00:13 (quit) dherman: Quit: dherman
00:17 (join) racket
00:17 racket: [racket] plt pushed 1 new commit to master: http://bit.ly/pU3QpS
00:17 racket: [racket/master] Better fix for the previous bug. - Guillaume Marceau
00:17 (part) racket
00:26 (quit) misterm: Ping timeout: 258 seconds
00:31 (join) dherman
00:32 (join) B4R0N
01:01 (quit) elliottcable: Quit: rage
01:02 (join) elliottcable
01:06 (quit) cmpitg: Ping timeout: 240 seconds
01:08 (quit) dherman: Quit: dherman
01:11 (part) B4R0N: "Leaving"
01:27 (join) EgnartsMs
01:36 (nick) elliottcable -> Guest82573
01:37 (nick) Guest82573 -> elliottcable
01:43 (quit) EgnartsMs: Quit: Page closed
01:54 (join) racket
01:54 racket: [racket] plt pushed 1 new commit to master: http://bit.ly/mUSgqO
01:54 racket: [racket/master] Fixed some missing links in the documentation. - Guillaume Marceau
01:54 (part) racket
02:08 (quit) Daemmerung: Quit: Smoove out.
02:36 (join) Daemmerung
02:47 (quit) jonrafkind: Ping timeout: 252 seconds
03:20 (join) noelwelsh
04:16 (quit) rgrinberg_: Read error: Connection reset by peer
04:47 (join) tfb
05:10 (quit) DT``: Read error: Connection reset by peer
05:11 (join) DT``
06:10 (join) MayDaniel
06:15 (quit) Daemmerung: Read error: Connection reset by peer
06:18 (quit) Penten: Remote host closed the connection
06:21 (join) lucian
06:26 (join) Eren
06:26 Eren: hello
06:36 (join) mceier
06:37 (quit) MayDaniel: Read error: Connection reset by peer
07:12 noelwelsh: hi
07:16 dsp_: hi folks. does this look particularly slow? http://www.technoanimal.net/siege.txt
07:16 dsp_: some benchmarks from a racket+continue service
07:17 dsp_: the gap between longest transaction and shortest transaction is worrying but i'll try on some beefier hardware later
07:20 dsp_: i suspect this is just GC pauses :(
07:25 noelwelsh: Yeah GC pauses shouldn't be that large, unless you have lots and lots of RAM
07:26 noelwelsh: We see a few secs max with 1-2GB RAM
07:26 dsp_: this machine has 300mb of ram
07:26 dsp_: not sure if that might be the issue
07:27 noelwelsh: That ain't a lot of RAM. Maybe you're hitting stack space?
07:27 noelwelsh: Are you making heavy use of continuations?
07:27 dsp_: not at all
07:27 noelwelsh: Ok, good ;)
07:27 dsp_: in fact, i'm only using continue since it does everything for me wrt concurrency
07:27 dsp_: i don't use any of the real features
07:27 noelwelsh: Yeah. Continuations can blow memory use if you don't take care.
07:28 dsp_: i just use it to grab input that is posted
07:28 dsp_: and check if it's in json format
07:28 dsp_: if it is, it goes and unwraps it and hits db and wraps result up and sends back to client
07:28 noelwelsh: But, 300MB RAM total doesn't give much to spend between a DB machine and a Racket process
07:28 noelwelsh: We usually give PostgreSQL a few hundred meg at least.
07:28 noelwelsh: Depends on use cases, I suppose.
07:28 dsp_: this is just a test machine
07:29 noelwelsh: But I'd take a look at memory usage and see if you're hitting swap
07:29 dsp_: the memory free is usually in the hundreds of megs
07:29 noelwelsh: Oh, ok.
07:30 noelwelsh: Well I don't think GC pauses should lead to 27s responses.
07:30 noelwelsh: That faaaaaar too long.
