#clojure log - Jun 17 2012

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

0:01 loliveira: and i just checked. there is no ring.server.leiningen in my source code.

0:01 xeqi: loliveira: what is noir 1.2.2.1-patch ?

0:04 loliveira: xeqi: https://github.com/loliveira/noir/blob/master/project.clj#L4

0:06 xeqi: you're requiring hiccup 0.3.7 in that project.clj, which is not compatible with 1.0.0

0:06 loliveira: xeqi: and... https://github.com/loliveira/compojure/commit/ef382f5e1679aec90ee5fe088047247827cd1846

0:06 xeqi: bingo

0:11 xeqi: changed to [hiccup "0.3.7"] and error persists. =/

0:12 xeqi: which error?

0:13 arrdem: anyone know a good tutorial for building websites with Clojure?

0:13 loliveira: java.io.FileNotFoundException: Could not locate

0:13 +hiccup/page_helpers__init.class or hiccup/page_helpers.clj on classpath

0:15 xeqi: what version of lein are you using?

0:17 loliveira: Leiningen 1.7.1 on Java 1.6.0_31 Java HotSpot(TM) 64-Bit Server VM

0:18 xeqi: what version of hiccup ends up in lib/?

0:22 loliveira: 0.3.7

0:27 akhudek: arrdem: the clojure noir tutorials should help

0:28 unless you want to try clojurescript as well

0:28 clojurescript is not well documented at the moment

0:29 xeqi: loliveira: your custom noir uses hiccup 0.3.7. that means ring-devel 1.0.2, ring 1.0.2, ring-server 0.2.1, and lein-ring 0.6.7 are the lastest you can use

0:31 loliveira: xeqi: makes a lot of sense.

0:35 i removed ring-server "0.2.3", now i back with the original error. =)

0:37 Could not locate ring/server/leiningen__init.class or ring/server/leiningen.clj on classpath

0:38 xeqi: what version of lein-ring ?

0:39 loliveira: :dev-dependencies [[lein-ring "0.7.1"]]

0:41 xeqi: with lein 1.7.1 that should be in :plugins

0:42 loliveira: same error.

0:42 arrdem: akhudek: thanks I'll check it out.

0:42 xeqi: try 0.6.7

0:49 loliveira: smae error with: 0.6.7 0.6.6 0.6.5 0.6.4 0.6.3 0.6.2 0.6.1

0:49 but

0:49 different error with 0.6.0

0:50 http://pastebin.com/jNC1K9eE

0:58 jayunit100__: technomancy: oops i just issued a pull request to leiningan that was a bit of a mistake. Sorry sir ! I meant to simply edit the readme in the main site to link to the stable version rather than the main version.

6:08 _ulises: morning

6:15 atc--: hey, I've come across a peculiarity that my inexperience can't explain. Can someone help me understand what I'm doing wrong?

6:15 http://pastebin.com/8yJhyeV1

6:18 raek: atc--: a string can be used as a sequence of characters, but you try to use it as a sequence of strings

6:18 atc--: raek, I feel stupid

6:18 raek, so I just need to (int ch) the chars instead of sequencing them to strings

6:19 raek: (defn is-period? [c] (or (= \. c) (= \, c)))

6:19 atc--: raek, question is though, why is it getting characters when ltrs is a secquence of strings

6:19 raek, ok

6:19 raek: ah, didn't see that

6:20 (the (map str (seq w)) thing

6:20 atc--: see what I mean?

6:20 ltrs is defined as a sequence of strings

6:20 rightly or wrongly

6:21 so given I'm probably pointlessly converting to a seq of strings

6:21 raek: atc--: also you are calling filter on a single element

6:21 atc--: there is still a bit of oddness in the behaviour

6:21 raek, so?

6:21 raek: which happened to be a string, which is seqable in your case

6:21 atc--: yeah

6:21 but a seq of chars

6:22 I wanted (granted, naively) a seq of strings

6:22 hence the let bindings to (map str (seq w))

6:22 so I can convert to an ascii val later on

6:22 raek: I would do (+ (if (is-capital? (first ltrs)) 1 0) (count (filter is-period? (rest ltrs))))

6:22 atc--: yeah

6:22 I'm just curious why they're chars and not strings when the filter calls is-capital>

6:22 *?

6:22 clojurebot: * is just for when you are lazy and sloppy

6:23 raek: I don't see why having the individual characters as strings simplify things

6:23 atc--: raek, it doesn't

6:23 raek: first you stringify them with str and then you extract the char again with (.charAt c 0)

6:23 atc--: raek, it's making it worse. I'm changing that

6:23 raek, I just want to understand why they're chars at runtime when my understanding suggests they should be strs

6:24 raek: atc--: filter operates on sequences of elements, but you gave it a single element

6:24 atc--: raek, right, but that element is a str, no?

6:24 raek: you could do (filter is-capital? [(first ltrs)])

6:25 atc--: yes, but you want to pass a sequence of strings to that filter, not a sequence of chars

6:25 atc--: right

6:25 so

6:25 raek: since your is-capital? expects a string, not a char

6:25 atc--: does (first ltrs) return a one-element seq of chars or strings?

6:25 raek: no

6:25 atc--: yes, but ltrs is a seq of strings, so when I (first ltrs) I get a one-element seq of str, now?

6:26 no, I mean, not "now"

6:26 raek: if a is ("foo" "bar"), a seq of strings, then (first a) is "foo", a string

6:26 atc--: agreed

6:26 raek: in the first case you give filter something like "foo", in the second case you give filter something like ("foo" "bar")

6:27 atc--: yet is-capital? expects a string to -- badly done, granted -- convert to a character to then convert to an int

6:27 No matching method found: charAt for class java.lang.Character ----- that's the error at runtime

6:27 I shoujld've clarified that further

6:27 earlier, rather

6:28 raek: atc--: do you understand why is-capital? receives a character?

6:28 atc--: raek, I'm sorry, I don't.

6:28 I am new to clojure so forgive my confusion.

6:28 raek: do you know that (filter f coll) runs the function for f for each element in coll?

6:28 atc--: yes

6:28 raek: so what type is coll in the first case?

6:29 atc--: seq of str, with one element?

6:29 raek: no, a single string

6:29 atc--: OH

6:29 raek: ltrs is a seq of strings

6:29 atc--: I should've RTFM'd on first

6:29 yeah

6:30 so,

6:30 raek: anyway, using characters rather than one-char-strings should make your code shorter

6:30 atc--: (first ltrs) gives me "h"

6:30 and that is then passed to is-capital?

6:30 correct?

6:30 so

6:30 raek: no

6:30 atc--: ok

6:30 raek: "h" is treated as a sequence of things

6:31 remember, filter expects a sequence in that position

6:31 atc--: yeah

6:31 raek: and strings can be treaded as sequences in clojure

6:31 atc--: and I'm giving it a string

6:31 I get you

6:31 raek: ,(seq "hello")

6:31 clojurebot: (\h \e \l \l \o)

6:31 atc--: gotcha

6:31 thanks for explaining and for the patience!

