#clojure log - Mar 16 2013

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

0:00 yedi: that's all good an well, but I want my app to work locally as well

0:00 so I want to be able to set the properties locally

0:02 jeremyheiler: yedi: Are you using lein?

0:02 yedi: to deploy to beanstalk, yes

0:02 to do everythinf*

0:03 jeremyheiler: If you're using lein to run the application locally, I *think* you can set :jvm-opts in your project.clj to set -D options.

0:03 tyler__: is there a clojure lib for running popular shell commands like `ls` `ps` etc but not from the shell, a native clojure version of it

0:05 ToxicFrog: I have caused the best error

0:05 2013-03-16 00:06:08,544 ERROR - (nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil)

0:06 tyler__: heh

0:06 jeremyheiler: tyler__: Well, you could use fs to emulate ls.

0:06 * tyler__ does not grok the meaning of fs

0:06 amalloy: tyler__: you want conch, by Raynes

0:06 jeremyheiler: tyler__: Actually, there is fs.core/list-dir

0:07 tyler__: Raynes makes some good stuff

0:07 Raynes: amalloy: He said specifically a native Clojure version of it though, so fs is what he wants in this case.

0:07 http://github.com/Raynes/fs

0:07 tyler__: ^

0:07 amalloy: oh. i filtered that, because it's a crazy thing to want

0:07 tyler__: what im *really* looking for is anything that outputs datastructures

0:07 it cnabe java

0:07 Raynes: lol

0:07 tyler__: can be*

0:08 amalloy: nobody is going to reimplement an operating system in clojure

0:08 tyler__: seems like an honest ambition if someone were to do it

0:08 heh

0:09 yedi: jeremyheiler: -D options?

0:09 jeremyheiler: tyler__: You're trollin'

0:09 lol

0:09 tyler__: im not saying they *should*

0:09 im just saying i would cheer them if they did

0:09 jeremyheiler: yedi: Read up on java command line -D

0:09 amalloy: you'll just get the lisp-curse version of an operating system

0:09 "all these other lisp operating systems are kinda okay, but they're missing feature X and do Y in a really weird way, i'll just write my own"

0:10 jeremyheiler: yedi: Better yet, just type "java" in a terminal and it explains it.

0:10 yedi: oh cool, thanks jeremyheiler

0:11 tyler__: im trying to build a system that talks to a server over ssh to get a bunch of meta about the server, right now im using standard shell commands to get data, but each one of them has their own output format

0:11 and its a pain in the ass

0:12 jeremyheiler: yedi: np

0:12 tyler__: i found a good awk script that converts column output to json

0:13 but ls is pesudocolumns

0:13 heh

0:13 jeremyheiler: tyler__: Welcome to unix :-)

0:13 tyler__: jeremyheiler: which is why i said a clojure os wouldn't be so bad

0:39 technomancy: so... for syme. I want to set up DNS entries for instances so you can do `ssh syme@technomancy.syme.in`

0:39 but the problem is there isn't a 1:1 mapping from usernames to instances

0:39 I could launch a bunch of instances

0:40 so the question is: is it worth subdividing it so that you have to do something like `ssh syme@technomancy.3.syme.in` so they can all be addressed? or should the DNS names only be set up for one instance per user?

0:59 jetkoten: hi all

1:43 yunfan: nice , i am in lispcn-2013-beijing

1:44 people all talk clojure

3:13 xorola: Hi guys, what are plans for improving clojure documentation? does anybody thought/discussed about it?

3:22 john2x: any getting started tutorial for lib-noir? or would webnoir.org's tutorial be viable?

3:45 mjc: or perhaps luminus

5:04 maio: technomancy: is it useful to have multiple active instances?

5:09 technomancy: btw what about project.user.syme.in?

5:32 mpenet: anyone knows how to type a merged map of 2 other possible types in core.typed (if that's possible)?

5:34 ex a merge of (HMap {:a '1}) (HMap {:a '2}) where these 2 are aliased, so if possible avoiding repetitions

5:38 mindbender1: how can I get cljsc to copy my js lib to the public directory and update deps.js(main.js) to point to it rather than point to it's orginal location which is not available to the webserver?

5:39 I'm hoping someone understands

6:59 tdignan: In the clojure repl, how do I enter a tab for purely aesthetic/indentation purposes?

7:12 pepijndevos: tdignan, you can't? Maybe by not using readline? I don't know.

7:39 tdignan: pepijndevos: ah, interesting

7:42 clojure -r works

7:42 but if I run 'clojure' or 'lein repl'

7:42 the tab goes straight to completion

9:12 ravster: morning, all

9:14 skelternet: (send-off ravster (top morning))

9:15 john2x: compojure, ring newbie here. how do I assign a ring handler with compojure's GET macro?

9:17 ravster: john2x: (GET "uri" [] handler)

9:17 I nearly always just send the entire request object to the handler.

9:19 john2x: ohh cool, thanks. I was doing it with (handler).

9:21 hmm but then how do I pass the arguments from a uri like "/user/:id"?

9:21 I'm geting ArityException with (GET "/user/:id" [id] handler)

9:21 ravster: https://github.com/weavejester/compojure/wiki/Routes-In-Detail

9:22 john2x: do you have wrap-params?

9:22 and what are the arguments your handler is expecting?

9:23 john2x: ravster: no wrap-params (yet. not sure when i'd use it). handler is (defn handler [request id] …)

9:39 ah so it get's included in (:params request). cool

9:42 mpenet: ambrosebs: Hi, could you have a look at this gist, it something I didn't manage to express so far. I am probably missing something obvious https://gist.github.com/mpenet/5175786

9:42

9:44 ambrosebs: mpenet: you want to express a type that has an :a but not a :b?

9:45 I guess that would solve it

9:47 mpenet: Something like (U (HMap {:a 1} :without [:b]) (HMap {:b 2} :without [:a]) '{:a 1 :b 2}) would be nice

9:47 The reason I don't have this currently is that this is sort of half assed row polymorphism, which I would like to fully support.

9:49 mpenet: not really, it can have a key :b but if it's the case it must match the contraint

9:49 I will have to wrap this in another type I think

9:49 ambrosebs: Wait, is :b optional?

9:49 mpenet: I would like to only allow {:a 1} {:b 2} or {:a 1, :b 2}, but not {:a 0 :b 4} or {:d 234}

9:50 yes

9:51 the use case is for a korma'isn library, where I want statements (maps) to only allow certain clauses (also maps)

9:52 right now I use an union, it's fine if the user only supply maps with keys that are typed, but not if some unknown key creeps in

9:52 ambrosebs: Yes, it's not possible currently to express this. Specifically the third case is a subtype of '{:a 1} which has no constraint on :b.

9:53 mpenet: right, no big deal, worse case scenario I'll use macros, or just copy paste some code :)

9:54 ambrosebs: mpenet: yes, this is a classic case for row polymorphism.

9:55 mpenet: I have to read about that! Thanks for the help.

9:56 ambrosebs: mpenet: Now that I think about it, this is a fatal flaw of :optional.

9:57 well it's not unsound, but it's not very useful.

