#clojure log - Jan 15 2009

The Joy of Clojure
Main Clojure site
Google Group
List of all logged dates

3:21 Lau_of_DK: Good morning gents

3:31 joma: is there a clojure command for writing to file or i have to go to Java?

3:33 yangsx: joma: clojure.contrib.duck-streams/spit

3:38 joma: god

3:38 spit slurp etc, please stop

3:53 if i need to do : (if (action) then action else something_else

3:53 how can I write that without a let

3:54 and without doing the action 2 times?

3:54 like in python

3:54 no not like in python

3:57 rottcodd: (or (action) (something-else))

4:01 sleven: I mean break a string and keep intendation

4:04 bergermeister: (first (filter apply [action something-else])

4:20 AWizzArd: clojurebot: max people

4:20 clojurebot: max people is 127

4:21 hiredman: /list #clojure

4:21 er

4:58 blackdog: i just upgraded to HEAD and my clj script doesn't want to find my -cp classpath anymore anyone else with this?

5:09 Lau_of_DK: your clj script ?

5:10 blackdog: my bash starting script

5:10 it hasn't changed, points to the same places

5:10 i'll try going back in svn a bit see if it starts working again

5:13 hiredman: do you use clojure.main?

5:13 blackdog: yes

5:13 started last week or so, but it's been fin

5:13 e

5:14 karmazilla: my script works fine: http://gist.github.com/47350

5:14 hiredman: have you checked the --help output recently? the options may have changed

5:14 blackdog: checking

5:48 mikem`: hi, is there a way to obtain the current version (ie: the SVN build) of clojure at the REPL?

5:49 dhaya: mikem`: No way currently.

5:50 mikem`: dhaya: thanks. not a problem :) just wanted to double-check it's running the latest SVN

6:45 i'm having trouble with chapter 9 of Programming Clojure, in the function create-snippets on page 209, create-table is not found. do I somehow need to require/use hsqldb first?

6:49 dhaya: mikem`: create-table is in clojure.contrib.sql

6:52 Lau_of_DK: mikem`: Do your self a favor and get familiar with: http://github.com/Lau-of-DK/clojureql/tree/master

6:52 mikem`: dhaya: ok, that was easy. thanks :)

6:53 Lau_of_DK: looks interesting. i'll bookmark that and take a look again when i'm doing some real work

6:53 thanks

6:54 Lau_of_DK: np

7:05 sleven: is Programming Clojure available for free as pdf or soemthing?

7:06 Fib: sleven: There's a beta PDF but it's not free

7:08 http://www.pragprog.com/titles/shcloj/programming-clojure

7:12 mikem`: Lau_of_DK: just got to a mention of clojureql in the book :)

7:44 blackdog: interesting when i have clojure.jar in my -Djava.ext.dirs path, HEAD svn does not find my other classes on the -cp classpath, if I add clojure.jar to -cp too everything works fine

7:45 it used to work with all the jars including clojure.jar in -Djava.ext.dirs

7:46 AWizzArd: And for me it's even more strange. In Slime I have to put everything that I already have on my classpath also into swank-clojure-extra-classpaths, because my classpath itself is completely ignored.

7:47 rhickey: blackdog: "does not find" mean import doesn't work?

7:47 blackdog: mmm, i mention this as it was a breaking change for me

7:47 yes

7:47 rhickey: blackdog: what OS/JVM?

7:47 blackdog: import fails when clojure.jar is in -Djava.ext

7:48 (build 1.6.0_11-b03)

7:48 ubuntu

7:48 i don't know what svn version i had that was working

7:49 i probably updated about 2 weeks ago

7:49 today it failed

7:51 that's import on other .clj files fail, clojure itself loads fine

10:01 durka: what does clojure.xml/startparse-sax do?

10:02 it has no documentation and looks like it could call any of four methods on SAXParser

10:03 is the idea to proxy DefaultHandler and pass that?

10:05 drewolson: hey all, should the "Getting Started" page on clojure.org be updated to show launching the repl with clojure.main?

10:18 Chouser: durka: the idea is to you to provide some other sax-compatible parser

10:18 durka: I use it for tagsoup

10:22 as I understand it, it's currently impossible to create a class with genclass that calls its own (or its super-classes') methods when it is constructed.

10:22 durka: because genclass doesn't look for an implementation until runtime?

