have to nag Gregor to get the proxy working again 02:15:44 i'll try too, if you share my opinion on that 02:16:02 which i think is the case 02:16:10 i think it would be nice to get some basic searches working 02:16:29 i'd like to go further 02:16:45 although lambdabot has a few relevant commands 02:17:10 and just be able to write my own services as embedded extensions 02:17:18 parked elsewhere 02:17:39 that was just a test in first case 02:18:03 -!- Vorpal has quit (Ping timeout: 245 seconds). 02:18:18 imagesearch is indeed not that relelvant in this channel 02:18:23 i think 02:18:24 :) 02:22:10 -!- zzo38 has joined. 02:22:43 If I have $z=12^xy^2$ in natural numbers, then what is an algorithm to figure out x and y when given z? 02:23:35 (Assume y is nonzero, but x is possible to be zero.) 02:24:12 you mean (12^x)(y^2) or 12^(xy)^2 or ...? 02:24:23 Fiora: the former in tex 02:24:29 ah 02:24:43 this way we wouldn't need to have more bots joining the channel and so on.. just a clean interface as defined by a webservice.. the output goes to stdout and can be piped elsewhere.. but then we should agree on some conventions on alias/script names so we won't get lost (if we not already are) 02:25:24 zzo38: something like while (z % 12 == 0) { z /= 12; x++ } y = sqrt (z) 02:26:04 oerjan: I am not sure that works. 02:26:40 oh hm 02:27:12 And it is needed one for computers that doesn't have square root and division and so on 02:27:24 zzo38: well ok you get _one_ solution that way, others will be given by x - 2n and y*144^n 02:27:47 oerjan: Yes, OK 02:28:05 zzo38: oh hm right there's the question of whether x or x+1 makes the remaining z a square 02:28:15 Yes 02:28:20 That is what I meant. 02:28:48 zzo38: but anyway starting with the above, the rest is a matter of finding the square root of the remaining part of z or 12*z 02:29:41 int y; 02:29:41 int x = 0; 02:29:41 while(1) 02:29:41 { int s = isqrt(z); if(s*s == z) {y = s; break;} if(z%12 != 0) goto fail; x++; z /= 12; 02:29:44 } 02:30:27 Fiora: oh hm right, that would work to halt as early as possible instead 02:30:41 to give the _largest_ solution for y instead 02:31:33 Fiora: in fact you could improve that by only testing twice, since if it doesn't become a square after dividing by 12 and wasn't one already, then it never will become one 02:32:02 makes sense 02:34:00 zzo38: so if you don't have a solution with x==0 or x==1 then you don't have any at all 02:34:46 but if y is divisible by 12 you will get additional ones 02:35:45 oerjan: O, OK. I may need to change some things then, because I need to support at least up to x==2 02:36:43 if x==2, y==n are solutions, then x==0, y==12*n will also be 02:37:54 OK. 02:38:04 I can see that. 02:45:48 --help display this help and exit 02:45:49 --version output version information and exit 02:45:49 --bibliography output recommended readings and exit 02:46:09 what program is that 02:46:24 pi 02:51:35 version 3.14159265358979 02:52:29 it is a program that outputs pi to a specified accuracy? 02:53:00 is it TeX who uses successive approximations of pi as version numbers? 02:53:06 Yes. 02:53:14 Yes, TeX uses that. 02:53:26 I currently have TeX version 3.1415926. 02:53:31 zzo38: What does TeXnicard use? 02:53:44 When Knuth is dead, the final official version shall be updated to version $\pi$ exactly. 02:53:52 shachaf: Just ordinary version numbers. 02:54:05 METAFONT uses approximations of the base of natural logarithms. 02:54:26 I should write a program whose version number approximates 2. 02:54:35 zzo38: you don't just call it e? 02:55:01 quintopia: It is called e, yes. 02:55:06 > (exp.exp.exp.exp)1 02:55:08 Infinity 02:55:24 monqy: what's the verdict on free monads 02:55:28 hi 02:55:34 hi 02:55:51 Penrose uses upright rather than italic "e" for the base of natural logarithms and I agree with him; so it would be ${\rm e}$ in TeX. 02:55:55 > (exp.exp.exp.exp) 1 :: CReal 02:56:01 mueval: ExitFailure 1 02:56:01 mueval: Prelude.undefined 02:56:12 Same with imaginary unit, and derivative operator. 