9:57 augustl: portland! Gonna be gooood..

9:58 torn: hi there, i've installed the latest java jdk 1.7, and leiningen, and created a new lein project

9:59 when i do 'lein repl' it downloads a bunch of jars into my .m2 repository but then dies with "java.lang.RuntimeException: Unable to resolve symbol: defn in this context2

9:59 -2*

9:59 any ideas?

10:00 mpenet: ambrosebs: I will have to use optional anyway I think, some statements have clauses that are required and some optional, I will have to make some tradeoffs, I have to run now, but I will try this later today/tonight

10:01 torn: I do get a few maven errors - Could not find artifact slingshot:slingshot:pom:0.10.2 in central, cheshire:pom:3.1.0, clojure-complete:jar:0.2.1, cd-client:jar:0.3.4, clj-http:jar:0.3.5

10:01 augustl: torn: you didn't change anything in the output of "lein new"?

10:01 torn: nope nothing changes

10:01 changed*

10:01 ambrosebs: mpenet: sure, be aware :optional keys may not be of the type stated in the :optional map.

10:01 augustl: tomoj: lein -v?

10:02 err, torn

10:02 torn: Leiningen 2.0.0-preview10 on Java 1.7.0_17 Java HotSpot(TM) 64-Bit Server VM

10:02 mpenet: ambrosebs: noted

10:03 torn: i'm thinking maybe maven central doesn't have the jar versions that my version of leiningen wants to download?

10:03 augustl: torn: that's not the latest version, but still weird :)

10:03 torn: I installed it using 'Chocolatey' package manager for windows

10:04 what is the latest version?

10:04 i can try a lein upgraade

10:06 augustl: tomoj: 2.0.0

10:07 torn: without the -preview10 i guess

10:07 hmm gives me a 403 when I try to self-install to 2.0.0

10:10 augustl: I've never used leiningen on windows so I should probably shut up ;)

10:13 torn: hmm ok i've uninstalled chocolately package manager

10:13 and installed lein by hand using the lein.bat file

10:14 cleared out my m2 repo also

10:14 this time 'lein repl' has no dependency issues with maven, but i still get the "Unable to resolve symbol: defn in this context" issue

10:15 augustl: seems like clojure.core isn't loaded properly (wild guess)

10:16 torn: ah it seems the version of lein installed with chocolately didn't make the project properly

10:16 or, aha! maybe calling my project 'clojure' was a bad idea (tm)

10:17 yedi: why is paredit indenting this all wierd: https://gist.github.com/yedi/5176574

10:17 torn: anyone here use Light Table IDE?

10:18 augustl: torn: ah :)

10:19 and leiningen by default creates yourproject.core :)

10:19 torn: yup

10:19 yedi: oh whoops, updated: https://gist.github.com/yedi/5176574

10:19 hyPiRion: torn: yeah, that's fixed by technomancy/leiningen#1017, so that bug will be gone by 2.1.0

10:19 lazybot: lein new should reject projects named "clojure" -- https://github.com/technomancy/leiningen/issues/1017 is closed

10:19 hyPiRion: heh

10:19 augustl: any early birds in Portland wanna meet up for some food?

10:20 torn: closed 21 days ago, i guess the version installed with chocolately didn't have that fix in

10:20 ok repl seems to work

10:20 now I need to get my head around light table, see if i can run a repl in that

10:30 anyone here run lein-light with light table?

10:31 yedi: does anyone have experience setting up RDS with elastic beanstalk

10:31 i'm having issues figuring out how I should be going about it

10:35 slightly related question: will running lein beanstalk deploy update the app on beanstalk? or will it terminate and redeploy?

11:02 torn: oh man light table is pretty neat

11:21 technomancy: maio: heh; missed you last night. so yeah, I think there's really no reason you can't re-use instances; better to manually bootstrap an existing instance for a new project than wait for a new one to boot.

11:22 maio: OTOH, do we enforce this limitation? if not then we have to be clever and fall back to the IP if someone launches a second instance without halting the first.

11:22 the problem with username.project.syme.in is that project is multi-segmented

11:23 so it'd have to be technomancy.technomancy.syme.syme.in which is even more annoying than the IP =)

11:24 Pure-Loulou: .(#((first '(and a b)) % %2) false true)

11:24 technomancy: I guess we could assume each user is not going to hack on multiple forks of a project I guess

11:24 Pure-Loulou: why this returns true? any help?

11:24 clojurebot: A bug report (or other request for help) has three parts: What you did; what you expected to happen; what happened instead. If any of those three are missing, it's awfully hard to help you.

11:25 hyPiRion: Pure-Loulou: '(and a b) is a list with three symbols, and symbols are IFns

11:26 Pure-Loulou: IFns?

11:26 hyPiRion: ,('and false true)

11:26 clojurebot: true

11:26 hyPiRion: ,('and {'and :hello} true)

11:26 clojurebot: :hello

11:26 hyPiRion: So they are like keywords, maps, sets and vectors

11:26 you can call them with functions

11:26 /s/with/as/

11:26 Pure-Loulou: ,(#((eval (first '(and a b))) % %2) false true)

11:26 clojurebot: #<Exception java.lang.Exception: SANBOX DENIED>

11:27 Pure-Loulou: how to wright this so it works?

11:28 hyPiRion: You could do (eval (list (first '(and a b)) false true))

11:28 since and is a macro, you cannot use it as a function itself

11:28 Pure-Loulou: ok:) thx

11:45 thorwil: with a nrepl server provided by immutant, a (use 'midje.repl) in user leads to: FileNotFoundException project.clj (No such file or directory) java.io.FileInputStream.open (FileInputStream.java:-2)

11:45 (trying to follow https://github.com/marick/Midje/wiki/Repl-tools)

11:47 hyPiRion: thorwil: Are you using leiningen for project management?

11:47 thorwil: hyPiRion: yes

11:51 so far i relied on "lein midje" if i wanted to run all tests, otherwise relying on midje-check-fact-near-point

11:51 jonasac: #(fn %) => #(fn &%) is it possible ?

11:52 thorwil: but now "lein midje" fails to resolve a dependency, while "lein immutant run" works fine. so i'm investigating how to run all tests from nrepl

11:54 and fail to find a way to make load-facts available :/

12:01 jcrossley3: thorwil: it's possible that midje is making some assumptions about the current working directory of the jvm

12:02 try 'lein immutant run' after cd'ing to the directory containing your project.clj

12:03 piranha: octagon: hi? :)

12:03 octagon: piranha: hey! how's it going?

12:03 thorwil: jcrossley3: that's already the case. i never run it from any other dir

12:04 piranha: octagon: well, still haven't figured out problem with state not updating in the middle of change

12:04 octagon: that's actually what I wanted to ask, if you had a chance to look at it :))

12:04 jcrossley3: thorwil: hmm. what version of midje? is this a public app i can clone and try?