10:22 Chouser: no, because the init function is called without access the 'this'

10:23 durka: that seems unhelpful

10:26 joma: if i want to do something sideeffectful X times, what do I use?

10:26 rhickey: (doc repeatedly)

10:26 clojurebot: Takes a function of no args, presumably with side effects, and returns an infinite lazy sequence of calls to it; arglists ([f])

10:27 rhickey: (doc dotimes)

10:27 clojurebot: bindings => name n Repeatedly executes body (presumably for side-effects) with name bound to integers from 0 through n-1.; arglists ([bindings & body])

10:28 Chouser: I've claimed before that this is because 'init' is run before any 'this' object has been created. But upon further reflection, I don't think that's possible.

10:29 So now I'd guess that providing 'this' would be a mis-feature in some way, or that it was simple overlooked.

10:29 simply

10:30 walters: Chouser: in the JVM the constructor is restricted to only calling code without reference to "this" before invoking the superclass constructor

10:30 Chouser: typically though what i do in Java at least is define a private static method which does all the computation, and just call that

10:31 Chouser: hm, that's interesting.

10:31 walters: Chouser: maybe genclass could be split into pre-init and post-init?

10:32 Chouser: that's what I was thinking originally. :init could be left as-is, but there could be a new :post-init function

10:32 rhickey: Chouser: init is called before ctor, and provides args for it

10:32 Chouser: called before the super's ctor, but from inside the gen'ed class's ctor, right?

10:33 the problem is if there is stateful initialization required by the superclass for which it has only methods, not ctor args.

10:33 poorly designed superclass if you ask me, but they exist.

10:34 dhaya: ,(if (Boolean. "false") "true" "false")

10:34 clojurebot: "true"

10:34 dhaya: That looks like a bug.

10:34 Chouser: Currently the only solution is to have a fn that you must always use to 'new' your gen'ed class, so that it can do some init after 'new'. Seems dangerous.

10:36 rhickey: Chouser: post-init is possible, but there's no safe-to-use this to pass to init, since the object is allocated but not yet constructed (init being one step in that construction)

10:37 Chouser: ok. I was about to post a :post-init feature request to the issues page.

10:37 rhickey: go for it

10:37 durka: ,(Boolean. "false")

10:37 clojurebot: false

10:37 durka: ,(Boolean. "true")

10:37 clojurebot: true

10:41 rhickey: dhaya: any object that is not nil or Boolean.FALSE is true. Sometimes there will be a conversion for you, but if you are dealing with boxed Booleans not coming from Clojure or known to come from Boolean.valueOf, then you'll need to call .booleanValue on them

10:52 dhaya: rhickey: Ok. I take it there is some reason that clojure itself cannot do that for me?

10:58 It also seems to me to be inconsistent with (= (Boolean. "false") false) returning true.

10:58 rhickey: dhaya: yes, the rest of the planet doesn't want if (and everything built on if) slowed down for that special case

10:59 dhaya: is this a real problem or a theoretical one?

10:59 dhaya: rhickey: I just stumbled on it while taking input from stdin and passing it directly to a if after boxing.

11:00 rhickey: how did you get a Boolean ?

11:00 Chouser: eval?

11:00 rhickey: you are boing yourself?

11:00 dhaya: yes.

11:01 I was boxing myself.

11:01 rhickey: public Boolean(boolean value)

11:01 Allocates a Boolean object representing the value argument.

11:01 Note: It is rarely appropriate to use this constructor. Unless a new instance is required, the static factory valueOf(boolean) is generally a better choice. It is likely to yield significantly better space and time performance.

11:01 public static Boolean valueOf(boolean b)

11:01 Returns a Boolean instance representing the specified boolean value. If the specified boolean value is true, this method returns Boolean.TRUE; if it is false, this method returns Boolean.FALSE. If a new Boolean instance is not required, this method should generally be used in preference to the constructor Boolean(boolean), as this method is likely to yield significantly better space and time performance.

11:01 from the JavaDocs, please use valueOf

11:02 this will return Boolean.FALSE, which will test consitional false in if

11:02 conditional

11:02 dhaya: Ah. I see the distinction now. Thanks.

11:02 rhickey: np

11:03 asbjxrn: How can I from an agent action run two calculations in parallell?

11:03 rhickey: asbjxrn: send to 2 other agents

11:04 asbjxrn: but I want to wait for them to finish before eg. adding them again, and I can't run await?

