00:02 (quit) francisl: Quit: francisl
00:08 (quit) spiderweb: Quit: ERC Version 5.3 (IRC client for Emacs)
00:13 (join) ambrosebs
00:14 mithos28: What is the difference between raco test -s main file.rkt and racket -e "(require (submod \"compiler.rkt\" main))"
00:15 (quit) dnolen: Ping timeout: 252 seconds
00:19 mithos28: And the answer is heap pressure causing a garbage collection invalidating my callback
00:29 (join) vu3rdd
00:29 (quit) vu3rdd: Changing host
00:29 (join) vu3rdd
00:41 (join) neilv
00:42 (quit) hash_table: Ping timeout: 244 seconds
00:45 (join) mye
00:47 mithos28: http://docs.racket-lang.org/raco/make.html?q=with-out&q=command-line#(def._((lib._setup/parallel-build..rkt)._parallel-compile-files))
00:48 mithos28: No where in that documentation does it say that that function will call exit if there is an error
00:51 (join) sw2wolf
00:52 (part) sw2wolf: "ERC Version 5.3 (IRC client for Emacs)"
00:52 (quit) ambrosebs: Ping timeout: 252 seconds
01:21 (quit) mye: Quit: mye
02:07 (join) jeapostrophe
02:07 (quit) jeapostrophe: Changing host
02:07 (join) jeapostrophe
02:29 (quit) Twey: *.net *.split
02:37 (join) Twey
02:40 (quit) neilv: Quit: Leaving
02:47 ozzloy: is there a way i can do "raco test file.rkt" and have the (module+ test ...) stuff in "file-test.rkt" ?
02:48 mithos28: thats not how test works
02:48 ozzloy: and "raco test file.rkt" still runs those tests?
02:48 ozzloy: i'm glad you're up!
02:48 mithos28: just have file.rkt have  (module+ test (require "test.rkt"))
02:49 mithos28: Its not even midnight yet
02:49 ozzloy tries that
02:51 ozzloy: yeah, that's what i thought.  test.rkt can't see the stuff from file.rkt
02:51 ozzloy: and i could have test.rkt require file.rkt
02:51 ozzloy: but then i'd have to (provide (all-defined-out))
02:51 ozzloy: in file.rkt
02:51 mithos28: Oh do you not want file-test.rkt to be a module
02:52 mithos28: if you don't put a #lang line in file-test.rkt you can use include instead of require
02:52 ozzloy: i'm not sure what i want
02:52 ozzloy: maybe that's what i want
02:52 ozzloy tries
02:54 ozzloy: mithos28, that is what i want!
02:54 ozzloy: now i can put all my tests in some other file, but still do "raco test file.rkt" and everything will work nicely
02:54 mithos28: Why do you want the tests in a different file?
02:55 ozzloy: i have a lot of tests
02:55 mithos28: reasonable answer
02:55 ozzloy: i'm using lex/yacc
02:55 ozzloy: and there's a lot of test cases to cover very small amounts of actual code
02:56 ozzloy: i also suspect i could make my tests smaller
02:56 ozzloy: there's a lot of repeated stuff in each of the tests
02:56 ozzloy: it seems like i want macros, but i'm not confident enough in them to do that yet
02:58 ozzloy: hm...
03:00 ozzloy: http://pastebin.ca/2293772 mithos28 care to make suggestions?
03:01 mithos28: Well first use check-equal? and check-not-equal? instead of check-true (equal?
03:02 ozzloy: i would, but i have to use my own position-token-equal?
03:02 ozzloy: because position-tokens don't compare
03:02 mithos28: What are position-tokens?
03:02 ozzloy: i use check-equal? when i can though
03:03 ozzloy: they're a struct with a symbol, line, column, and offset
03:03 mithos28: Did you define them or did someone else?
03:03 (quit) jonrafkind: Ping timeout: 246 seconds
03:04 ozzloy: someone else.  http://docs.racket-lang.org/parser-tools/Lexers.html
03:04 ozzloy: they're actually a symbol, start-pos and end-pos
03:04 mithos28: And whats wrong with equal method?
03:04 ozzloy: and start-pos and end-pos are position structs, which have line, col, offset
03:05 ozzloy: rudybot, init
03:05 rudybot: ozzloy: your sandbox is ready
03:05 ozzloy: rudybot, (require parser-tools/lex)
03:05 rudybot: ozzloy: Done.
