00:04 (join) dented42
00:20 (quit) Kaylin: Read error: Connection reset by peer
00:38 (join) jeapostrophe
00:53 (join) egslava
00:53 egslava: Hello!
00:53 egslava: I'm newbie in Racket, but it's look pretty.
00:53 egslava: And yeah, sorry for my English - I'm from Russia.
00:54 egslava: I'm interesting - is there Racket compilator?
00:54 (quit) jeapostrophe: Ping timeout: 244 seconds
00:54 egslava: Or tool,that able to run my racket-program as .exe or elf-file?
00:56 asumu: egslava: Yeah, look up the raco program.
00:57 asumu: egslava: http://docs.racket-lang.org/raco/exe.html
00:57 asumu: There aren't very many people on at this time of night, so ask questions on the mailing list if you don't have luck here.
00:57 asumu: I'm going to sleep soon myself, so good luck. Racket is fun. :)
01:06 egslava: Thanks for the answer!
01:06 egslava: Oh, really. We have 9.00 AM, I forgot, that other country yet sleepping.
01:13 egslava: Heh, I used Raco from DrRacket. But result binaries is too big :(
01:13 egslava: They can use about 12 MB. It is too lot. I think, I should use SBCL in the future.
01:13 egslava: But thanks a lot!
01:18 (join) egslava_
01:20 (quit) egslava: Ping timeout: 245 seconds
01:24 (quit) yoklov: Quit: bye!
01:26 (join) jonrafkind
01:33 (join) Shambles_
01:44 (join) dous_
01:46 (quit) dous: Ping timeout: 246 seconds
02:22 (join) dzhus
02:34 (quit) jonrafkind: Ping timeout: 260 seconds
02:41 (join) mmajchrzak
03:05 (quit) dous_: Remote host closed the connection
03:06 (join) dous
03:15 (join) Demosthenes
03:22 (quit) Demosthenes: Ping timeout: 264 seconds
03:29 (join) Demosthenes
03:45 (quit) dzhus: Remote host closed the connection
03:57 (join) Shvillr_
03:57 (quit) Shviller: Disconnected by services
03:57 (nick) Shvillr_ -> Shviller
04:05 (quit) dous: Remote host closed the connection
04:15 (quit) egslava_: Ping timeout: 245 seconds
04:21 (quit) mmajchrzak: Ping timeout: 245 seconds
04:29 (quit) dsp_: Read error: Operation timed out
04:31 (join) dsp_
04:41 (join) gciolli
04:46 (quit) gciolli: Ping timeout: 246 seconds
04:55 (join) gciolli
05:04 (join) dzhus
05:14 (quit) gciolli: Ping timeout: 252 seconds
05:23 (join) bluezenix
05:24 (join) _danb_
05:32 (join) gciolli
05:36 (quit) gciolli: Ping timeout: 252 seconds
05:54 (join) masm
06:10 (join) bitonic
06:14 (join) dous
06:38 (join) mceier
06:50 (quit) dzhus: Ping timeout: 244 seconds
07:03 (join) dzhus
07:03 (quit) jhemann: Ping timeout: 264 seconds
07:31 (quit) bluezenix: Quit: Leaving.
07:56 (join) djcb`````
07:58 (quit) djcb````: Ping timeout: 244 seconds
08:23 (join) mmajchrzak
09:12 (join) veer
09:31 (join) samth
09:38 (quit) dzhus: Remote host closed the connection
09:50 (join) jeapostrophe
10:12 (quit) duomo: Quit: Leaving...
10:15 (quit) mmajchrzak: Remote host closed the connection
10:25 (join) Skola
10:53 (join) gridaphobe
11:01 (quit) djcb`````: Remote host closed the connection
11:05 (join) bluezenix
11:11 (quit) bitonic: Ping timeout: 276 seconds
11:22 (join) djcb
11:26 (join) djcb`
11:28 (quit) djcb: Ping timeout: 264 seconds
11:39 (quit) veer: Quit: Leaving
11:41 (join) duomo
11:43 (quit) ssbr_: Ping timeout: 246 seconds
11:54 (join) RacketCommitBot
11:54 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://git.io/D1bAdw
11:54 RacketCommitBot: [racket/master] streamline chaperoned struct field reference - Matthew Flatt
11:54 (part) RacketCommitBot
12:14 (quit) Skola: Quit: Lost terminal
12:35 (join) gciolli
12:39 (quit) Demosthenes: Read error: Connection reset by peer
12:40 (join) Demosthenes
12:54 (quit) Demosthenes: Read error: Connection reset by peer
12:55 (join) Demosthenes
12:57 (quit) gciolli: Ping timeout: 244 seconds
13:01 (quit) jrslepak: Quit: What happened to Systems A through E?