11:05 Chouser: got an assert, did you? :-)

11:06 asbjxrn: 'send' called in an agent action is held until the action completes and the agent has gotten its new value.

11:09 asbjxrn: Right. So if what I want to do is ((task 1) (in parallel (task 2) (task 3)) (task 4)) How would I do that? Not sure I understand.

11:11 rhickey: asbjxrn: you could write an in-parallel fn using executors and futures, a la: http://clojure.org/concurrent_programming

11:12 Chouser: I guess you'd have to use something other than agents or await for one of thse steps.

11:13 your main task could be a regular Thread instead of an agent, or task 4 could wait on some blocking object rather than using 'await'

11:13 asbjxrn: But they're so nice. :)

11:16 Chouser: would pmap be appropriate for tasks 2 and 3?

11:16 asbjxrn: That test-stm function from the concurrent_programming page looks promising. Is there a lot of overhead in setting up/tearing down such a pool?

11:20 Chouser: is clojurebot pointing at the wrong repo? or is that feature just not working at the moment?

11:20 clojurebot: latest?

11:20 clojurebot: latest is 1207

11:21 rhickey: http://code.google.com/p/clojure/source/list

11:22 Chouser: clojurebot: latest is 1215

11:22 clojurebot: Alles klar

11:22 Chouser: heh. hm.

11:22 karmazilla: clojurebot: latest?

11:22 clojurebot: latest is 1215

11:36 cgrand: Hello! format, name and namespace are not hinted {:tag String}. Do I open an issue for this?

11:37 rhickey: cgrand: sure

11:38 cgrand: ok

11:53 lisppaste8: rhickey pasted "call-par" at http://paste.lisp.org/display/73650

11:54 asbjxrn: :)

11:57 rhickey: better name needed

11:57 danlarkin: papply? you'd have to change it a little though

11:59 rhickey: danlarkin: requiring the last arg be a seq would also require it be realized (i.e. not parallel)

12:00 cgrand: hmm... should force be made to work on Futures?

12:00 Chousuke: so it evaluates the arguments in parallel and applies op to them... hmm.

12:00 rhickey: cgrand: I would worry about the possibly hidden blocking

12:01 Chouser: pinvoke

12:02 * rhickey just got .Net chills

12:02 rhickey: http://en.wikipedia.org/wiki/Platform_Invocation_Services

12:03 Chouser: Mmm... DLLs

12:03 so yummy

12:03 * rhickey shudders

12:04 rhickey: I don't want MS to sue me

12:05 Chouser: seque only uses one worker thread.

12:05 rhickey: something indicating that the args are parallel would be good

12:05 Chousuke: parallel-eval-args is too long :(

12:06 Chouser: nm, I was thinking there was more similarity between call-par and seque than there is.

12:07 it could return a vector, in which case 'peval' is pretty close.

12:07 danlarkin: but is that too close to peeval?

12:30 sp00oon: how about simply: parallelize

12:32 I'm new to lisp & concurrent mind spaces so not sure if that has other connotations...

12:35 Chousuke: sp00oon: that sounds like a higher order function that takes a function and returns a parallelised version of it.

12:35 shoover``: is it a side-effects thing? doparallel

12:37 lisppaste8: rhickey annotated #73650 with "pvals, pargs, using pool" at http://paste.lisp.org/display/73650#1

12:37 rhickey: shoover``: no, not side effects, collecting results

12:39 danlarkin: it's sortof reduce-like behavior

12:42 cgrand: Slighty different: par-args (or par-list or par-seq or...) and (apply f (par-args (expr 1) (expr 2))?

12:52 shoover``: cgrand's approach seems more general. it's an apply, and the args are eval'ed in parallel. (apply f (peval a b c))

13:04 danlarkin: nooo not peeval

13:06 rhickey: cgrand: are you talking about: http://paste.lisp.org/display/73650#1 ?

13:08 pvals fn, could do plist as macro, leaving applying out of it

13:11 jkantz: in clojure, what's the idiomatic equiv of common lisp case?

13:12 Chouser: I don't know CL, but you might look at cond or condp

13:21 jkantz: ok thanks, condp will work

13:32 cgrand1: rhickey: I'm sorry I didn't see your annotation. pvals (fn) and plist (macro) are fine. My point was to leave apply out of teh macro.

14:28 hiredman: you guys see the "Problem using fn macro" thread?

