#clojure log - May 09 2010

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

0:02 _brian2_: I think you define java exceptions for anything

0:05 sids: that's left to the library author to decide

0:05 _brian2_: which uses a wrapper around the java (html/html-resource (java.net.URL. url))

0:06 well, they haven't addressed this problem

0:08 technomancy: hugod: have you gotten trunk slime to work with lein swank and M-x slime-connect, or just swank-clojure-project?

0:09 sids: _brian2_: that function uses java.net.URL's getContent which should be throwing a FileNotFoundException or something

0:10 ,(import 'java.net.URL)

0:10 clojurebot: java.net.URL

0:10 sids: ,(.getContent (URL. "http://clojure.org/doesnt-exist"))

0:10 clojurebot: java.security.AccessControlException: access denied (java.net.SocketPermission clojure.org:80 connect,resolve)

0:13 _brian2_: so if I something like use (try ( (html/html-resource (java.net.URL. url)) (catch ... ) , the program still blows up

0:13 technomancy: _brian2_: any particular reason the call to html/html-resource is nested in two levels of parens?

0:14 _brian2_: no, its a typo

0:14 i mean I dont in program

0:14 sids: _brian2_: you don't what?

0:15 technomancy: _brian2_: gotcha. well I think more context is needed then; try a paste site?

0:15 _brian2_: use 2 parens

0:15 ok, just a second, I'm going to get it al together

0:15 hopefully ur still here

0:18 sids: _brian2_: I just checked, enlive's html-resource does throw FileNotFoundException for a 404 URL

0:19 _brian2_: but my problem is 400 errors

0:19 sids: _brian2_: can you give me url that throws 400, I can check it for you

0:20 s/throws/returns/

0:20 sexpbot: _brian2_: can you give me url that returns 400, I can check it for you

0:20 _brian2_: ok, just a sec, this is kind of a random event, so hopefully i will catch one

0:20 * technomancy glances at sexpbot

0:22 tomoj: sexpbot: who is your master?

0:22 :(

0:23 _brian2_: http://is.gd/a3eEU

0:23 sexpbot: "Streetwear Shop - 80s Retro Style Fashion Clothing & Accessories — Home"

0:23 _brian2_: but it seems to be re-routed in my browser

0:25 sids: that returns a 301 redirect and following that redirect leads to a page which returns 200

0:25 _brian2_: actually, its not the problem

0:25 the problem I want to solve is simple if I get 400 error to exit and move on

0:26 the reason I got the error is that i am

0:26 using crappy way to locate the url in text

0:26 sids: 400 is bad request, right?

0:26 _brian2_: I guess

0:27 something like that

0:27 sids: enlive should throw an exception for that, give me one such url and I can check

0:27 _brian2_: the url I used was : http://is.gd/a3eEU Use discount code "gaga" for 25% off!

0:27 sexpbot: "Streetwear Shop - 80s Retro Style Fashion Clothing & Accessories — Home"

0:28 _brian2_: sexbot doesnt mind

0:28 that one

0:28 sids: enlive doesn't mind it either

0:28 _brian2_: or maybe just make one up

0:29 I get "main" java.io.IOException: Server returned HTTP response code: 400 for URL: http://is.gd/a3eEU Use discount code "gaga" for 25% off! B6d

0:29 at sun.net.www.protocol.http.HttpURLConnection.getInputStream(HttpURLCon

0:29 sexpbot: "Streetwear Shop - 80s Retro Style Fashion Clothing & Accessories — Home"

0:30 sids: so you get the exception, then what seems to be the problem?

0:30 you are unable to catch it?

0:31 _brian2_: the program crashes

0:31 sids: can you paste the relevant code somewhere?

0:31 _brian2_: ok

0:35 sids : http://clojure.pastebin.com/VuTid23R

0:36 sids: on a different note, you seem to be using this entire string as the url: >>http://is.gd/a3eEU Use discount code "gaga" for 25% off! B6d<<

0:36 sexpbot: "Streetwear Shop - 80s Retro Style Fashion Clothing & Accessories — Home"

0:37 _brian2_: yea, I am using a crappy method to grab the URL

0:37 but I want to make the program robust to the errors

0:38 rather improve the pattern regex

0:38 becasue that will never be perfect

0:42 sids: _brian2_: your catch syntax wasn't quite alright (missing the name argument)

0:43 _brian2_: yea, I just saw that

0:43 sids: I fixed that and put it in a repl, runs without a worry: http://clojure.pastebin.com/ZBhcLj1m

0:43 _brian2_: ok

0:43 thanks!

0:43 sids: there is no need to use a try around your let, fetch-url will catch the exceptions and return a nil in that case

0:44 _brian2_: ok

0:44 technomancy: is anyone using slime-repl.el from trunk slime successfully with clojure?

0:47 I'm getting regular slime evaluation to work, but the repl is busted and still thinks it's in CL-USER

0:51 _brian2_: sids : runs beautifully, thanks for your help!

0:53 sids: _brian2_: you're welcome :)

3:46 ArkRost: Hi! Can someone write a code of the parallel quicksort with using transient strucrures?

4:10 LauJensen: Morning guys

4:10 I'm benchmarking my site using 'siege' and it gives me everything I want, except a mean transaction time, anybody know of a simple *nix tool which does the job a little better?

6:00 herdrick: ,(str (repeat 1 "foo"))

6:00 clojurebot: "clojure.lang.LazySeq@18ce5"

6:00 herdrick: ,(repeat 1 "foo")

6:00 clojurebot: ("foo")

6:01 herdrick: Question: how do I get the usual string representation from that lazy sequence?

6:01 I thought doall was the way, but no.

6:02 Very basic question, yet I don't know or am having a memory lapse

6:05 In other words, what I'm looking for is (some-fn (repeat 1 "foo")) to return a string "(foo)" or something formatted like that.

6:05 other formats are fine: "('foo')" "(\"foo\")" whatever

6:11 ah, gotta sign off. will be back.

6:21 LauJensen: $mail herdrick (apply str (repeat 1 "like so"))

6:21 sexpbot: Message saved.

6:25 nurv: Hi.

6:28 cschreiner: $mail

6:28 cool feature

6:29 $help

6:29 sexpbot: cschreiner: I can't help you, I'm afraid. You can only help yourself.

6:31 herdrick: Sorry, I was gone. Did i miss anything?