12:05 octagon: piranha: i did "make" but i didn't get anything but js files

12:05 thorwil: jcrossley3: just updated to midje 1.5.0

12:05 octagon: piranha: i'd like to be able to experiment with the actual app live

12:05 piranha: octagon: if you open 'build/index.html', it opens it in browser

12:05 octagon: piranha: oh ok, cool

12:05 piranha: 'make' is just shortcut for me to not type 'lein cljsbuild auto main'

12:06 thorwil: jcrossley3: https://github.com/thorwil/tlog but i just *now* pushed the changes to project.clj

12:06 octagon: piranha: i think your problem is maybe related to creating cells out of order, which is a weird issue i've run into in my own apps

12:07 piranha: octagon: oh... not sure, I really was a bit lost in traceback there, it's still a bit hard to debug clojurescript

12:07 octagon: piranha: like i've found that especially with formula cells, the dependencies don't get set up correctly if the cells they depend on don't exist when they're created

12:07 piranha: octagon: ah, but in this case it shouldn't be a dependency, it's just assignment to cell somewhere in a function

12:08 octagon: piranha: it's because when you create a formula it's trying to set up the dependency graph ranks, and if dependencies don't exist then weirdness might result

12:08 piranha: hm, doesn't look like a case to me

12:08 maio: technomancy: oh yeah good point about the hostname - i missed that. but yeah I think we could assume that person wouldn't work on different forks

12:20 technomancy: maio: yeah, sounds good. as long as it can fall back to showing the IP when a specific instance has its subdomain deregistered we'd be good

12:20 Glenjamin: Hi guys, does anyone know if there's a non-trivial compojure app that's open-sourced?

12:20 I'm trying to get started but the examples I can find are quite light

12:20 technomancy: Glenjamin: oh yeah absolutely; funny you should ask: https://syme.herokuapp.com

12:20 well... it's still on the simple side. but quite useful =)

12:21 Glenjamin: the biggest OSS compojure app is Clojars

12:22 Glenjamin: i'm mostly trying to figure out how much logic to put in the def-routes definition vs spinning out functions, where to put things, and how to test effectively without running every case end-to-end

12:22 maio: technomancy: btw what are you using these hostnames for? why not just IP?

12:24 technomancy: btw here is the status page I'm working on - http://maio.cz/tmp/syme/Screen%20Shot%202013-03-16%20at%205.23.04%20PM.png

12:44 mye: What's the canonical way to join to longs and get the bigint that represents? Do I have to join up the individual bytes? (New to JVM bit twiddling..)

12:53 Guest76401: hello

12:53 Is it okay if I pose a begginer macro question?

12:54 I want to make a macro that returns a function.

12:54 This is what I've got:

12:55 (defmacro rotate-base [code]

12:55 `(fn [piece]

12:55 (let [s (dec (count piece))]

12:55 (map #(let [[x y] %] ~code) piece))))

12:55

12:55 (rotate-base '(list y (- s x)))

12:55 This does not work. Why?

12:57 momomomomo: Guest76401: That's a terrible way to ask for help haha. Paste your code into a http://gist.github.com/ and explain where you think it might be failing

12:58 xeqi: Guest76401: hmm, looks like what you want could just be a function

12:58 Guest76401: Thank you momo...

12:58 xeqi: well, almost

12:59 Guest76401: xeqi: the reason for macro is that i've got diffrent code snippets

13:01 https://gist.github.com/anonymous/5177314

13:02 tomoj: "explain where you think it might be failing" -- "Can't use qualified name as parameter" ?

13:03 arcatan: Guest76401: you can look at what that expands into with (macroexpand '(rotate-base '(list y (- s x))))

13:03 Guest76401: yes i have tried macroexpand

13:04 (macroexpand-1 '(rotate-base (list y (- s x))))

13:04 (clojure.core/fn [user/piece] (clojure.core/let [user/s (clojure.core/dec (clojure.core/count user/piece))] (clojure

13:04 e/map (fn* [p1__1__2__auto__] (clojure.core/let [[user/x user/y] p1__1__2__auto__] (list y (- s x)))) user/piece)))

13:04 momomomomo: Guest76401: Try editing the gist you've created. It's hard to parse code in an irc client.

13:04 arcatan: Guest76401: piece becomes user/piece, but that's not what you want. you'll want it to have a new unique name i.e. use gensym. rename piece to #piece

13:04 Guest76401: sorry, piece -> piece#

13:05 Guest76401: ok thanks arcatan

13:05 tomoj: basically you're trying to do something evil (unhygienically bind x/y, letting them leak out into the code the macro-user writes) and clojure is preventing you

13:05 arcatan: also listen to tomoj :)

13:05 Guest76401: momo... thank you for the comments i will ask question in a better way in the future

13:06 thanks arcatan and tomoj, will try/read about it

13:06 tomoj: consider writing it as a higher order function instead of a macro

13:06 perhaps: (defn base-rotater [f] (fn [piece] (let [s (dec (count piece))] (map (fn [[x y]] (f s x y)) piece))))

13:07 Guest76401: yes i did higher-order function previously, just thought macro was the way to do things

13:07 tomoj: (base-rotater (fn [s x y] (list y (- s x))))

13:08 SegFaultAX: Guest76401: As a rule of thumb: if it can be written as a function, write it as a function. Try to stave off converting it to a macro unless you're absolutely sure that's what you want.

13:08 I also happen to think tomoj's curried approach is nicer.

13:10 Guest76401: ok

13:11 Im new to macros but have done functional-programming previously, thought macros was the thing to do often.

13:11 I'm finished with this question, I have learned somethings :).

13:12 SegFaultAX: Guest76401: I think most people would agree that macros aren't usually the first thing you should jump to. But sometimes they are absolutely perfect for what you're trying to do.

13:12 Guest76401: ok

13:12 SegFaultAX: Guest76401: While you're learning, it might be useful to adopt the previously mentioned guiding principle of defaulting to functions as much as possible.

13:12 Guest76401: yea

13:12 tomoj: maybe there is a function Duration -> Int which tells you the number of consecutive heads you need to see from a fair coin to say "OK, I really need a macro" given the amount of time you're been learning clojure

13:13 drorbemet: Hi, what does a cross behind a name of a liningen plugin mean exactly?

13:13 Guest76401: tomoj: ok

13:14 SegFaultAX: drorbemet: What do you mean?

13:14 drorbemet: e.g. lein-autotest† Start Lazytest's ... @ https://github.com/technomancy/leiningen/wiki/Plugins

13:15 xeqi: drorbemet: "Plugins marked with † have not been confirmed to work with Leiningen 2." from the top of the page

13:16 SegFaultAX: drorbemet: It says it right at the top

13:16 drorbemet: Ah, thanks :-)

13:16 Now I saw it too ..

13:22 Actually I am looking for the current way to do TDD with automatic execution on change of the source code. There are so many projects to choose from: expectations, Midj, speclj, lein-autotest, lein-autoexpect ... I want to configure Eclipse and Emacs for "autotest".

13:24 I am trying to get expectations and lein-expectations to run with Eclipse under Windows but I still havn't figured out the configuration.

13:24 sorry I meant lein-autoexpect

13:24 clojure-new: Hello, i'm getting "Warning: profile :base not found." with lein from git, how can i fix it?