14:28 XD

14:28 * hiredman giggles

14:31 hiredman: I think the reply "if this gets fixed" really makes the thread

14:35 bitbckt: hiredman: "you're doing it wrong"

14:35 hiredman: bitbckt: yeah

14:38 gnuvince: This Mark Volkmann is... weird.

14:42 Do you guys refer to x as in (let [x 0] x) as a variable or as a binding?

14:43 bitbckt: gnuvince: binding

14:44 as in, "x is locally bound to..."

14:45 gnuvince: Thanks

14:46 When I speak of Clojure (or Haskell for that matter) to people used to languages like C or Java, they always ask "why is it called a variable if it can't change?"

14:46 I just realized that binding is probably a better term that would also be less confusing

14:46 bitbckt: I think it's easier when discussing such things with them to use "variable."

14:46 that's the term their used to

14:47 but binding is more accurate, I think

14:47 gnuvince: But like I said, they buck at the "you can't modify it?" part

14:47 bitbckt: yeah... that's a pain ;-)

14:47 rhickey: gnuvince: I've been saying a "local"

14:48 bitbckt: rhickey: that's a solid alternative

15:01 drewr: How do you get to a static member in an inner class?

15:01 walters: $

15:01 drewr: Foo$Bar/baz => No such namespace: Foo$Bar

16:00 danlarkin: drewr: you have to import Foo$Bar

16:00 sorry for being an hour late

16:01 drewr: danlarkin: Yeah. Strange that having the package doesn't work for me. E.g., clojure.lang.LispReader$RegexReader/stringrdr

16:04 Lau_of_DK: /query chouser

16:16 Anybody know what the maximum filesize is, that I can upload view a HTML-form ?

16:18 danlarkin: depends on the server

16:18 I think

16:18 which probably means it doesn't depend on the server, since I seem to be wrong more than right

16:19 drewr: :-)

16:21 Lau_of_DK: Hmmm Dan :)

16:21 So its a matter of setting some MAX_FILE_SIZE var somewhere, and then its up to the server after that?

16:39 grosours: hi

16:39 Lau_of_DK: yo

16:40 durka: Lau_of_DK: are you using apache?

16:40 Lau_of_DK: Yep

16:41 durka: as i remember there is max_file_size, but also max_upload_size and max_post_size

16:41 since files are uploaded by POST requests

16:42 Lau_of_DK: There is - I just couldnt work out if the system gave an upper limit - which I think it does, just dont know what that is

16:46 stuhood: hey gang

16:46 Lau_of_DK: Hey Stu

16:47 stuhood: is it necessary to add type hints to the 'this' parameter for functions used in gen-class?

16:47 danlarkin: hints are never necessary :)

16:48 Although! going by the rule I've set out, they are indeed necessary

16:48 stuhood: haha

16:49 hiredman: Lau_of_DK: clojureql do postgres yet?

16:49 Lau_of_DK: hiredman: I suspect that it does, but I have not tested yet.

16:54 danlarkin: Lau_of_DK: how's the create-table stuff handled, since there's :mysql and :sqlite specific handlers

16:54 Lau_of_DK: sqllite doesnt have type for columns, it has types for cells

16:55 Mysql, postresql, mssql all have one type per column

16:55 danlarkin: ah mmhmm

17:00 stuhood: the trouble i'm running into with hints is that if you hint the 'this' parameter, the compiler cannot find the class definition, since it hasn't been gen-class'd yet... chicken and egg?

17:01 Chouser: that's not right. The class fully exists before you get to the method function definitions.

17:01 stuhood: i'm declaring the functions before executing gen-class... is that wrong?

17:01 hiredman: postgresql's alter table has a slightly different syntax than what clojureql is generating

17:01 Chouser: doesn't mean the classname is imported yet though -- have you tried using the full package and class name.

17:02 stuhood: yes, full package and class name

17:02 Chouser: stuhood: oh, yep. that's backwards. (ns my.package.fooclass (:gen-class ...)) (defn -mymethod ...)

17:02 stuhood: gotcha. because gen-class is just creating stubs, right?

17:02 cool beans... thanks.

17:03 Chouser: yep, the gen'd class doesn't go looking for method implementations until as demanded at runtime.

19:00 rhickey: anyone on 1215 yet?