6:31 raek: they are treated as sequences of their characters

6:31 atc--: yeah

6:31 which I should be doing in the first place

6:31 raek: so is-capital? receives \h, not "h"

6:31 atc--: but was just curious why my already bad code was misbehaving

6:32 yeah

6:32 got you now

6:32 thanks :)

6:32 I've simplified it all anyhow

6:32 so just using chars

6:32 raek: but since (first ltrs) is always one element, you could do (if (is-capital? (first ltrs)) 1 0)

6:33 no need to loop over each element of a seq if the seq is always of length one

6:33 atc--: raek, yeah

6:33 raek: but if you really want that, then you need to make a seq of one element: (filter is-capital? [(first ltrs)])

6:34 atc--: right

6:34 I think we both agree I don't want that! :D

6:34 unless I'm trying to right superfluous code, which I'm not

7:55 si14: anyone who wrote big web-apps with CLJS here?

8:45 antoineB: hello

8:45 ohpauleez: Hi antoineB

9:30 si14: looks like there is a bug in CLJS — ::foobar in ns "test" is actually :user/foobar instead of :test/foobar

9:33 gfredericks: si14: I don't think that's a CLJS bug; that's a byproduct of cljs using clojure's reader

9:34 antoineB: how can i extract from a list of number the frequency of each number?

9:34 i need some sort of mutable map

9:34 gfredericks: ,(frequencies [2 3 3 2 4 2 5])

9:34 clojurebot: {2 3, 3 2, 4 1, 5 1}

9:34 gfredericks: antoineB: mutable map..?

9:35 raek: antoineB: no, you can use recursion instead of mutation (or use an already available function in the standard library)

9:35 antoineB: i don't know the frequencies function thanks

9:36 raek: how?

9:36 gfredericks: antoineB: if you wanted to implement it yourself you should be able to use reduce

9:36 raek: ,(loop [m {}, coll [2 3 3 2 4 2 5]] (if (empty? coll) m (recur (update-in m [(first coll)] (fnil inc 0)) (rest coll)))

9:36 clojurebot: #<ExecutionException java.util.concurrent.ExecutionException: java.lang.RuntimeException: EOF while reading>

9:36 raek: ,(loop [m {}, coll [2 3 3 2 4 2 5]] (if (empty? coll) m (recur (update-in m [(first coll)] (fnil inc 0)) (rest coll))))

9:36 clojurebot: {5 1, 4 1, 3 2, 2 3}

9:38 raek: antoineB: it's not the map itself you need to update, it's which map is the current one

9:38 gfredericks: ,(reduce (fn [m x] (update-in m [x] (fnil inc 0))) {} [2 3 3 2 4 2 5])

9:38 clojurebot: {5 1, 4 1, 3 2, 2 3}

9:38 cshell: hah,t hat's what I was just oging to type!

9:38 raek: loop/recur and fn/recur provide the basic way of "updating" variables in clojur

9:40 the most common iteration patterns have already been factored out into higher order functions like map, filter, and reduce

9:42 antoineB: raek: gfredericks: tanks

9:57 http://pastebin.com/B4y1sDT7

9:58 it is a function which give the higher element (in term of frequencie) from a frquencies sequence

9:58 is there better way to do it?

9:59 raek: ,(max-key second {2 3, 3 2, 4 1, 5 1})

9:59 clojurebot: {2 3, 3 2, 4 1, 5 1}

9:59 raek: ,(apply max-key second {2 3, 3 2, 4 1, 5 1})

9:59 clojurebot: [2 3]

10:00 raek: antoineB: also, in you most recent paste you could keep the two components of m as separate variables

10:00 antoineB: ,(max-key second (frequencies '(1 2 3 2 5 2 3)))

10:00 clojurebot: {1 1, 2 3, 3 2, 5 1}

10:01 raek: ,(apply max-key second (frequencies '(1 2 3 2 5 2 3)))

10:01 clojurebot: [2 3]

10:02 raek: antoineB: max-key expects the elements as separate arguments

10:02 antoineB: ok

10:03 raek: i don't understand what you say about my paste

10:03 raek: antoineB: sorry, I was reading your code wrong

10:07 si14: what am I doing wrong? https://gist.github.com/cb2458433ee72b72417d

10:07 this code falls to infinite loop

10:07 printing "assoc!" all the time

10:08 jjido: er, reify

10:08 si14: yeah

10:08 I'm trying to mimic backbone with it's models that emit messages when some field is changed

10:09 Chousuke: your assoc method implementation calls itself recursively

10:10 si14: Chousuke: yeah, I know. there is a similar example in clojuredocs for Clojure, don't think that that one loops too

10:10 http://clojuredocs.org/clojure_core/clojure.core/reify here it is

10:11 raek: antoineB: I was thinking about something like this: https://gist.github.com/2944661

10:11 Chousuke: that one doesn't call seq on this, it calls seq on f

10:12 piranha: hmm.. how do I supply clojurescript externs to cljsbuild?

10:14 si14: Chousuke: yeah, I see now, thank you. but if I call (-assoc map k v) in that example I miss reified version of map; how can I avoid loop and keep reified version?

10:14 Chousuke: si14: call the function that creates the reified map with the updated map as parameter

10:15 si14: that avoids calling the method, but creates a new reified thing for you

10:15 si14: Chousuke: yeah, got it. Thanks a lot!

10:17 antoineB: raek: i don't understand the meaning of the "let"s specialy the second vector parameter

10:32 gfredericks: antoineB: you're talking about line 2 in raek's gist?

10:32 ArvinJA: hello there

10:32 Trying out overtone and am new to clojure

10:33 there are some clj files

10:33 How do I run them outside of REPL?

10:33 si14: Chousuke: if I reify one of protocols, others are lost?

10:33 Chousuke: si14: lost?

10:35 ArvinJA: seems like "java -cp clojure.jar clojure.main scriptname.clj" should work

10:35 si14: Chousuke: https://gist.github.com/209b5b7fead972e01296

10:35 antoineB: gfredericks: i am talking about the 2 "let" sexpression

10:35 i don't understand the bindings

10:36 so line 2 and 8

10:37 gfredericks: antoineB: it is using destructing; the expectation on line 2 is that (first coll) is a seq with at least 2 elements

10:37 ArvinJA: E:\overtone\examples>java -cp clojure.jar clojure.main blues.clj

10:37 Error: Could not find or load main class clojure.main

10:37 antoineB: gfredericks: ok

10:37 gfredericks: (let [[first-value first-freq] (first coll)] ...) is equivalent to (let [foo (first coll), first-value (first foo), first-freq (second foo)] ...)

10:38 si14: Chousuke: looks like I'm missing something :)

11:04 dnolen: ping

11:42 dnolen: ohpauleez: ping, yesterday - no protocol method -hash error, is that still a problem, we have default now in master

11:42 si14: pong

11:43 si14: dnolen: hi. first of all — recently saw a record of your talk on dispatch, it's amazing :)

11:44 ohpauleez: dnolen: I haven't seen it for awhile now, it was my most common error

11:44 dnolen: ohpauleez: k good to hear - yeah should no longer occur

11:44 si14: thanks!

11:45 si14: dnolen: and here is a question: how would you implement a modification of map that calls something on each modification of it's field? I'm speaking about CLJS

11:46 dnolen: si14: hard to say without knowing what you are trying to achieve.

11:46 si14: I've tried something like this https://gist.github.com/cb2458433ee72b72417d , but reify returns a new class; extend-type is nice, but I can't call a "super-method" from there, can I?

11:47 dnolen: si14: there's no inheritance of any kind in CLJS, if that's what you mean.

11:47 si14: dnolen: I'm trying to build something like Backbone, because currently I'm not aware of any kind of useful framework in CLJS.

11:48 dnolen: si14: so I take it you're thinking about the data binding problem?