02:57:24 However, the lowercase Greek in Computer Modern is only italic, and I am not entirely sure how I might change them to be able to work with not italic lowercase Greek as well, for pi and so on, but I might be able to try. 02:57:32 zzo38: then you can just say e instead of "the base of natural logarithms" because we will know what you mean AND...you don't have to type as many words! 02:58:03 quintopia: I can type fast. 02:58:27 me too. "fast" 02:59:12 oerjan: Can you unkarma me? 02:59:23 The whole mwahahaha deal. 03:02:11 zzo38: Does TeXnicard support TeXnicolour? 03:02:39 shachaf-- 03:02:51 I did play Dungeons&Dragons game today; making screaming fire and greasy floor did help to earn some time, a bit. I also put the wand I stole underneath the door they were going to break down, before teleporting to outside; therefore it activated when they broke the door. 03:03:24 ^ul ((@karma- shachaf)S:^):^ 03:03:25 @karma- shachaf@karma- shachaf@karma- shachaf@karma- shachaf@karma- shachaf@karma- shachaf@karma- shachaf@karma- shachaf@karma- shachaf@karma- shachaf@karma- shachaf@karma- shachaf@karma- shachaf@karma- shachaf@karma- shachaf@karma- shachaf@karma- shachaf@karma- shachaf@karma- shachaf@karma- shachaf@karma- shachaf@karma- s ...too much output! 03:03:25 shachaf@karma-'s karma lowered to -1. 03:03:31 oops 03:03:32 good job 03:03:44 ^ul ((@karma- shachaf )S:^):^ 03:03:45 @karma- shachaf @karma- shachaf @karma- shachaf @karma- shachaf @karma- shachaf @karma- shachaf @karma- shachaf @karma- shachaf @karma- shachaf @karma- shachaf @karma- shachaf @karma- shachaf @karma- shachaf @karma- shachaf @karma- shachaf @karma- shachaf @karma- shachaf @karma- shachaf @karma- shachaf @karma- shachaf @kar ...too much output! 03:03:45 shachaf's karma lowered to 72. 03:03:56 er wait 03:04:03 ^ul ((shachaf-- )S:^):^ 03:04:03 shachaf-- shachaf-- shachaf-- shachaf-- shachaf-- shachaf-- shachaf-- shachaf-- shachaf-- shachaf-- shachaf-- shachaf-- shachaf-- shachaf-- shachaf-- shachaf-- shachaf-- shachaf-- shachaf-- shachaf-- shachaf-- shachaf-- shachaf-- shachaf-- shachaf-- shachaf-- shachaf-- shachaf-- shachaf-- shachaf-- shachaf-- shachaf-- shac ...too much output! 03:04:12 @karma shachaf 03:04:12 shachaf has a karma of 40 03:04:13 shachaf: No; if reading a DVI file it will use its own commands to determine colours, or it will generate the nodes directly without reading an external DVI in which case it also uses its own commands for colors. 03:04:13 i think you can decrease karma multiple times w/ one message using @@ 03:04:39 elliott: well the last one worked 03:04:59 (Reading external DVI works (it calls TeX to typeset and then reads the created DVI); internal typesetting is not yet completely implemented but when it is, it will also be able to do that with many possible effects) 03:05:40 zzo38: dvi? hdmi is the future these days 03:06:28 shachaf: It is a different kind of DVI. (Anyways, HDMI is no good, so I prefer Digi-RGB, but this is not relevant here anyways.) 03:10:27 -!- fungot has quit (Ping timeout: 248 seconds). 03:10:38 -!- FireFly has quit (Excess Flood). 03:11:35 -!- FireFly has joined. 03:31:40 -!- Phantom_Hoover has quit (Read error: Connection reset by peer). 04:00:05 -!- oerjan has quit (Quit: Good night). 04:07:22 -!- FireFly has quit (Excess Flood). 04:08:21 -!- FireFly has joined. 04:09:16 ...ponies are not just baby horses. 04:09:24 `quote ponies 04:09:26 No output. 04:09:29 @quote ponies 04:09:29 No quotes match. That's something I cannot allow to happen. 04:10:10 If you add "...ponies are not just baby horses." then it will match. 04:23:57 "'Oompa Loompas' sought by police over Norwich assault" 04:26:34 They should add functions like http://msdn.microsoft.com/en-us/library/windows/desktop/aa366713(v=vs.85).aspx to glibc 04:27:06 oompa loompa doopity dunction 04:27:11 you do not talk about adding libc function 04:27:33 But how will I know if my pointers are valid? 