19:02 hiredman: I have a jar built, but I haven't done anything with it

19:03 I think I used the 1215 jar file when I was trying to talk to postgresql with clojureql

19:17 durka: what hiredman said

19:41 durka_: is anyone using yourkit on a >1 cpu machine?

19:41 i'm trying to figure out why activity monitor reports a 2x higher cpu usage than yourkit does, for the same clojure program at the same time

19:44 burkelibbey: Probably activity monitor goes up to 200% for what I presume is a Core 2 Duo.

19:44 </guess>

19:47 durka_: well, it only goes up to 102% or so, but i am wondering if that is more-or-less evenly split across the core 2 duo

19:47 and if that was the case i would wonder why yourkit only looks at one cpu

19:52 pjb3: rhickey: I'm interested in learning more about how Clojure works under the hood, any recommendations on things to read to learn about the JVM, JIT, Classloaders, etc.?

19:53 rhickey: pjb3: I don't know that there's a single place with all this stuff - Java Concurrency in Practice is a must read, though

19:55 pjb3: rhickey: Yeah, I have JCIP, only read the first few chapters, I suppose I should get through that

20:00 Cark: how do i replace an item in a list based on some predicate ?

20:00 there must be some way

20:00 durka: map?

20:01 Cark: mhh yes that's what i'm doing right now

20:01 let's say i have a list of users like this {:name "cark" :foo "bar"}

20:02 and i only need to change the user "cark in that list"

20:02 or update its value :foo from "bar" to "baz" really

20:03 lisppaste8 : url

20:03 lisppaste8: To use the lisppaste bot, visit http://paste.lisp.org/new/clojure and enter your paste.

20:03 Chouser: Cark: assoc ?

20:03 (doc assoc)

20:03 clojurebot: assoc[iate]. When applied to a map, returns a new map of the same (hashed/sorted) type, that contains the mapping of key(s) to val(s). When applied to a vector, returns a new vector that contains val at index. Note - index must be <= (count vector).; arglists ([map key val] [map key val & kvs])

20:04 Cark: let me prepare an example =P

20:05 Chouser: sure

20:07 lisppaste8: cark pasted "untitled" at http://paste.lisp.org/display/73679

20:08 Cark: see the map form, i wonder if there is some library function to do that

20:09 Chouser: do you always look things up by :name ?

20:09 Cark: well in this case yes

20:10 Chouser: or to head in a different direction, are your hashes unique and their order in the list unimportant?

20:10 Cark: but this cold be a more general function

20:10 the order in the list is important

20:10 Chouser: yeah, I think it's not all that "general" anymore, but we may be able to tighten up that example a bit...

20:11 durka: perhaps an "assoc-if"?

20:11 Cark: ah that sounds promising =P

20:12 durka: i wasn't saying that exists, i was suggesting that someone write one :p

20:12 Chouser: but that really assumes an ordered seq of hashes -- not something I've seen very often.

20:12 durka: well, isn't that what defstruct is designed for

20:12 which would imply that it is expected

20:13 Cark: what i'm using it for : i have a list of column in a csv file, and the cv-import stuff needs to assign column numbers with field names

20:13 csv-import

20:14 Chouser: durka: structs are just maps, and maps are supported. There's a library of functions to work with unordered sets of maps, but not ordered seqs of them.

20:14 durka: i see

20:14 Cark: anyways if you guy can't tell me about this function, that means it doesn't exist, and it's trivial to write one ...thanks !

20:15 Chouser: hm, and since the list itself changes, we can't use an array-map.

20:17 (map second (vals (assoc-in (zipmap (map :name users) (indexed users)) ["rick" 1 :foo] "baz")))

20:17 lisppaste8: durka annotated #73679 with "untitled" at http://paste.lisp.org/display/73679#1

20:18 durka: eek

20:18 Chouser: hm. durka's is better than mine.

20:18 durka: but pretty specific to this example

20:19 yours looks like it adds a lot of complexity, but maybe because i don't know what a zipmap is

20:19 Chouser: yeah, just a failed approach.

20:19 hm, not even correct. It needs a 'sort' in there somewhere. skip it.

20:20 durka: your solution is better in every way

20:20 durka: i think the more general answer to this question is "sql"

20:20 Chouser: or would be if at least one of the argument names wasn't a builtin function.

20:21 durka: actually all of them are

20:21 i think

20:21 Chouser: yeah, I know. don't do that.