11:48 si14: dnolen: yeah. the idea is pretty simple: just generate an event (using Closure's PubSub, for example) every time when some field is changed

11:49 ohpauleez: si14: I do that almost exact thing in Shoreleave's PubSub system

11:49 si14: it can be achieved with atom and watcher, but unfortunately the only way to check what particular field was changed is comparing old and new map, I don't think that it's reasonable

11:49 ohpauleez: but I have a Function deftype that I abuse

11:50 it's a generic decorator for shaping functions, at the heart of it

11:50 si14: ohpauleez: what's Shoreleave?

11:50 ohpauleez: si14: A group of utilities to make CLJS apps

11:51 with a focus on security, HTML5 feature sets, and CLJS's capabilities

11:51 Embedded workers, a removing package, and a pub sub system to declaratively bind it all together

11:51 dnolen: si14: there's a few people interested in this problem space. Including Kevin Lynagh (C2), Chris Granger (Light Table), Kovas Boguta (Session)

11:51 ohpauleez: We've all taken slightly different approaches

11:52 but they all complement each other

11:52 I don't think any of us have the final answer

11:52 dnolen: si14: I think a new official abstraction is warranted if people can agree upon something.

11:52 si14: ohpauleez: found your repo at github. is there any example of it?

11:53 dnolen: ohpauleez: yes, but it seems like a missed opportunity to create a new reference type, data structure, protocol - no idea what it should look like of course. But it seems like we have enought smart people to figure this out :)

11:53 ohpauleez: dnolen: For sure

11:54 si14: I'm cutting shoreleave up right now. Then I'm moving in the few example apps I have

11:54 Tutorspree has built some serious client-side apps

11:54 si14: ohpauleez: there is a nice thing called "todo mvc" here: http://addyosmani.github.com/todomvc/

11:54 ohpauleez: I worked on that example last night

11:55 si14: the problem is that I didn't find anything even close to https://github.com/addyosmani/todomvc/blob/master/architecture-examples/backbone/js/todos.js

11:55 ohpauleez: I'm going to do a TODO app and a client-side only searching app

11:55 si14: in terms of conciseness at least.

11:55 ohpauleez: I feel like I've nailed the conciseness

11:56 Also you have to realize a lot of these client-side only apps make no assumptions about security

11:56 you need to bolt it all on as an after thought

11:56 it's baked into Shoreleave

11:56 si14: ohpauleez: can you share your results?

11:56 ohpauleez: I mean TODO

11:57 ohpauleez: It'll be pushed as soon as I finish cutting Shoreleave up and putting up the new repos (in the Shoreleave organization on github)

11:59 dnolen: I think we're all close on the binding piece - but definitely something worth talking about over beers on Wednesday

11:59 dnolen: ohpauleez: sounds good!

11:59 si14: kinda sad. we are just in the middle of the "maybe we should just rewrite it to Coffee instead of messing with framework invention" moment, so your example can change the final decision.

12:00 dnolen: si14: I will say I'm not a fan of any of the JS MVCs, Backbone.js included.

12:00 ohpauleez: I also agree with that si14 ^

12:00 si14: dnolen: me neither, but sometimes we just need some stuff done.

12:00 dnolen: si14. for sure.

12:01 si14: dnolen: at least I wouldn't try CLJS if I like JS :)

12:01 ohpauleez: si14: Are you in the NYC area?

12:01 si14: ohpauleez: nope, Russia

12:01 unfortunately :)

12:01 ohpauleez: Ahh, if you can hold off for a little longer, I'll have Shoreleave examples up and probably a screencast of a build out

12:02 But that's two or so weeks out

12:02 si14: ohpauleez: I'm sorry if I was a little too persevering :)

12:03 ohpauleez: haha nah

12:07 si14: ohpauleez: did you see "static" templates in Enfocus?

12:07 sorry, "compiled"

12:07 ohpauleez: si14: Yeah - definitely cool stuff

12:08 I use enfocus in all of my CLJS apps, but I haven't used the compiled templates

12:08 For me, my apps look like this:

12:09 Build some static HTML and CSS. Write a bunch of pure functions. Wire the functions up with my pub sub system. Use enfocus' listeners as entry points into the pub sub system. Use Enfocus' actions as exit points in the pub system

12:10 SL's pub sub allows for Local Storage to be published too, so that solves a lot of problems

12:10 si14: ohpauleez: why not Closure's PubSub?

12:11 ohpauleez: You can subscribe to atoms, workers, and functions

12:11 si14: It's built on top of that, but as a implementation of a protocol

12:11 because you might need cross doc or an xpc bus

12:11 and Closure's pub sub isn't crossdoc

12:11 also it only supports strings as topics

12:12 and I want to use it to declaratively bind my entire app

12:12 So two protocols - one to describe buses. One to describe publishables

12:12 And you can extend the system ala carte to anything that comes along

12:13 si14: ohpauleez: sounds great. bad for us that we didn't start like 2 month later :)

12:14 ohpauleez: si14: It's definitely the wild west

12:14 "wild west" with CLJS

12:19 bobry: wild west indeed :)

13:06 jayunit100_: how integrated is the clojurescript ui tools. Do we have to use jquery ? Id like to do some GWT style development with clojurescript if possible (i.e. have it do my layouts and stuff for me, convention over configuration style).

13:08 I mean, I know its just a JS generator, at its core, but I'm assuming the goal is to wrap JS in a lisp style of GWT, like A DSL for JS web apps. But maybe I'm wrong.

13:11 cshell: do you mean that since GWT compiles to JS, that CLJS is similar as it takes Clojure and compiles to JS?

13:13 dnolen: jayunit100_: you have Google Closure if you want a widget library. being like GWT is a non-goal.

13:13 cshell: yeah, I am missing the GWT/CLJS connection

13:16 jayunit100_: cshell: yeah, thats what i meant.

13:16 but also -

13:16 that GWT looks "good" out of the box :) Convention over configuration.

13:18 So, clojurescript and GWT are BOTH compilers to JS. But GWT is convention over configuration, I believe, whereas ClojureScript is not. I'd like a ClojureScript framework that was convention over configuration, like GWT, SproutTools, ExtJS….

13:19 of course, if the only goal of CLJS is to wrap JS, than my question is quite silly.

13:19 cshell: My understanding is that CLJS helps you write javascript code via clojure - if you want those things, you could just leverage the EXTJs libraries

13:19 gelvaos: hi!

13:19 cshell: I tihink the goal is to wrap js and then you can leverage the native libraries

13:19 call the extjs libraries from clojurescript

13:20 you could build a clojure idiomatic wrapper library over ext-js

13:20 jayunit100_: cshell: thats a good idea.

13:20 cshell: just like they do with all the java libraries in the jvm

13:20 gelvaos: does anybody had experience with setup and usage of hypermedia.video.OpenCV in clojure?

13:23 jayunit100_: arg web guis got so complicated while i was away

13:31 ohpauleez: jayunit100_: Rather than fight upstream, I'd take some time and see how people are generating CLJS ui's now

13:31 you might like another approach better

13:32 otherwise, you could generate JS with GWT, make an externs file, and wrap it all up with CLJS. But that'd be a headache

13:33 jayunit100_: ohpauleez: yeah, i see what you mean. where are the best CLJS apps. Are there any good CLJS app templates out there ?