03:05 ozzloy: (require rackunit)
03:05 ozzloy: rudybot, (require rackunit)
03:05 rudybot: ozzloy: Done.
03:05 mithos28: If they don't compare equal that is likely a bug
03:06 ozzloy: rudybot, (define a-start (make-position 0 0 0))
03:06 rudybot: ozzloy: Done.
03:07 ozzloy: rudybot, (define a-end (make-position 0 0 0))
03:07 rudybot: ozzloy: Done.
03:07 mithos28: You have a lot of tests which are the same setup work, you should abstract that out
03:07 ozzloy: this is taking more time than i thought, i'll write a simple thing and pastebin it
03:08 mithos28: I understand the issue
03:08 mithos28: Its a bug
03:08 ozzloy: oh cool
03:08 ozzloy: cuz that would be really awesome to use check-equal?
03:09 mithos28: You should write a function tokenize-vtjson-string that opens it up, counts lines, and calls tokenize-vtjson
03:09 ozzloy: oh genius
03:11 (join) asvil
03:11 mithos28: If for many of the examples you are not going to be using more than 1 column for the positions than you should find a way to not have to say column 1
03:12 asvil: Hi all, are there owners of Windows with more than one cores? Do stable racket releases work for you?
03:12 mithos28: column -> row
03:12 mithos28: asvil: Sorry, don't use Windows
03:12 mithos28: ozzloy: Also automatically compute offset
03:13 mithos28: Until the fix goes in, you should write a check-position-equal? and a check-position-not-equal?
03:15 (quit) myx: Quit: ушёл
03:16 ozzloy: mithos28, i wrote a position-equal? and position-token-equal?, are they good starting points for fixing the problem?
03:16 ozzloy: i'd like to write a patch like that
03:17 ozzloy: i've submitted a few doc spelling fixes, but never actual code
03:18 mithos28: ozzloy: Maybe, but the fix may be to just add a transparent on the struct definition
03:18 ozzloy: mithos28, i was wondering if it would be that, or if there was a reason to keep it ... opaque?  and to write a struct compare method thing
03:19 mithos28: idk
03:19 ozzloy: there's a way to attach a "this is how to check if i'm equal?" to structs, rigth?
03:19 ozzloy: right*
03:21 ozzloy: so i guess that's over my head for now
03:21 ozzloy: i'm not sure what you mean by "you are not going to be using more than 1 column for the positions"
03:21 mithos28: column->row
03:22 mithos28: in all your examples you only have on row
03:22 mithos28: so make it easy to write that case out
03:23 (quit) zuurr: Quit: computer sleeping
03:24 (join) soegaard
03:24 ozzloy: mithos28, you mean all of my lexed text is only 1 line?
03:25 ozzloy: so i could eliminate that part?
03:25 mithos28: in your test cases yes
03:25 mithos28: have a helper which always fills in the row as 0/1
03:25 ozzloy: yeah, that's a good idea
03:26 ozzloy: even if i did add multiple lines, i could do it the long way for those
03:29 ozzloy: mithos28, thanks!
03:29 ozzloy: mithos28, are you going to file a bug report, or should i?
03:30 mithos28: you should, I'm about to head to bed
03:30 ozzloy: heh, me too actually
03:30 ozzloy: i'll fill one out tomorrow.  i'll write up a minimal test case too
03:30 ozzloy: thanks again, and goodnight
03:50 (quit) mithos28: Quit: mithos28
03:51 (quit) karswell_: Remote host closed the connection
04:01 (join) karswell_
04:05 (quit) jackhammer2022: Quit: Textual IRC Client: http://www.textualapp.com/
04:11 (quit) karswell_: Remote host closed the connection
04:12 (quit) jeapostrophe: Ping timeout: 260 seconds
04:21 (join) karswell_
04:57 (join) bitonic
05:53 (join) cdidd
06:02 (join) errstr
06:19 (join) nathanpc
06:22 (nick) otterdam -> mrsotterdam
06:22 (quit) karswell_: Remote host closed the connection
06:24 (join) MayDaniel
06:30 (join) bjz
06:32 (join) karswell_
07:26 (quit) bitonic: Ping timeout: 256 seconds
07:27 (join) kofno
07:29 (join) jao
07:29 (quit) jao: Changing host
07:29 (join) jao
07:38 (quit) soegaard: Quit: soegaard
08:34 (join) masm
08:37 (quit) karswell_: Remote host closed the connection
08:38 (join) mizu_no_oto
08:42 (join) Blkt
08:47 (quit) kofno: Remote host closed the connection
08:47 (join) karswell_
09:05 (quit) karswell_: Remote host closed the connection
09:15 (join) karswell_
09:19 (join) mye
09:21 blubberdiblub: is there something in Racket's library that can represent ranges/intervals? in particular, i'm looking for something that has a lower bound and an upper bound, with the lower bound being at least a fixnum and the upper bound at least a fixnum or infinity. Sequences are no good match, since I don't need the individual members, need to extract the lower and upper bound as easily as possible and need to do an intersection between ranges/intervals. d
09:21 blubberdiblub: o i need to implement that myself?