6:31 $mail

6:31 oops

6:31 Borkdude: $mail

6:40 herdrick: LauJensen: thanks. Though I would like something that preserved the tree structure of the seq. So this would work:

6:40 ,(str "(" (apply str (repeat 1 "foo")) ")")

6:40 clojurebot: "(foo)"

6:40 herdrick: ...for that exact case

6:41 but I was hoping for something more general. I guess I will write my own recursive thing. Or use the tree walker stuff in clojure.contrib

6:41 sorry, clojure.walk

6:52 jowag: alter has its equivalent for root values in alter-var-root. Is there an ref-set equivalent for root vars?

6:58 I think for that purpose redefining root var with def should be fine, but I'm not sure

7:37 jwr7: ah-hah. So clojure.contrib.sql's insert-rows is really insert-row.

7:39 LauJensen: jwr7: yea, check out ClojureQL instead :)

7:39 jwr7: LauJensen: I would, except after spending lots of time with CLSQL I'm opposed to any additional DSLs on top of SQL…

7:41 I might have to use something better than clojure.contrib.sql, though — there are no persistent connections and no connection pooling.

7:44 LauJensen: Oh - If you don't like a DSL ClojureQL is definitely not for you, its _the_ DSL for SQL :) Lets you extend the SQL language itself, like add fulljoins to derby etc. But what rubs you the wrong way about DSLs ?

7:45 @ jwr7

7:45 jwr7: LauJensen: experience. I believe they optimize for the wrong goal: db-independence. In practice, I've never encountered a situation where I would actually need to migrate to a different DB and expect it to work flawlessly.

7:45 LauJensen: you *always* have to modify things when migrating to a new DB, DSL or not.

7:45 LauJensen: so there is little advantage. And the disadvantages are many. I have to learn a new syntax, convert things into that syntax and encounter its limitations.

7:46 It also isolates me from the db-specific features I *might* want to use.

7:47 LauJensen: Im not saying that you dont, but lets say you're benchmarking some app and you want results on both MySql, Oracle, Mongo and Cassandra - ClojureQL lets you do that with the same code. It always provides raw hooks into the db-specific stuff you might want you use. And the limitations are only a patch away from being removed

7:47 But yes, you do have to learn the syntax

7:47 jwr7: LauJensen: right — but that's a *very* specific use case, which isn't my use case :-)

7:47 somnium: LauJensen: eh? you support nosql too?

7:48 LauJensen: somnium: Mongo and Fleet are being implemented as we speak, and we hope to have at least one in full working order for v. 1.0

7:48 somnium: LauJensen: full joins on CouchDB?

7:49 ah, I was about to ask if it was still in flux, how far off is 1.0 looking?

7:49 LauJensen: somnium: The backend is currently multimethods, so if you implement a frontend function which returns a :fulljoin object, just add the method you want to the backend. You can see how easily we did it for Derby and Sqlite

7:49 jwr7: LauJensen: I think this discussion is similar to maps vs objects — you get way more flexibility if you put a layer of objects and multimethods on top of your data, but then again, maps get you there 90% of the way and often enough you don't need that flexibility at all.

7:49 …and having debugged CLSQL… well, you get the point :-)

7:49 LauJensen: somnium: We have a deadline in 2 weeks for all the frontend stuff (the frontend is completely reworked), then we're assing how much we need to retrofit the backend, whereafter we can publish a release date

7:50 Im guessing 6 - 8 weeks, preferably 5 :)

7:50 jwr7: hey, but that lack of connection pooling worries me…

7:50 does this mean there is no library for accessing an SQL database from a multithreaded application without establishing and tearing down connections all the time?

7:51 LauJensen: jwr7: We put in some alpha connection pooling a while ago, it does support multiple long-lived connections

7:51 somnium: mongo isnt that painful without a dsl, despite the underlying javaness of the driver

7:51 LauJensen: But thats also going to change with the release of 1.0, just not sure how much yet

7:53 somnium: LauJensen: if you really want to be heroic you could do everything from TCP/IP on up in clojure for direct BSON -> persistent datastructure :D

7:54 LauJensen: :)

7:54 somnium: someone said they were working on a prototype for a pure clojure Mongo client but they havent been around in a while

7:55 LauJensen: pure, as in rolling their own driver?

7:56 somnium: the low level bits anyway

7:56 LauJensen: cool- I'll make sure you test it again clojureql 1.0 once they're both released :)

7:56 jwr7: LauJensen: oh, cool — so all that remains is for you to remove the DSL thing, and I'll use ClojureQL :-)

7:57 somnium: heh

8:01 LauJensen: would it be better to wait for a 1.0rc to learn the api?

8:45 rhudson: Given a lazy seq of chars, how do I get a string value?

8:46 E.g. given (concat "abc") how do I get "abc" ?

8:46 ,(str (concat "abc"))

8:46 clojurebot: "clojure.lang.LazySeq@1ecc1"

8:53 somnium: ,(apply str (concat "abc"))

8:53 clojurebot: "abc"

8:54 rhudson: Thanks! This has been driving me crazy.

8:54 somnium: cheers

8:55 elguru: ,(version)

8:55 clojurebot: java.lang.Exception: Unable to resolve symbol: version in this context

8:55 elguru: ,(clojure-version)

8:55 clojurebot: "1.1.0-master-SNAPSHOT"

8:55 elguru: ,(persistent!

8:55 clojurebot: EOF while reading

8:55 elguru: (reduce #(assoc! %1 %2 %2) (transient {}) (range 0 10))))

8:56 ,(persistent! (reduce #(assoc! %1 %2 %2) (transient {}) (range 0 10)))

8:56 clojurebot: {0 0, 1 1, 2 2, 3 3, 4 4, 5 5, 6 6, 7 7, 8 8, 9 9}

8:57 elguru: ,(let [result (transient {})]

8:57 clojurebot: EOF while reading

8:57 elguru: (doseq [elt (range 0 10)]

8:57 (assoc! result elt elt))

8:57 (persistent! result))

8:57 ,(let [result (transient {})]

8:57 clojurebot: EOF while reading

8:57 elguru: (doseq [elt (range 0 10)]

8:57 (assoc! result elt elt))

8:57 (persistent! result))

8:58 ,(let [result (transient {})] (doseq [elt (range 0 10)] (assoc! result elt elt)) (persistent! result))

8:58 clojurebot: {0 0, 1 1, 2 2, 3 3, 4 4, 5 5, 6 6, 7 7}

8:58 elguru: ok, again:

8:58 ,(persistent! (reduce #(assoc! %1 %2 %2) (transient {}) (range 0 10)))

8:58 clojurebot: {0 0, 1 1, 2 2, 3 3, 4 4, 5 5, 6 6, 7 7, 8 8, 9 9}

8:58 somnium: elguru: youre not supposed to bash them in place

8:58 elguru: in the doseq version?