13:08 (join) jhemann
13:14 (quit) cdidd: Ping timeout: 276 seconds
13:18 (join) basepi
13:19 basepi: so I'm trying to go through a list like '(1 2 3 4 (5 6) 7) and make it into '(1 2 3 4 5 6 7).  Is there any way to do this elegantly?
13:19 (join) jrslepak
13:19 basepi: I only want to unwrap a single level, so '(1 2 3 4 (5 (6 7))) would become '(1 2 3 4 5 (6 7))
13:21 samth: rudybot, append-map
13:21 rudybot: samth: haesbaert, use SRFI 1's APPEND-MAP.
13:21 samth: rudybot, eval append-map
13:21 rudybot: samth: your sandbox is ready
13:21 rudybot: samth: ; Value: #<procedure:append-map>
13:21 basepi: thanks, i'll look into that.
13:22 samth: rudybot, (append-map (lambda (e) (if (list? e) e (list e))) '(1 2 3 4 (5 6) 7))
13:22 rudybot: samth: ; Value: (1 2 3 4 5 6 7)
13:22 samth: rudybot, (append-map (lambda (e) (if (list? e) e (list e))) '(1 2 3 4 ((5 6) 7)))
13:22 rudybot: samth: ; Value: (1 2 3 4 (5 6) 7)
13:22 samth: basepi, ^
13:23 basepi: wow awesome, thanks.
13:24 (quit) Demosthenes: Ping timeout: 252 seconds
13:26 (join) Kaylin
13:28 (quit) gridaphobe: Ping timeout: 245 seconds
13:40 (quit) jeapostrophe: Ping timeout: 265 seconds
13:41 (join) gridaphobe
13:43 (join) dnolen
13:45 (join) jonrafkind
13:46 basepi: thanks again, samth.  been struggling against this part of my assignment for hours, and a modified version of your code above solved it perfectly.
13:49 (quit) karswell: Read error: Operation timed out
14:01 (quit) _danb_: Ping timeout: 265 seconds
14:16 (join) yoklov
14:26 (join) Demosthenes
14:32 (join) djcb``
14:34 (quit) djcb`: Ping timeout: 244 seconds
14:38 (quit) Demosthenes: Ping timeout: 240 seconds
14:39 (join) Demosthenes
14:43 (quit) Demosthenes: Client Quit
14:49 cky: Quick question: are places effectively like message-passing in other concurrent systems? (I'm thinking of Erlang in specific.) A quick skim seems to suggest it but I don't want to get the wrong idea.
14:49 (join) errstr
15:21 (quit) duomo: Quit: Leaving...
15:22 (join) jeapostrophe
15:29 (join) duomo
15:42 (quit) Kaylin: Quit: Leaving.
15:58 (quit) jeapostrophe: Ping timeout: 260 seconds
16:04 (join) karswell
16:39 (quit) jamessan: Quit: Coyote finally caught me
16:43 (join) jamessan
16:49 (quit) bluezenix: Quit: Leaving.
17:14 (join) Kaylin
17:14 (join) bluezenix
17:15 chandler: cky: Place channels are a form of message-passing communication, but the comparison to Erlang is perhaps imperfect.
17:25 (join) kingless
17:28 rapacity: there's also support for shared memory vectors between places
17:33 (quit) kingless: Read error: Connection reset by peer
17:33 (join) kingless_
17:37 (quit) kingless_: Client Quit
17:38 (quit) yoklov: Quit: computer sleeping
17:39 (join) yoklov
17:42 (quit) yoklov: Client Quit
17:48 (quit) mceier: Quit: leaving
17:52 errstr: hello all.
17:52 errstr: in the definition for 'hash-update' it says the 'updater' can be 'any/c', which seems to be any value. I'm trying to do the following, but it complains about the arity...
