00:09 asumu: adu: anything in particular that prompted that? :)
00:10 (join) adu_
00:12 (quit) adu: Ping timeout: 245 seconds
00:12 (nick) adu_ -> adu
00:38 (quit) mithos28: Quit: mithos28
00:44 (quit) dto: Remote host closed the connection
00:48 (quit) Fare: Ping timeout: 272 seconds
00:49 (quit) dauterive: Quit: Leaving
00:53 (join) ambrosebs
00:53 (join) adu_
00:55 (join) mithos28
00:57 (join) bjz
00:58 (join) soegaard
01:15 (quit) asvil: Ping timeout: 245 seconds
01:35 (join) Kaylin
01:38 (quit) mceier: Quit: leaving
01:40 (quit) soegaard: Quit: soegaard
01:42 (quit) jao: Ping timeout: 256 seconds
01:44 (quit) kvda: Quit: z____z
01:49 (quit) adu_: Quit: adu_
01:55 (quit) Kaylin: Read error: Connection reset by peer
02:32 (join) mceier
02:36 (join) Kaylin
02:45 (quit) mithos28: Quit: mithos28
02:51 (quit) ASau: Quit: I be back.
03:01 (join) mithos28
03:06 (quit) jonrafkind: Ping timeout: 272 seconds
03:08 (quit) racycle: Quit: racycle
03:18 (join) noam__
03:21 (quit) noam_: Ping timeout: 256 seconds
03:29 (quit) mithos28: Quit: mithos28
03:31 (join) egnarts-ms
03:41 (join) hkBst
03:42 (quit) sirdancealo2: Ping timeout: 264 seconds
03:47 (quit) Nisstyre-laptop: Ping timeout: 245 seconds
03:50 (join) jao
03:51 (quit) jao: Changing host
03:51 (join) jao
03:57 (join) rohni
03:58 (join) sirdancealo2
04:00 (join) bitonic
04:02 (join) basdirks
04:04 (quit) ambrosebs: Remote host closed the connection
04:16 (quit) Kaylin: Read error: Connection reset by peer
04:24 (quit) jao: Ping timeout: 252 seconds
04:39 (quit) sirdancealo2: Ping timeout: 256 seconds
05:13 (quit) vipjun: Ping timeout: 260 seconds
05:16 (join) vu3rdd
05:16 (quit) vu3rdd: Changing host
05:16 (join) vu3rdd
05:38 (join) ambrosebs
05:50 (quit) rohni: Quit: Leaving.
05:54 (join) mye
05:55 (join) sirdancealo2
06:28 (quit) bitonic: Ping timeout: 256 seconds
06:30 (quit) sirdancealo2: Ping timeout: 245 seconds
06:41 (quit) vu3rdd: Read error: Connection reset by peer
06:43 (join) sirdancealo2
06:43 (join) vu3rdd
06:43 (quit) vu3rdd: Changing host
06:43 (join) vu3rdd
06:50 (quit) vu3rdd: Read error: Connection reset by peer
06:54 (join) rohni
06:59 (quit) yacks: Ping timeout: 255 seconds
07:02 (quit) sirdancealo2: Ping timeout: 245 seconds
07:02 (join) didi
07:03 (part) troydm: "What is hope? That all of your wishes and all of your dreams come true? (C) Rau Le Creuset"
07:06 (join) ambroseb_
07:06 (quit) ambrosebs: Ping timeout: 255 seconds
07:12 (join) yacks
07:13 (join) sirdancealo2
07:19 (join) jeapostrophe
07:19 (quit) jeapostrophe: Changing host
07:19 (join) jeapostrophe
07:33 (quit) ambroseb_: Ping timeout: 276 seconds
07:35 (join) vu3rdd
07:35 (quit) vu3rdd: Changing host
07:35 (join) vu3rdd
07:38 (join) bitonic
07:38 (quit) sirdancealo2: Ping timeout: 245 seconds
07:38 (join) sizz_
07:38 (quit) sizz: Ping timeout: 245 seconds
07:39 (join) mye_
07:40 (quit) mye: Ping timeout: 240 seconds
07:40 (nick) mye_ -> mye
07:41 (join) vipjun
07:53 (join) cdidd
08:00 (join) sirdancealo2
08:01 (join) vkz
08:10 (quit) vkz: Quit: vkz
08:10 (join) mizu_no_oto
08:12 (join) ambrosebs
08:16 (join) asvil
08:18 (join) marcin
08:19 (quit) marcin: Client Quit
08:22 (quit) cored: Ping timeout: 256 seconds
08:24 (join) cored
08:24 (quit) cored: Changing host
08:24 (join) cored
08:27 egnarts-ms: are Racket hash tables reasonably lightweight ? for example, can they be used like dictionaries in Python or plain objects in JS ?
08:28 didi: egnarts-ms: Try it. :^)
08:29 egnarts-ms: i mean, when there are pretty many of small hashtables (3-10 key-value pairs)
08:29 egnarts-ms: didi: i will :)
08:29 didi: egnarts-ms: For small tables, I use assocs. But that's just me.
08:30 (join) Fare
08:31 egnarts-ms: didi: yeah, this is traditional Lisp approach. but you know, what's wrong with alists is that you always worry whether they're still appropriate (not too long)
08:32 egnarts-ms: in other words, they're not scalable, to my mind
08:32 didi: egnarts-ms: I say they will always be.
08:32 didi love some assoc in the morning
08:32 egnarts-ms: didi: what's the need for hashtables then ?
08:33 didi: egnarts-ms: You'll have to ask someone more knowledgeable than me. I like both.