13:33 (client only, i already have a backend)

13:34 ohpauleez: jayunit100_: ClojureScript One, Take a look at Kevin Lynagh's C2 demos, dig through some of Chris Granger's gists

13:34 Take a look at the enfocus demos

13:36 jayunit100_: ohpauleez: ok thanks

13:36 ohpauleez: np

13:42 uvtc: How does one get started with editing the confluence wiki? I've created an account there, but there doesn't seem to be any way to edit pages.

13:44 ohpauleez: uvtc: You need to request for a bump in permissions on the dev mailing list

13:44 uvtc: ohpauleez, Does editing the confluence wiki require signing/mailing-in the CA?

13:44 dnolen: uvtc: yes

13:45 uvtc: Thanks!

14:25 muhoo: c2?

14:27 gfredericks: d3!

14:29 muhoo: gfredericks: you sunk my battleshiip!

14:29 wtf is c2? probably the most ungoogleable abbrev evah

14:31 gfredericks: oh I see. I did not look far enough back in the conversation

14:32 muhoo: nm, found it http://keminglabs.com/c2/

14:33 gfredericks: perhaps it is an inelastic compute cloud?

14:34 muhoo: naw, data visualization, clojure version of d3

14:34 so your guess was really good

14:54 pipeline: for some reason it bothers me

14:54 that slime still tells me that lemonodor fame is only a hack away

14:54 since lemonodor has so long left us

15:21 gfredericks: writing clojure.core/map in core.logic takes some thinks

15:26 lynaghk: si14: there's a C2-implementation of todo mvc here: https://github.com/lynaghk/c2-demos/tree/master/todoMVC

15:27 Haven't had time to give serious thought to doing more knockout.js-like two-way data binding abstraction between cljs structures and DOM elements (e.g., same interface for reading/writing DOM checkboxes or text areas or some such)

15:28 this demo is basically single direction binding from data -> UI, with manual event handlers that alter the data in response to user actions.

15:34 Jack57: ciao

15:50 jayunit100: hmmm

15:50 No matching field found: contains for class java.lang.String

15:50 ,(.contains "AA" "A")

15:50 clojurebot: true

15:50 gfredericks: ,(.contains "AA" :foo)

15:50 clojurebot: #<ClassCastException java.lang.ClassCastException: clojure.lang.Keyword cannot be cast to java.lang.CharSequence>

15:51 gfredericks: ,(.contains "AA" :foo :bar)

15:51 clojurebot: #<IllegalArgumentException java.lang.IllegalArgumentException: No matching method found: contains for class java.lang.String>

15:51 jayunit100: I would assume that there should always be a (.contains…)

15:51 its a JVM method, not a clojure one.

15:51 gfredericks: jayunit100: are you passing the wrong number of args?

15:52 jayunit100: must be an extra / missing paren !

15:52 gfredericks: good catch :)

16:36 ohpauleez: lynaghk: chopped it like a surgeon: https://github.com/shoreleave

16:39 pipeline: i have a fencepost error that is just killing me

16:39 and i can't seem to see what i'm doing wrong

16:39 http://cljbin.com/paste/4fde404be4b0815a95f5a8fb

16:39 confusingly if i drop the "if" and just use "take 4" i get my desired result so i'm really puzzled about my bad recursion here

16:41 brehaut: pipeline: random probably unrelated comment, i its more idiomatic to test (seq collection) rather than (empty? collection)

16:41 pipeline: and it will let you ditch the 'nil' clause by switching to a when too

16:42 pipeline: yep that is how the original function in clojure.core works

16:42 but i don't understand why my independently mangled version is fencepostin'

16:45 brehaut: pipeline: why are you def'ing a fn ?

16:46 pipeline: brehaut: so i can copy and paste out of emacs and into 4clojure

16:46 brehaut: ah right

16:46 pipeline: in my interactive repl i like to have names for things

16:49 brehaut: now that i have solved it with when-let, as in the source, i see how my solution SHOULD have worked

16:49 brehaut: "(if (not-empty c)" with no false return value

16:49 brehaut: pipeline: you are fence posting because if you take the rest of [1] you get nil, and a head of 1

16:49 lynaghk: ohpauleez: damns.

16:50 ohpauleez: I need to merge in the local storage stuff, but it'll all be up on clojars soon. Plus the super project: shoreleave

16:50 brehaut: pipeline: i you are cheating by looking at the source, i guess you can see that in the real version, the cons is outside the when

16:50 lynaghk: ohpauleez: one suggestion; if you're going to maintain remotes, you should decouple them from noir. At the very least, make sure you talk with Chris about the shared namespace "noir".

16:50 pipeline: brehaut: yeah i discovered that problem also

16:50 ohpauleez: Already decoupled

16:50 and already talked to chris

16:51 lynaghk: ohpauleez: but it isn't hard at all to make it accessible from just ring or compojure, and that would make it much more general and useful to a lot of people (myself included)

16:51 ohpauleez: I'll define a shoreleave-remote-compojure

16:51 just for you :)

16:52 brehaut: pipeline: in future, it can help to remember that fn bodies have implicit do's

16:52 pipeline: so you can jam a (prn …) in before your actual behavior

16:53 pipeline: yeah a common-lisp-ism i appreciate

16:55 lynaghk: ohpauleez: Don't go copy/pasting into a whole new repo just for me; my suggestion would be to sit down and think about how you could make the same codebase work with both. Also in the docs mention why this should be used over fetch (you patched some XSS stuff, yeah?)

16:56 ohpauleez: I mention CSRF and remote-ns

16:56 I'll make it more explicit, thanks

17:02 bbloom: ohpauleez: looks like a nice collection of libraries you got there

17:02 ohpauleez: Thanks man. Docs and demo projects are just around the corner

17:02 bbloom: ^

17:02 bbloom: ohpauleez: you using this stuff in prod?

17:03 ohpauleez: yeah, it was built in parallel as we were building apps

17:03 so, in C/Unix fashion … but because of protocols, it's all bottom-up in true Lisp fashion

17:04 bbloom: ohpauleez: cool. https://github.com/shoreleave/shoreleave-core/blob/master/src/shoreleave/common.cljs#L13 does that work with maps of more than 32 keys?

17:04 ohpauleez: it needs a lot of scrubbing up though

17:05 clj->js is crutch - so I would only use it if you really have to

17:05 bbloom: ohpauleez: was just skimming the code. looks bugged to me, it relies on the internal implementation of ObjMap, which gets promoted to a PersistentHashMap after some number of updates or size

17:06 ohpauleez: Where at?

17:06 You mean -strobj?

17:06 bbloom: yeah

17:07 ohpauleez: That doesn't work on PersistentHashMap?

17:07 (I've never encountered a moment where that failed)

17:08 but I also don't use it very much

17:08 What would be the generic protocol call replacement?

17:08 just str?

17:09 bbloom: try this:

17:09 (.-strobj (into {} []))

17:09 vs (.-strobj {})

17:09 you'll notice the former is nil

17:09 you can check for strobj as an optimization, if you want

17:10 but really you should (doseq [[k v] map] ...)

17:11 ClojureScript:cljs.user> (type (loop [m {} i 0] (if (< i 10) (recur (assoc m (str i) i) (inc i)) m)))

17:11 cljs.core.ObjMap

17:11 ClojureScript:cljs.user> (type (loop [m {} i 0] (if (< i 100) (recur (assoc m (str i) i) (inc i)) m)))

17:11 cljs.core.PersistentHashMap

17:12 ohpauleez: Ahh there you have it

17:12 Ok, I'll make a ticket, thanks man

17:13 bbloom: np

17:18 ohpauleez: bbloom: https://github.com/shoreleave/shoreleave-core/compare/master...cljsjs

17:19 uh, variable names, hang on

17:19 bordatoue: which one is good for emacs inferior lisp or swank clojure

17:20 ohpauleez: bbloom: ok, fixed. Look good to you?