07:30 dsp_: yeah :/
07:30 dsp_: and another time, when i was stress testing, racket just stopped responding entirely
07:30 dsp_: not fun
07:31 noelwelsh: Hmmm... not something I've seen
07:32 noelwelsh: Did Racket segfault, or just hang?
07:32 dsp_: just hung
07:32 dsp_: was very odd indeed
07:32 noelwelsh: If you can get a reproducible setup I expect the core Racket maintainers would be very interested.
07:32 dsp_: i'll try my best
07:32 noelwelsh: Matthew Flatt is extremely (overly, almost) generous with his time.
07:33 dsp_: i think the next step for me is to just try this on a beefier machine
07:33 dsp_: it's just odd that i get requests going through with response times in fractions of seconds while others take 12+
07:34 dsp_: i'd have expected them to be somewhat symmetric
07:34 dsp_: might be cpu-bound. will try and report back if there are changes :)
07:38 noelwelsh: try to get a log of cpu and mem usage every few secs
07:39 noelwelsh: there a zillion tools that do this
07:43 (join) cmpitg
07:44 dsp_: heh well when moving it to another machine and using the old one just for postgres i managed to get postgresql-connect: sorry, too many clients already (SQLSTATE 53300)
07:44 dsp_: so that's a step up :)
07:45 dsp_: max_connections was only set to 100 though
07:52 dsp_: oh here's something that i've hit
07:52 dsp_: force: reentrant promise fun-p
07:53 dsp_: perhaps i am doing something i'm not supposed to
07:54 dsp_: no idea what it means :(
07:59 (quit) cmpitg: Ping timeout: 255 seconds
07:59 noelwelsh: is that the complete error message?
07:59 noelwelsh: hitting up the lunch-a-tarium
08:00 dsp_: yep
08:00 (join) cmpitg
08:01 lisppaste: dsp pasted "reentrant promise fun-p" at http://paste.lisp.org/display/123263
08:02 dsp_: i'm not using promises in my own code, so i suspect this is either from continue or another library i am using
08:02 dsp_: only seems to happen under stress
08:09 dsp_: and there's a crash! thread: the custodian has been shut down: #<custodian>
08:10 dsp_: continue is no longer listening
08:11 dsp_: seems that comes from scheme_arg_mismatch, though not 100%
08:12 dsp_: (just googled it)
08:17 (quit) hyko: Read error: Connection reset by peer
08:17 (join) hyko
08:23 (quit) cmpitg: Read error: Connection reset by peer
09:11 dsp_: triggered it again.
09:13 (join) tzhuang
09:16 tzhuang: hey guys
09:17 tzhuang: anybody here code with vi?
09:23 dsp_: also another possible bug: output-file: tcp-write: error writing (Broken pipe; errno=32)
09:25 lisppaste: dsp pasted "simple test code" at http://paste.lisp.org/display/123264
09:26 dsp_: this taken from the continue example, stressing it with siege causes those tcp-write messages to fill the console
09:32 noelwelsh: The broken pipe probably means the web server closed a connection due to excessive time to respond or similar.
09:32 noelwelsh: dunno about the others.
09:32 noelwelsh: also: EMACS
09:36 dsp_: yeah, that might be the cause. the server was sometimes taking ages to respond again
09:36 dsp_: cpu-bound i believe
09:36 dsp_: but yeah, the other stuff is more worrying, since it just dies completely
09:37 (join) svk_
09:37 dsp_: perhaps i should post on the list.
09:41 svk_: Hi, I'm having some trouble trying to embed Racket, I get a crash inside scheme_free_immobile_box()
09:41 svk_: I've tracked it down to GC_get_GC() returning NULL in GC_free_immobile_box(), which seems to be assuming it shouldn't.. is that a tip-off for whatever I'm doing wrong?
09:43 svk_: This would be 5.1.1 on Linux, the 3m .so
09:45 noelwelsh: svk: I suggest posting that to the list.
