00:05 (quit) flaggy_: Ping timeout: 258 seconds
00:08 (join) RacketCommitBot
00:08 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://git.io/HS3frg
00:08 RacketCommitBot: [racket/master] "typed-scheme" -> "typed-racket". - Eli Barzilay
00:08 (part) RacketCommitBot
00:30 (quit) jeapostrophe: Quit: jeapostrophe
00:34 (join) RacketCommitBot
00:34 RacketCommitBot: [racket] plt pushed 2 new commits to master: http://git.io/D8td5Q
00:34 RacketCommitBot: [racket/master] Some more "typed-scheme" -> "typed-racket". - Eli Barzilay
00:34 RacketCommitBot: [racket/master] Remove uses of `scheme' in db docs. - Eli Barzilay
00:34 (part) RacketCommitBot
00:42 (quit) realitygrill: Read error: Connection reset by peer
00:46 (join) realitygrill
00:49 (join) yoklov
01:03 (quit) yoklov: Quit: Leaving.
01:05 (join) yoklov
01:22 (quit) yoklov: Quit: Leaving.
01:23 (join) yoklov
01:57 (quit) yoklov: Quit: Leaving.
02:01 (join) MayDaniel
02:06 (quit) MayDaniel: Read error: Connection reset by peer
02:18 (quit) Lajla: Quit: Phallic Illuminatus
02:22 (quit) realitygrill: Quit: realitygrill
02:34 (join) Lajla
02:38 (quit) dwaynecrooks: Ping timeout: 240 seconds
02:51 (join) dwaynecrooks
02:52 (join) hkBst
03:00 (quit) dwaynecrooks: Ping timeout: 260 seconds
03:05 (quit) jonrafkind: Ping timeout: 240 seconds
03:16 (join) bluezenix
03:26 (join) mceier
03:59 (quit) bluezenix: Ping timeout: 240 seconds
04:38 (join) noelw
05:15 (quit) mceier: Ping timeout: 264 seconds
05:21 (quit) eli: Ping timeout: 240 seconds
05:31 (join) mceier
05:53 (join) masm
06:22 (join) Burlingk
07:31 (quit) Lajla: Read error: Connection reset by peer
07:31 (join) Lajla
07:32 (join) JuanDaugherty
07:46 (join) neilv
08:35 (join) jeapostrophe
08:54 (join) mceier_
08:55 (quit) mceier: Disconnected by services
08:55 (nick) mceier_ -> mceier
08:56 (join) dwaynecrooks
08:58 (join) realitygrill
09:22 (quit) neilv: Quit: Leaving
09:43 (join) tfb
09:47 (join) RacketCommitBot
09:47 RacketCommitBot: [racket] plt pushed 6 new commits to master: http://git.io/5H8L0g
09:47 RacketCommitBot: [racket/master] added a note about the blue dot - Robby Findler
09:47 RacketCommitBot: [racket/master] added a #:dialog-mixin argument to get-file, put-file, get-file-list, and get-directory - Robby Findler
09:47 RacketCommitBot: [racket/master] added a #:dialog-mixin argument to message-box and related functions - Robby Findler
09:47 (part) RacketCommitBot
10:01 (quit) noelw: Quit: noelw
10:04 (join) noelw
10:13 (quit) Burlingk: Quit: Leaving
10:23 (join) neilv
10:26 (nick) samth_away -> samth
10:36 (quit) realitygrill: Ping timeout: 245 seconds
10:37 neilv: i usually like to put an entire module into one file.  but messing about with multiple separate phases pushes me to split into multiple files
10:37 samth: neilv: really?