13:25 hyPiRion: clojure-new: Ignore it for now

13:25 or use 2.0.0 until we've released 2.1.0 :)

13:25 clojure-new: hyPiRion: Ok. :)

13:26 hyPiRion: It should be out by tuesday-wednesday

13:30 lynaghk: drorbemet: the latest version of lein-midje has autotest built into it---you can just run "lein midje :autotest" and you'll get automatic source reloading and test execution

13:32 drorbemet: Oh, that's great! Then I'll take that. Thanks for the hint :-)

13:41 SegFaultAX: lynaghk: I hadn't looked at midje very closely until you brought it up. (I've been using clojure.test up til now). Midje looks pretty sweet.

13:50 ToxicFrog: So, I have a macro with a type annotation on some of its arguments.

13:50 But it will still happily be called on the wrong things.

13:50 What is the appropriate way to add type checking to the macro? Just (assert) that the types are correct?

13:51 mpenet: ambroseb_: is there a way to express a native java array of X (with All) to handle "[LSomething;" or do I need to fallback to Ljava.lang.Object ?

13:53 seems convoluted, I will use Ljava.lang.Object + Any in the Fn body for now

13:55 or better pass something sequential to this function and kill some ugly in the process

13:55 ambroseb_: mpenet: use (Array Something)

13:56 Does LObject actually work?

13:56 mpenet: I didn't try yet, still drafting

13:56 oh I missed Array

13:57 ambroseb_: mpenet: There's (Array wr), (Array2 w r), and (ReadOnlyArray r), which is an alias for (Array2 Nothing r)

13:57 It's a solution to array covariance while preserving static soundness.

13:58 mpenet: nice

13:58 ambroseb_: eg. use ReadOnlyArray when you have a generic action on arrays.

13:58 tomoj: ToxicFrog: what happens when you don't assert and they pass the wrong thing?

13:58 ambroseb_: mpenet: I'm hoping to elaborate on that idea in a paper I'm coauthoring.

13:59 "Hey, array covariance doesn't have to suck"

13:59 OneFourSeven: Ambrose, you were in GSOC 2012?

14:00 ambroseb_: OneFourSeven: Yes.

14:00 mpenet: ambroseb_: for my map issue of earlier, since I never have more than 2 or 3 clauses possible, I think ll just create multiple arities with intersections I think, listing the possible combos

14:03 ambroseb_: mpenet: Ah, yes that's probably the correct strategy.

14:03 mpenet: But the most specific clauses go first. ie. {:a 1 :b 2}

14:03 mpenet: right

14:03 ambroseb_: Good thinking :)

14:03 mpenet: :) learning!

14:04 rebcabin: why doesn't (-> [1 2 3] reverse (partial apply vector)) produce [3 2 1}? instead seems to produce (partial apply vector)

14:04 ambroseb_: If it works out, please post the solution as a response to my google group post.

14:05 mpenet: Will do. I am typing another libary first, it's way smaller and I need to play a bit with All, but I'll go back to the other one later or tomorrow latest.

14:07 bbloom: rebcabin: you're doing (partial (reverse [1 2 3]) apply vector)

14:07 ,(macroexpand '(-> [1 2 3] reverse (partial apply vector))

14:07 )

14:07 clojurebot: #<RuntimeException java.lang.RuntimeException: EOF while reading>

14:08 bbloom: ,(macroexpand '(-> [1 2 3] reverse (partial apply vector)))

14:08 clojurebot: (partial (clojure.core/-> [1 2 3] reverse) apply vector)

14:08 bbloom: rebcabin: try ##(macroexpand '(->> [1 2 3] reverse (partial apply vector)))

14:08 lazybot: ⇒ (partial apply vector (clojure.core/->> [1 2 3] reverse))

14:09 rebcabin: ahh, i see

14:10 thanks bbloom and clojurebot

14:11 bbloom: rebcabin: lazybot helped out too :-P i can never remember which meta characters ,&## go to which bot

14:12 rebcabin: (let [f (partial apply vector)] (-> [1 2 3] reverse f)) ; closer to my erroneous expectation

14:12 ambroseb_: If anyone missed it, there's a new core.typed mailing list https://groups.google.com/forum/?fromgroups#!forum/clojure-core-typed

14:20 tyler__: ohpauleez: you're a portland guy?!

14:20 ohpauleez: tyler__: I am now, mmhmm

14:21 Previously I've lived in Boston, Philly, NYC, San Francisco and Eugene, OR

14:21 I grew up in NH

14:21 tyler__: ohpauleez: oh nice we should have you hang out sometime. I work at Little Bird out of PIE we are converting to a clojure stack

14:21 ohpauleez: tyler__: I'm a PIE mentor

14:21 tyler__: ohpauleez: nice

14:21 ohpauleez: always more than happy to help with a Clojure migration

14:21 :)

14:22 I really enjoyed the Little Bird demo during demo day

14:22 tyler__: thnx, heh yeah Marshall knows how to give a presentation

14:22 ohpauleez: tyler__: You guys were in Rails before? Or Python?

14:23 tyler__: ohpauleez: ruby, not rails. i hate rails

14:23 we are still half ruby

14:23 we have a services model with a message queue

14:23 so its about half ruby services, half clojure services now

14:23 ohpauleez: tyler__: Ah yes, very common. That's what I was just about to say

14:23 tyler__: maybe still more ruby than clojure

14:24 ohpauleez: we should totally meet up for beer/coffee. And definitely let me know if you guys need any advice on anything

14:24 I've done the startup thing for awhile now ;)

14:25 tyler__: ohpauleez: you going to clojurewest? (if going? (meet you me))

14:26 ohpauleez: tyler__: I'll be there; Speaking Monday just before Rich

14:26 tyler__: woot

14:26 ohpauleez: http://clojurewest.org/sessions#degrandis

14:29 ToxicFrog: tomoj: the macro runs just fine and I get really confusing errors at runtime because someone passed in a string rather than a regex.

14:30 torn: i'm teaching myself clojure, current playing around with namespaces

14:30 is (str (ns-name *ns*)) a common thing to do?

14:30 or is there a shorthand to get the name of the current namespace, for when I'm creating namespaced symbols or keywords

14:34 amalloy: torn: not very common, no. namespaced symbols are very rarely needed, and namespaced keywords are fairly rare as well as being easier to get another way

14:34 tyler__: ,(namespace ::foo)

14:34 clojurebot: "sandbox"

14:34 amalloy: try: ::foo, for the keyword foo in the current namespace

14:37 tyler__: ,(dir sandbox)

14:37 clojurebot: nil

14:37 tyler__: heh

14:38 TimMc: ,(-> dir var meta :added)

14:38 clojurebot: nil

14:40 torn: yep i saw ::foo does the current ns

14:40 i guess there's no equivalent for ''foo

14:40 I'd have to (symbol (str (ns-name *ns*)) "foo")

14:41 mindbender1: I've written a js source that extends closure and I want to make the compiler copy this file to :output-dir as it does with cljs source and jarred sources

14:41 torn: although why i'd want to, i have no idea -- I don't really know what symbols are used for yet - am guessing dynamic evaluation or something

14:41 mindbender1: is this currently possible?