20:21 ,[test map key val]

20:21 ,(prn [test map key val])

20:21 clojurebot: [#<core$test__4053 clojure.core$test__4053@1fd9b49> #<core$map__3371 clojure.core$map__3371@fee225> #<core$key__3235 clojure.core$key__3235@c57f88> #<core$val__3238 clojure.core$val__3238@de6570>]

20:22 durka: of course, core/assoc names it params map/key/val too

20:22 Chouser: shame on it, too, then. :-)

20:29 lisppaste8: cark annotated #73679 with "untitled" at http://paste.lisp.org/display/73679#2

20:29 Cark: there you go !

20:30 lisppaste8: Chouser annotated #73679 with "map-if" at http://paste.lisp.org/display/73679#3

20:30 Chouser: hey!

20:31 I still screwed up the usage, though. sheesh.

20:32 Cark: ahyes, stay consistent put the coll param at the end

20:38 Chouser: well, if there's a simpler approach, I'm not seeing it.

20:38 Cark: yep that's good enough =P

20:38 thanks !

20:45 Chouser: (assoc-in (vec users) [((apply hash-map (mapcat reverse (indexed (map :name users)))) "rick") :foo] "baz")

20:47 Cark: i'm not sure i would understand the meaning of that after a couple of days

20:47 or minutes even

20:58 durka: so we take the names out, number them in order and reverse those so there is a list of name/number pairs. then flatten that out with the mapcat so there is a list of name, #, name, #, etc. we make a hash-map out of that, which is keyed by names and valued with numbers, and then look up "rick" to get the index of rick's record in the original list. with that index in hand, we create a vector (which is like a hashmap from indices to s

20:58 well, now i understand it

20:59 lisppaste8: Chouser annotated #73679 with "indexed seq" at http://paste.lisp.org/display/73679#4

20:59 durka: not sure my brain dump is coherent for anyone else

20:59 Chouser: :-)

20:59 durka: you got it.

20:59 durka: does that end up turning the original list into a vector

21:00 or does assoc-in return a list

21:02 Chouser: a vector

21:03 so, not strictly the same as the rest.

21:03 All these are making the original change-rick-to-baz look pretty good. :-)

21:03 durka: heh

21:08 Cark: now i'll feel responsible for any delay in textjure =(

21:18 Chouser: heh. no, I'm pretty sure I'm resposible for any dealys in textjure.

21:20 * durka regrets that next week he returns to school and will have much less time available to hack/lurk in #clojure

21:48 danlarkin: le sigh

21:49 Caused by: java.io.FileNotFoundException: Could not locate calliope/conf/urls__init.class or calliope/conf/urls.clj on classpath

21:49 but it _is_ on my classpath

22:00 Hm. very tricky, it was running with -Djava.ext.dirs

22:00 I guess changes how classpath look ups work? who knows

22:01 durka: that's strange...

22:01 classpath magic is black magic indeed

22:02 KevinAlbrecht: I have a program which uses "require" to include another clojure file, but when I use (load-file) from the Repl, it won't load the required file... what's the solution?

22:04 danlarkin: KevinAlbrecht: what's the exception it throws?

22:04 likely the file isn't on your classpath

22:04 durka: sounds like the opposite problem

22:04 if require works

22:04 KevinAlbrecht: It throws a not on classpath exception

22:05 danlarkin: oh, right.. wait so which one works, KevinAlbrecht, require or load-file?

22:05 KevinAlbrecht: Ah, I get it now... so you need to add the current directory to your classpath

22:06 require works when I'm running the program as a script, but when

22:06 I load the program on the Repl, the require fails

22:09 thanks, danlarkin

22:10 I really hate classpath nonsense

22:11 durka: it seems like a great idea -- you just import packages, and java translates the names into directories and magically finds the class files

22:11 but... it doesn't seem to work in practice

22:12 it's the portable equivalent of dll hell

22:12 banisterfiend: Yo dawg, I heard you like recursion so we put a Yo dog, I heard you like recursion so we put a Yo dog, I heard you like recursion so we put a Yo dog, I heard you like recursion so we put a Yo dog...

22:13 durka: StackOverflowError

22:14 at: net.freenode.irc.clojure

22:14 Caused by: banisterfiend

22:15 banisterfiend: hehe

22:15 durka: ... ? more

22:16 * durka had to go hunting for an ? to copy and paste

22:16 durka: but apparently it is option-5

22:16 danlarkin: ?