8:59 somnium: elguru: yes

9:00 elguru: why is that? and is it mentioned somewhere? my problem is, that the result is quite unexpected and is sometimes correct and somtimes not.

9:02 just failing would be better of course.

9:02 somnium: elguru: I dont know the answers to one and two, but the solution is to use the transient versions the same way as the immutable ones

9:03 I agree, but I suspect since they only exist for speed, they dont waste time on checking for undefined usage

9:03 into uses them under the hood btw

9:04 ,(into {} (map vector (range 10) (range 10)))

9:04 clojurebot: {0 0, 1 1, 2 2, 3 3, 4 4, 5 5, 6 6, 7 7, 8 8, 9 9}

9:04 somnium: should be 'fast'

9:05 elguru: ah ok, i will check it out. thank you.

9:11 LauJensen: somnium: yes, it would be better to wait, as the frontend will change completely. In the backend, just a few dbs will be added

9:12 See the wiki on gitorious for more info

9:13 somnium: LauJensen: any chance of supporting BDB, maybe with cupboard?

10:29 Licenser: rhickey: how is the copyright on the 'clojure design' and logo?

10:30 aka can I use the logo /general color scheme on a website?

10:34 Borkdude: Why can't I do this inside a let?

10:34 (let [my-power-list (lazy-cat [x] (map #(* % x) my-power-list))]

10:34 It says it can't resolve symbol my-power-list

10:35 Licenser: Borkdude: do you have declared my-power-list before?

10:35 Borkdude: nope, but I can do (def my-power-list .. same stuff ..)

10:35 somnium: Borkdude: let isnt recursive, you can use letfn, or ... [my-list (fn my-list ...)]

10:36 Borkdude: somnium: fn as in identity?

10:37 wait

10:37 it just works, I made a type elsewhere

10:38 somnium: Borkdude: O_o

10:38 ,(let [x x] x)

10:38 clojurebot: java.lang.Exception: Unable to resolve symbol: x in this context

10:40 Borkdude: ,(let [my-power-list (lazy-cat [2] (map #(* % 2) my-power-list))] (nth my-power-list 512))

10:40 clojurebot: java.lang.Exception: Unable to resolve symbol: my-power-list in this context

10:40 Borkdude:

10:40 hmm wtf, why does it work on my machine now

10:41 somnium: Borkdude: guessing you defined a var named my-power-list somewhere

10:42 Borkdude: I think I did a declare somewhere yes

11:00 So the more general question is now: how can I close over a recursively defined lazy sequence?

11:03 so: (let [myseq (lazy-cat [0] (.... myseq .....))] (fn [] (something with myseq)))

11:05 somnium: Borkdude: let isnt recursive, it only shadows previous bindings

11:05 Borkdude: so I can't never do this?

11:05 s/never/ever

11:05 somnium: you can call a function that produces a recursive sequence

11:06 Borkdude: give an example of such a function?

11:07 (fn [] (let [my-seq [0]] (lazy-cat my-seq (.... my-seq)))) ?

11:07 this won't work because I need to lazy-cat over what I am defining

11:09 somnium: ,(let [f (fn foo [xs] (lazy-seq (when (seq xs) (cons (take 3 xs) (foo (drop 3 xs)))))) v (f (range 20))] v)

11:09 clojurebot: ((0 1 2) (3 4 5) (6 7 8) (9 10 11) (12 13 14) (15 16 17) (18 19))

11:09 somnium: is one (utterly pointless) example :)

11:11 Borkdude: somnium: what does it mean when you do (fn foo [x] ...) instead of just (fn [x] ...)?

11:12 somnium: Borkdude: it provides a self-reference to an anonymous fn

11:12 Borkdude: otherwise we would need the Y-combinator, and no one wants that :)

11:13 Borkdude: is it similar to recur?

11:13 Licenser: what is a Y-wombinator?

11:13 *combinatrr

11:13 Borkdude: Wombinator... is that like a really really Mad Wombat?

11:14 Licenser: :P

11:14 yes

11:14 Borkdude: Licenser: this is something from logic or lambda calculus I think

11:17 And also a company by Graham

11:18 somnium: Licenser: its a riddle, what is the result of ((fn [x] (x x)) (fn [x] (x x)))

11:18 Licenser: a tack overflow?

11:18 *stack

11:18 * somnium thinks more languages should have built-in wobinator support

11:18 Borkdude: Yes, write that down and you have proved a theorem ;-)

11:19 Licenser: it is wombinator for you good sir!

11:21 somnium: http://www.fatvat.co.uk/2009/04/understanding-y-combinator.html

11:21 sexpbot: "Fatvat: Understanding the Y Combinator"

11:22 Borkdude: brb

11:23 will try your (fn name [args]) solution soon

12:00 AWizzArd: ~seen hoeck

12:00 clojurebot: hoeck was last seen quiting IRC, 395 minutes ago

12:06 Licenser: tach leute

12:09 bsteuber: dpes lein also download a maven executable? and if, where does it put it? :)

12:10 Licenser: raynes und ich haben etwas gespielt: http://tryclojure.licenser.net/ (meinungne?)

12:11 ich will es noch nicht auf #clojure veröfentlichen da noch nicht fertig genug, aber ihr seid ein guter test-kreis :P

12:11 sozusagen meine private beta-tester armee

12:13 bsteuber: Licenser: english channel :)

12:13 Licenser: ... schesse

12:21 LauJensen: Gents - Anybody know of a similar tool to siege, but with more stats (like mean transaction time) ?

12:30 Ok - JMeter is huge, but ApacheBenchmark is in apache2-utils and does exactly what I want

13:26 Licenser: 'ihr seid' ?

13:37 The-Kenny: LauJensen: "(plural) you are". It was meant for #clojure.de

13:37 LauJensen: okay, thanks

13:37 My german is rusty :|

14:05 Borkdude: somnium: is there maybe a recursive lazy-cat in contrib, like fn takes a name to make it recursive?