14:44 tomoj: ToxicFrog: then, yeah. normally in clojure you just let the user figure out the confusing errors, because they wrote incorrect code so it's their fault :). but asserting in some cases to make debugging easier is appropriate. especially at macro-time since you don't really have to worry about slowing things down with checks

14:45 tyler__: ohpauleez: we are considering utilizing clojurescript but one of our concerns that that, thus far, our frotn end guys have been mostly html css wizards with no real CompSci concepts. So if we migrate to clojurescript we are feaful we might lock ourselves out of good front end people

14:45 s/that that/is that/

14:46 * tyler__ can't type today

14:46 ohpauleez: tyler__: That hasn't been my experience nor Kevin Lynagh's experience

14:47 mindbender1: tyler__: it doesn'tstop your frontend guys from doing their painting job

14:47 ohpauleez: great front end people still cook up baller HTML and CSS - CLJS frees them from having to muck with JS's idiosyncrasies and allows engineers to more easily jump in on the client side

14:47 tyler__: they do a lot of the app logic stuff too

14:47 to animate the dom and such

14:47 ohpauleez: I think you'd find you'd start questioning more where things get done

14:47 tyler__: ah

14:48 tomoj: meaning, what, we get more work and the frontend guys get less?

14:48 ohpauleez: CLJS also doesn't exclude the use of JS. I think more people just realize it's easier to express what they want in CLJS once it's in place

14:48 tomoj: s/guys/people/

14:48 tyler__: (inc tomoj)

14:48 lazybot: ⇒ 6

14:48 ToxicFrog: tomoj: yeah, that's what I'm after here - I'd rather have a compile-time error because they used the macro incorrectly than get a runtime error much, much later with an unclear origin

14:49 ohpauleez: tyler__: all in all, it comes down to your team, your culture, and the goals of your project

14:49 Clojure and CLJS aren't always the best tools for the job - but they are pretty amazing tools

14:50 antares_: Neocons 1.1 is released: http://blog.clojurewerkz.org/blog/2013/03/16/neocons-1-dot-1-0-is-released/, plans for the future: http://blog.clojurewerkz.org/blog/2013/03/16/the-future-of-neocons/

14:50 tyler__: if i can mix cljs and vanilla js pretty easilly that could solve our dilemma

14:51 wjlroe: does anyone know if there's an equivalent of lein-ring (with namespace reloading in development) for aleph?

14:51 torn: anyone here find Light Table's line selection / instarepl cursor stuff messes up sometimes?

14:51 mindbender1: I find cljs is best suited for data manipulation on the client

14:52 ohpauleez: tyler__: It's very easy to mix

14:52 mindbender1: the view is best left for an oo framework like closure

14:52 ohpauleez: the one caveat is advanced compilation - if you use it on your CLJS, you're going to need an externs file

14:52 tyler__: mindbender1: i can see how a clojure system could be good for data manipulation ;)

14:53 tomoj: mindbender1: something about the view is inherently suited to an oo framework like closure? or just cljs view libraries aren't there yet?

14:54 mindbender1: even Rich Hickey said that

14:54 tomoj: the former?

14:54 mindbender1: you can't touch a mouse so there's an object

14:55 torn: does idiomatic clojure use camelCase or underscores or dashes for variable / function naming?

14:55 tomoj: where'd rich talk about this?

14:55 wjlroe: torn: dashes

14:55 mindbender1: simplemadeeasy

14:55 tomoj: thanks

14:55 ohpauleez: mindbender1: I'm skeptical of the claims you're making - it is *very* dependent on the type of application you are creating

14:55 tomoj: seems obvious you need identities, obviously wrong (to me) that OO is well-suited

14:56 inherently

14:56 torn: wjlroe: ty - even for function names?

14:56 wjlroe: torn: yup

14:56 mindbender1: I'm not. you only need to show me solid examples

14:56 tyler__: torn: dash all the things

14:56 ohpauleez: ie: perhaps you want to make transitions on the UI, like you would make transitions in data. In which case a Pub/Sub // pipeline // event system is what you want

14:56 egghead: just made it into portland for clj-west :)

14:56 torn: having done a lot of js recently, dashes in things scare me haha

14:56 egghead: lol torn

14:56 ohpauleez: and surely functional approaches to those systems are concise and easy to use

14:56 torn: but i guess dashes by themselves aren't s-expressions so it's all good

14:57 tyler__: having done a lot of js lately, js scares me

14:57 egghead: torn: that's the problem with infix!

14:57 torn: meh js isn't that bad once you get around the prototypal inheritance and closures and function invocation

14:57 egghead: did you mean foo minus bar or to have a ref named foo-bar :)

14:57 mindbender1: I find I just have to embrace both ways of thinking

14:57 ohpauleez: egghead: Welcome to PDX!

14:57 egghead: cheers ohpauleez, using the heck out of your guide

14:58 ohpauleez: egghead: Please do. Everyone should feel free to ping me with questions, requests, etc

14:58 tyler__: egghead: welcome to town

14:58 ohpauleez: damn u beat me

14:58 ohpauleez: I literally live a stone's throw from the venue

14:58 Just me and Fred Armisen, drinking beers

14:58 torn: PDX?

14:58 tyler__: ohpauleez: did you see them filiming grim there last week?

14:58 ohpauleez: torn: Portland, OR

14:59 tyler__: y i no can type

14:59 ohpauleez: tyler__: They shoot Grim all over the place, and it totally messes everything all up haha

14:59 they shut down my entire block like a month ago

14:59 tyler__: heh

14:59 ohpauleez: which ultimately meant, my dog had no where to shit

14:59 haha

15:21 clojure-new: (sql/with-connection (db-connection) (sql/with-query-results results ["select * from bla"] results))

15:21 This code produces: java.sql.SQLException: You can't operate on a closed ResultSet!!!

15:22 clojure.java.jdbc -- sql

15:22 What's wrong with it?

15:22 gfredericks: clojure-new: with-query-results

15:22 it expects you to fully process the resultset before returning

15:22 vec is the easiest way to do that

15:23 so return (vec results) instead of results

15:23 results is a lazy sequence that reads from the ResultSet as needed

15:23 clojure-new: gfredericks: Yay, thanks!

15:23 It works now.

15:23 gfredericks: so you can't realize it out of the w-q-r scope

15:24 antares_: clojure-new: if you need some background on that lazy sequences stuff: http://clojure-doc.org/articles/language/laziness.html

15:24 clojure-new: antares_: Thank you, i'll read it.

15:33 tomoj: is query the replacement for deprecated with-query-results?

15:38 gfredericks: oh wqr is deprecated too? what are the resultset-openness semantics of the new function?

15:54 mpenet: jimduey: Did you use TFn already? I am trying to create an heterogenous type "fn" for Sequential, failing so far

16:12 jimduey: Not yet. I'm off trying to figure co vs contra-variant.

16:12 mpenet: ^

16:13 mpenet: jimduey: I think I may have found a bug, or that could just be me misusing it.

16:13 tyler__: how do i do something like '(+ 1 num) but actually interpolate the value of num into there?