09:48 svk_: The appropriate list being users, or dev? In any case I want to investigate a little further first, it's probably some newbie mistake
09:55 noelwelsh: users, probably
09:56 (join) EgnartsMs
09:56 svk_: Okay, thanks
10:08 noelwelsh: dsp_: the reentrant promise fun-p error comes from deep within racket (collects/racket/private/promise.rkt)
10:08 noelwelsh: I can't find anything in the DB driver forcing promises
10:10 noelwelsh: So it is almost certainly coming from the web-server guts
10:10 noelwelsh: (request-structs.rkt)
10:12 dsp_: i am using a few other libraries however
10:12 dsp_: neil/json-parsing and dherman/json
10:12 dsp_: but i don't know why they'd be using promises
10:12 noelwelsh: Hmmm...
10:13 noelwelsh: The web-server call that uses promises is request-bindings/raw
10:13 noelwelsh: I don't know about the others
10:14 dsp_: just checked and json libs don't use it, so yes must be in the web-server
10:17 dsp_: i'll post to users list about that and the thread: the custodian has been shut down: #<custodian> error i think
10:18 dsp_: appreciate the help
10:35 (quit) EgnartsMs: Quit: Page closed
11:15 (join) cmpitg
11:21 (join) Daemmerung
11:22 (join) misterm
11:29 (join) masm
11:32 (join) egnarts-ms
12:03 (join) jonrafkind
12:03 (join) dherman
12:03 (join) racket
12:03 racket: [racket] plt pushed 1 new commit to master: http://bit.ly/o7exSN
12:03 racket: [racket/master] sk requested empty scenes with optional background color - Matthias Felleisen
12:03 (part) racket
12:06 (quit) noelwelsh: Quit: noelwelsh
12:06 (join) racket
12:06 racket: [racket] plt pushed 2 new commits to master: http://bit.ly/q6XvY7
12:06 racket: [racket/master] Switch tcsh(!) and bash scripts to use /bin/sh. - Eli Barzilay
12:06 racket: [racket/master] Added Guillaume's gmail address to mailmap - Eli Barzilay
12:06 (part) racket
12:13 (quit) tfb: Read error: Connection reset by peer
12:13 (join) racket
12:13 racket: [racket] plt pushed 1 new commit to master: http://bit.ly/noYGGx
12:13 racket: [racket/master] Shift typo. - Eli Barzilay
12:13 (part) racket
12:15 (join) racket
12:15 racket: [racket] plt pushed 4 new commits to master: http://bit.ly/qvzzVE
12:15 racket: [racket/master] rename jobqueue and workqueue to work-queue - Kevin Tew
12:15 racket: [racket/master] Fix timeout for place-channel.rkt test - Kevin Tew
12:15 racket: [racket/master] fix class names - Kevin Tew
12:15 (part) racket
12:55 (quit) misterm: Read error: Operation timed out
13:01 (join) racket
13:01 racket: [racket] plt pushed 1 new commit to master: http://bit.ly/nv45Rn
13:01 racket: [racket/master] Corrected an error message that said it wanted a real, but actually - Stephen Bloch
13:01 (part) racket
13:04 (quit) cmpitg: Read error: Connection reset by peer
13:09 (join) racket
13:09 racket: [racket] plt pushed 1 new commit to master: http://bit.ly/nalUYY
13:09 racket: [racket/master] Corrected the grammar in check-arg and check-result to follow the - Stephen Bloch
13:09 (part) racket
13:12 (join) tauntaun
13:14 (join) MayDaniel
13:16 (quit) egnarts-ms: Quit: Page closed
13:31 (join) ISF_
13:43 (quit) MayDaniel: Read error: Connection reset by peer
14:03 (join) anRch
14:29 (join) mithos28
14:47 (quit) dherman: Read error: Connection reset by peer
14:47 (join) dherman
14:58 (join) littlebobby
15:06 (quit) dherman: Quit: dherman
15:31 (quit) anRch: Quit: anRch
15:55 (quit) mithos28: Quit: mithos28
15:59 (join) didi
16:06 _p4bl0: Hi there, anybody knows where I could find a pdf of the paper "Macro writers bill of rights" by Dybvig [Friedman Feschrift '07] which appear in the "Macros matter" talk by Robert Findler ?