08:35 egnarts-ms: didi: :) okay. i just incline to believing that Python/JS-like approach is really good: you need key-value mapping and you use dictionaries/objects, and that's it
08:35 egnarts-ms: didi: no extra philosophy or design decision
08:35 didi: egnarts-ms: You have that too with Racket.
08:35 didi: egnarts-ms: And freedom to choose.
08:36 egnarts-ms: didi: (though i'm not sure that, say, Python dictionaries are well-optimized)
08:36 didi: egnarts-ms: Also, there is `dictionary'.
08:36 egnarts-ms: didi: ... for large number of entries
08:37 didi: egnarts-ms: AFAIK, Python's dictionaries are hash tables.
08:37 didi: egnarts-ms: Observe that the key has to be hashble.
08:37 egnarts-ms: didi: yes, they are
08:38 egnarts-ms: didi: BTW, Racket `dictionary' is not first-class
08:38 egnarts-ms: https://github.com/plt/racket/blob/master/collects/racket/private/dict.rkt
08:38 didi: egnarts-ms: What do you mean?
08:38 egnarts-ms: `dictionary' is one of 1) hashtable, 2) assoc, ... (something else)
08:38 didi: egnarts-ms: Ye.
08:39 didi: It's an abstraction.
08:40 egnarts-ms: didi: i think i just should measure performance with Racket hashtables
08:40 didi: egnarts-ms: That's the spirit.
08:46 (join) RacketCommitBot
08:46 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://git.io/59FcYQ
08:46 RacketCommitBot: racket/master e6dc9c2 Robby Findler: use clipping in the recent commit to actually test bounding boxes
08:46 (part) RacketCommitBot
08:46 (join) Nisstyre-laptop
08:46 (join) zyoung
08:56 (join) RacketCommitBot
08:56 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://git.io/JZfphQ
08:56 RacketCommitBot: racket/master 74c3c13 Matthias Felleisen: added a void contract to fix setup
08:56 (part) RacketCommitBot
08:58 (quit) mizu_no_oto: Quit: Computer has gone to sleep.
08:59 (join) mizu_no_oto
09:01 (quit) mizu_no_oto: Client Quit
09:04 (join) vkz
09:06 (quit) Nisstyre-laptop: Ping timeout: 260 seconds
09:15 (join) anRch
09:17 (join) soegaard
09:17 (join) mizu_no_oto
09:21 (quit) Twey: Excess Flood
09:22 (join) Twey
09:28 (join) myx
09:39 (quit) jeapostrophe: Ping timeout: 245 seconds
09:40 soegaard: asumu: The Racket version of this task looks fishy:   http://rosettacode.org/wiki/Modular_Arithmetic
09:49 (join) jeapostrophe
09:49 (quit) jeapostrophe: Changing host
09:49 (join) jeapostrophe
09:56 (join) sstrickl
09:56 (quit) sstrickl: Changing host
09:56 (join) sstrickl
10:04 (join) racycle
10:04 asumu: soegaard: hmmm, does it need another `mod`?
10:06 asumu: egnarts-ms: you can just use dictionary functions (e.g., dict-ref, etc.) and then it doesn't matter if you use alists or hashtables.
10:06 asumu: (or rather, you can switch whenever it's convenient then)
10:06 (quit) cored: Quit: leaving
10:14 (quit) anRch: Quit: anRch
10:19 (quit) soegaard: Quit: soegaard
10:19 (quit) hkBst: Quit: Konversation terminated!
10:21 (join) mithos28
10:46 (join) neyuh
10:50 (quit) neyuh: Ping timeout: 245 seconds
10:54 (quit) mceier: Quit: leaving
11:04 (quit) sirdancealo2: Ping timeout: 255 seconds
11:06 (join) dyoo
11:10 (quit) didi: Ping timeout: 255 seconds
11:14 (join) mye_
11:16 (quit) yacks: Read error: Connection reset by peer
11:16 (join) cored
11:17 (quit) basdirks: Ping timeout: 245 seconds
11:17 (quit) mizu_no_oto: Quit: ["Textual IRC Client: www.textualapp.com"]
11:17 (join) yacks
11:18 (quit) yacks: Max SendQ exceeded
11:19 (join) yacks
11:19 mithos28: samth, stamourv: Would either of you have time to look at my pull requests? I think I addressed all of stamourv's points.
11:20 (quit) ambrosebs: Remote host closed the connection
11:21 stamourv: mithos28: Planning to do it today.
11:22 (join) Agent-P
11:22 Agent-P: Hello
11:22 Agent-P: I need Help with "JAVA"
11:22 mithos28: stamourv: Great
11:23 stamourv: Agent-P: You may want to try #java for that.
11:23 mithos28: Agent-P: This is the #racket room, you may have better luck in a java specific one.
11:23 Agent-P: #java Cannot join channel (+i) - you must be invited
11:23 Agent-P: I must be Invited to join the Chanel !
11:23 Agent-P: So if someone is on the #java channel please invite me !
11:23 (join) sirdancealo2
11:24 stamourv: Agent-P: Maybe there's a ##java ?
11:25 dyoo: what's the java question, out of curiosity?
11:25 mithos28: stamourv: Do you know why subtype and cgen have totally different code even though they need to do almost exactly the same thing
11:26 dyoo: agent-p: … and does your question have a relationship to racket in any way?  That might make it ok :)
11:27 mithos28: I'm trying to figure out how I can use prototypes (in odersky's sense) to make inference better
11:27 Agent-P: well
11:27 mithos28: and it would require a function almost exactly like subtype but with a slight twist
11:27 Agent-P: I copied this anywhere but no one seems to be interested ...