16:16 hyPiRion: tyler__: `(+ 1 ~num)

16:16 ,(let [num 10] `(+ 1 ~num))

16:16 clojurebot: (clojure.core/+ 1 10)

16:16 tyler__: hyPiRion: thanks, still getting the macro-y stuff down

16:16 hyPiRion: Well, that will namespace the symbols too.

16:28 tyler__: is parallelism directly proportional to concurrency on the JVM since threads run on different cores?

16:34 ohpauleez: tyler__: Not necessarily

16:34 bbloom: shall we do this again? ^^

16:34 bbloom: ohpauleez: heh. dammit. i guess i need to write that blog post

16:35 tyler__: bbloom: that sounds like a good blog post

16:35 ohpauleez: We just need to create a blog: "Things we've said in #clojure that need to be properly archived"

16:35 tyler__: speaking of which i just wrote: http://tyler808.wordpress.com/2013/03/16/its-time-for-a-new-unix/

16:35 ohpauleez: tyler__: Since it seems like you have a general idea, I'll give you the quick and dirty

16:36 bbloom: tyler__: http://clojure-log.n01se.net/date/2013-02-26.html#16:17

16:36 tyler__: irclogasblog.tumblr.com

16:36 heh

16:36 that actually is a proper blog: "web log"

16:37 literally

16:37 ohpauleez: concurrency is just handling/doing more than one thing at the same time, potentially swapping between things. Parallelism is *actually* doing multiple things at exactly the same time. In the most precise definition, it comes in two flavors: data parallelism and task parallelism.

16:38 hyPiRion: http://clojure-log.n01se.net/date/2012-12-17.html#22:30 <- my proudest moment in this channel

16:38 ohpauleez: dp - you're doing a single task, in parallel across some chunk of data

16:38 that's the most common

16:38 tp - you're doing a bunch of different tasks in parallel, most likely on the same pool of data

16:38 llasram: hyPiRion: It was something to be proud of :-)

16:38 ohpauleez: then there are some analogies: dinner parties, schools and classrooms

16:39 hyPiRion: hehe

16:39 tyler__: i get the difference between the two, i was asking does the amount of parallelism go up as concurrency goes up

16:39 ohpauleez: on the JVM, you could do async concurrency, of you could use OS-level threads. Assuming you have an SMP kernel, those threads will hit different cores, but that's not always a promise

16:39 tyler__: or you you top out at a certain point

16:40 ohpauleez: oh no, the JCM will hit all cores if it's able to

16:40 JVM

16:40 tyler__: like if i have 4 cores and 5 threads is the fifth thread considered "parallel"?

16:40 bbloom: dammit ohpauleez! stop bringing up data vs task parallelism! didn't we agree that this was orthogonal!?

16:40 ohpauleez: hahaha

16:40 antares_: tyler__: only if your CPU supports hyperthreading (core i7 all do, I think)

16:40 ohpauleez: I still feel it's important

16:41 tyler__: core+virtual cores are a hard limit on true parallelism for anything

16:41 antares_: tyler__: if not, then no. But you are bound by hardware, not JVM limitations most of the time.

16:41 ohpauleez: so you can have as many concurrent things going as you like

16:41 but you are capped on true parallel things

16:42 antares_: hey ohpauleez

16:42 ohpauleez: bbiab

16:43 tyler__: so why even discuss parallelism vs concurency if parallelism is just a function of concurrency + # of cores

16:44 ohpauleez: tyler__: Because we have distributed systems, GPUs, etc

16:44 antares_: tyler__: for example, because not all runtimes provide parallelism (CRuby, CPython do not) and not all of them do it equally well

16:44 tyler__: yeah but we all use java here amiright? ;)

16:44 antares_: GPUs are also a factor, too. Currently JVM won't use GPUs but there is a project that (hopefully) will make that possible in the next few years.

16:45 tyler__: antares_: ah, gpu is like a cpu on your graphics card right? im not much of a hardware guy

16:45 antares_: tyler__: more or less

16:46 bbloom: hyPiRion: nice.

16:47 gfredericks: how's the weather in portland?

16:47 tyler__: hyPiRion: wow thats amazing, i am not worthy

16:48 gfredericks: its been nice, a little overcast today but its been sunny

16:48 its finally getting warmer

16:48 hyPiRion: tyler__: It's not so difficult, really. It's just a lot of trial and error

16:48 You should've seen my history with lazybot before I made it work

16:48 gfredericks: tyler__: man these forecasted temperatures are so high...

16:49 tyler__: actually next week looks kinda shite according to google now

16:50 gfredericks: shite == wet?

16:50 tyler__: and colder

16:50 gfredericks: it's 30 here

16:50 tyler__: last week we had the door open because it was so hot

16:51 gfredericks: well now I know not to pack my scarf

16:51 hyPiRion: tyler__: wow, you Americans and your weather

16:51 tyler__: hyPiRion: how is weather an american thing?

16:52 hyPiRion: tyler__: Mostly because the only time I hear about weather from another country, it's America :]

16:52 tyler__: correlation not causation ;)

16:53 hyPiRion: tyler__: Well, I meant it as in "Dang, you guys are lucky with the weather", not as in "talking about weather is an American thing"

16:53 gfredericks: hyPiRion: our lake just thawed

16:54 tyler__: hyPiRion: aaaah i understand

16:54 hyPiRion: gfredericks: Where are you from?

16:54 tyler__: hyPiRion: "american weather" is not a thing we have a huge landmass with hugely different climates

16:55 last week i learned that all of the UK is about the size of Oregon heh

16:55 gfredericks: hyPiRion: Chicago

16:55 hyPiRion: tyler__: yeah, I know. I'm extrapolating based on one datapoint :p

16:56 * gfredericks tries to cram the UK into lake michigan

16:56 clojure-new: We have weather in Russia too! Sometimes it's a pain.

16:56 -7C ATM.

16:56 gfredericks: clojure-new: is it a single kind of weather?

16:56 in all of russia?

16:57 bbloom: tyler__: i just checked wikipedia. oregon is slightly larger, heh

16:57 amalloy: gfredericks: in small countries like that, that's not uncommon

16:57 antares_: gfredericks: standardized by the government, too

16:57 gfredericks: haha

16:57 bbloom: UK: 243,610 km^2 vs OR: 255,026 km^2

16:58 hyPiRion: gfredericks: Wow, I just realized Chicago is on about the same latitude as Madrid

16:58 gfredericks: europe is falsly north

16:58 clojure-new: bbloom: RUSSIA: 17,075,400 km2 vs UK: 243,610 km^2 vs OR: 255,026 km^2

16:59 17,075,400 km2 of weather.

16:59 tyler__: russia is a country though, not a provence/state

16:59 gfredericks: all of it the same

16:59 and all of it in metric

16:59 tyler__: is the US the last place to refuse to use metric?

17:00 antares_: tyler__: not at all. it is accompanied by burma, see.