09:35 greghendershott: blubberdiblub: Maybe data/interval-map ?  http://docs.racket-lang.org/data/interval-map.html
09:36 greghendershott: I haven't tried it, just recalled having seen it in the docs and went looking for it now.
09:43 blubberdiblub: greghendershott, thanks, but it looks like it has a completely different purpose. i don't need any mapping functionality. it also needs to be continuous and really needs a way to represent an infinite upper bound
09:50 greghendershott: blubberdiblub: Oh OK. So you want something like (struct from upto) but you were hoping not to write the intersection yourself?
09:50 blubberdiblub: greghendershott, the intersection and infinity handling, yes
09:50 (join) masm1
09:50 (quit) mye: Quit: mye
09:51 (quit) masm: Ping timeout: 256 seconds
09:52 (quit) mizu_no_oto: Quit: Computer has gone to sleep.
09:52 blubberdiblub: greghendershott, anyway, i could write it myself. i should probably read the guide on modules first, tho
09:53 greghendershott: blubberdiblub: I think so. BTW there is a +inf.0 constant. Just keep in mind it's float not integer, in case that matters.
09:54 (join) kofno
09:54 blubberdiblub: greghendershott, yes, thanks, i'm aware of that, tho ;) i think i prefer something like 'inf or 'infinity for my case
10:00 (join) mye
10:03 (quit) masm1: Ping timeout: 245 seconds
10:08 (join) soegaard
10:17 (join) masm
10:21 (quit) karswell_: Remote host closed the connection
10:26 (quit) masm: Ping timeout: 248 seconds
10:28 (join) zyoung
10:31 (join) karswell_
10:43 (join) spiderweb
10:55 (join) jeapostrophe
11:10 (join) mizu_no_oto
11:13 (quit) karswell_: Remote host closed the connection
11:15 (quit) mizu_no_oto: Ping timeout: 248 seconds
11:19 (quit) Nisstyre: Quit: Leaving
11:22 (quit) jao: Ping timeout: 245 seconds
11:23 (join) karswell_
11:51 (join) masm
11:52 (quit) karswell_: Remote host closed the connection
11:55 (join) hash_table
11:57 Haffe: Hmmm.
11:57 Haffe: It sounds like intervall arithmetic.
12:00 (join) anRch
12:02 (join) mithos28
12:03 (join) karswell_
12:04 (quit) soegaard: Quit: soegaard
12:13 (join) soegaard
12:16 (quit) spiderweb: Quit: ERC Version 5.3 (IRC client for Emacs)
12:17 (quit) jeapostrophe: Ping timeout: 260 seconds
12:33 (quit) anRch: Quit: anRch
12:34 (quit) karswell_: Read error: Connection reset by peer
12:44 (join) karswell_
12:47 (quit) karswell_: Remote host closed the connection
12:47 (join) anRch
12:49 (quit) masm: Ping timeout: 255 seconds
12:52 (join) mizu_no_oto
12:59 (join) karswell
13:42 (quit) mithos28: Quit: mithos28
13:42 (join) masm
13:42 (quit) soegaard: Quit: soegaard
13:45 (quit) anRch: Quit: anRch
13:52 (join) soegaard
14:06 (join) Nisstyre-laptop
14:12 (quit) kofno: Ping timeout: 245 seconds
14:16 (join) kofno
14:17 (join) jao
14:17 (quit) jao: Changing host
14:17 (join) jao
14:21 (quit) mye: Quit: mye
14:32 (join) netrino
14:36 (quit) kofno: Remote host closed the connection
14:37 (join) nejucomo
14:44 (join) mithos28
14:52 (quit) mizu_no_oto: Quit: ["Textual IRC Client: www.textualapp.com"]
14:52 (join) kofno
14:58 (quit) dca: Ping timeout: 255 seconds
15:01 (join) dca
15:01 (join) francisl
15:03 (quit) mithos28: Quit: mithos28
15:04 (quit) francisl: Client Quit
15:13 (quit) kofno: Remote host closed the connection
15:14 (join) h11r
15:27 (quit) soegaard: Quit: soegaard
15:28 h11r: Hi, I've got a quick question about polymorphism and racket. Some time ago I asked myself if it is possible to extend racket datatypes in the way you can extend classes in CLOS by adding a dynamically dispatched method outside of the data-types definition. The intend is to extend an existing type (struct, class or whatever) by an additional interface. Someone gave me the link to the really interesting paper ' Modular Object-Oriented Programming wit
15:29 (join) francisl
15:31 offby1: Did you really describe that as a "quick" question?