11:27 Agent-P: no it does not have any relation with Racket ...
11:27 Agent-P: I need help .. It's about Decorators in Java .. I am trying to program a game but I get Failures and one Error ! Please Help ... !
11:27 Agent-P: this is my Test :
11:28 Agent-P: http://pastebin.com/GMSv6wee
11:28 Agent-P: and this is my Class
11:28 Agent-P: http://pastebin.com/0uyR794B
11:28 dyoo: agent-p: darn.  Completely out of scope for me.  I don't think I can help; I havne't played with decorators.
11:28 stamourv: mithos28: I don't know. I haven't spent much time in the inferencer.
11:29 mithos28: ok, I knew sam would be the better person to ask but its hard being online at the same time as you east coasters
11:29 Agent-P: well
11:29 Agent-P: I'm here
11:30 Agent-P: All time !
11:30 Agent-P: I just need to find a solution for this :)
11:30 mithos28: Agent-P: But I doubt you have an answer to my question.
11:30 stamourv: mithos28: samth's also traveling a lot these days.
11:31 Agent-P: which question ?
11:31 mithos28: Yeah, he said
11:31 (quit) dyoo: Quit: dyoo
11:31 mithos28: Agent-P: Why subtype and cgen have totally different code even though their algorithims are almost identical
11:32 Agent-P: That wasn't my idea ...
11:33 mithos28: Right which is why I doubted you having an answer to it. :)
11:33 Agent-P: It's a project i don't have the right to change anything .. I know it's weird but I have to live with that !
11:33 (quit) egnarts-ms: Ping timeout: 245 seconds
11:33 Agent-P: yep :/
11:37 (join) dyoo
11:37 (quit) Agent-P: Read error: Connection reset by peer
11:37 (quit) rohni: Quit: Leaving.
11:40 (join) didi
11:50 stamourv: mithos28: I think I've looked at all at the ones you updated.
11:50 stamourv: Looking at the new ones now.
11:50 mithos28: Great, those two should be simple. Unlike the poly dots one which is a much larger overhaul
11:53 (join) jbclements
11:55 (quit) zyoung: Ping timeout: 245 seconds
11:56 (join) zyoung
11:56 (quit) jeapostrophe: Ping timeout: 255 seconds
11:58 (quit) vu3rdd: Ping timeout: 272 seconds
12:02 (join) jeapostrophe
12:03 (join) basdirks
12:03 (join) Aune
12:03 (join) mceier
12:04 stamourv: mithos28: Right, I didn't feel comfortable reviewing that one.
12:05 stamourv: asumu: Could you have a look at it? You've been working with polydots more than I have.
12:05 asumu: stamourv: sure
12:05 mithos28: and breaking them :P
12:06 mithos28: http://bugs.racket-lang.org/query/?cmd=view%20audit-trail&pr=13576
12:07 asumu: That's not a recent regression is it? It doesn't seem to work in 5.3 either.
12:07 mithos28: Oh, not you then. I assumed it was related to you fixing the scoping
12:08 mithos28: stamourv: Have you added optimizer tests recently?
12:08 asumu: I don't think so. My guess is that has never worked. :p
12:08 asumu: (although I did break some nested Alls with the scope change)
12:09 stamourv: mithos28: I think so, why?
12:09 mithos28: because I'm getting diffs now from them when trying to run the integration tests on all of these commits
12:10 mithos28: If you added them thats good, it means that just more things need to be cleaned up with the check-below commit
12:10 mithos28: if you hadn't that would imply an interesting interaction between my commits
12:11 sstrickl: mithos28: shouldn't that technically be (lister . args)?  Though that syntax doesn't seem to work at all with define:
12:12 mithos28: except that it looks bad, vector-length is doing a known-length optimization when it should. Well I'll work on that commit some more
12:12 stamourv: Ok, I checked. Yes I did add new optimizer tests. The latest were added on Feb 20.
12:12 stamourv: sstrickl: Nice to see you here, I was planning to email you today.
12:13 sstrickl: stamourv: oh, well, then I chose wisely
12:13 stamourv: sstrickl: But I guess I can just talk to you here instead. :)
12:13 sstrickl: stamourv: what's up?
12:13 stamourv: It's about `define/contract'.
12:13 mithos28: sstrickl: To annotate an identifier in TR you do (var : Type).  (lister . (var : Type)) reads equivalent to (lister var : Type)
12:13 (quit) basdirks: Ping timeout: 250 seconds
12:13 sstrickl: okay, what specifically about it?
12:13 stamourv: I've been working on contract-aware profiling, and noticed something problematic about `define/contract' in one of my example programs.
12:13 sstrickl: mithos28: oh, right, sorry, it's been long enough since I've worked with TR syntax that I'd forgotten.
12:14 stamourv: Basically, `define/contract' applies its contract every time the `define/contract'ed function is called.
12:14 stamourv: This can get *very* expensive.
12:14 sstrickl: stamourv: Oh, I bet it can.
12:14 stamourv: In the example I had (a game that Matthias wrote), contract *application* time accounted for about 20%.
12:15 stamourv: Now, `define/contract' needs to do that because the defined identifier can be mutated, so it may need to re-apply the contract, in case there's a new value there.
12:15 sstrickl: stamourv: yep
12:15 stamourv: So in a sense, this means that `define/contract' applies the contract to the *location*, not the *value*.
12:16 stamourv: Now, if we were to change it to contract the value, we'd only need to apply the contract once.
12:16 stamourv: Of course, that would mean that, if the identifier is mutated, the contract "goes away".