17:00 gfredericks: a coalition of the willing

17:01 * drewc is just north of USA, uses metric, and lives in the second largest country, so weather is interesting indeed

17:01 gfredericks: man I'm already old enough to be muttering political innuendo that the youth will not understand

17:02 * technomancy refuses to refuse to use metric

17:02 amalloy: tyler__: there are like three other countries, with a combined GDP of like $10/year

17:02 tyler__: heh

17:02 amalloy: as i recall

17:03 * drewc aslo lives on a sailboat, so uses proper nautical miles, fathoms, cables, etc etc

17:03 * gfredericks invents improper nautical miles

17:03 drewc: and of course, the speed is measured in knots

17:03 * arrdem is young and foolish enough to use metric. also enjoys saying "click"

17:04 hyPiRion: Oh, I like that

17:04 "three clicks from here"

17:04 gfredericks: man I have no idea how we measure the speed of my apartment

17:04 drewc: gfredericks: I think they have already been invented, they are called 'miles'

17:04 gfredericks: will have to ask the landlord

17:05 technomancy: gfredericks: it's called a flat

17:05 gfredericks: technomancy: I guess that helps it be more aerodynamic?

17:05 technomancy: now you're catching on

17:05 drewc: (or often "land mile" or even "statute mile")

17:05 tyler__: gfredericks: around 1125kph

17:05 gfredericks: http://geography.about.com/library/faq/blqzearthspin.htm

17:05 heh

17:05 clojurebot: No entiendo

17:06 Frozenlock: technomancy: as a mechanical engineer, I want to express how much I hate you for not using metric.

17:06 gfredericks: I wonder if velocity is a torsor

17:06 arrdem: (inc Frozenlock)

17:06 lazybot: ⇒ 2

17:06 * gfredericks just discovered torsors last week

17:06 hyPiRion: Frozenlock: You read technomancy wrong :)

17:06 * tyler__ just discovered torsors today

17:06 technomancy: Frozenlock: what? I use metric

17:06 hyPiRion: "refuses to refuse to use metric" -> metric

17:06 Frozenlock: Oh

17:06 damn

17:06 gfredericks: hyPiRion: well that's a bit of a stretch

17:06 hyPiRion: gfredericks: true

17:06 gfredericks: hyPiRion: he might just be wishy washy

17:06 Frozenlock: Double negative got me -_-

17:07 technomancy: only the language nerds are allowed to hate me (for the double negative)

17:07 gfredericks: technomancy: only fake language nerds would complain about a double negative

17:07 bbloom: technomancy: as an english major (not really) i what to express how much i hate you for using a double negative

17:07 tyler__: i don't know nothing

17:07 bbloom: damn, typed that too slow to make the joke in time

17:07 slowly

17:07 gfredericks: I think I remember seeing a quadruple negative in the wild once

17:07 Frozenlock: bbloom: still found it funny :)

17:07 * bbloom is not an english major

17:07 technomancy: http://wondermark.com/434/

17:08 tyler__: i luled

17:09 * gfredericks finally adds wondermark to his google reader, which is a cool new way to keep up with your favorite blogs and sites

17:11 arrdem: gfredericks: i c u trollin

17:11 * arrdem is using ring and ab to benchmark his PC and feels bad

17:11 Frozenlock: While we are on the news, I read somewhere that EA's lastest game was awesome, and the always-online requirement wasn't a problem, at all.

17:12 tyler__: so i decided to take a peek at core.logic and now my brain is melting...

17:12 Frozenlock: are you being facetious?

17:13 arrdem: ,(str "(" (rand-nth ["inc" "dec"]) " Frozenlock)")

17:13 clojurebot: "(inc Frozenlock)"

17:13 arrdem: ,(print ",(" (rand-nth ["inc" "dec"]) " Frozenlock)")

17:13 clojurebot: ,( dec Frozenlock)

17:13 AimHere: tyler__, no. He was just reading a report of an EA press release

17:13 arrdem: bother

17:14 * Frozenlock never forgave EA for ruining Ultima Online

17:14 arrdem: ,(print (str ",(" (rand-nth ["inc" "dec"]) " Frozenlock)"))

17:14 clojurebot: ,(dec Frozenlock)

17:14 * arrdem gives up

17:14 Frozenlock: You mean the 'straight answers from Lucy' one?

17:14 AimHere: &(print (str ",(" (rand-nth ["inc" "dec"]) " Frozenlock)"))

17:14 lazybot: ⇒ ,(dec Frozenlock)nil

17:14 AimHere: ,(print (str "&(" (rand-nth ["inc" "dec"]) " Frozenlock)"))

17:14 clojurebot: &(dec Frozenlock)

17:15 arrdem: so do we have to quine thi?

17:15 *this

17:15 clojurebot -> lazybot -> clojurebot?

17:15 AimHere: Lazybot prefixing it's output with the arrow makes me glum

17:16 But it's probably for the best, all things considered

17:16 * arrdem contemplates a quine which will alternate between Lazybot and Clojurebot

17:19 tyler__: woah just found: https://github.com/jonase/kibit/

17:19 that looks awesome

17:20 hyPiRion: arrdem: that's impossible by now. Clojurebot ignores lazybot now

17:20 I think

17:20 $timer 0:0:1 ,(println "foo")

17:20 lazybot: Timer added.

17:20 ,(println "foo")

17:20 clojurebot: foo\n

17:21 tyler__: or not

17:21 hyPiRion: ,(print "&(println "foo")")

17:21 clojurebot: #<CompilerException java.lang.RuntimeException: Unable to resolve symbol: foo in this context, compiling:(NO_SOURCE_PATH:0:0)>

17:21 hyPiRion: ,(print "&(println \"foo\")")

17:21 clojurebot: &(println "foo")

17:21 hyPiRion: the other way around it was

17:21 arrdem: Q_Q

17:24 Frozenlock: Anyone here using compojure could confirm if googlebot can access their site?

17:24 MikeSeth: if I do (defrecord Character ...) Clojure complains about java.lang.Character already defined. Is this because java.lang classes are imported into every namespace?

17:25 arrdem: Frozenlock: confirmed

17:26 Frozenlock: Grrr Googlebot, why u fail on me...

17:29 antares_: MikeSeth: most likely

17:30 tomoj: (ns-unmap *ns* 'Character)

17:30 I can't remember why I wanted :import-exclude

17:30 didn't think of that workaround till after I forgot :(

17:31 MikeSeth: tomoj: is this a "normal" thing to do?

17:31 tomoj: I guess (:import-java :exclude [...])

17:31 (that's hypothetical)

17:31 no, it's not normal

17:31 MikeSeth: hmmm

17:32 tomoj: it's rare enough for people to alter their ns map in any other way besides ns and def

17:32 but if you really want it to be called Character, doing that at the beginning of your ns top-level wouldn't cause much real trouble afaik

17:33 hmm

17:33 though that will unmap the record every time after

18:01 hiredman: /win 15

18:04 bbloom: lynaghk: i think you're gonna like what i'm cooking up in the rewrite department :-)

18:11 lynaghk: bbloom: I'd love to see it over a beer

18:11 bbloom: and then not over a beer, to make sure I can understand =P

18:12 bbloom: lynaghk: i'm basically porting https://code.google.com/p/kiama/wiki/Rewriting to clojure, but with a focus on operating on trees of maps

18:12 lynaghk: it's basically a combinator library for rewriting strategies

18:35 dnolen: i too keep finding myself wanting reduce-until

18:35 ambrosebs: mpenet: need help with TFn?