15:31 offby1: yes, you did
15:31 h11r: :D
15:32 h11r: Depends on the answer ;)
15:34 (join) soegaard
15:35 h11r: I ask because by this feature of CLOS Common lisp allows to extend the base system in a rather easy way and racket with its emphasize on language design seems like a candidate for a similar feature. But I didn't find one yet and I'm about to decide whether to use CL, Clojure or Racket for my next project. So I may have missed something
15:38 soegaard: h11r: Are you looking for something like this? http://docs.racket-lang.org/reference/struct-generics.html
15:45 h11r: Yes, in principle. But I have to define the methods for the struct inside the struct definitions, am I right? So I tried something like (struct <struct-name> () #:super struct:<struct-name> #:methods ... ) to override the original definition by a struct definition that implements the new interface. But racket/set seems not to export the struct-id (It fails with 'struct:set: undefined;' at least). Is there another way to use the generics for this pu
15:45 h11r: I'll be back in about 15 minutes because now is my last chance to get some food for the weekend ;)
15:46 (quit) francisl: Quit: francisl
16:05 h11r: So, I'm back. Any Ideas :) ?
16:11 (quit) hash_table: Ping timeout: 255 seconds
16:12 (join) zuurr
16:13 (join) RacketCommitBot
16:13 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://git.io/TJfczA
16:13 RacketCommitBot: racket/master 0b09538 Robby Findler: improve the performance of typing a close paren
16:13 (part) RacketCommitBot
16:17 soegaard: h11r: Try asking your question on the mailing list.
16:23 h11r: soegaard: Yes, I'll give it a try. And thanks anyway.
16:26 h11r: soegaard: But maybe you can tell me if there is a way to get the type of an object in form of a symbol or identifier so that I can build some kind of dispatch table on my own. I know the predicates but ids could come in handy in this case.
16:27 soegaard: Maybe http://docs.racket-lang.org/reference/inspectors.html?q=describ#(def._((quote._~23~25kernel)._object-name))
16:27 soegaard: ?
16:29 soegaard: If you are working with structs, then see: http://docs.racket-lang.org/reference/structprops.html?q=struct&q=struct&q=describ
16:29 rudybot: http://tinyurl.com/cog7b2r
16:29 (join) dyoo
16:32 (join) dzhus
16:34 h11r: Yes, but either I don't get the inspectors or they have the same problem as my struct-replace approach: The racket primitives seem to not provide these.
16:34 h11r: The structure type properties seem promising and I'll take a look at them. Thanks for the tip.
16:37 dyoo: h11r: can you use struct-info to get at this information?
16:37 dyoo: Let me see if I can code up an example...
16:41 (join) bitonic
16:41 h11r: (struct-info (set 'a 'b)) returns #f
16:43 soegaard: rudybot: (struct? (set 'a 'b))
16:43 rudybot: soegaard: your sandbox is ready
16:43 rudybot: soegaard: ; Value: #f
16:43 soegaard: Sets are not represented as structs.
16:44 dyoo: h11r: here's a small example of struct-info: https://gist.github.com/4299528
16:44 (join) mizu_no_oto
16:45 dyoo: The thing is that the reflective API will not open up the abstraction unless it's allowed to, either with the structs being transparent, or the current inspector having high enough authority to look.