12:16 sstrickl: stamourv: if there was a nice way of ensuring that there's no mutations _within_ the region (it's easy to do without), then we could always allow for marking contracted identifiers as immutable/mutable, and default functions to immutable and normal identifiers to mutable
12:17 sstrickl: stamourv: yep, exactly
12:17 sstrickl: (I've been thinking about this for a while, since obviously you're not the only person who runs into this issue :/)
12:17 mithos28: Couldn't you just reapply the contract on set!
12:17 stamourv: Checking for mutability could fix that, but that would require local-expand, which is probably not desirable here.
12:17 sstrickl: mithos28: that's the part that makes it easy for without, but remember that you can mutate the value within the region as well
12:18 sstrickl: what you might be able to do, though, is create a set!-transformer for the identifier within the region, that does the right things, but I've never tried to see if I can actually do that
12:18 mithos28: sstrickl: Yes that would be what I was thinking
12:18 stamourv: sstrickl, mithos28: Right. That sounds doable.
12:19 stamourv: And that should preserve the semantics, too.
12:19 sstrickl: right
12:19 stamourv: Now, I'm not sure which semantics makes more sense though, contracting the location or the value.
12:19 sstrickl: so yeah, I've had that idea rolling around my head, but things have been busy enough around here that I haven't had a chance to make that change, sorry :/
12:20 stamourv: I'm kinda tempted to give it a try.
12:20 sstrickl: go for it!  I'll be more than happy to look at what you do and check it before you commit it
12:20 stamourv: I've been playing in the contract system a fair amount, so maybe I could do it.
12:20 mithos28: This is one of the reasons I love Racket. That something like this is even possible.
12:20 stamourv: I'll look into it, maybe later today.
12:20 stamourv: mithos28: Agreed!
12:20 sstrickl: stamourv: fantastic, thank you :)
12:21 mithos28: and the best part is that at runtime you pay no cost for it if it is not mutated
12:21 stamourv: sstrickl: If that doesn't work, then the future Contract Coach will at least warn you about it.
12:21 sstrickl: stamourv: you're a good man and thorough
12:21 stamourv: :)
12:21 sstrickl: stamourv: excellent, though I hope it does because I'd rather not make people even LESS likely to use them except in times of extreme need
12:22 sstrickl: (it does work, rather)
12:22 stamourv: Agreed.
12:22 mithos28: Yes, I really want contracts on by default in TR.
12:22 mithos28: The issue is that they are just too slow currently
12:22 stamourv: But even if it doesn't work, the Contract Coach wouldn't recommend removing the contract, just turning them into provide/contract.
12:23 sstrickl: stamourv: yeah, of course
12:23 stamourv: Which, with submodules, can basically end up giving you similar code.
12:23 stamourv: mithos28: chrdimo is looking into it, actually.
12:23 stamourv: The new option contracts could be a solution.
12:23 mithos28: what are option contracts, I have seen some commits about them
12:23 sstrickl: I've seen his commits about those go by, but haven't read them enough to get the idea about it
12:24 mithos28: Yeah Google is not helpful
12:24 mithos28: I'm not looking for stocks
12:24 stamourv: They're basically contracts that you can check partially, or not at all, and still have a reasonable blame story.
12:24 sstrickl: oh, interesting
12:25 stamourv: So, IIUC, chrdimo's plan would be to enable the TR contracts by default, but either disable the checking, or only check the first n times the boundary is crossed.
12:25 sstrickl: is there a particular venue he's planning on targeting that work with? I'd be interested in seeing the theory writeup
12:25 stamourv: It wouldn't provide as much checking, but it would still catch some bugs, and would help contracts stay in sync with the code, which is currently a big problem.
12:26 stamourv: sstrickl: IIRC, he wrote about them in is thesis.
12:26 sstrickl: stamourv: yeah, agreed.  I remember doing TR work in the past where we'd turn on contracts then realize how badly they'd fallen behind
12:26 sstrickl: oh, then I should check that out.
12:26 mithos28: stamourv: I thought half the issue was that we didn't want the contract require in production TR
12:26 (join) jonrafkind
12:27 (join) RacketCommitBot
12:27 RacketCommitBot: [racket] plt pushed 4 new commits to master: http://git.io/qb4C2Q
12:27 RacketCommitBot: racket/master 9f01d26 Eric Dobson: Cleanup kw/opt lambda annotations.
12:27 RacketCommitBot: racket/master d59a600 Eric Dobson: Make kw/opt lambda special checks work with other expected values....
12:27 RacketCommitBot: racket/master ddf0c39 Eric Dobson: Reduce unnecessarily complicated restrict case.
12:27 (part) RacketCommitBot
12:27 stamourv: mithos28: IIUC, option contracts with checking turned off should be pretty low overhead.
12:28 stamourv: chrdimo is looking into it now, so we'll know the answer soon, I hope.
12:28 stamourv head out for lunch.
12:28 mithos28: ok, I have a optimizer question now that I have gotten the other code in
12:28 stamourv: s/head/heads/
12:28 stamourv: mithos28: Is it quick?
12:28 mithos28: not really
12:28 mithos28: I'll catch you another day
12:29 stamourv: I'll be back in 1h or so.
12:29 mithos28: I'll be at work by then
12:29 stamourv: Or you can just paste it here, and I can answer it when I'm back.
12:29 stamourv: Or you can send me an email.
12:29 mithos28: Its more back and forth
12:29 (quit) dyoo: Quit: dyoo
12:29 (join) dauterive
12:29 stamourv: Ok, whenever you want.