17:20 bbloom: ohpauleez: lgtm

17:21 bordatoue: can anyone advise on the pros and cons for inferior lisp and swank clojure mode

17:24 antares_: bordatoue: compared to what?

17:24 swank kinda uses inferior lisp approach, no?

17:25 and at least for clojure, swank is more widely used than bare inferior lisp

17:31 bordatoue: antares_: thanks, I just wanted to know which one was widely supported for clojure emacs

17:31 mattmoss: ,postwalk

17:31 clojurebot: #<CompilerException java.lang.RuntimeException: Unable to resolve symbol: postwalk in this context, compiling:(NO_SOURCE_PATH:0)>

17:32 antares_: bordatoue: clojure-mode's clojure-jack-in uses swank-clojure under the hood

17:32 mattmoss: ,clojure.walk/postwalk

17:32 clojurebot: #<CompilerException java.lang.RuntimeException: java.lang.ClassNotFoundException: clojure.walk, compiling:(NO_SOURCE_PATH:0)>

17:32 bbloom: ,(ns-aliases *ns*)

17:32 clojurebot: {}

17:33 antares_: mattmoss: http://clojuredocs.org/clojure_core/clojure.walk/walk

17:33 mattmoss: antares_: I'm looking at postwalk there now, trying to understand the example.

17:33 On http://clojuredocs.org/clojure_core/clojure.walk/postwalk

17:34 If I use their example on [:a :b], I get [2 [[0 :a] [1 :b]]]

17:34 antares_: yeah, the example there is confusing

17:34 mattmoss: But the same example on {:a :b}, I get [3 {2 [[0 :a] [1 :b]]}]

17:34 Trying to understand the reason for the difference.

17:34 bbloom: ,(filter #(.contains (str %) "walk") (ns-refers *ns*))

17:34 clojurebot: ()

17:35 bbloom: mattmoss: apparently not available in clojurebot :-)

17:35 antares_: mattmoss: because iteration over maps vs vectors is slightly different, perhaps?

17:35 mattmoss: antares_: I guess... I suppose I need to dig into the source for walk/postwalk and see where the difference occurs.

17:36 It just seems like it's doing an extra evaluation for the map compared to the vector.

17:36 Or, perhaps, one less for the vector compared to the map.

17:37 Bronsa`: ,(do (in-ns 'clojure.walk) postwalk)

17:37 clojurebot: #<CompilerException java.lang.RuntimeException: Unable to resolve symbol: postwalk in this context, compiling:(NO_SOURCE_PATH:0)>

17:37 antares_: mattmoss: that's because iterating (e.g. doseq) over a map yields pairs and for vectors, only values

17:37 Bronsa`: &(do (in-ns 'clojure.walk) postwalk)

17:37 lazybot: java.lang.ClassNotFoundException: clojure.walk

17:37 Bronsa`: derp

17:37 mattmoss: lol

17:37 Bronsa`: i was pretty sure it was going to work

17:38 mattmoss: antares_: I guess that makes sense... iteration over a map "adds" an extra form, in a sense... the pairing

17:40 Not that any form is truly being added, but it just seems like that because [:a :b] looks so much like {:a :b} in the editor.

17:41 replcated: Let me start by saying I'm not flamebaiting. Did anyone here that really loves and makes regular use of clojure start out hating it? I'm struggling to get beyond what-the-hell-was-he-thinking and actually start on the project I had in mind.

17:43 Raynes: I might have, but Rich Hickey has been very clear in numerous presentations and talks what the hell he was thinking. :>

17:43 zomg: replcated: I didn't like the syntax

17:43 I always found lisp syntax stupid

17:43 Still do kind of, but I can live with it :P

17:46 Raynes: It still surprises me that people concern themselves with parentheses.

17:46 brehaut: ive never loved a new syntax for a programming language i dont understand

17:46 replcated: Raynes, I understand. I've actually watched a number of his talks.

17:47 I'm actually coming from a CL background, so the parens don't bother me.

17:47 zomg: Raynes: it's just that every time I move some lines of code or such I always need to check them when in most other languages you don't need to bother with such things

17:47 Of course paren matching in the editor helps with that but it's still something you need to do

17:47 brehaut: zomg: no its not

17:47 Raynes: I don't have to do it.

17:47 replcated: zomg: A "real" editor helps a lot.

17:47 Raynes: paredit does it for me.

17:48 zomg: I guess you're using emacs or something then

17:48 brehaut: zomg: structural editing removes the need completely

17:48 Raynes: Even counterclockwise has some paredit support these days.

17:48 brehaut: CCW has had paredit support for at least 2 yeard

17:48 Raynes: And besides, with paren matching, I don't understand how you have to pay any more attention than you would with another language.

17:48 zomg: Using vim myself, and tbh it's mostly an issue when I'm being lazy and using dd and not d% or such :P

17:49 Raynes: When I used Vim, I used paredit.vim and lots of da(

17:49 I actually do the same thing in Emacs because I use evil-mode

17:51 antares_: hey Raynes

17:51 Raynes: antares_: Hey thar

17:52 antares_: Raynes: pretested pull requests on travis-ci.org are on for your projects (fs, bultitude, conch — everywhere we could find a travis file)

17:52 Raynes: antares_: clojail is a flatland project

17:53 That's probably why you didn't see it.

17:53 Anyways, awesome!

17:53 Now I just need people to send me pull requests.

17:53 antares_: Raynes: turning it on for clojail, too

17:54 I have forgotten about it

17:56 Raynes: can you please log in on travis-ci.org so we can sync your repos?

17:56 flatland/clojail is not in our database

17:57 Raynes: antares_: Logged in.

17:58 antares_: Raynes: do you see flatland/clojail on your profile page?

17:58 if you do, that's all we need

17:58 Raynes: Yes.

17:58 None of these are 'on' though, because I added all the hooks manually (because this page was broken before).

17:59 antares_: Raynes: that's fine

17:59 as long as our database has the repo

19:19 brooksbp: hello

19:19 newbie question

19:19 is there a way to conj to the end of a map?

19:20 I want to do this: (conj {:a 1} [:b 2]) => {:a 1 :b 2}

19:21 Raynes: You can't rely on the ordering of maps.

19:27 brooksbp: Raynes: I see.. I guess it doesn't matter in my case. It's just easier to read when I evaluate the map in repl to view contents.. would like it to be ordered

19:27 gfredericks: brooksbp: clojure maps don't really have an 'end'; semantically they're unordered

19:28 brooksbp: I'm guessing it's for implementation efficiency?

19:29 bbloom: brooksbp: and because, semantically, what order do you want?

19:29 ,(doc sorted-map-by)

19:29 clojurebot: "([comparator & keyvals]); keyval => key val Returns a new sorted map with supplied mappings, using the supplied comparator."

19:29 bbloom: brooksbp: if you just want orderd keys for debugging sake, you can use something like:

19:29 brooksbp: semantically I don't care

19:30 bbloom: ,(into (sorted-map) (conj {:a 1} [:b 2]))

19:30 clojurebot: {:a 1, :b 2}

19:34 brooksbp: I am coming from an imperative language background. Does clojure have some way to use debug flags in expressions? Something to the effect of #ifdef DEBUG ... #endif

19:35 Like... if I wanted to write that function, but have two cases where debug would use (sorted-map) version and non-debug wouldn't

19:35 Would that require passing around a debug var to every stinkin function?