16:49 (quit) nejucomo: Ping timeout: 276 seconds
16:51 dyoo: h11r: here's another example that explicitly plays with inspectors: https://gist.github.com/4299589
16:52 h11r: oh. I didn't realize that define-struct does not define a struct? ... But this kind of illustrates the whole issue. With all this different ways of representing data it seems really difficult to implement an efficient displatch mechanism.
16:52 h11r: And thanks for the examples.
16:53 soegaard: ? define-struct does define structs.
16:53 soegaard: It just happens that sets are implemented in C.
16:53 soegaard: I think.
16:53 h11r: (define-struct a ())
16:53 h11r: (struct? (make-a)) => #f
16:55 soegaard: huh?
16:56 dyoo: soegaard: struct? is poorly named.  It's a predicate that returns true if the value is inspectable via struct-info.
16:56 soegaard: Okay.
16:58 soegaard: rudybot: (define-struct a() #:transparent) (struct? (make-a))
16:58 rudybot: soegaard: (struct? (make-a)) => #f
16:58 dyoo: rudybot: (define insp (current-inspector))
16:58 rudybot: dyoo: your sandbox is ready
16:58 rudybot: dyoo: Done.
16:59 dyoo: rudybot: (define child-insp (make-inspector insp))
16:59 rudybot: dyoo: Done.
16:59 soegaard: Fwiw that evaluated to #t in DrRacket.
16:59 bartbes: it does?
16:59 dyoo: rudybot: (current-inspector child-insp)
16:59 rudybot: dyoo: Done.
17:00 bartbes: oh, toy mean the transparent one?
17:00 bartbes: *you
17:00 soegaard: yes
17:00 dyoo: rudybot: (define-struct a ())
17:00 rudybot: dyoo: Done.
17:00 dyoo: rudybot: (struct? (make-a))
17:00 rudybot: dyoo: ; Value: #f
17:00 dyoo: rudybot: (current-inspector insp)
17:00 rudybot: dyoo: Done.
17:00 dyoo: rudybot: (struct? (make-a))
17:00 rudybot: dyoo: ; Value: #t
17:01 (quit) bitonic: Ping timeout: 264 seconds
17:01 dyoo: DrRacket has to be doing something with inspectors to make the results different from base Racket.
17:02 bartbes: the transparent one worked for me in the cli
17:03 (join) anRch
17:08 h11r: I mean you could wrap all the primitives in tagged types (like lists) as they do in the generic programming examples in sicp, but this is a lot of work that you may have to do for every project your on. And since this idea is not too exotic compared to the other ideas in racket, i thought there would be a build in way. But i discovered the multimethods planet package and I am reading the docs at the moment. This could be a solution for the problem.
17:11 dyoo: h11r: yeah.  As far as I know, things that dispatch in Racket have a structure like the following: handle the base types via explicit case analysis, and expose a general protocol for hooking into the dispatch.  e.g. the serialization protocol http://docs.racket-lang.org/reference/serialization.html  explicitly knows how to handle base Racket types, with a hook (serializable-struct) to extend the dispatch.
17:13 (join) jackhammer2022
17:14 (join) kofno
17:16 (join) Rajah
17:16 (quit) Rajah: Client Quit
17:16 (join) Rajah
17:19 h11r: I wonder how much effort it would be to build some kind of generic type function that words on all objects (lisp objects not object-system objects) and returns a id that can be ordered. Than you could build some dispatch mechanism using balanced trees that would perform well. At the moment an idea would be to store the predicates in the dispatch table but than you would have to try all predicates in sequence until one succeeds.
17:21 h11r: I mean the system should require such type info anyway.
17:22 dyoo: h11r: I suspect we can do this, but we have to be willing to break abstractions.  It should all be technically accessible http://docs.racket-lang.org/inside/im_values_types.html , using the FFI
17:30 (quit) soegaard: Quit: soegaard
17:30 (join) bitonic
17:30 h11r: Using the FFI for this kind of thing seems a bit brittle. But if this gives a neat solution and the solution would be integrated into standard racket than i suppose it would work. The question is how well you can incorporate user defined structs into this solution.
17:33 (join) Kiloneie
17:33 Kiloneie: hi
17:35 Kiloneie: Anyone here ?