10:37 neilv: er, i mean, split a package into multiple modules
10:37 samth: ah yes
10:37 samth: that is very true
10:37 (quit) hkBst: Remote host closed the connection
10:37 (quit) jao: Remote host closed the connection
10:39 neilv: i has a suspicion that "define-for-syntax" is trying to trick me into doing it the painful way
10:39 samth: `define-for-syntax' is good for small things, but when things get big, you want a separate module
10:42 (join) RacketCommitBot
10:42 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://git.io/uKFajg
10:42 RacketCommitBot: [racket/master] add forgotten file - Robby Findler
10:42 (part) RacketCommitBot
10:42 neilv: i have a new idea for a useful programmer's utility.  you provide it with a set of source files and some set of tests to run, and it removes unnecessary "require" clauses from the source files
10:43 samth: neilv: fortunately, ryan already implemented that
10:43 neilv: the "check syntax" info alone is enough to remove unnecessary "require" clauses in most of my code, but not in some of it that gets fancy with syntax
10:44 samth: http://docs.racket-lang.org/macro-debugger/index.html#%28part._.Checking_requires%29
10:44 rudybot: http://tinyurl.com/4xnj7eg
10:44 neilv: does he use check-syntax-like info, or do it empirically?
10:44 samth: it instruments the expansion process
10:44 samth: so it's better than what check-syntax does
10:45 samth: if you're using a development version w/ xrepl, then there's a simple command for using it at the repl
10:45 samth: http://pre.racket-lang.org/docs/html/xrepl/index.html?q=xrepl#%28xrepl._check-requires%29
10:45 rudybot: http://tinyurl.com/3uofquj
10:46 neilv: so if you provide test cases as part of the input, then it should be smart enough to recognize requires that are needed only when your library is used and cannot be determined statically?
10:46 samth: neilv: i don't understand
10:50 neilv: it's hard to formulate a concise example in this little irc prompt
10:51 samth: the input is a module path, so it can't include test cases
10:51 samth: in general, i don't see how test cases relate at all
10:53 neilv: say "mymodule" provides a macro, and you want to eliminate unnecessary "require" in "mymodule"'s source.  but one of the requires in "mymodule" could actually be used sometimes because input to the macro provided by "mymodule" can result in a reference to an identifier provided by that require
10:54 neilv: that isn't a great example
10:54 neilv: i'm trying to suggest that racket language leaves theoretical and even not-unlikely cases in which you can't determine all the dependencies statically
10:55 (join) anRch
10:56 neilv: i definitely do have a big use for jay's check-requires, however. just not my immediate one
10:56 samth: neilv: yes, it's possible to write a macro like that
10:57 samth: also, requires which are only for effect are missed
10:57 samth: also*2, it's ryan, not jay
10:57 neilv: anyway, that's why i originally said test cases too. because then the tool would put the burden on the user to cover what the tool cannot
10:58 neilv: oops, my bad. i have a jay fixation
10:58 neilv: for the record, thank you for check-requires, ryan. it looks great.
10:58 (quit) jeapostrophe: Quit: jeapostrophe
10:59 samth: almost any macro in the category you describe, though, is a bad idea, i think
10:59 noelw: you scared off jay
10:59 (join) yoklov
11:12 (quit) yoklov: Quit: Leaving.
11:56 neilv: matthew and eli sure did spend a lot of effort on scribble
11:58 (join) jeapostrophe
12:00 (quit) anRch: Read error: Connection reset by peer
12:00 (join) anRch
12:02 (quit) anRch: Client Quit
12:11 (nick) noelw -> noelw_away
12:11 (quit) noelw_away: Quit: noelw_away
12:13 (join) MayDaniel
12:13 samth: neilv: yes, yes they did
12:21 (join) anRch
12:24 (join) bluezenix
12:25 (quit) MayDaniel: Read error: Connection reset by peer
12:32 (quit) tfb: Quit: sleeping
12:35 (join) MayDaniel
12:37 (join) realitygrill
12:44 (join) jonrafkind
12:46 (quit) bluezenix: Quit: Leaving.
12:54 (quit) anRch: Quit: anRch
13:07 (quit) masm: Ping timeout: 240 seconds
13:28 (join) masm
13:32 (join) bluezenix
13:33 (join) fooo
13:36 (quit) jeapostrophe: Quit: jeapostrophe
13:38 (nick) fooo -> eli
13:39 (quit) eli: Remote host closed the connection
13:39 (join) eli
13:39 neilv: in a scribble defproc, is it permissible to use typed racket types rather than contracts?