19:40 Hodapp: somehow I'd think you could make use of a global.

19:50 miclorb: I am trying to look up what "->" really does - not the easiest thing to search for - anyone got a link to the doc?

19:50 bhenry: ,(doc ->)

19:50 amalloy: &(doc ->)

19:50 lazybot: ⇒ "Macro ([x] [x form] [x form & more]); Threads the expr through the forms. Inserts x as the second item in the first form, making a list of it if it is not a list already. If there are more forms, inserts the first form as the second item in second form, etc."

19:51 clojurebot: "([x] [x form] [x form & more]); Threads the expr through the forms. Inserts x as the second item in the first form, making a list of it if it is not a list already. If there are more forms, inserts the first form as the second item in second form, etc."

19:53 amalloy: see also http://clojuredocs.org/clojure_core/clojure.core/-%3E

19:53 arohner: miclorb: (-> a (foo 1) (bar 2) (baz 3)) = (baz (bar (foo a 1) 2) 3)

19:55 miclorb: amalloy: arohner: thanks !

19:55 I can see will see will take some practice to internalise, but looks really handy

19:55 amalloy: though tbh lines 15-38 of the first example on that page should be thrown in the trash

19:57 arohner: amalloy: what's wrong with that example?

19:57 amalloy: (-> foo :x :y :z) is *not* the same as (((foo :x) :y) :z)

19:57 it is the same as (:z (:y (:x foo))), which just happens to usually evaluate to the same result in most clojure programs

19:58 arohner: right. though that seems to be a bug w/ line 34 then, not a problem with having the example

19:58 amalloy: further, -> is only known as the "thrush operator" by people who read about thrushing and think it's cool. thrush is a function, which does something sorta similar to what -> does as a macro, in most cases

19:58 arohner: and IMO, if (foo :x) is not the same as (:x foo) in your program, you have other bugs

19:59 amalloy: i also

19:59 er

19:59 arohner: (= x nil)

19:59 gfredericks: arohner: or you are using records

19:59 amalloy: or foo is nil

19:59 or a record

19:59 gfredericks: can you use (:x foo) for a j.u.Map?

20:00 ,(let [m (java.util.Map.)] (.put m :foo 7) (:foo 7))

20:00 clojurebot: #<CompilerException java.lang.IllegalArgumentException: No matching ctor found for interface java.util.Map, compiling:(NO_SOURCE_PATH:0)>

20:00 gfredericks: ,(let [m (java.util.HashMap.)] (.put m :foo 7) (:foo 7))

20:00 clojurebot: nil

20:00 gfredericks: ,(let [m (java.util.HashMap.)] (.put m :foo 7) (:foo m))

20:00 clojurebot: 7

20:00 amalloy: as a matter of taste, i also don't want anyone to be told that you should use -> to get values from a nested data structure - that's exactly what get-in is for, and -> doesn't make it readable

20:00 but i realize that's not a universally-shared opinion

20:01 arohner: amalloy: yes. :-)

20:01 * gfredericks universally shares amalloy's opinion

20:01 arohner: I think -> is more readable once you're used to it, and I like the flexibility of being able to use things like first & last in the -> expr

20:02 amalloy: arohner: i love ->. but not for mundane data-structure drilling

20:02 brehaut: i like -> for unpacking stuff like that only if im chucking an (or …) as the last term

20:02 amalloy: brehaut: (get-in m [a b c] default)

20:02 i think

20:02 brehaut: oh true

20:02 amalloy: &(doc get-in)

20:02 lazybot: ⇒ "([m ks] [m ks not-found]); Returns the value in a nested associative structure, where ks is a sequence of keys. Returns nil if the key is not present, or the not-found value if supplied."

20:02 * brehaut goes and rewrites code

20:03 gfredericks: but there's no substitute for (-> m :foo :bar fn1 fn2 (try (catch Exception e e)))

20:03 arohner: I still prefer (-> m a b c (or default))

20:03 amalloy: doesn't work if your value is nil or false

20:03 arohner: shrug. doesn't happen often enough for me to say "never use ->" for nested structures

20:04 brehaut: amalloy: who are you replying to there?

20:04 amalloy: to arohner

20:04 i didn't say never use it. i use it when it's appropriate. but i don't want it held up as a primary example usage of ->

20:05 but i doubt if i'll ever make any changes to clojuredocs, so my opinion doesn't matter much

20:05 brehaut: oh phew. my use of (-> … (or …)) has fns as well as keywords

20:06 bbloom: brooksbp: regarding your question about a debug flag. you can just use any old var: (def foo 123)

20:14 gfredericks: you could use a macro to do the debug testing at compile time

20:15 akhudek: interesting, it seems that sqlkorma insert statements need select permissions in the database

20:24 has anyone encountered this? Do I really need to drop down to raw jdbc just to do a pure Insert?

20:26 bbloom: akhudek: what query gets generated?

20:26 akhudek: bbloom: using sql-only prints out a query that is just a normal looking insert.

20:26 it works direclty in psql

20:27 but not via korma

20:27 if I add select permission to the user, then the korma version starts working

20:27 bbloom: akhudek: have you tried logging all queries on your server to see if any ADDITIONAL queries get run? i know that rails's active record often queries extra stuff, like table shapes

20:28 akhudek: I'll try that

20:28 bbloom: akhudek: can also look at the norma source more carefully for calls to whatever it is that does the query execution

20:28 korma* dopey osx autocorrect

20:29 akhudek: ah, in the logs it has a "RETURNING *" at the end

20:29 I'm assuming that's the problem

20:30 bbloom: akhudek: probably :-)

20:30 akhudek: now to figure out how to get korma not to do add that!

20:30 bbloom: akhudek: curious: why don't you allow select permissions? are you exposing the sql db directly without any sort of API or anything in between?

20:31 akhudek: it's a table that logs transactions in a write only fasion

20:31 gfredericks: https://www.refheap.com/paste/3202

20:31 bbloom: akhudek: that i assumed

20:31 gfredericks: ^ a with-foo macro-writing macro

20:32 akhudek: it aggregates data from separate databases; the databases are separate for security concerns

20:32 bbloom: akhudek: ok gotcha

20:32 gfredericks: heh. nice.

20:33 gfredericks: bbloom: dangit I don't think it works

20:33 bbloom: gfredericks: now make it support multiple different arities ;-) you'll need a multi-arity macro-defining-macro

20:33 gfredericks: bbloom: it's supposed to

20:33 you should be able to (def-wrapper-macro [x y z func] (prn x y z) (func))

20:33 but I think that fails

20:33 * gfredericks debugs

20:34 bbloom: gfredericks: i mean (def-wrapp-macro ([x] (prn x)) ([x y] (....

20:34 gfredericks: bbloom: I can't even get my head around what that means

20:35 bbloom: gfredericks: heh nevermind then

20:35 gfredericks: bbloom: also it might not make sense here

20:35 certainly the resulting macro is already vararged

20:35 since it wraps them up in the body func

20:36 yeah I'm going with that doesn't make sense here

20:36 bbloom: gfredericks: *looks again* oh yeah you're right :-P

20:38 akhudek: if anyone ever needs this, you can disable returning of results by simply removing the :results key from the query record

20:39 kmicu: gfredericks: what are these 3 asterisks?