12:31 sstrickl: stamourv: BTW, I'll try to stay on, so if you do start working on the stuff we discussed and need questions answered, feel free to ping me or /msg me, send me an email if I've disappeared
12:31 (join) dyoo
12:36 (join) soegaard
12:36 soegaard: asumu: No. The problem is that (f 10) is computed without using modular exponentiation. It ought to use modexpt instead of expt.
12:39 (quit) myx: Ping timeout: 260 seconds
12:40 (join) ASau
12:47 (join) Kaylin
13:01 (quit) Kaylin: Read error: Connection reset by peer
13:06 (quit) dauterive: Quit: Leaving
13:12 (quit) ASau: Ping timeout: 264 seconds
13:12 (join) snorble_
13:15 (quit) asvil: Ping timeout: 245 seconds
13:17 (join) ASau
13:19 (quit) Aune: Quit: Hath Deprated
13:20 (join) Aune
13:20 (join) basdirks
13:25 (quit) basdirks: Ping timeout: 256 seconds
13:29 (quit) jeapostrophe: Ping timeout: 272 seconds
13:30 (join) rich1
13:32 (join) RacketCommitBot
13:32 RacketCommitBot: [racket] plt pushed 2 new commits to master: http://git.io/4-m2rg
13:32 RacketCommitBot: racket/master 2170e17 Robby Findler: tidy up gui...
13:32 RacketCommitBot: racket/master 4997800 Robby Findler: make-traversal's optional argument is now ignored...
13:32 (part) RacketCommitBot
13:37 (join) basdirks
13:46 rich1: as an example, given...
13:46 rich1: (define (logb n b) (/ (log n) (log b)))
13:46 rich1: (define (log256 n) (logb n 256))
13:47 rich1: should log256 be a define or a macro - alternatively, as a define will racket automatically optimise it away
13:48 didi: rich1: Procedure.
13:48 rich1: didi: gotcha.
13:49 jbclements: rich1: I claim that logb may very well be inlined, leading to the same performance you'd get from a macro; was that your question?
13:50 rich1: jbclements: yes - you put it much clearer
13:50 rich1: out of interest why does racket binding of log have arity 1, but the rnr6 binding have 2 to allow for custom base?
13:52 (join) anRch
13:52 jbclements: rich1: I'm guessing it's for symmetry with "exp", but you might well respond that there's then no analog to "expt". I'm guessing it was personal preference on mflatt's part.
13:54 rich1: righto
13:57 stamourv: rich1: If you want to check whether your helper function is being inlined, you can use the Optimization Coach DrRacket plugin.
13:58 (join) DanBurton
14:00 rich1: stamourv: nice, it also said that logb was inlined as well as log256
14:01 rich1: does racket base inlining on the size of expression or "deterministicness" ?
14:02 rich1: found an entry in the manual, reading
14:02 stamourv: Size of function body is the main factor.
14:04 didi: rich1: Be careful with macros like that. If you find a bug, you'll have to eval all the procedures that uses it.
14:05 didi: Something like this will be impossible:
14:05 didi: rudybot: (define (foo) 42)
14:05 rudybot: didi: Done.
14:05 didi: rudybot: (define (bar) (foo))
14:05 rudybot: didi: Done.
14:05 didi: rudybot: (bar)
14:05 rudybot: didi: ; Value: 42
14:05 didi: (define (foo) 7)
14:05 didi: rudybot: (define (foo) 7)
14:05 rudybot: didi: Done.
14:05 didi: rudybot: (bar)
14:05 rudybot: didi: ; Value: 7
14:06 asumu: soegaard: ah, right. The issue was that the problem statement explicitly said that wasn't allowed.
14:06 asumu: So maybe I should just remove the entry.
14:06 soegaard: It is a rather strange task description.
14:09 soegaard: If only modexpt support a modulus of +inf.0 ...
14:10 soegaard: Since the task is to demonstrate that operators can be redefined, it would we fine to define new versions of +,-, and expt in a submodule and use them instead.
14:10 rich1: didi: are you saying if bar was a macro, that the 2nd application of bar would be 42 ?
14:10 didi: rich1: Yeap.
14:11 asumu: soegaard: yeah, that's a good point. Or even just (define + mod+)
14:11 didi: rich1: No.
14:12 soegaard: Yeah, but then f doesn't work for general numbers.
14:12 didi: rich1: I mean, If `foo' was a macro.
14:13 soegaard: Actually I can't see that one can solve the task if numbers  in Z/pZ are represented as plain numbers.
14:14 rich1: didi: ok cool, good catch.
14:14 jbclements: didi: … but #lang racket doesn't allow the redefinition of an identifier anyway.
14:14 soegaard: "It should behave the same way with normal and modular integers" implies that one can look at an integer and see what type it is.
14:15 jbclements: rich1: the classic problem with a macro is that it can't then be passed as a function.
14:15 didi: jbclements: Sorry, I don't know what you mean.
14:15 rich1: jbclements: e.g. to map?
14:16 asumu: soegaard: I find this to be a problem with Rosetta in general. Their task statements are often silly.
14:16 soegaard: Yes!
14:16 didi: rich1: Yes, like `map'.
14:16 didi: Some wrap macros inside lambdas, but...
14:17 rich1: starting to see why the style guide says use define :)
14:17 rich1: much simpler
14:22 (quit) yacks: Quit: Leaving
14:24 (join) serhart
14:27 (quit) Aune: Quit: Hath Deprated
14:28 didi: jbclements: Oh, I see what you meant. Right.
14:32 (join) jeapostrophe
14:32 (quit) jeapostrophe: Changing host
14:32 (join) jeapostrophe
14:39 (quit) anRch: Quit: anRch
14:53 (quit) jeapostrophe: Read error: Operation timed out
14:59 (quit) bitonic: Ping timeout: 252 seconds
15:03 (join) jao
15:03 (quit) jao: Changing host
15:03 (join) jao
15:04 (quit) rich1: Quit: Leaving.