13:40 offby1: ooh oooh
13:51 (join) yoklov1
13:53 samth: neilv: it's just an arbitrary racket form, so you can use whatever you want
13:54 neilv: that was my plan.  just sanity checking on whether that's bad form (no pun)
13:55 (quit) mceier: Quit: leaving
13:57 samth: neilv: i certainly don't think so
13:57 samth: i'm rather a typed racket fan :)
13:58 neilv: ok. but just to be safe, you should probably still wear a red sox cap in some parts of town
13:59 offby1: who knew
14:01 (quit) dwaynecrooks: Ping timeout: 240 seconds
14:04 (join) mceier
14:07 (join) jeapostrophe
14:08 (quit) jeapostrophe: Client Quit
14:13 (join) dwaynecrooks
14:18 (join) RacketCommitBot
14:18 RacketCommitBot: [racket] plt pushed 3 new commits to master: http://git.io/LmfRyg
14:18 RacketCommitBot: [racket/master] Actually run the strictness-tests. - Eli Barzilay
14:18 RacketCommitBot: [racket/master] Moved `net/base64' code from unit to module. - Jon Zeppieri
14:18 RacketCommitBot: [racket/master] Moved `net/cgi' code from unit to module. - Jon Zeppieri
14:18 (part) RacketCommitBot
14:35 (join) RacketCommitBot
14:35 RacketCommitBot: [racket] plt pushed 2 new commits to master: http://git.io/-JGOTg
14:35 RacketCommitBot: [racket/master] Use better language for lang-info - Sam Tobin-Hochstadt
14:35 RacketCommitBot: [racket/master] Increase time limit. - Sam Tobin-Hochstadt
14:35 (part) RacketCommitBot
14:43 neilv: people really shouldn't use ancient rhel/centos in labs of students who don't run only ancient software
14:48 ohwow: word
14:54 (join) anRch
15:13 neilv: "if" requiring an else clause just helped me out again
15:14 neilv: the change to "if" was worth being non-backward-compatible
15:15 (join) RacketCommitBot
15:15 RacketCommitBot: [racket] plt pushed 2 new commits to master: http://git.io/MbO2BQ
15:15 RacketCommitBot: [racket/master] when the filename changes, re-run the "what language are we in" code to - Robby Findler
15:15 RacketCommitBot: [racket/master] more adjustment of the drracket test suites to avoid using os-given focus information - Robby Findler
15:15 (part) RacketCommitBot
15:19 ohwow: :]
15:20 ohwow: the separation of if and when is nice indeed
15:23 (join) RacketCommitBot
15:23 RacketCommitBot: [racket] plt pushed 13 new commits to master: http://git.io/QMCckA
15:23 RacketCommitBot: [racket/master] Moved `net/dns' code from unit to module. - Jon Zeppieri
15:23 RacketCommitBot: [racket/master] Moved `net/ftp' code from unit to module. - Jon Zeppieri
15:23 RacketCommitBot: [racket/master] Moved `net/head' code from unit to module. - Jon Zeppieri
15:23 (part) RacketCommitBot
15:27 (join) cb`
15:32 (join) RacketCommitBot
15:32 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://git.io/DOBuOA
15:32 RacketCommitBot: [racket/master] make memory accounting and places work together - Matthew Flatt
15:32 (part) RacketCommitBot
15:34 neilv: ohwow: i think i was the person who originally asked for "when" and "unless", and now i never use them. i use "if", "and", "or" for boolean branching on one expression. sometimes "cond" as well, if i want to use the non-#f value and it's not already in a variable
15:35 ohwow: D8
15:38 offby1: I use when alla time.
15:38 offby1: "Unless" more or less never.
15:40 (join) jeapostrophe
15:42 (quit) jeapostrophe: Client Quit
15:47 neilv: offby1: have you seen these idioms? they are one of several reasons why i don't use "when" anymore. http://paste.org/pastebin/view/38019
15:48 ohwow: hm
15:49 ohwow: yeah but the whole point of (when) is that you don't have an else clause
15:49 neilv: it sometimes beats a (let ((VAR EXPR)) (when VAR CODE ...)) , and you can add more "cond" clauses then
15:49 ohwow: hm
15:49 neilv: you usually have else clauses in functional programming
15:49 ohwow: Yeah (when) is more for side-effects
15:49 ohwow: i think
15:50 offby1: GOD I HATE PASTE SITES
15:50 neilv: i'm not too fond of them either, but they're handy for irc sometimes
15:50 offby1: neilv: sorry, I'm too lazy and distracted to read that paste closely enough to understand it
15:51 offby1: they're quite handy, but I hate that they tart up your data with links, ads, etc.