20:39 gfredericks: kmicu: ...three asterisks?

20:40 kmicu: gredericks: ***gfredericks debugs

20:40 gfredericks: oh; that must be your IRC client

20:40 kmicu: try typing "/me tries this out"

20:41 * kmicu irc noob

20:41 akhudek: https://gist.github.com/2946182

20:42 bbloom: akhudek: does that work?

20:42 akhudek: yep

20:42 bbloom: cool! glad you got it :-)

20:42 bordatoue: when i load clojure-jack-in it never load clojure.core or clojure.repl , is there any reason for it

20:45 hello

20:45 tmciver: bordatoue: it *should* load clojure.core, but not clojure.repl.

20:46 bordatoue: tmciver: you are right it loaded clojure.core but not clojure.repl

20:47 tmciver: as a result i don't have doc fn working

20:47 tmciver: bordatoue: yeah, it's a pain but you always have to 'use' clojure.repl manually.

20:47 bordatoue: tmciver: then wat is the point of this clojure-jack-in

20:47 tmciver: bordatoue: I think there might be a way to get it to 'use' it automatically but I haven't spent the time to figure out how.

20:49 bordatoue: tmciver: thanks, i will investigate this

20:49 tmciver: bordatoue: there's a lot of benefit. Check out https://github.com/technomancy/swank-clojure if you haven't already.

20:50 bordatoue: tmciver: any idea how you test your private method

20:50 tmciver: bordatoue: for example, I often C-c C-l a file to load the current buffer into the repl and then C-c M-p to change the repl's ns to that file's ns.

20:50 bordatoue: tmciver: private fns from repl

20:51 tmciver: thats handy

20:51 tmciver: bordatoue: Yep. And M-. is another good one.

20:51 bbloom: bordatoue: you can also use the var object, which is invokable

20:51 (#'somens/somefn arg1 arg2)

20:51 bordatoue: tmciver: when i am creating a executable jar, should i explicity mention gen-class in all clj file or just the project.clj file

20:54 tmciver: bordatoue: I don't know much about it but I think you can get away without AOT but using TimMc's lein-otf: https://github.com/timmc/lein-otf

20:55 bordatoue: tmciver: what does M- do, i am using aquamacs and can't see any binding to it

20:55 tmciver: bordatoue: mind the period: M-.

20:55 bordatoue: tmciver: oh..

20:56 tmciver: bordatoue: Jumps to a definition of the var under point.

21:26 calvados: is there a way to get request headers in ring without using middleware ?

21:26 weavejester: calvados: In Ring? They're in the request map

21:27 calvados: (fn [req] (do-something-to (:headers req)))

21:27 calvados: yes but how to get req without middleware

21:27 weavejester: calvados: It's the argument passed to the handler

21:27 brehaut: calvados: in _ring_ or some other lib on top like compojure?

21:28 calvados: actually what i want to do is

21:28 weavejester: That was going to be my next question :)

21:28 calvados: i want to change the content-type dynamicly i will check Request:headers:accept part and will change content-type

21:29 weavejester: calvados: What library or framework are you using? Are you just using Ring straight off?

21:30 calvados: so lets say a got (defn - response [resposne] {:status 200 :body "something" headers: *i-need-to-get-request-headers-here* })

21:30 yes ring stuff

21:31 i just need to get request in (response [response])

21:31 weavejester: calvados: The request is passed as the argument to the handler function

21:31 calvados: So your handler looks like (defn handler [request] …) right?

21:31 calvados: The headers are in the request map passed to the handler function.

21:31 calvados: hm

21:31 brehaut: calvados: you might be overthinking this

21:32 calvados: well i am quite new

21:32 brehaut: calvados: look down at "Request Map" https://github.com/mmcgrana/ring/blob/master/SPEC

21:32 calvados: my brain still refuses to get it :)

21:32 weavejester: calvados: In ring you have a request, pass it to a function called a handler, which returns a response.

21:33 calvados: yes

21:33 i mean i got more than one handler because of routes thought

21:33 each route goes different handler i beleive

21:33 brehaut: what library are you using for your routes?

21:34 ring doesnt do routes itself

21:34 calvados: :ring {:handler project./map-routes}

21:34 it should be ring

21:35 brehaut: calvados: ring is the lowest level api in the clojure web stack. lots of things are part of the ring ecosystem, but are not ring itself

21:36 weavejester: calvados: Routes might indicate you're using Compojure

21:36 calvados: (defroutes map-routes)

21:36 weavejester: calvados: That's Compojure

21:36 calvados: ok

21:36 yes it is sorry

21:36 weavejester: calvados: You can use destructuring to access the request

21:36 brehaut: calvados: you might find http://brehaut.net/blog/2011/ring_introduction helpful

21:37 weavejester: calvados: Like (GET "/" request …) or (GET "/" {:as request} …) or (GET "/" [:as request] ...)

21:39 calvados: weavejester: than i need to write it to every route thought

21:39 weavejester: calvados: That's usually an indication you want middleware

21:39 calvados: weavejester: i used middleware but the problem

21:39 weavejester: calvados: Doing some operation to multiple routes = middleware

21:40 cshell: isn't middleware analogous to a filter?

21:41 calvados: yes i used middle ware lets says client a visited page content-type changed to a but he got nil , client b visited and he got content-type a and it content-type set to b , when client a revisits he gets content-type response as b

21:41 weavejester: cshell: It's more like something that modifies the behaviour of a function in a general way.

21:41 calvados: i thought it is not the correct place to modify it, because i am modifying and next request its getting active

21:41 let me paste bin

21:42 mmarczyk: dnolen: ping

21:44 weavejester: calvados: The request is immutable. If you alter it for one handler, it's not going to be altered for another.

21:44 All data structures in Clojure are immutable.

21:46 cshell: after middleware returns, is there any code that runs after middleware that could alter the result?

21:46 brehaut: cshell: any other middleware that wraps that middleware

21:46 cshell: or does middleware wrap handlers on both entry and exit

21:47 calvados: http://pastebin.com/1thhJE9a

21:47 cshell: brehaut: ah okay, thanks!

21:47 brehaut: cshell: it pays not to draw to stark a line between middleware and handlers

21:47 cshell: ah

21:47 calvados: weavejester: i guess its not the best way i am using but i got it so far

21:47 cshell: calvados: you shouldn't def inside a function

21:47 use let

21:47 that will slow you down a lot

21:47 brehaut: cshell: with something like compojure its possible that your handlers are middelware themselves

21:48 weavejester: calvados: Yes, don't def inside functions - you're using Clojure like an imperative language.

21:48 cshell: brehaut: I'm using noir - have been injecting the friend middleware in

21:48 calvados: weavejester: i told i am newbie

21:48 brehaut: cshell: my noir knowledge is extremely limited

21:49 cshell: brehaut: I considered moving back to plain ring, but it seems like noir is just a bunch of helper functions that help to use ring

21:50 weavejester: calvados: I'd like to help, and usually I would, but I'm almost out of time. I need to get to bed!

21:50 calvados: But in a nutshell...

21:50 brehaut: cshell: theres a big drop off to go back to 'plain ring' but stepping back to plain compojure isnt a huge change. i suspect you could still use a bunch of noir's helpers from compijure anyway

21:50 calvados: weavejester: ok thanks anyway, i need to search abit more.

21:51 brehaut: cshell: however, i would recommend learning about how ring and compojure work as they are the underpinnings anyway, and they arent hard to grasp

21:51 weavejester: calvados: If you get rid of the defs and use lets instead, that should be your starting point...