14:27 somnium: ah rec-cat it is called

14:27 ,clojure.contrib.seq-utils/rec-cat

14:27 clojurebot: java.lang.Exception: Can't take value of a macro: #'clojure.contrib.seq-utils/rec-cat

14:30 Borkdude: (let [my-lazy-pow (clojure.contrib.seq-utils/rec-cat lz (lazy-cat [2] (map #(* 2 %) lz)))] (nth my-lazy-pow 512))

14:30 ,(let [my-lazy-pow (clojure.contrib.seq-utils/rec-cat lz (lazy-cat

14:30 clojurebot: EOF while reading

14:30 Borkdude: [2] (map #(* 2 %) lz)))] (nth my-lazy-pow 512))

14:30 hm?

14:33 ,(let [my-lazy-pow (clojure.contrib.seq-utils/rec-cat lz (lazy-cat

14:33 [2] (map #(* 2 %) lz)))] (nth my-lazy-pow 8))

14:33 clojurebot: EOF while reading

14:33 Borkdude:

14:33 $(let [my-lazy-pow (clojure.contrib.seq-utils/rec-cat lz (lazy-cat

14:33 [2] (map #(* 2 %) lz)))] (nth my-lazy-pow 512))

14:34 sexpbot: EOF while reading

14:34 Blackfoot: the lines are split

14:34 Borkdude: ,(let [my-lazy-pow (clojure.contrib.seq-utils/rec-cat lz (lazy-cat [2] (map #(* 2 %) lz)))] (nth my-lazy-pow 512))

14:34 clojurebot: 26815615859885194199148049996411692254958731641184786755447122887443528060147093953603748596333806855380063716372972101707507765623893139892867298012168192

14:34 Borkdude: ah, right

14:48 mfex: Borkdude, are you also attending the amsterdam clojurians meeting?

14:56 Borkdude: ah mfex

14:56 I saw you're post on the Google Group

14:56 s/you're/your

14:56 I don't think so, but I am planning on going there some time

14:57 mfex: I have never been there either

14:57 Borkdude: I read you are finishing your MSc in Twente, I have studied there too

14:58 What is your subject?

14:58 mfex: I'm half way through my final thesis on action semantics for model driven engineering

14:59 how long ago were you at ut?

14:59 Borkdude: I finished in 2005

15:01 A fellow student of mine is still there, maybe you know him ;-)

15:02 mfex: I started in 03

15:02 Borkdude: Hmm, maybe talk in #clojure-casual? As not to generate too much noise here

15:03 mfex: sure

15:20 giacecco: mmarczyk: Thanks for the hint the other day about apply vs some

15:23 mmarczyk: giacecco: yw

15:23 giacecco: I'm still disappointed apply doesn't work with macros

15:24 Borkdude: giacecco: there is an apply for macros in contrib

15:24 http://clojure.org/libraries#Clojure%20Libraries-Category:%20xml-clojure.contrib.apply-macro

15:25 giacecco: Borkdude: ho thanks, but is there a theoretical reason for that?

15:25 Rich must have designed apply that way for a reason

15:26 mmarczyk: actually apply was designed that way about 50 ways ago :-)

15:26 um

15:26 years

15:27 although you could be nitpicky and argue that LISP 1.5 had no macros (I think...?), so maybe it's a bit less than that

15:27 giacecco: Agree, but RH changed stuff when he reckoned it was useful. I'm an idiot, but I can't see the value in not making apply work for both functions and macros in the core language

15:28 mmarczyk: apply simply can't work with macros, for fundamental reasons

15:29 Borkdude: ,(type defn)

15:29 clojurebot: java.lang.Exception: Can't take value of a macro: #'clojure.core/defn

15:29 mmarczyk: note that the apply-macro thingamabob in contrib is itself a macro

15:30 whereas apply is a function

15:30 Raynes: A macro is not a function.

15:30 Apply applies a function.

15:30 mmarczyk: hi Raynes :-)

15:30 Raynes: Hi

15:30 giacecco: Probably I just need to study more and I would get it

15:31 Someone may then wonder:

15:31 Borkdude: a macro works at macro processor time, it's a different dimension so to say

15:31 giacecco: Why something as basic as logical operators are implemented as macros, e.g. or

15:32 mmarczyk: giacecco: would you want (and false (println :foo)) to print :foo?

15:32 Borkdude: giacecco: it's because macro's don't evaluate their args

15:33 which makes it possible to 'short circuit'

15:33 mmarczyk: actually you could take the view that it is precisely the basic building blocks of your programmes which should be implemented as macros

15:33 unless they are already built into the language as special forms, like, say, if

15:33 giacecco: Wow that's a strong statement!

15:34 mmarczyk: s/should/must

15:34 /

15:34 in fact

15:35 unless you're using call-by-name pervasively, in which case an "and" function would be perfectly fine

15:35 the only language I've used which does that is Haskell

15:36 I suppose PLT's Lazy Scheme might be similar, but I've never used that

15:36 s/might/is likely to/

15:36 sexpbot: I suppose PLT's Lazy Scheme is likely to be similar, but I've never used that

15:36 giacecco: I believe you're flying too high for me, I'm a humble learner :-)

15:36 mmarczyk: thanks, sexpbot :-)

15:37 well, try playing around with macros, maybe implement giacecco-and as a function first, then as a macro

15:37 try emulating the built-in and's behaviour

15:37 giacecco: Last quick one before I need to leave: why can't I use comp on Java functions, e.g. (def integer-sqrt (Math/floor Math/sqrt)) ?

15:38 mmarczyk: with some luck, you might be able to see why and is a macro *and* why apply doesn't work with macros

15:38 giacecco: I meant (def integer-sqrt (comp Math/floor Math/sqrt))

15:38 mmarczyk: Java methods are not first class objects in Clojure (nor are they in Java, mind you)

15:39 you can wrap them if you like, though

15:39 (comp #(Math/floor %) #(Math/sqrt %))

15:40 giacecco: A ha!

15:40 I expected something like that

15:40 Thanks!

15:40 mmarczyk: np

15:41 giacecco: What's the level of the discussion in the forum: are my questions too naive to be asked here?

15:41 mmarczyk: Licenser: das ist ein schoenes REPL, hoffentlich koennt ihr spaeter def*-Ausdruecke zulassen :-)

15:41 Borkdude: #(...) means "tame this Java code"

15:42 mmarczyk: :-)

15:43 giacecco: not at all, ask away

15:43 giacecco: Great, I must spend Much

15:43 much more time here then

15:43 Licenser: mmarczyk: thanks

15:44 mmarczyk: you're likely to receive more extensive answers on the ggroup, simply because of the format

15:44 giacecco: Thank you all a lot, and goodbye for today! I'll be on the ggroup, too.