04:29:59 BOOL WINAPI DoesFunctionHaltExA(HFUNPTR fpFunctionToTellWhetherItHaltsOrNot) 04:30:52 shachaf: you call fork() and dereference the pointer in a child process and then call wait() and see if the child died with SIGSEGV 04:30:55 IT IS TRIVIAL 04:31:39 But then it isn't portable to Windows. 04:32:00 monqy: edwardk is generalizing all of lens :'( 04:32:12 +findOf :: (RepresentableProfunctor p, Profunctor q, Comonad (Rep p)) => Overloading p q (Accessor (Endo (Maybe a))) s t a b -> p a Bool -> q s (Maybe a) 04:32:16 Is it portable to Cygwin though? 04:32:41 How does Cygwin implement fork()? 04:32:48 I know it does something but now how well it works. 04:32:55 Slowly, I think. 04:37:42 shachaf: I have no clue if you're serious or not about that type signature 04:38:06 lightquake: That's currently in lens HEAD 04:38:10 D: 04:38:16 20:37 lens/master 98245d6 Edward Kmett: more generalization, because that last patch wasn't general enough 04:38:29 Let's see what this latest one is. 04:38:33 that's an absolutely unreadable type signature 04:38:47 +imapMOf_ :: (Profunctor q, Monad m) => Overloading (Indexed i) q (Accessor (Sequenced m)) s t a b -> (i -> a -> m r) -> q s (m ()) 04:38:50 +iconcatMapOf :: Profunctor q => Overloading (Indexed i) q (Accessor [r]) s t a b -> (i -> a -> [r]) -> q s [r] 04:38:54 +ifindOf :: Profunctor q => Overloading (Indexed i) q (Accessor (Endo (Maybe a))) s t a b -> (i -> a -> Bool) -> q s (Maybe a) 04:38:57 +itoListOf :: Profunctor q => Overloading (Indexed i) q (Accessor (Endo [(i,a)])) s t a b -> q s [(i,a)] 04:39:00 Not too bad. 04:39:32 lightquake: I'm assuming this is an "it'll get worse before it gets better" sort of thing. 04:39:37 I have no clue how to read those type signatures 04:39:41 * kmc screams in terror 04:39:59 i can read them syntactically but i have no idea what all the classes and one-letter variables mean 04:40:12 yeah, that's what I meant 04:40:36 OK, the simple version is, uh... 04:40:45 Well, p and q are both just (->), normally. 04:40:53 Overloading is this one thing. 04:41:11 -!- hagb4rd has quit (Ping timeout: 252 seconds). 04:41:13 type Overloading p q f s t a b = p a (f b) -> q s (f t) 04:41:31 OK, you know what, this is hopeless. 04:41:45 In edwardk's defense, he has several "simple" versions of each type signature next to the "real" version. 04:42:01 -- 'findOf' :: 'Traversal'' s a -> (a -> 'Bool') -> s -> 'Maybe' a 04:42:16 yeah, but the errors would be awful 04:42:41 This is edwardk's "expansion" phase. 04:42:41 that oompa loompa story is like some terrible crossover of Willy Wonka and A Clockwork Orange 04:42:53 Where he generalizes everything and later figures out what sticks. 04:43:38 the perils of a rapidly-developed library, I guess 04:43:59 lightquake: Do you understand lens signatures in general? 04:44:10 not in this general! 04:44:13 Like Applicative f => (a -> f b) -> (a,a) -> f (b,b)? 04:44:15 -!- hagb4rd has joined. 04:44:47 i understand that you use the functor context so that you can run it with both Identity and Const and combine a getter and a setter 04:44:50 but not much beyond that 04:45:20 Well, that signature is just like mapM for tuples, right? 04:45:21 @ty mapM 04:45:22 Monad m => (a -> m b) -> [a] -> m [b] 04:45:40 right 04:47:45 So it's like that, except with Applicative. 04:48:03 Now we want to generalize that in various directions. 04:48:08 what does applicative get y ou? 04:48:08 Also seems, as if you have ((->) Bool) instead of [] 04:48:23 lightquake: All you need is Applicative, so why require Monad? 04:48:32 (You can't make Const a Monad, anyway.) 04:48:40 right, i meant, why isn't functor enough 04:48:55 Functor is enough when you're traversing exactly one value. 04:48:56 Is mapM a functor? 04:49:06 But both traverses two values. 04:49:15 lightquake: Try to write the function. You need (<*>). :-) 04:49:22 both f (x,y) = (,) <$> f x <*> f y 04:49:54 Anyway, so one direction we can generalize this is that we want indexing. 