22:16 oooo

22:16 fun

22:16 banisterfiend: do you know of any programming langauges that use unicode in their keywords etc?

22:16 i heard some functional languages are thinking about using an actual lambda symbol

22:16 danlarkin: fortress

22:16 Guy Steele's new language

22:17 hiredman: banisterfiend: there is a hack where you can use '?' for 'lambda' in ruby

22:18 banisterfiend: hiredman: really? how do i find out about it?

22:18 hiredman: http://www.oreillynet.com/ruby/blog/2007/10/fun_with_unicode_1.html

22:18 very simple

22:18 (very silly)

22:19 durka: ,(let [? 3.14] (* 2 ?))

22:19 clojurebot: 6.28

22:20 hiredman: ? shoudl be a lazy seq of the digits of ?

22:20 arohner: There's a scheme hack for emacs that replaces (lambda with (<lambda symbol>

22:21 it's display only, but still cool

22:21 banisterfiend: hiredman: is that just ruby 1.9 or 1.8 too?

22:21 hiredman: banisterfiend: 1.8

22:22 dunno if it even works on 1.9

22:22 banisterfiend: cool

22:22 well proc in 1.9 is actually a Proc.new so should probably change the alias to a lambda

22:24 hiredman: ,(let [f (? [x] x)] :a)

22:24 clojurebot: :a

22:26 hiredman: oh

22:26 duh

22:26 the nyquil is kicking in

22:26 ,(let [f (? [x] x)] (f :a))

22:26 clojurebot: :a

22:33 hiredman: ,(((S K) K) :a)

22:33 clojurebot: :a

22:35 durka: oh no not that language :(

22:40 drewr: danlarkin: http://groups.google.com/group/clojure/msg/f65dd874f779b15c

22:40 Not sure how much that's related to your classpath issue.

22:40 danlarkin: ahh ha

22:40 interestingk

22:42 durka: (defn sgn [x] (if (>= 0 x) 1 -1))

22:42 (defn ? [n] (.doubleValue (apply + (take n (map #(/ 4 %) (iterate #(* (sgn %) (+ 2 (Math/abs %))) 1))))))

22:42 (? 359)

22:42 3.14437816350691

22:43 above that Doubles don't cut it

22:44 lisppaste8: durka pasted "pi" at http://paste.lisp.org/display/73683

22:44 durka: Ratios seem to be rather higher precision ^

22:45 hiredman: you may consider your ? lazy seq request beaten to death

22:45 Cark: hey ... (float your-very-long-number) returns NaN

22:46 hiredman: that's not a lazy-seq of digits

22:46 durka: Cark: yeah, float can't hold that many digits

22:46 above (? 359) it does that

22:47 Cark: that's the problem with not having the full numeric tower

22:47 durka: hiredman: true. but it is a sum of a lazy seq of the terms of 4*arctan(1) :)

22:49 Cark: lispworks says your approximation is 3.1405926

22:49 better but not quite it yet

22:57 durka: is it a bug that take seems to hang on exceptions?

23:06 yangsx: what is the perl equivalent for s/ab([0-9]+)cd/[\1]/ ? I mean how to refer to the captured group in re-gsub, for instance

23:06 equivalent for the Perl*

23:07 danlarkin: you can use re-seq

23:07 or re-groups

23:08 yangsx: can re-seq or re-groups be used in re-gsub's replacement?

23:10 hiredman: yangsx: java has its own back reference syntax

23:10 $1

23:12 yangsx: hiredman: thanks, exactly what I'm looking for. I'm very unfamiliar with Java.

23:12 hiredman: ,(.replace "123nd" "(\\d+)nd" "$1ND")

23:12 clojurebot: "123nd"

23:12 hiredman: er

23:13 ,(.replaceAll "123nd" "(\\d+)nd" "$1ND")

23:13 clojurebot: "123ND"

23:21 durka: hiredman: if lazy hexadecimal digits are satisfying enough, http://en.literateprograms.org/Pi_with_the_BBP_formula_(Python) should suffice. i don't understand it, but maybe i'll come back to it later if i'm bored again...

23:56 Chouser: i just wrote a function with a parameter named "key", and tried to call clojure.core/key in the function, and it took me about 15 minutes to find the cause of the exception

23:57 so there i am

23:57 let this be a lesson to me

23:59 Chouser: :-) yes

Logging service provided by n01se.net