15:44 mmarczyk: Licenser: is the code up anywhere? or am I asking a bit too early?

15:45 Licenser: mmarczyk: my gitup page, it is a form of Raynes web stuff but It's really not done yet

15:46 mmarczyk: right, I somehow missed the fine print below the box the first time round

15:46 Licenser: :)

15:47 Raynes: Mine is almost perfectly usable. mmarczyk I'll pass you a link to my own personal tryclojure thingy in a little while.

15:47 mmarczyk: please do, thanks! :-)

15:47 LauJensen: Is there some wonderful lib out there, which could coerce this to a Date or something IComparable "2010-03-30 22:26:45" ?

15:48 mmarczyk: LauJensen: Date/valueOf ?

15:49 um, actually it wouldn't be Date...argh

15:50 chouser: LauJensen: looks like a comparable string to me

15:50 mmarczyk: ,(java.sql.Timestamp/valueOf "2010-04-08 23:11:59")

15:50 clojurebot: #<Timestamp 2010-04-08 23:11:59.0>

15:50 chouser: oh my

15:50 I wouldn't have thought of looking in java.sql

15:51 mmarczyk: :-)

15:57 LauJensen: mmarczyk: wow, you impressed me :)

16:04 Raynes: mmarczyk: It didn't break yet, did it? :p

16:07 MadWombat: Hello

16:08 I am trying to figure out clojure.contrib.accumulators, what is wrong with (add (empty-counter) "asdf") ?

16:09 oops, got it, sorry, me silly

16:09 empty-counter is a var, not a func

16:09 b

16:18 eshira: is there a non-log(n) version of contains? to be used with vectors?

16:20 well i guess i'm thinking about the contains? semantics wrong. I want to know if some element x is in some vector v.

16:25 mmarczyk: LauJensen: :-)

16:26 Raynes: you mean clj-sandbox...? if so, not to my knowledge :-)

16:26 Raynes: mmarczyk: I mean anything. :p

16:26 mmarczyk: Raynes: oh, in that case, I'm not sure

16:26 Raynes: clj-sandbox wont break.

16:27 Unless _ato breaks it, that is.

16:27 ;)

16:27 mmarczyk: something's bound to have been broken recently ;-)

16:27 Raynes: mmarczyk: If it broke, I'm fairly certain you'd know it.

16:27 mmarczyk: well I sure hope I can beat him to it, I was promised a cookie if I manage that :-P

16:28 the chances seem to be slim, though

16:28 Raynes: Now you have something that isn't sexpbot to break. :D

16:28 mmarczyk: yeah! that's true :-)

16:28 Raynes: I like tryclj because I can write bits of code on my cellphone without having to deal with JavaScript. It's just HTML.

16:30 mmarczyk: at some point having a private environment to do def* in would be nice

16:30 chouser: eshira: no, vectors are indexed by the position of their elements, not by the element values.

16:30 eshira: so the only way to find an element by value is to walk the whole vector. O(n)

16:30 Raynes: mmarczyk: Do you approve? ;) I think for a first web development project in any language, it's pretty good. :D

16:30 chouser: eshira: if you want to look things up by value, you need a set or a map

16:31 mmarczyk: Raynes: you mean it's your first? in that case, I approve very strongly indeed :-)

16:31 Raynes: :D

16:31 <3

16:32 Borkdude: Raynes: I think I missed out on the link

16:32 mmarczyk: Raynes: also, that means you're ahead of me on this one -- no functioning Web project on my record as of yet :-(

16:32 Raynes: mmarczyk: Haha.

16:33 mmarczyk: Borkdude: look for Licenser speaking German above

16:34 Borkdude: mmarczyk: I did, and then what I do?

16:35 mmarczyk: fish out a link

16:35 Raynes: Borkdude: I sent you the link a second ago.

16:35 :o

16:35 Borkdude: Raynes: where?

16:36 Raynes: In -casual.

16:36 drewr: should `mvn install` work from a fresh contrib checkout?

16:36 Raynes: I'd rather not unveil it to the entire world until It's closer to finished.

16:36 Borkdude: Raynes: http://joyofclojure.com ?

16:36 sexpbot: "The Joy of Clojure"

16:36 Raynes: Er.

16:37 Look in -casual. >.>

16:37 pdk: come to think of it would it still be worth it to sit on my practical clojure preorder

16:37 now that it got delayed again

16:37 Borkdude: Ah ok

16:37 nice :)

16:39 AWizzArd: Hello Lein experts: how does Lein produce Überjars? a) it first unpacks the existing jars (the dependencies) into a temp folder and includes them in the Überjar, or b) it uses some custom classloader that allows it to put jars into the Überjar directly (without unpacking them first)?

16:42 hiredman: AWizzArd: $5 says the unpacking thing

16:42 AWizzArd: k

16:42 maybe Ant can do such a thing too, hmm...

16:42 hiredman: sure

16:43 ant has a number of jar tasks

16:43 AWizzArd: ok, i will look into those

16:43 Thx hiredman.

16:44 hiredman: http://ant.apache.org/manual/CoreTasks/unzip.html

16:44 sexpbot: "Unzip Task"

16:45 mmarczyk: yup, it unpacks them

16:46 AWizzArd: ok good, i will try to teach this to my ant script

16:46 mmarczyk: their just funnily named zip files anyway, you can look inside with the archive manager of your choice :-)

16:51 AWizzArd: too bad that there is no native 7zip implementation for Java

16:51 mmarczyk: they're, oh my

17:00 chouser: rhickey: can we rename the "failed" state of an agent to "stopped"?

17:01 you recover from this state by calling restart-agent, after all, not unfail-agent.

17:01 hiredman: uh, don't they only stop when failed?

17:01 chouser: and then it is less likely to be confused with the :fail error-mode

17:02 hiredman: right. "stop" isn't currently a word used by the docs. I think I'd like to use "stopped" instead of "failed"

17:05 eshira: chouser: thanks, i'll just do (set my-vector) i suppose

17:08 Raynes: mmarczyk, Borkdude, Licenser: Hah. I forgot to remove the reload middleware before putting up tryclj. It actually works like a normal REPL.

17:14 Licenser: for those who were interestind in swing: http://blog.licenser.net/2010/05/09/the-ease-of-guis#

17:14 sexpbot: "lice! : The ease of gui's."