04:50:11 For instance, mapMWithIndex :: (Int -> a -> f b) -> [a] -> f [b] 04:50:16 (For Applicative f.) 04:50:33 But that's not so great because it's not compatible with regular lenses. 04:50:55 so wait 04:51:10 obviously if f is Identity, that just turns into a plain mapWithIndex 04:51:16 Right. 04:51:17 what does it do if f is Const? 04:51:29 Oh, instance Monoid m => Applicative (Const m) 04:51:49 Where pure _ = Const mempty; Const mf <*> Const mx = Const (mf <> mx) 04:51:57 shachaf: You can also make instance Monoid m => Alternative (Const m) 04:52:02 oh, so you wind up accumulating the result 04:52:04 So when you use Const, it never actually rebuilds the structure. 04:52:19 But you can use a monoid to decide how to accumulate, right. 04:52:35 > foldOf both ("hello","there") 04:52:36 "hellothere" 04:52:43 That uses the [] monoid instance. 04:53:02 right 04:53:14 > foldOf both (Add 2, Add 8) 04:53:15 Not in scope: data constructor `Add'Not in scope: data constructor `Add' 04:53:19 Sum 04:53:22 ah 04:53:43 > getConst $ both (\x -> Const (Sum x)) (1, 2) 04:53:44 Instead of [] and Int perhaps you could make it a class of a type which can be indexable by some other type? Such as, indexable (x ->) by x 04:53:45 Sum {getSum = 3} 04:53:57 So that's how both works. 04:54:10 Now, the trick with indexing is a little bit of overloading. 04:55:02 clsas Indexable i p where ...; indexedMapM :: (Indexable Int p, Applicative f) => p a (f b) -> [a] -> f [b] 04:55:08 Indexing has two instances. 04:55:18 One of them is (->); when you use that instance, it just works without the index. 04:55:40 The second is newtype Index i a b = Indexed { runIndexed :: i -> a -> b } 04:55:49 Er, newtype Indexed 04:56:08 slow down a bit 04:56:23 so you have an instane Indexable Int (->) that just ignores the index 04:56:30 But it is a category like (Kleisli ((->) i)) 04:56:39 Right. 04:56:56 presumably this is actually Indexable i (->) 04:57:03 Yes. 04:57:05 then what's the instance for Indexed look like? 04:57:23 The class is actually class Indexable i p where indexed :: p a b -> i -> a -> b 04:57:43 So instance Indexable i (->) where indexed f i x = f x 04:57:44 I assume 04:58:18 That way any indexed traversal can degrade to a regular traversal. 04:58:28 and for Indexed, indexed = runIndexed 04:58:35 Right. 04:59:21 ok 04:59:40 So that's one generalization. 04:59:52 It turns out this "p" thing is actually pretty useful. 05:00:28 There's a class Profunctor p where lmap :: (a -> b) -> p b c -> p a c; rmap :: (b -> c) -> p a b -> p a c 05:00:45 I can explain that class further if you like. :-) 05:01:17 * elliott prefers the dimap presentation 05:01:19 Oh, so the way we write indexed both is both f (x,y) = (,) <$> indexed f 0 x <*> indexed f 1 y 05:01:28 OK, elliott has volunteered to take over. 05:01:30 * shachaf is out. 05:02:10 aw, come back 05:02:16 * shachaf is in. 05:02:20 But elliott has to explain it with dimap 05:02:34 Is there any relation to representable functor? 05:02:35 I'll do that in ten years 05:02:59 zzo38: We have a lot of (co)representable profunctors. 05:03:48 lightquake: OK, does that class make sense? 05:03:53 (It probably doesn't.) 05:03:57 Profunctor? sort of 05:04:21 It's a thing which is covariant in one argument and contravariant in the other. 05:04:43 In a sense p a b "consumes" things of type a and "produces" things of type b 05:04:49 rmap = fmap 05:05:00 (->) is the classic example. 05:05:01 as in given (Profunctor p) you have (Functor (p a)) for all a (but you can't express this) 05:05:07 I can see that (->) fits 05:05:30 And (->) is a category 05:05:39 Is it related to being a category? 05:05:41 Not all instances of Profunctor are categories. 05:06:02 so what's another instance of Profunctor? 05:06:07 Indexed :-) 05:06:12 shachaf: It is what I thought. 05:06:15 figured :) 05:06:24 Also there are a bunch of others. 05:06:32