17:53 errstr: something equivalent to this: (hash-update (hash) 'x (+ 4 1))
17:53 chandler: what version are you using? the 5.2.1 documentation says the updater is (any/c . -> . any/c)
17:54 errstr: rather ... (hash-update (hash) 'x (+ 4 1) 0)
17:54 errstr: 5.2
17:54 chandler: might be a fixed doc bug
17:55 errstr: so, I'm doing it wrong or I need to upgrade?
17:55 (quit) jonrafkind: Ping timeout: 245 seconds
17:55 chandler: rudybot: eval (hash-update #hash{(x . 1)} 'x add1)
17:55 rudybot: chandler: your sandbox is ready
17:55 rudybot: chandler: ; Value: #hash((x . 2))
17:55 chandler: The updater has to be a procudure which receives the existing value and returns a new value
17:55 chandler: "procedure"
17:56 chandler: rudybot: eval (hash-update #hash{(x . 1)} 'y add1 0)
17:56 rudybot: chandler: ; Value: #hash((x . 1) (y . 1))
17:57 errstr: sorry this probably the dumbest question you'll have, but, add1 just adds 1, how do I add any value to the existing one?
17:58 chandler: (lambda (v) (+ v 42)) for instance
17:58 errstr: ah, there we go, thanks. I really appreciate it
17:58 chandler: no problem, hope that helps.
17:59 errstr: yes, it does. thank you.
18:03 samth: cky, places are much like erlang
18:03 samth: with the major difference being that places are effectively separate vms, with separate heaps
18:03 chandler: aren't erlang processes separate heaps too?
18:03 samth: whereas erlang threads are very lightweight
18:04 samth: chandler, no, i don't think so
18:04 samth: chandler, they could be, but i don't think they are
18:04 chandler: I'm pretty sure they are, and that messages are copied between processes
18:05 samth: chandler, well, it all depends what you mean by separate
18:05 chandler: oh, they're not as separate as I thought
18:05 chandler is reading more
18:05 samth: if you mean "there are no pointers from the data owned by one erlang process to another"
18:06 samth: then yes
18:06 chandler: oh, it looks like there's just an exception for a specific type
18:06 samth: if you mean "the erlang/otp vm maintains a separate heap space for each thread" then no, i think
18:07 chandler: I think it does, actually
18:07 samth: there's a paper about places that will know more than me (and probably better compare to erlang)
18:07 samth: http://www.cs.utah.edu/plt/publications/dls11-tsffd.pdf
18:09 chandler: Ah; there's a shared heap for messages, and messages can be allocated directly into the shared heap. Smart.
18:17 (join) yoklov
18:44 (quit) jrslepak: Quit: This computer has gone to sleep
18:56 (quit) gridaphobe: Remote host closed the connection
19:06 (quit) dnolen: Ping timeout: 250 seconds
19:36 (join) Fare
19:37 (join) dnolen
19:53 (join) jonrafkind
19:58 (join) jrslepak
20:02 cky: chandler, samth: Thanks for the neat discussion about places vs Erlang! I'll read that paper.
20:10 (quit) jhemann: Ping timeout: 260 seconds
20:13 (quit) bluezenix: Quit: Leaving.
20:25 (join) lewis1711
20:30 (quit) masm: Quit: Leaving.
20:36 (quit) jrslepak: Quit: This computer has gone to sleep
20:38 (join) jules
20:39 (join) jhemann
20:43 jules: I have a question about contracts. Contracts are for letting two pieces of code that don't trust each other interoperate. For example if Alice wrote a function and Bob uses that function, then Alice wraps the function with her error projection that protects her against malicious inputs given by Bob. Similarly, Bob wraps the function given by Alice with his error projection to protect himself against malicious returned values.
20:43 jules: Is this the right way to think about contracts?
20:46 jules: If so, then it seems to me that the order in which the error projections are applied in the paper "Contracts as Pairs of Projections" in the function guard3 should be the other way around?
20:55 (join) jrslepak
20:58 (join) cdidd
21:00 chandler: jules: I can't answer that question, but I think given the holiday weekend you might want to send that to the user list instead of asking on IRC...
21:03 jules: chandler: ah, right, easter :) thanks for the suggestion, I'll ask it on the mailing list
21:29 asumu: jules: I don't remember all the details of that paper but I can take a look...