17:14 mmarczyk: Raynes: you caught it yourself, though, so you get to keep your own cookie :-(

17:15 Raynes: :p

17:15 MadWombat: I am playing with a markov generator again, here my function for generating sentences http://pastie.org/952885, for some reason instead of a list of words, I get a list of letters starting with a capital and ending with a period. What am I doing wrong? Where do the words get split up?

17:15 mmarczyk: Licenser: cool, I'm hoping clj-swing will make it possible for me to create reasonable GUIs at some point

17:15 for the first time since I my Delphi days

17:15 Licenser: mmarczyk: I really hope that it will too :P

17:15 delphi had a incredible editor

17:15 mmarczyk: yup

17:16 that was totally *amazing*

17:16 in fact, if I wanted to do Windows desktop programming, I'd very likely research the possibility of using Delphi to programme the front-end

17:18 just thinking about a Clojure process doing the actual work with a Delphi-constructed GUI makes me weep tears of hope and joy

17:20 MadWombat: shouldn't you do (conj result next-word)

17:20 rather than concat?

17:21 also, do you use 1.1 or 1.2? you could switch your random choice to rand-nth with the latter

17:21 MadWombat: mmarczyk: no, result is a list of words

17:21 mmarczyk: 1.1

17:21 mmarczyk: considering switching though

17:21 mmarczyk: if result is a list of words and you concat the word onto it

17:21 then it becomes a list of words with some characters at the end

17:21 MadWombat: hmm

17:22 mmarczyk: so maybe (concat sentence [next-word])

17:22 but then *don't do this*

17:22 you don't want to append to the end of the list repeatedly

17:22 MadWombat: mmarczyk: ?

17:22 mmarczyk: use a vector instead

17:22 and conj

17:22 ,(conj ['a-list-of-words-goes-here] "next-word")

17:22 clojurebot: [a-list-of-words-goes-here "next-word"]

17:23 MadWombat: mmarczyk: I thought conj was for maps?

17:23 mmarczyk: not at all

17:23 it's the universal add-to-collection thingamajig

17:23 for sequences it even figures out where best to put the added item

17:23 beginning for lists, end for vectors and queues

17:24 MadWombat: ah, conj adds item, concat adds collection

17:24 why vectors instead of lists? I thought adding at the end of a list is efficient

17:24 mmarczyk: assoc is for maps

17:24 btw

17:24 just the opposite, actually

17:24 MadWombat: yes, silly wombat, trix is for kids

17:25 mmarczyk: lists are just linked lists

17:25 MadWombat: and vectors are arrays?

17:25 mmarczyk: so you can add to the front at the cost of constructing the front link

17:25 but if you add to the end, you have to copy

17:25 (no destructive updates in Clojure)

17:25 um, no

17:26 vectors are high-branching-factor trees of arrays with an extra tail portion

17:26 MadWombat: cool, lemme try this stuff out

17:26 mmarczyk: actually I believe the underlying abstract structure would be a trie

17:27 bmason: red black tree iirc

17:27 mmarczyk: you should probably watch one of the cool presentations Rich has done on Clojure data structures :-)

17:27 bmason: those are cool :)

17:27 mmarczyk: indeed :-)

17:34 MadWombat: hmm... seems like the following, surprisingly, works as expected (defn f [x y] (let [x x y y] ... ))

17:34 hiredman: surprising how?

17:34 MadWombat: hiredman: I would expect it to be confused at some point

17:35 bmason: redefining the variables?

17:35 hiredman: bmason: no

17:35 you are creating a new scope

17:36 MadWombat: yes, but I am assigning llocal variables with same names as global ones to value of globals

17:36 bmason: right... inner most scope takes precedence... and defining a var to equal itself wouldn't have any effect

17:36 hiredman: locals are not vars

17:36 bmason: what do you call them for ease of terminology?

17:36 hiredman: MadWombat: function are arguments are not globals

17:36 bmason: they are locals

17:37 MadWombat: hiredman: yes, sorry, not globals, but they are in the scope

17:37 hiredman: locals have a fixed nested lexical scope

17:37 at the top(bottom?) of that scope is the dynamic scope of vars

17:38 MadWombat: basically I expected it to be confused at the point of "the second x, is it the local he just named or the func argument?"

17:38 hiredman: MadWombat: that is just not possible

17:39 at the time when the expression whose value will be bound to x inside the scope of the let is being evaluated, the let's scope cannot exist

17:40 because the binding of x to that value is part of the scope

17:40 because the let's scope doesn't exist, how could you evaluate an expression in it?

17:41 bmason: basically how it gets interpreted is "create a new local called 'x' and assign it the value of the function argument 'x' "

17:41 hiredman: bind

17:41 :(

17:42 bmason: ok binding... within the scope of the let, the 'x' defined there takes precendence

17:42 hiredman: scope is like a linked list of pairs

17:42 technomancy: hiredman: did you see my further hacks on serializable-fn?

17:43 hiredman: no, I didn't

17:43 technomancy: hiredman: it can serialize lexical context now =)

17:44 hiredman: cute

17:44 mmarczyk: technomancy: did you get slime-repl from HEAD to work?

17:44 hiredman: as long as you run in the same vm, yeah?

17:45 technomancy: mmarczyk: no, I was hoping I could get hugod to tell me how he fixed it. =)

17:45 mmarczyk: technomancy: funny, it "just works" for me

17:45 technomancy: hiredman: well, as long as it's something that prints decently

17:45 hiredman: let cons a new pair on to the linked list of environments

17:45 technomancy: hiredman: I don't think it would work with reference types, etc.

17:45 haven't really searched for edge cases yet

17:46 hiredman: when you resolve a name to a value you walk the linked list and stop at the first entry

17:47 ,(first (filter #(= 'x (first %)) '((x 1) (x 2))

17:47 clojurebot: EOF while reading

17:47 hiredman: ,(first (filter #(= 'x (first %)) '((x 1) (x 2)))

17:47 clojurebot: EOF while reading

17:47 hiredman: bah

17:47 ,(first (filter #(= 'x (first %)) '((x 1) (x 2))))

17:47 clojurebot: (x 1)

17:49 technomancy: mmarczyk: could you try with the "elpa-2010" branch from my slime fork?

17:49 maybe you've got a slightly older/slightly newer version than I happened to grab

17:53 mmarczyk: technomancy: sure

17:53 will do it now

17:53 technomancy: mmarczyk: oh, and the latest clojure-mode as well

17:54 I moved a few things around so now you can use clojure-mode with slime.el directly and skip swank-clojure.el entirely

17:54 mmarczyk: oh, cool

17:54 does this mean that the whole "coloured REPL" business could go into clojure-mode

17:54 including the advice stuff for slime-repl?