15:51 offby1: http://ix.io is the only site I know of that doesn't do that by default, but it's slow
15:51 neilv: lisppaste is offline
15:52 (quit) MayDaniel: 
15:53 offby1: what ohwow said: "when" is for side-effects.
15:54 (join) tfb
15:55 ohwow: hpaste is good
15:55 ohwow: http://hpaste.org
15:55 ohwow: hm
15:55 ohwow: it doesnt support lisp anymore
15:55 ohwow: strange
15:55 offby1: I don't want _any_ sort of support.
15:55 ohwow: there is also a fpaste.org and http://paste.call-cc.org/
15:55 offby1: I just want: plain text in, plain text out.
15:56 ohwow: well
15:56 ohwow: i like syntax hilighting
15:56 offby1: I do too, when it's done correctly (namely: by Emacs, or a dedicated IDE)
15:57 ohwow: well
15:57 ohwow: sometimes i am on irc when i am not coding anything/dont have emacs open/chatting from the phone
15:58 ohwow: but you certanly have a point
15:58 offby1: I do.  Usually my hat hides it, though
16:07 (quit) anRch: Quit: anRch
16:34 ohwow: what happened to leftparen?
16:38 neilv: there, i'm still not getting procedure type info from tr or contracts, but at least i can get the args from some of the more common patterns  http://i.imgur.com/DQZ72.png
16:40 neilv: for api documentation, i'd like to you to only have to give the procedure name, and it supplies the full scribble contracts with arg names for prototype and result
16:40 ohwow: tr?
16:41 neilv: but if you want to specify the scribble contracts, or mcfly can't infer them, then you can
16:41 neilv: typed racket
16:41 ohwow: ah
16:42 neilv: when mcfly knows it's missing information, it will do the best it can to give you a valid manual, and then add another red warning to the end of the document about it
16:44 neilv: so, you are encouraged to use tr or contracts in your planet package, and when you do, adding documentation is a lot easier
16:47 neilv: and the mcfly development tool will be able to generate info.rkt in addition to the .scrbl file. so you just edit one racket file, put in the "doc" forms, run the mcfly tool, and perhaps you then have all the files you need to release a planet package
16:48 offby1: ooh
16:48 offby1: maybe I'll finally make a proper package outta all my misc stuff
16:48 neilv: kinda like how people used to release emacs add-ons.  you wrote one .el file, followed some conventions for metadata in the file, and then just posted that file
16:49 neilv: the barrier to releasing a planet package is quite a bit bigger than for emacs package
16:49 offby1: indayd
17:05 (join) jeapostrophe
17:11 (quit) jeapostrophe: Quit: jeapostrophe
17:13 neilv: nice. all that quasisyntax/loc paid off
17:14 neilv: the error has the right source location info. http://i.imgur.com/yfHn3.png
17:15 offby1 <3 neilv
17:15 offby1: how much do we owe you?
17:16 neilv: i think we have to pay the plt people first, before we get to me
17:18 offby1: then the emacs people
17:19 offby1: then I have to feed the cat
17:19 offby1: pay alimony
17:19 offby1: _then_ we'll get to you
17:21 neilv: i have learned that iphone developers can make $200+/hr, and big-city law firm associates can bill close to $1k/hr (plus whatever they make from the firm)
17:21 neilv: i think racket developers should make somewhere in between there
17:21 jonrafkind: i make $2/hr
17:21 jonrafkind: working on racket
17:22 neilv: if you're a grad student, you're supposed to be an indentured servant
17:22 neilv: when you graduate, and learn iphone programming or go to law school, then you make the big bucks
17:23 jonrafkind: i learned android programming last week
17:23 jonrafkind: i should write some stupid apps and make some passive income
17:23 neilv: it's hard to make money on android apps selling them yourself. probably make more as work-for-hire
17:25 offby1: I am clearly not interested in learning Android programming, despite having thought that I was.
17:25 offby1: If I _were_, I'd have done something about it by now.
17:25 neilv: you're waiting for whalesong, and doing web-based mobile thin-client apps
17:25 offby1: yes!  that's it.