21:51 cshell: brehaut: yeah, that's a good idea - i've been going through their apis in the last couple of days

21:51 i'm so close to getting friend to work

21:52 brehaut: cshell: good plan. i think looking at how routes and GET etc interact is a great place to start. it embodies the spirit of compojure without having to grok a huge amount of code

21:52 calvados: weavejester: yes but test is out of it so i cant use let, as i read let is local

21:52 cshell: calvados: oh no

21:53 weavejester: calvados: Yes, but there's nothing you're doing that isn't local

21:53 cshell: brehaut: thanks!

21:53 * gfredericks just realized namespace names could be non-alphanumeric

21:54 gfredericks: ,'<o>.<o>

21:54 clojurebot: <o>.<o>

21:54 brehaut: cshell: weavejester might disagree ;)

21:55 cshell: brehaut: he doesn't ahve time to, tonight :)

21:56 ibdknox: some enterprising individual could fairly easily create the wrap-noir-around-compojure middleware :)

21:57 cshell: what would the code for that look like? ;)

21:57 weavejester: Goodnight all

21:57 brehaut: night weavejester

21:57 cshell: night, wj

21:59 brehaut: ibdknox: i cant imagine that would be very much code?

21:59 ibdknox: yeah, not much at all

21:59 brehaut: ibdknox: is there an noir equivalent to the context macro in compojure?

21:59 ibdknox: no

22:00 though that's another thing that could be added fairly easily

22:00 not as easily as wrapping compojure though

22:00 brehaut: heh, added that would effectively make it trivial to drop compojure in

22:01 ibdknox: yeah

22:10 akhudek: ibdknox: is noir-cljs not intended to be used with uberjar or war files for production use?

22:10 ibdknox: I guess it could be

22:11 xeqi: is that just (fn [h1 h2] (fn [r] (if-let [res (h1 r)] res (h2 r)))) ?

22:12 ibdknox: akhudek: you'd have to do some trickery to make it work

22:12 Raynes: ibdknox: What is context vs :base-url?

22:12 ibdknox: Raynes: potentially nothing, I didn't know about context

22:13 Raynes: until much later

22:13 Raynes: ibdknox: BTW, did you see https://github.com/Raynes/refheap/issues/87

22:14 brehaut: ibdknox: i think context was new around the time you released noir

22:14 ibdknox: Raynes: what's the canonical host thing for?

22:14 brehaut: I think it came a bit after

22:14 brehaut: in that sort of ballpark anway

22:14 ibdknox: yeah

22:14 if it serves the same purpose, I'm all for consolidation

22:14 brehaut: it wasnt particurly well known about when it was out too

22:15 Raynes: ibdknox: For web apps that have numerous hosts. For example, refheap is refheap.com, www.refheap.com, and refheap.herokuapp.com. It's a way to redirect to whatever the canonical host is (in refheap's case, www.refheap.com) at an application level, since you can't do it at a server level in Heroku (and I imagine other scenarios).

22:15 xeqi: Raynes: friend also has a similar force-ssl fn https://github.com/cemerick/friend/blob/master/src/cemerick/friend.clj#L46

22:16 ibdknox: Raynes: I see

22:16 Raynes: ibdknox: Note that I'm not claiming any of this to be useful in general, just some guy mentioned it.

22:16 So I figured it worth asking what you thought.

22:16 ibdknox: Raynes: they all sound reasonable to include in noir. They could use comments though :)

22:17 Raynes: Of course. :P

22:17 xeqi: His is bigger than mine.

22:17 * Raynes slaps his knee and guffaws.

22:17 ibdknox: that is an epic name

22:18 Raynes: Mine also handles the case where the scheme is there too.

22:19 xeqi: yah, thats his require-scheme

22:20 akhudek: ibdknox: I think I'll just switch over to cljsbuild for now since it supports pre-compilation. Noir-cljs is pretty nice during dev work though.

22:20 ibdknox: akhudek: hm?

22:20 xeqi: just mentioning since it looks like that concept is useful to move to a higher spot

22:20 ibdknox: if you run the server in prod mode

22:20 akhudek: and then uberjar it will be precompiled to advanced

22:21 akhudek: I had to disable aot compilation for it to work though

22:21 also, it needs to compile before it packages the jar

22:22 I suppose I could run it once in prod mode, re-enable aot, then run uberjar

22:22 ibdknox: yeah, admittedly not something I ever do

22:22 I just use lein to run my apps

22:25 brehaut: ibdknox: what a cowboy :P

22:32 akhudek: ls

22:32 whoops

22:36 arohner: how do I write a literal plus in a regex?

22:36 brehaut: [+] ?

22:36 \+ ?

22:37 arohner: those aren't working for me

22:37 neither is \\+

22:37 I'd like something like #"^[a-zA-Z0-0/_-+]+$"

22:37 brehaut: , (re-find #"\+" "a+b")

22:37 clojurebot: "+"

22:38 arohner: java.util.regex.PatternSyntaxException: Illegal character range near index 14 ^[a-zA-Z0-0/_-+]\++$

22:38 brehaut: ah

22:38 its the hyphen

22:38 arohner: oh right

22:38 brehaut: it defines a range unless its the first character in the character class

22:39 arohner: right, forgot about that

22:39 thanks!

22:40 brehaut: no problem

22:57 * gfredericks scans the reasoned schemer for a reverso

23:07 ghadishayban: does anyone know how to force AOT to be 1.6 JDK compatible, when running 1.7?

23:07 is there a lein flag?

23:08 antares_: ghadishayban: yes

23:08 ghadishayban: what version of lein?

23:08 ghadishayban: 2-preview

23:08 antares_: :javac-options ["-target" "1.6" "-source" "1.6"]

23:09 ghadishayban: that affects clojure AOT as well? I thought it's just for .java

23:09 antares_: it is only for Java sources

23:09 ghadishayban: oh

23:10 I uberjarred a project using OpenJDK 1.7, ran it on a 1.6 hadoop cluster...had a bad time

23:11 received checksum errors on map input files, all sorts of unknown badness. took a while to track down

23:11 gfredericks: does the clojure compiler actually vary its bytecode output?

23:12 ghadishayban: that's the thing. i doubt it, doesn't clojure use an old version of ASM?

23:12 * gfredericks doesn't know much

23:12 antares_: ghadishaybangh: from what I see, lein only sets up some system properties Clojure compiler uses

23:13 so maybe there is a property for that

23:13 ghadishayban: actually, I remember doing something similar and I think :javac-options helped me (but I also have java sources in that codebase)

23:16 ghadishayban: interesting (i am not using any java in this project)

23:16 well it's running like a champ now that I figured it out

23:16 antares_: ghadishayban: what did you do? did :javac-options help?

23:17 ghadishayban: i compiled on 1.6 and deployed *that* (update-java-alternatives in ubuntu)

23:17 antares_: ghadishayban: yeah but that's a workaround, not a solution :)

23:18 ghadishayban: exactly

23:18 gfredericks: the difference between a workaround and a solution is how tired you are

23:19 ghadishayban: i was about to give up on hadoop forever =)

23:19 the job was only calling (read-string) on about 150GB, feeding it into Cascalog and sorting/reducing...

23:20 this was my baptism with hadoop, it seems to me to be help duct tape

23:21 i'd much rather work using only the clojure reader for my purposes

23:39 antares_: gfredericks: :)

Logging service provided by n01se.net