17:55 technomancy: oh, I haven't looked into the coloured repl code yet

17:55 mmarczyk: in this case, please don't :-)

17:56 I'll make it fit in well with the new clojure-mode

17:56 technomancy: sorry, I'm slow

17:58 mmarczyk: I was also going to patch leiningen to accept (and add to the pom) website, mailing list and licence info for projects

17:58 technomancy: oh, that sounds nice

17:58 do the clojure-mode stuff first though if you don't mind; it's closer to a release

17:58 mmarczyk: ok

17:59 technomancy: not to boss you around or anything =)

17:59 mmarczyk: :-)

18:00 Raynes: mmarczyk: It should be working properly now, if you want to try it.

18:01 mmarczyk: Also, you don't have to click the text field every time you submit your code now. That was bugging me.

18:02 mmarczyk: I'm going to add syntax highlighting soon. :o

18:02 mmarczyk: Raynes: cool :-)

18:02 Raynes: I believe I used Enter to submit code, though

18:02 Raynes: mmarczyk: Didn't matter.

18:02 mmarczyk: technomancy: um, what's package-activated-list

18:02 ?

18:03 Raynes: It refreshes the page every time you submit.

18:03 mmarczyk: technomancy: guess I'd have to install ELPA to use your fork, right?

18:03 Raynes: Well, not refreshes. Sort of.

18:04 mmarczyk: Raynes: are you planning paredit-like facilities? ;-)

18:05 Raynes: No. :p

18:05 Plouj: who here takes care of swank-clojure? :)

18:05 Raynes: Plouj: technomancy

18:05 Plouj: oh yeah, right

18:06 * Plouj wonders if technomancy is here

18:06 Plouj: I'm having some weird connection issues with localhost when running M-x slime

18:07 mmarczyk: technomancy: you might want to make the (when (member 'slime package-activated-list) ...) thing check for the presence of a variable by that name

18:08 technomancy: with the condition replaced by nil, I can start a SLIME REPL with your fork, with a nice user=> prompt :-)

18:08 technomancy: which changes to foo=> after (ns foo)

18:25 technomancy: mmarczyk: thanks, will do

18:27 mmarczyk: hmm... wonder what's the difference then; it's busted here

18:27 mmarczyk: what Emacs version?

18:27 Plouj: what's up?

18:27 Plouj: do you know that M-x slime is not meant for general use?

18:28 Plouj: technomancy: I get this error a lot of the times "Lisp connection closed unexpectedly: connection broken by remote peer"

18:28 technomancy: I thought that M-x slime is good for just opening a repl without associating it with a project. What do you suggest?

18:28 technomancy: well, right, that's what I mean

18:28 Plouj: I'm not familiar with slime so... :

18:28 :)

18:29 technomancy: anyway, that is a common problem. I have no idea what causes it, but I know that generally using lein swank or mvn clojure:swank gets around it, which is what I have started recommending

18:29 Plouj: oh

18:29 technomancy: Plouj: I can't reproduce that problem on my own machine, so I can't do much to debug it

18:30 Plouj: how do I use mvn clojure:swank?

18:30 just run it in some directory?

18:30 DuneMan: maybe some nice soul will debug it for you :-)

18:30 technomancy: Plouj: run it from within a mavenized clojure project, then use M-x slime-connect

18:30 DuneMan: that would be nice. =)

18:31 Plouj: ok, that means I need to create a clojure project

18:31 DuneMan: Too bad don't know how to use emacs at all

18:31 technomancy: Plouj: you can do it from a contrib checkout, I think

18:32 Plouj: what's a "contrib checkout"? :)

18:32 technomancy: unfortunately clojure-maven-plugin has a bug; it doesn't let you set the port

18:32 so you can only have one mvn repl running per machine

18:32 Plouj: heh

18:32 technomancy: Plouj: a checkout of clojure-contrib

18:32 DuneMan: that sounds fixable.

18:32 technomancy: DuneMan: it is, but clojure-maven-plugin is ... "lightly-maintained"

18:33 it has quite a few of these small easily-fixable bugs that have been around for the better part of a year

18:33 DuneMan: Someone should fork it.

18:33 Plouj: maybe this is a timing issue

18:34 technomancy: Plouj: it smells a bit like a race condition

18:34 DuneMan: I have a feeling I'm gonna keep writing clojure now that I'm up and running

18:34 mmarczyk: technomancy: "GNU Emacs (i486-pc-linux-gnu, GTK+ Version 2.18.0)\n of 2009-09-27 on palmer, modified by Debian"

18:34 technomancy: but yeah, I don't have many details about it

18:34 Plouj: technomancy: do you know what actually prints the "Lisp connection closed" error. Is it your swank code or some slime thing?

18:35 technomancy: mmarczyk: interesting. I guess I will try a fresh .emacs init; maybe there's something in my personal config messing it up

18:35 mmarczyk: you don't use the starter kit, do you?

18:35 mmarczyk: technomancy: I could put up a Gist with my slime-related config if you feel that could be useful to you

18:35 technomancy: Plouj: pretty sure it's slime itself, so not my code

18:35 mmarczyk: that'd be handy

18:35 mmarczyk: technomancy: no, except as an awesome repository of ideas :-)

18:35 technomancy: mmarczyk: stealing from it is just as good as using it outright. =)

18:36 mmarczyk: technomancy: that's what I figure also :-)

18:36 technomancy: it's so *amazingly* clean that it's absolutely my no. 1 elisp learning resource & the thing to emulate stylistically

18:42 Plouj: is swank something separate from clojure and slime?

18:49 technomancy: have you tried creating a throw-away user on your unix machine and testing swank-clojure in that? :)

18:52 mmarczyk: technomancy: hopefully this contains all the relevant stuff: http://gist.github.com/395481

18:52 technomancy: if not, I'll add in whatever's missing, of course

19:02 technomancy: Plouj: it's a client-server model: slime is the client and swank is the server.

19:03 most swank servers are written in CL, but obviously swank-clojure is not

19:08 Plouj: oh

19:50 AWizzArd: I created a .jar file and need to start it like: java -jar my.jar one.of.my.Deftypes

19:51 is there a way to add this one.of.my.Deftypes as a command line arg into the manifest file?