15:17 (join) elazul
15:19 (quit) vkz: Quit: vkz
15:21 (quit) soegaard: Quit: soegaard
15:24 (quit) didi: Ping timeout: 245 seconds
15:26 (join) didi
15:41 (join) rich1
15:42 (join) soegaard
15:44 soegaard: Nice examples: http://srfi.schemers.org/srfi-13/srfi-13.html#string-replace
15:47 stamourv: soegaard: That's a horribly unintuitive function/interface, though.
15:48 stamourv: It's a bad sign that the best documentation is the function body itself.
15:48 (join) bitonic
15:49 soegaard: The convention follows substring.
15:49 stamourv: Yes, but with 4 magic numbers.
15:49 soegaard: Well, the last two are optional :-)
15:50 stamourv: Yes, but I'm not sure the extra flexibility is worth the extra confusion.
15:50 stamourv: Removing the implicit substring on s2 would make the function a lot clearer, IMO.
15:51 (quit) jonrafkind: Ping timeout: 255 seconds
15:51 soegaard: True.
15:52 soegaard: I ended up reading srfi 13 when searching for string-prefix? and string-suffix?
15:53 soegaard: Eli added some useful string functions to racket/string, but there some nice additional ones in srfi 13.
15:56 rich1: how can i remove duplicates from a list? make another but check for each element if it's already in there?
15:56 soegaard: An amusing aside:  string-prefix? is startsWith in Python.   Guess which language came from academia :-)
15:56 Fare: remove-duplicates ?
15:57 rich1: Fare: knew i was missing something - search for unique, distinct though
15:57 rich1: s/search/searched
15:58 soegaard: Is there a blessed way of adding additional keywords to the documentation?
15:58 Fare: rich1: actually, that's a CL name -- dunno if scheme and/or racket has that name
15:58 Fare: rich1: that said, you realize that checking for duplicates in a list is in O(n), don't you?
15:58 stamourv: soegaard: There's a scribble command for it, I forget what it's called. Lemme look it up/
15:59 (join) mizu_no_oto
16:00 dyoo: rich1: http://docs.racket-lang.org/reference/pairs.html?q=remove-duplicates#%28def._%28%28lib._racket%2Flist..rkt%29._remove-duplicates%29%29
16:00 rudybot: http://tinyurl.com/bz76d6a
16:00 stamourv: rudybot: (require scribble/base)
16:00 rudybot: stamourv: Done.
16:00 stamourv: rudybot: doc index
16:00 rudybot: stamourv: http://docs.racket-lang.org/scribble/base.html#(def._((lib._scribble%2Fbase..rkt)._index))
16:00 stamourv: rudybot: doc section-index
16:00 rudybot: stamourv: http://docs.racket-lang.org/scribble/base.html#(def._((lib._scribble%2Fbase..rkt)._section-index))
16:00 stamourv: soegaard: ^
16:00 soegaard: stamourv: thanks
16:02 rich1: Fare: won't it be O(n) for the single duplicate check - making it larger for the whole list case ?
16:02 (quit) acarrico: Ping timeout: 245 seconds
16:03 Fare: rich1: well, the trick is to convert the list to a hash-table, then do the check in linear time in the second pass
16:04 Fare: of course, you could argue that the hash-table is log n rather than constant time.
16:04 Fare: but yes, naive remove-duplicates could be O(n^2)
16:04 Fare: I don't know how naive the racket implementation is
16:04 rich1: Fare: nice. i really do stand on the shoulder of giants. i was thinking more of sorting then looping as the sort complexity has to be taken into account
16:04 dyoo: Fare: it uses a hashtable
16:05 dyoo: Fare: if the equality is one of the blessed ones, like equal? or eq?
16:05 soegaard: For small lists the O(n^2) appoach is faster.
16:06 dyoo: and it does that if the lists are short enough.  See: http://git.racket-lang.org/plt/blob/HEAD:/collects/racket/list.rkt#l233
16:06 rich1: for this case i'll be dealing up to 16k elements
16:07 Fare: does racket use dynamic code generation to determine the break-even point depending on processor architecture, memory pressure, etc.?
16:07 Nisstyre: rich1: is there some reason it has to be a list in the first instance?
16:07 (join) dauterive
16:07 dyoo: Fare: the threshold in the code appears to have been derived empirically
16:07 rich1: Nisstyre: no, i just need to read 1 element per line from a file - perform an action on each of them, then filter duplicates from that result
16:08 Fare: interesting that it's set at 40. I would have guessed between 16 and 32.
16:08 soegaard: Surprising to see let-syntax to pop up in that context.
16:08 Fare: I suppose hashing is quite expensive.
16:08 asumu: soegaard: well string-starts-with? is rather long. ;)
16:09 soegaard: ;-)
16:09 (join) rohni
16:14 (join) noam
16:16 (join) Kaylin
16:17 (quit) noam__: Ping timeout: 245 seconds
16:21 (join) jonrafkind
16:48 (quit) rohni: Quit: Leaving.
16:51 (quit) mizu_no_oto: Ping timeout: 276 seconds
16:53 (join) mizu_no_oto
16:59 stamourv: sstrickl: I'm close to having a patch.
16:59 stamourv: It's not very pretty though.
16:59 sstrickl: stamourv: well, still, an ugly patch is a patch that can be cleaned up if needed, but better than no patch!
17:00 stamourv: Right.
17:00 stamourv: One slight issue is that it adds a hidden keyword to `with-contract'.