16:07 _p4bl0: (All I can find are a video talk and slides, but not a paper)
16:10 (join) mithos28
16:12 (nick) samth_away -> samth
16:15 samth: _p4bl0: there is not such paper, thus it's hard to find :)
16:15 samth: *no
16:20 _p4bl0: oh, then I guess R. Findler was refering to the talk, but I really didn't get that impression when watching the video
16:20 _p4bl0: (see http://www.youtube.com/watch?v=GBpfOpk-ZBU at 17:36)
16:20 _p4bl0: anyway, thanks !
16:22 _p4bl0: yup, just found a paper which quotes "the macro-writer's bill of right" and it's the video talk which is listed in the references.
16:22 _p4bl0: I'll watch the video then :-)
16:24 (join) dherman
16:30 samth: the talk is recommending
16:38 (quit) mithos28: Quit: mithos28
16:40 (join) racket
16:40 racket: [racket] plt pushed 1 new commit to master: http://bit.ly/njeCOW
16:40 racket: [racket/master] Corrected signature of scene+line to match examples and actual behavior - Stephen Bloch
16:40 (part) racket
16:43 _p4bl0: samth: sorry to ask, but I'm not sure what you mean. "the talk is recommending" means that the talk recommends other stuff or that you recommend it (but phrased in a way that make it sound like an absolute truth value because the talk is so good)? (or yet something else maybe? ^^)
16:44 samth: sorry, that was just a typo
16:44 samth: the talk is recommended (by me)
16:44 _p4bl0: ah okay
16:44 _p4bl0: thanks
17:07 (join) racket
17:07 racket: [racket] plt pushed 10 new commits to master: http://bit.ly/nYSUsk
17:07 racket: [racket/master] TR: Produce better error messages with values. Closes PR11772. - Eric Dobson
17:07 racket: [racket/master] Working on improving contracts in TR. - Eric Dobson
17:07 racket: [racket/master] Fixes for inference. Closes PR10937. - Eric Dobson
17:07 (part) racket
17:18 samth: eli: can you add me to the list of people who get email about pull requests?
17:22 (join) racket
17:22 racket: [racket] plt pushed 3 new commits to master: http://bit.ly/p24hGS
17:22 racket: [racket/master] New `xrepl' collection. - Eli Barzilay
17:22 racket: [racket/master] Add a new `#:dont-re-require-enter' flag for `enter!', to avoid - Eli Barzilay
17:22 racket: [racket/master] Use the new `set-port-next-location!' instead of the previous hack. - Eli Barzilay
17:22 (part) racket
17:26 jonrafkind: xrepl? fail
17:29 (join) mithos28
17:31 (join) racket
17:31 racket: [racket] plt pushed 15 new commits to release: http://bit.ly/pNmad3
17:31 racket: [racket/release] Fixed some more error messages. - Stephen Bloch
17:31 racket: [racket/release] fix bad contract - Robby Findler
17:31 racket: [racket/release] rebuild Mac x86 libraries to work with 10.4 - Matthew Flatt
17:31 (part) racket
17:32 jonrafkind: hm, seems the racket bot only prints the first 3 commit messages..
17:39 (quit) mithos28: Quit: mithos28
17:54 jonrafkind: raco setup: running: version/version.scrbl
17:54 jonrafkind: Warning World!
17:56 (join) asumu
18:00 jonrafkind: eli, do you think the shorthands for xrepl could work like gdb where the shortest unique string succeeds?