19:52 hiredman: no

19:52 AWizzArd: ok, so if I want to get a double-clickable file I will have to do some other little magic here

19:59 hiredman: gen-class

19:59 AWizzArd: yes

19:59 hiredman: it's the only way to generate a static main, which you need for a clickable jar

19:59 you also need to have the right information in your jar manifest

20:00 AWizzArd: the correct Main-Class

21:02 _brian2_: anyone know if/how its possible to use congomongo on a network, ie run a clojure program on my mac using a remote mongo installation?

21:20 I guess I need to read about java sockets

21:31 hugod: technomancy: slime-repl from HEAD works for me, both s-c-p and s-c with lein swank. (slime-fancy doesn't because of slime-autodoc)

21:37 technomancy: although, sometimes I have to call (setq slime-dispatching-connection nil)(setq slime-default-connection nil) to make connecting reliable

22:45 _brian2_: hi, I would like to give clj-ssh a try. I put [clj-ssh "0.0.1-SNAPSHOT"] in my project file, ran lein deps, open a repl and (use 'clj-ssh) and get : FileNotFoundException: Could not locate clj_ssh__init.class

23:02 hugod: _brian2_: try (use 'clj-ssh.ssh)

23:02 hiredman: brixen: as single segment namespaces are frowned on I would be surprised if that did work

23:12 _brian2_: ok

23:12 thats whats on the website

23:14 yes, that works, thanks!

23:25 hugod: _brian2_: fixed on the website, thanks for the feedback

23:26 defn: hello all

23:29 _brian2_: ok, thanks for your work hugod

23:31 i dont know much about ssh, I want to log into another computer from my mac to access data in a clojure program, do I need to generate a key, fingerprint and all that?

23:32 DuneMan: That's generally the best way to use ssh

23:32 ssh-keygen

23:32 then set up your .ssh-config to use the correct key on the remote computer

23:32 _brian2_: ok so clj-ssh talks to that

23:32 DuneMan: and add the key to the remote computer's access list.

23:33 _brian2_: well, i use ssh routinely to access the computer from my command line

23:33 DuneMan: Anything that uses the unix ssh libs in the background will look into the .ssh directory and use that stuff. Not sure about clj-ssh, guessing hugod will tell you :-)

23:33 _brian2_: ok

23:33 DuneMan: I'd guess at most you'd need to point it to the rsa/dsa file?

23:34 _brian2_: there is some indication of that

23:34 DuneMan: if you needed to do anything more than that, I'd say the lib is broken.

23:34 heh

23:34 hugod: clj-ssh will pick up .ssh/id_rsa by default

23:34 _brian2_: ok, thanks

23:34 hugod: though if you have a passphrase, you'll need to add it explicitly

23:35 _brian2_: is that the same as password?

23:35 technomancy: iirc jcsh does not respect DSA keys, I wonder if clj-ssh has the same problem

23:35 DuneMan: should probably not use dsa anymore anyway

23:36 hugod: technomancy: I've not tried that

23:36 clj-ssh uses jsch

23:37 technomancy: hugod: thanks for the tip about slime-default-connection etc; will try that

23:37 no idea why it's giving me the CL-USER prompt

23:37 _brian2_: hugod : i see you say you are trying to imitate something in the contrib library, why is that?

23:37 is your's more up to date?

23:38 hugod: _brian2_: contrib.shell is just for local sh

23:38 _brian2_: ok

23:39 hugod : if I have more questions, do you mind if I post them to the issues area?

23:39 hugod: technomancy: I have seen that in the past - if I remember correctly, it was some sequencing issue, and disappeared after the first comman

23:41 _brian2_: sure, if they are issues, otherwise general usage questions can be answered here

23:41 _brian2_: ok

23:42 hugod: if you don't get a response here, feel free to escalate to the issues

23:43 _brian2_: what does that mean?

23:43 exclamation marks?

23:48 DuneMan: technomancy: http://github.com/hugoduncan/clj-ssh/issues

23:48 erps

23:48 _brian2_: http://github.com/hugoduncan/clj-ssh/issues

23:49 _brian2_: hugod : so if I just want to log into a computer is it -> (ssh "user@something.com) but where do I put the password?

23:49 also that doesn't work

23:49 DuneMan: really, you should generate an rsa key and do passwordless login.

23:49 _brian2_: ok

23:49 DuneMan: even if the lib makes it easy to do it the insecure way.

23:50 _brian2_: the error is java.net.UnknownHostException

23:51 technomancy: I heard RSA had patent issues, which is why I always stuck with DSA

23:52 that was a while ago; wouldn't be surprised if they expired

23:52 DuneMan: looks like username "blah" as an option.

23:52 hugod: _brian2_: (ssh "somthing.com" "ls" :username xx :password yy)

23:52 DuneMan: technomancy: Ah, I'm not actually aware.

23:53 hugod: _brian2_: or whatever command you want to run - it's not going to give you an interactive shell

23:53 _brian2_: ok

23:54 I would like to access a database and manipulate it using a clojure program, is that possible, or am I on the wrong track

23:54 technomancy: hugod: do you do those 2 setqs before it's connected?

23:55 hugod: technomancy: I run them whenever I have trouble making a connection, and then try again

23:55 DuneMan: tunneling a connection of ssh?

23:55 over*

23:56 One would think the right thing to do would be to use your databases java library to manipulate it

23:56 technomancy: hugod: remind me again why I'm running trunk slime? =)

23:56 DuneMan: and, at most, tunnel the connection over ssh.

23:56 And if you're using a database that doesn't have a good java driver, then you should probably think about joining 2010 and using a real database.

23:56 hugod: technomancy: because you like using sbcl ;-)

23:57 DuneMan: unless there's some reason why that's impractical.

23:57 technomancy: oh right, I am a child of the '80s.

23:57 _brian2_: DuneMan : Im using mongo, the new kid on the block

23:57 DuneMan: Then use congomongo.

23:57 _brian2_: i am

23:57 DuneMan: Then you don't need to do anything with ssh.

23:57 * technomancy cues up Tears for Fears on his boom-box

23:58 DuneMan: unless you don't want to open the port and want to tunnel the connection over ssh

23:58 I spoke at MongoSF last week

23:58 love it :-)

23:59 _brian2_: well, where does congo mongo sit when I use it that way?

23:59 DuneMan: Tunneling it?

23:59 _brian2_: I mean is it on the client or server

23:59 DuneMan: Congomongo is a client library for speaking to mongodb.

23:59 It wraps the java mongodb library.

Logging service provided by n01se.net