17:00 stamourv: You're already using one, but the one I add could be abused a bit more.
17:01 stamourv: The real solution would be to have a helper macro that takes these that `with-contract' expands into.
17:01 stamourv: That would take a bit of work to get the error messages right, so I'm not going to do it now.
17:01 (join) rohni
17:01 (join) RacketCommitBot
17:01 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://git.io/b9sSNA
17:01 RacketCommitBot: racket/master 1ed7e8c Burke Fetscher: Redex: generation for judgment-form side-conditions...
17:01 (part) RacketCommitBot
17:02 sstrickl: stamourv: Yeah.  Well, wait, to make sure, which hidden keyword are you talking about? I thought I documented all the keywords in existence already
17:03 stamourv: No, there's also `#:region'.
17:03 stamourv: Which only seems to be used from `define/contract'.
17:04 sstrickl: I thought I'd documented that, but if not it could be.
17:04 sstrickl: Yeah, the whole point is to be able to give a more user-friendly name to the sexp used for naming the blame object (so define/contract uses `(function ,name) or `(definition ,name) depending on if it's a function-style def)
17:04 sstrickl: so anything else that expands into with-contract is more than welcome to use it as desired
17:05 sstrickl: so I guess it should be documented, I'll have to put that on a list somewhere
17:05 stamourv: Ok.
17:05 stamourv: The keyword I added is not really meant to be used publically.
17:06 stamourv: It specifies an identifier that shouldn't be contracted in all cases.
17:06 stamourv: Anyway, sending  you the patch now.
17:06 sstrickl: Okay, thanks, will take a look
17:07 stamourv: The keyword I'm talking about is `#:define/contract-id'.
17:07 stamourv: I haven't run the tests yet, doing it now.
17:07 stamourv: Tried it on one example and it worked.
17:09 stamourv: Actually, fixing the interface to avoid exposing my hook shouldn't be too hard.
17:09 stamourv: I think I can get it done before happy hour.
17:09 sstrickl: Okay, I'll check out the patch as-is, and if you have an updated one to throw at me later, I'll look it over also.
17:09 (join) acarrico
17:10 stamourv: Sounds good.
17:10 stamourv: The interface changes shouldn't affect the semantics.
17:10 eli: soegaard: Which other ones?  -- All I have on my todo list ATM is `string-index'.
17:11 soegaard: string-prefix? and string-suffix? and string-contains?
17:11 eli: soegaard: (re srfi-13, that is.)
17:11 sstrickl: stamourv: you're testing this with mutation within the contracted region as well, right?
17:12 sstrickl: stamourv: or is your solution only protecting against mutation from without?
17:13 stamourv: sstrickl: Haven't thought about that.
17:13 stamourv: It should be the same as before.
17:13 stamourv: I think.
17:13 sstrickl: stamourv: ... that's why I mentioned it earlier, and why I haven't taken the route you did with d/c.
17:13 (quit) basdirks: Ping timeout: 260 seconds
17:14 stamourv: Ok, maybe I misunderstood earlier.
17:14 sstrickl: No, rewrapping on access or mutation from outside (and leaving the original value uncontracted) avoids when the server mutates that value as well when the client does.
17:14 stamourv: I see.
17:15 stamourv: So in that case, we'd need 2 set!-transformers, one for the inside, that wraps only on mutation, and one for the outside, that's the current one.
17:15 stamourv: Right?
17:15 (quit) bitonic: Read error: Operation timed out
17:16 (quit) jao: Ping timeout: 248 seconds
17:16 sstrickl: Kinda.  The problem is that case then causes contracts to be checked when the inside mutates it, and then the inside accesses it.
17:16 sstrickl: (that is, when values are flowing purely internally)
17:16 sstrickl: Now you see why there hasn't been a quick fix to this :)
17:17 stamourv: Mmm, I see.
17:17 sstrickl: What you _can_ do, though, is keep two values in sync, one used for the inside, one used for the outside, and every time there's a mutation internally, change both values, but with the external one getting the contract appropriately.
17:17 (join) bitonic
17:18 stamourv: How would you do that?
17:18 sstrickl: And then on the outside, you don't have to wrap on access (win!), but you do have to change both values appropriately with appropriate contract wrapping on mutation.
17:18 sstrickl: hmm, I'll outline my thoughts over email, okay?
17:18 sstrickl: or maybe I can find the old discussion of this with Robby and forward it
17:18 sstrickl: since I outlined the solution there also
17:19 (join) anRch
17:20 stamourv: Sounds good.
17:20 sstrickl: stamourv: found it, forwarding now
17:20 stamourv: Thanks.
17:21 sstrickl: stamourv: No, thank _you_ for putting time in on this :)
17:21 stamourv: Ok, I think I fixed the interface.
17:21 stamourv: Well, I'm procrastinating. ;)
17:21 sstrickl: stamourv: hey, productive procrastination is useful sometimes.
17:21 stamourv: I'm writing an experience report for ICFP with Neil ⊥, but writing doesn't work today.
17:21 stamourv: Agreed.
17:33 (quit) anRch: Quit: anRch
17:33 (quit) serhart: Quit: Leaving.
17:35 (join) kvda
17:36 (join) RacketCommitBot
17:36 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://git.io/wqlFJg
17:36 RacketCommitBot: racket/master d9c81a5 Robby Findler: declare failing test should not have failed...
17:36 (part) RacketCommitBot
17:39 (join) RacketCommitBot
17:39 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://git.io/Tx9Rbg
17:39 RacketCommitBot: racket/master eee43eb Robby Findler: like d9c81a593, but this time only 4 pixels...