21:29 (quit) dented42: Quit: Computer has gone to sleep.
21:32 (part) lewis1711
21:37 asumu: jules: that looks right to me.
21:37 asumu: In the flat case, the order doesn't really matter.
21:38 asumu: In the higher-order case, the client projection should be applied last as it does the function wrapping last.
21:38 asumu: That way, it runs the wrapper of the client (i.e., the argument check) first.
21:45 jules: The server is the guy providing the value, right?
21:46 jules: and the server-projection is the projection that protects the server's integrity?
21:47 jules: However, in the paper, the definition of ho3 is the other way around: the car is called the server-projection, but it protects the integrity of the *client*
21:51 asumu: jules: The server is the guy providing the value, but in the case of a ho contract, that is the function.
21:52 asumu: It's not so much the integrity of the server, but its obligations.
21:52 asumu: The server (providing the function) is obliged to make the function return the right result.
21:52 asumu: Thus, the server projection will check the result.
21:52 jules: I see, so the server-projection protects the client?
21:52 asumu: In a sense, yes.
21:53 asumu: It's about obligation rather than who it protects.
21:53 asumu: (although I suppose it's always both)
21:53 jules: hmm yes
21:55 jules: suppose we did not have a contract library at hand (yet) but Alice wrote a function Int -> Int. She does not want to give Bob the chance to abuse the function, so before she gives it to bob she puts a wrapper around it that checks that the arguments that bob passes are ints.
21:56 jules: Then Bob receives this function, and he puts a wrapper around it that checks that it always returns an int (otherwise his code might misbehave)
21:57 jules: In other words, Alice calls client-projection of (ho int int) on it, and then bob calls server-projection of (ho int int) on it
21:57 jules: so (server-projection (client-projection f))
21:57 jules: but in the paper it is the other way around
21:58 jules: but perhaps this business of everyone protecting himself is not the right way to think about it?
22:02 asumu: Oh, hmm, actually maybe it'll work either way.
22:03 asumu: It will fail quicker in this order though, I think
22:04 asumu: Oh wait.
22:04 asumu: It might matter in the case where you have even more ho contracts.
22:04 jules: well, if this protection is the right way to think about it, then the client-projection does not provide adequate protection
22:04 jules: (the client-projection of ho)
22:05 jules: because if val is not a procedure, the client-projection (the one that alice calls) passes the value on to bob anyway
22:05 jules: so it might be, say, an array that bob can then proceed to modify, and presumably that was not alice's intention
22:06 (join) _danb_
22:06 asumu: Actually, shouldn't both projections check that the thing is a procedure?
22:07 jules: I don't know, but the one in that paper only signals an error one way
22:11 asumu: You mean ho_3?
22:11 jules: yes
22:11 asumu: ho_3 appears to do the procedure check in each projection.
22:11 jules: yes, but in the second case it does not raise an error
22:12 asumu: Ah, yes, I missed that.
22:14 asumu: Well, the order in the paper should be fine for that. Since the check only happens at the wrapping time.
22:15 asumu: And since it's call-by-value, the server projection is run first.
22:15 asumu: During wrapping.
22:15 asumu: So the procedure? check that produces blame will happen first.
22:17 jules: right, I can see that it works out in the end, but I still find it strange that the client's protection protects the server
22:17 jules: or in other words: that the server's obligation protects the server
22:18 asumu: Well, it makes sense that the procedure? check fails there since `val` being a function is an obligation of the server.
22:19 asumu: And since the client actually invokes the procedure, the server is protecting the client.
22:20 asumu: The procedure? check in the other case just makes the blame carry through (I think?)
22:20 asumu: Or maybe not, I have to think about that.
22:21 jules: if the server-projection always runs before the client-projection, then why is the check there?
22:21 jules: then the else-branch of the check is dead code?
22:26 jules: It is time to sleep...perhaps tomorrow I can understand the paper. Thanks a lot for your help!
22:27 asumu: jules: Yeah, no problem. Not sure about the else branch but I'll think about it. Good night.
22:30 (quit) jules: Ping timeout: 245 seconds
22:38 (quit) duomo: Quit: Linkinus - http://linkinus.com
23:39 (join) ofthelesser
23:43 (quit) samth: Ping timeout: 246 seconds