17:36 (quit) bitonic: Ping timeout: 264 seconds
17:36 (join) nejucomo
17:37 Kiloneie: hi
17:39 h11r: Hi
17:42 h11r: dyoo: The multimethod package is kind of useful but also has the problem with getting the type of an object. The solution in the package is that you have to pass the multimethod an signature-generator, a procedure that retrieves the types/signatures of the passed objects. And I can't come up with an implementation for this signature generator that isn't a single big cond expression => slow and cumbersome.
17:42 h11r: But I'll give it a rest and think about it. Maybe someone on the mailing list has an idea.
17:43 h11r: So, good night and thanks for your help and opinion.
17:43 h11r: Bye.
17:43 dyoo: h11r: yeah.  I suspect Asumu and the others have been thinking about this, with regards to generics
17:43 dyoo: good luck!
17:43 h11r: Thanks :)
17:44 (join) soegaard
17:48 (quit) h11r: Ping timeout: 264 seconds
18:13 (quit) anRch: Quit: anRch
18:15 (quit) dyoo: Quit: dyoo
18:17 (quit) masm: Ping timeout: 255 seconds
18:21 (quit) MayDaniel: Read error: Connection reset by peer
18:26 (join) francisl
18:28 Kiloneie: Why does noone talk here ?
18:30 jao prefers to whistle and sing
18:30 Kiloneie: o.O
18:30 Kiloneie: O.o
18:30 Kiloneie: O.O
18:31 (quit) Kiloneie: Quit: Page closed
18:31 (quit) francisl: Quit: francisl
18:33 (join) neilv
18:33 (quit) dzhus: Remote host closed the connection
18:37 (join) francisl
18:39 (join) masm
18:40 (quit) Rajah: Remote host closed the connection
18:41 (quit) mizu_no_oto: Quit: Computer has gone to sleep.
18:41 (join) spiderweb
18:47 (quit) soegaard: Quit: soegaard
18:52 (join) masm1
18:54 (quit) masm: Ping timeout: 248 seconds
19:07 (quit) Blkt: Read error: Connection reset by peer
19:15 (quit) francisl: Quit: francisl
19:19 (join) jonrafkind
19:41 (quit) cdidd: Remote host closed the connection
19:53 (quit) masm1: Quit: Leaving.
20:00 (join) dnolen
20:13 spiderweb: what is the name of racket's init file?
20:18 spiderweb: .racketrc
20:41 (join) mizu_no_oto
21:04 (quit) neilv: Quit: Leaving
21:06 (quit) spiderweb: Ping timeout: 256 seconds
21:16 (quit) mizu_no_oto: Quit: Computer has gone to sleep.
21:18 (join) mizu_no_oto
21:21 (join) sw2wolf
21:28 (quit) netrino: Quit: Ave!
21:38 (quit) jackhammer2022: Ping timeout: 245 seconds
21:41 (quit) mizu_no_oto: Quit: Computer has gone to sleep.
22:00 (join) jackhammer2022
22:02 (quit) nathanpc: Quit: Computer has gone to sleep.
22:07 (quit) nejucomo: Ping timeout: 276 seconds
22:11 (quit) jackhammer2022: Ping timeout: 245 seconds
22:12 (part) sw2wolf: "ERC Version 5.3 (IRC client for Emacs)"
22:16 (join) gcr`
22:16 gcr`: So if I do (define-cstruct _Name ()), the blank cstruct makes racket spit out the cryptic "internal error: ffi_prep_cif did not return FFI_OK"
22:16 gcr`: just letting others know in case they run into that
22:19 (join) ambrosebs
22:22 (quit) Demosthenex: Ping timeout: 244 seconds
22:24 (join) Demosthenex
22:37 (join) nathanpc
22:40 (join) jackhammer2022
22:44 (quit) nathanpc: Quit: Computer has gone to sleep.
22:49 (quit) jackhammer2022: Ping timeout: 264 seconds
22:54 (join) jackhammer2022
22:58 (quit) Shviller: Ping timeout: 245 seconds
22:59 (join) Shviller
22:59 (join) mizu_no_oto
23:07 (quit) mizu_no_oto: Quit: Computer has gone to sleep.
23:26 (join) spiderweb
23:38 (quit) kofno: Read error: Connection reset by peer
23:41 (join) kofno
23:50 (quit) spiderweb: Quit: ERC Version 5.3 (IRC client for Emacs)