17:39 (part) RacketCommitBot
17:42 (join) Kaylin1
17:42 (join) basdirks
17:44 (quit) rich1: Quit: Leaving.
17:45 (quit) Kaylin: Ping timeout: 240 seconds
17:49 jonrafkind: where are the pdf links on the newdocs page?
17:49 (quit) basdirks: Ping timeout: 255 seconds
17:49 jonrafkind: also changing the version brings you to the old docs style page
17:50 jonrafkind: oh wait i see, you have to click on one of the specific manuals, like the guide, to see the pdf link
18:11 (quit) soegaard: Quit: soegaard
18:24 (join) jeapostrophe
18:24 (quit) jeapostrophe: Changing host
18:24 (join) jeapostrophe
18:25 (quit) vipjun: Ping timeout: 245 seconds
18:25 (quit) zyoung: Ping timeout: 250 seconds
18:32 (join) mithos28_
18:37 (quit) mithos28: Read error: Connection reset by peer
18:37 (quit) cipher: Remote host closed the connection
18:37 (nick) mithos28_ -> mithos28
18:52 (quit) kvda: Quit: z____z
18:53 (quit) mizu_no_oto: Quit: ["Textual IRC Client: www.textualapp.com"]
19:01 (quit) dyoo: Quit: dyoo
19:03 (quit) bitonic: Remote host closed the connection
19:04 (join) bitonic
19:04 (quit) rohni: Quit: Leaving.
19:05 (join) Kaylin
19:07 (quit) Kaylin1: Ping timeout: 255 seconds
19:08 (join) ambrosebs
19:10 (quit) bitonic: Remote host closed the connection
19:10 (join) bitonic
19:12 (quit) jbclements: Quit: jbclements
19:18 (quit) elazul: Quit: Leaving
19:19 (join) jbclements
19:20 (quit) Mattykins: Remote host closed the connection
19:22 (join) Mattykins
19:25 (quit) didi: Quit: ERC Version 5.3 (IRC client for Emacs)
19:30 (join) DanBurto_
19:32 (quit) jonrafkind: Ping timeout: 256 seconds
19:34 (quit) DanBurton: Ping timeout: 272 seconds
19:36 (quit) eikonos: Ping timeout: 245 seconds
19:38 (join) eikonos
19:38 (quit) eikonos: Changing host
19:38 (join) eikonos
19:39 (join) mizu_no_oto
19:46 (quit) DanBurto_: Remote host closed the connection
19:47 (quit) jeapostrophe: Ping timeout: 260 seconds
19:49 (join) kvda
19:50 (join) DanBurton
19:55 (join) jao
19:55 (quit) jao: Changing host
19:55 (join) jao
20:08 (quit) DanBurton: Remote host closed the connection
20:13 (quit) jbclements: Quit: jbclements
20:23 (quit) bjz: Read error: Connection reset by peer
20:26 (quit) eikonos: Ping timeout: 250 seconds
20:34 (join) eikonos
20:34 (quit) eikonos: Changing host
20:34 (join) eikonos
20:47 (join) cipher
20:47 (quit) cipher: Changing host
20:47 (join) cipher
20:54 (quit) mye_: Quit: mye_
20:57 (quit) jao: Ping timeout: 248 seconds
21:06 (join) zyoung
21:13 (join) mye
21:16 (quit) mye: Client Quit
21:35 (quit) dauterive: Quit: Leaving
21:36 (join) jeapostrophe
21:36 (quit) jeapostrophe: Changing host
21:36 (join) jeapostrophe
21:38 (quit) ambrosebs: Remote host closed the connection
21:39 (quit) mizu_no_oto: Quit: ["Textual IRC Client: www.textualapp.com"]
21:43 (quit) pono: Ping timeout: 264 seconds
21:44 (join) pono
21:45 (quit) pono: Changing host
21:45 (join) pono
21:56 (quit) bitonic: Ping timeout: 245 seconds
22:01 (join) ambrosebs
22:11 (join) yacks
22:23 (join) jonrafkind
23:05 (quit) jeapostrophe: Ping timeout: 245 seconds
23:19 (quit) zyoung: Remote host closed the connection
23:43 (join) verahsa
23:44 verahsa: I'm having a small issue with figuring out how to execute two separate commands on either a cond expression or something similar... would someone be willing to lend me some advice? :)
23:49 Nisstyre: verahsa: paste your code to ideone.com
23:49 Nisstyre: (use the guile option but don't run it)
23:50 verahsa: http://ideone.com/HRt3Gg
23:51 Nisstyre: okay, so where is the problem here?
23:51 verahsa: in the lgrep function, i need to both evaluate sublist of a & the first element of b as well as recursively call lgrep on a and the remainder of b
23:51 verahsa: the (lgrep SL1 lst) should return multiple lines of the list "lst", not just the first
23:51 Nisstyre: verahsa: so you want to return both things?
23:52 verahsa: I'm wanting to evaluate sublist for each element, and return any true element
23:52 verahsa: i can't seem to get *both* to work
23:52 Nisstyre: this looks like a good job for pattern matching
23:52 Nisstyre: have you looked at that?
23:53 Nisstyre: http://docs.racket-lang.org/reference/match.html
23:53 verahsa: I looked at that earlier, but it wasn't cooperating for some reason.
23:54 verahsa: Ah.... yeah, I remember now... match only returns a one item. I'm matching lists, and have to return each list that matches the search list, not just one of them or the portion of them that matches.
23:55 (join) kirk
23:55 (nick) kirk -> ksefchik
23:57 ksefchik: Would anyone be able to help me out with racket crashing?