17:25 offby1 glances around nervously.
17:26 offby1: rudybot-on-android, any day now
17:30 (quit) bluezenix: Quit: Leaving.
17:36 (quit) offby1: Read error: Connection reset by peer
17:39 (join) offby1
17:39 (quit) offby1: Changing host
17:39 (join) offby1
17:45 neilv: @(racketinput (+ 1 2 3) #,(racketresult 6))
17:45 neilv: is that the right way to use racketinput and racketresult?
17:50 (join) dnolen
17:50 neilv: with the #,(racketresult ...) on a new line, i guess
17:52 samth: @examples[(+ 1 2 3)] is the right way to use them :)
17:53 neilv: heh
17:53 samth: eli: is there any way to change the sort order of the profile?
18:04 neilv: i'm not fond of how scribble wants you to format racket results. i liked texinfo's way, which did not have the ">" and used a double-arrow character between expression and result
18:05 neilv: for one thing, the texinfo lets you line up a series of expressions and their results together tightly, in a columnar format, so you can see how result changes with input, more easily
18:06 (quit) tfb: Quit: gone
18:07 neilv: http://www.neilvandyke.org/soundex-scheme/index.html#(def._((planet._main..ss._(neil._soundex..plt._1._3))._soundex))
18:07 neilv: that's an example, converted from the texinfo to scribble. normally the "==>" would be a double-arrow character
18:09 (join) jeapostrophe
18:12 samth: jeapostrophe: i assume this: http://drdr.racket-lang.org/23456/collects/tests/planet/run-all.rkt means you should have a drdr step to blow away the planet cache
18:17 (quit) jeapostrophe: Quit: jeapostrophe
18:17 (quit) wtetzner: Remote host closed the connection
18:21 (join) jeapostrophe
18:22 mithos28: samth: how is StructTop different from Struct?
18:23 samth: mithos28: for structs that don't have a subtyping relationship between their instanatiations
18:23 samth: then StructTop is the supertype of all of the instantiations
18:24 mithos28: ah so if they are mutable, or have mutable fields, say?
18:24 (quit) jeapostrophe: Client Quit
18:27 eli: samth: Huh?
18:28 samth: mithos28: that, or they have a type parameter that appears in both co- and contra-variant position
18:28 samth: for example
18:29 samth: eli: i'd like to sort the output of the `text-render' profile renderer by "self" time
18:29 samth: is that possible, or should i write my own renderer?
18:30 eli: Did you look at the docs?
18:30 samth: eli: yes
18:30 eli: Then you probably know more than I do.
18:31 samth: i'll take that as a no
18:31 samth goes to hack text-render
18:32 eli: samth: Looking at the sources too, I think that that's correct.
18:33 eli: But one thing that I do remember is that sorting it like that leads to pretty useless output, as I described in the recent email.
18:33 eli: Something better to do would be to visualize the bad self times so it's easy to find them.
18:33 samth: eli: the reason i want to do it is i have 400+ entries, and very few large self times
18:33 eli: (Well, even more useful are the things I described in that email.)
18:33 samth: so it's hard to find them
18:34 eli: So if you have some visual indication of the large self time, it would be easier to find them.
18:34 samth: yes, that would be nice
18:34 eli: For example some stupid "bar chart" with "###" indicating its size.
18:34 samth: in this case, most of the noticeable self times aren't even that big
18:34 samth: so i want to see them in order
18:34 eli: Then you can search interactively and enter ##s to get to the more interesting bits.
18:35 eli: Can you send me the output?
18:35 eli: (The current one.)
18:35 eli: And BTW, did you try the graphviz renderer too?
18:36 samth: punge.ccs.neu.edu/~samth/prof.out.txt punge.ccs.neu.edu/~samth/prof.out2.txt punge.ccs.neu.edu/~samth/prof.out3.txt
18:36 samth: no, i haven't tried that
18:36 (quit) mceier: Quit: leaving
18:37 samth: annoyingly, the functions that show up with big self times are `map' `andmap' and `ormap'
18:37 neilv: graphviz is often helpful out of the box. getting it to give a really good visual takes many hours of tuning
18:38 samth: eli: more helpful than better visualization would be knowing which core runtime function was on the top of the stack
18:39 samth: but that isn't really your department
18:41 neilv: @(define ==> (larger "\u21d2"))
18:41 neilv: @(racketblock (+ 1 2 3) #,==> #,(racketresult 6))
18:43 eli: samth: That (the "annoyingly" part) is exactly why sorting by self times is not helpful.