18:01 (join) mithos28
18:09 (join) dherman_
18:09 (quit) dherman: Read error: Connection reset by peer
18:09 (nick) dherman_ -> dherman
18:09 (nick) samth -> samth_away
18:18 (quit) mithos28: Quit: mithos28
18:20 (join) mithos28
18:22 (quit) mithos28: Client Quit
18:25 (join) racket
18:25 racket: [racket] plt pushed 1 new commit to master: http://bit.ly/oMMuSD
18:25 racket: [racket/master] Fix a few framework contracts to match code - Asumu Takikawa
18:25 (part) racket
18:33 (quit) ISF_: Ping timeout: 252 seconds
18:35 (join) mithos28
18:48 (join) ISF_
18:51 (join) racket
18:51 racket: [racket] plt pushed 1 new commit to master: http://bit.ly/pnihLN
18:51 racket: [racket/master] Fix unbalanced curly brackets. - Vincent St-Amour
18:51 (part) racket
18:55 (join) racket
18:55 racket: [racket] plt pushed 3 new commits to master: http://bit.ly/nCyvtf
18:55 racket: [racket/master] fix parsing for operators - Jon Rafkind
18:55 racket: [racket/master] enable #lang honu - Jon Rafkind
18:55 racket: [racket/master] fix sequences - Jon Rafkind
18:55 (part) racket
19:03 (quit) masm: Quit: Leaving.
19:04 (quit) mithos28: Quit: mithos28
19:05 (join) martinhex
19:13 (join) mithos28
19:24 (join) B4R0N
19:58 (quit) mithos28: Quit: mithos28
20:02 (quit) ISF_: Ping timeout: 260 seconds
20:03 (join) mithos28
20:09 (quit) DT``: Read error: Connection reset by peer
20:20 (join) DT``
20:21 (quit) dherman: Quit: dherman
20:22 (join) racket
20:22 racket: [racket] plt pushed 4 new commits to master: http://bit.ly/n3u5jJ
20:22 racket: [racket/master] fix `procedure-reduce-keyword-arity' checking - Matthew Flatt
20:22 racket: [racket/master] more info on 10.4 support - Matthew Flatt
20:22 racket: [racket/master] tweak JIT to improve and slightly generalize register tracking - Matthew Flatt
20:22 (part) racket
20:22 (join) dnolen_
20:24 (quit) mceier: Quit: leaving
20:33 (quit) jonrafkind: Ping timeout: 252 seconds
20:35 (quit) mithos28: Quit: mithos28
20:39 (quit) lucian: Remote host closed the connection
20:52 (join) ISF_
21:00 (join) mithos28
21:16 (join) Penten
21:36 (join) racket
21:36 racket: [racket] plt pushed 1 new commit to master: http://bit.ly/q3bPAV
21:36 racket: [racket/master] Made continuation-mark-set-first future-safe.  Fixed a rarely occurring bug with lightweight continuation capture for futures. - James Swaine
21:36 (part) racket
21:37 (quit) mithos28: Quit: mithos28
21:42 (join) misterm
21:44 (join) racket
21:44 racket: [racket] plt pushed 1 new commit to master: http://bit.ly/rnjV1f
21:44 racket: [racket/master] Deleted junk code that was commented out in the futures test suite - James Swaine
21:44 (part) racket
21:47 (join) mithos28
21:58 (quit) ISF_: Ping timeout: 252 seconds
22:34 (quit) tzhuang: Ping timeout: 252 seconds
22:58 (join) shofetim
23:30 (join) dherman
23:33 (join) racket
23:33 racket: [racket] plt pushed 3 new commits to master: http://bit.ly/qrdrFJ
23:33 racket: [racket/master] fix typo; eliminate "this" as a noun; otherminor improvements - Matthew Flatt
23:33 racket: [racket/master] remove obsolete reference to '#%mred-kernel - Matthew Flatt
23:33 racket: [racket/master] places: fix printing of symbol resolved module paths - Matthew Flatt
23:33 (part) racket
23:35 (quit) martinhex: Read error: Connection reset by peer
23:36 (join) ISF
23:38 (join) martinhex
23:39 (part) B4R0N: "Leaving"
23:51 (join) jonrafkind