18:43 eli: If `map', `andmap', etc are the big ones, then you want to know who called them the most.
18:44 eli: Or, who they called the most.
18:44 eli: samth: And here's another idea for giving you what you want:
18:44 eli: at the end of the profile report (same as it is now), show a list of function pointers sorted by self times.
18:44 eli: Where pointers are those "[N]" markers.
18:45 eli: (That will probably be easy to add.)
18:47 samth: eli: that would be all that i need
18:49 (join) wtetzner
18:51 eli: samth: Looking at your profile, `subtype*' is the biggest bottleneck, which is probably not surprising.
18:52 eli: And it's called mostly from `ormap' which in turn makes that a bottleneck too.
18:52 samth: yes, in the sense that it's the only thing in *my* code that appears slow
18:53 samth: eli: which profile are you looking at -- they turned out surprisingly different
18:53 samth: even though they're just the same file with more copies of the same code
18:53 eli: There are also `f1333' and `f1555', maybe they're contract checks?
18:53 samth: yeah, i haven
18:53 eli: I'm looking at #2.
18:53 samth: 't figured out which those are
19:05 samth: the other problem is that it's impossible to tell from the profile output what's slow about `subtype*' -- is it called too often, are some of the calls slow, are all the calls slow, etc
19:25 (quit) dwaynecrooks: Quit: Leaving
19:44 mithos28: samth: If I a have a parametric struct type S of one variable X. Is (S Integer) a subtype of (StructTop (S X))? From what you said I think they should be, but that doesn't match the current implementation.
19:45 samth: mithos28: you're right on both counts
19:45 samth: sadly, gotta run now
19:45 (nick) samth -> samth_away
19:45 samth_away: btw, i'll get to your pull req tomorrow
19:45 mithos28: k, hopefully there will be another patch in there by then
19:46 (quit) sethalves: Ping timeout: 252 seconds
19:46 (quit) em: Ping timeout: 264 seconds
19:58 (join) em
20:04 (join) RacketCommitBot
20:04 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://git.io/5nabjg
20:04 RacketCommitBot: [racket/master] pr# 12143 raco make -j 2, fix split-path's 'relative to byte-string conversion - Kevin Tew
20:04 (part) RacketCommitBot
20:05 neilv: when inside a codeblock in scribble, what's the right way to escape to a proportional, serifed font?
20:06 neilv: #,(italic ...) works
20:14 (join) jeapostrophe
20:17 (quit) jeapostrophe: Client Quit
20:19 (join) RacketCommitBot
20:19 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://git.io/IRYh-w
20:19 RacketCommitBot: [racket/master] Corrections to the guide's places entry - Matthew Flatt
20:19 (part) RacketCommitBot
20:23 (join) darkf
20:26 (join) jeapostrophe
20:44 (quit) jeapostrophe: Quit: jeapostrophe
20:45 (quit) masm: Quit: Leaving.
20:53 (join) sethalves
21:06 neilv: lindenlab, racket is ready to serve :)
21:19 (join) flaggy
21:22 (quit) yoklov1: Quit: Leaving.
21:25 (join) ryanaghdam
21:31 (join) RacketCommitBot
21:31 RacketCommitBot: [racket] plt pushed 3 new commits to master: http://git.io/nVl_Iw
21:31 RacketCommitBot: [racket/master] add missing require - Robby Findler
21:31 RacketCommitBot: [racket/master] add yet another dialog-mixin argument to a standard dialog, this time - Robby Findler
21:31 RacketCommitBot: [racket/master] more adjustments to try to get the drracket test suites running in drdr - Robby Findler
21:31 (part) RacketCommitBot
21:52 (quit) jonrafkind: Ping timeout: 252 seconds
22:11 (quit) ryanaghdam: Quit: Lost terminal
22:14 (quit) flaggy: Ping timeout: 276 seconds