18:36 wmax: how do i retrieve all the attributes in datomic?

18:36 bbloom: dnolen: before, i've wanted left-fold-until, but now i have a case where i want fold-with-an-abort-condition, ie un-ordered

18:36 jjttjj: anyone ever use fleet for templating? are there any example apps out there?

18:39 ohpauleez: bbloom: You know you can do that right now, right?

18:39 reduce supports a short circuit

18:39 bbloom: ohpauleez: how?

18:39 hiredman: ,(doc reduced)

18:39 clojurebot: "([x]); Wraps x in a way such that a reduce will terminate with the value x"

18:39 bbloom: bwha?

18:39 1.5?

18:39 ohpauleez: yeah

18:40 bbloom: glorious.

18:40 ohpauleez: let me dig up the example I just saw

18:42 mpenet: ambrosebs: I think I understand TFn, but this was not really the solution to my problem (still good to have learned about it). I sent you a privmsg earlier with a gist

18:43 ambrosebs: either I misunderstand Seq* or there is a bug somewhere, more likely the former

18:44 bbloom: ohpauleez: oh. this is amazing.

18:45 ohpauleez: bbloom: It's reduce right? I was trying to dig up the code I had

18:47 bbloom: ,(reduce (fn [acc x] (let [x' (* x x)] (if (> x' 10) (reduced x') x'))) (range))

18:47 clojurebot: 16

18:47 bbloom: ohpauleez: there's an example

18:47 ohpauleez: Ahhh there we go!

18:47 bbloom: ohpauleez: sadly, it doesn't seem to work with `reductions :-(

18:48 ohpauleez: reduced! right

18:48 it is indeed a pity

18:48 bbloom: ohpauleez: i am sooo submitting a patch to reductions....

18:48 tomoj: we still need a reductions reducer too

18:49 ohpauleez: tomoj: What would that even look like?!

18:49 haha

18:49 bbloom: well, if the reducers produces a linear reduction, it's easy, but otherwise, it could return a tree

18:50 ohpauleez: right

18:50 I was just thinking, well… you could write it in terms of the current reduce and build a seq

18:50 tomoj: no

18:50 no seqs :)

18:51 I dunno what you mean probably

18:52 what would return a tree?

18:52 ohpauleez: reductions in reducers, that runs non-linearly

18:52 because order isn't preserved

18:52 Frozenlock: Ok googlebot is definitely trolling me. Every proxies I tried can see the site, in addition to web-sniffer.net, but not googlebot -_-

18:52 ohpauleez: which you typically want in reductions

18:53 tomoj: what reducer doesn't preserve order?

18:54 to me that is pretty much the essence of a reducer, an order

18:54 ohpauleez: tomoj: how would you visual the results of reduction where the first three elements were reduced and the last two were, but no others

18:54 visualize**

18:55 … I suppose I'm thinking of fold

18:55 tomoj: ah, right, fold reductions is weirder

18:55 folds?

18:56 presumably you pass a combinef and reducef and you get the ordered results from the combinef as a reducer?

18:57 can't we move the short-circuiting code into Reduced?

18:58 instead of writing it into every reducer

18:59 perhaps not without CPS'd reducers?

19:07 wmax: how do i get the namespace of a namespace qualified keyword

19:09 bbloom: http://dev.clojure.org/jira/browse/CLJ-1185 boom.

19:09 devn: bam

19:09 * devn gives bbloom a mic to drop

19:10 technomancy: arg; I hate it when I realize two seconds too late that I could have used paredit-convolute-sexp

19:10 devn: what does that do?

19:11 technomancy: words cannot describe it

19:11 devn: haha

19:11 oh i see

19:12 that's uh, very specific

19:12 wmax: why don't anyone help me

19:12 &(namespace :foo/bar)

19:12 lazybot: ⇒ "foo"

19:14 devn: wmax: isn't that what you want?

19:15 tomoj: it's very common to have seq forms where some initial subseq is distinguished and provides context for the remaining subforms

19:15 wmax: that is exactly what i want

19:15 tomoj: paredit-convolute-sexp just reorders those kinds of operators

19:15 devn: wmax: :)

19:15 &*clojure-version*

19:15 lazybot: ⇒ {:major 1, :minor 4, :incremental 0, :qualifier nil}

19:16 Frozenlock: arrdem: May I ask if you are using some middleware? force-ssl, gzip, etc..?

19:17 arrdem: Frozenlock: it

19:17 's been a long time and that blog's codebase is still on noir.. lemme look tho

19:18 Frozenlock: yeah stock noir opaque as always

19:21 bbloom: ohpauleez: thanks much for letting me know about `reduced... now i wonder what else i'm missing in 1.5

19:22 ohpauleez: bbloom: totally welcome man

19:23 Frozenlock: arrdem: I just moved away from noir, thinking it might be what caused googlebot to choke :/

19:23 * bbloom goes to work changing some ugly loop/recur form into reduce and reduce-kv forms

19:24 gfredericks: tomoj: what's a seq form?

19:24 arrdem: Frozenlock: heh yeah I have a branch that's noir-free but it's stuck in me being too lazy to fix all the mistakes I made with my first Clojure project XP

19:24 tomoj: a form that is a seq

19:24 probably a list

19:24 gfredericks: like (if foo bar baz)?

19:24 tomoj: yeah

19:26 bbloom: looking at the impl of clojure.lang.Reduced, it occurs to me that IDeref is like an abstract pointer on steroids

19:27 tomoj: steroids?

19:27 bbloom: i wish that there was a way to opt-in to auto-deref like vars, but still allow do-not-deref via something like #'

19:27 but i'm not gonna think about that too much :-P

19:29 tomoj: auto-deref on invocation? or something crazier?

19:29 gfredericks: auto-deref like vars

19:29 bbloom: tomoj: on invocation is easy: implement IFn

19:29 gfredericks: you mention them and they get deref'd

19:30 tomoj: yeah, so what are you thinking? (foo bar baz) auto-derefs bar?

19:30 bbloom: yeah what gfredericks says, but it requires static knowledge

19:30 tomoj: oh, right

19:30 bbloom: the compiler knows that a name refers to a var. so for example of you refer to `inc it emits clojure.core.inc.deref()

19:30 tomoj: IFn is the "oh, actually do deref" case for vars

19:30 bbloom: but if you #'inc, then you get clojure.core.inc without the deref()

19:31 gfredericks: tomoj: wat?

19:31 bbloom: but i wish there was like a metadata hint you could put on symbols: (assert (= 5 (let [x ^:deref (delay 5)] x)))

19:31 tomoj: gfredericks: nevermind, I was confused

19:31 bbloom: or maybe ^:deref x

19:31 and then you could get at the delay by writing: #'x

19:32 gfredericks: bbloom: for local stuff symbol macros would solve it, right?

19:32 bbloom: gfredericks: hm, i guess it would

19:32 dnolen: bbloom: you already in portland?

19:33 bbloom: dnolen: no, i get in around 8pm pacific tomorrow

19:33 how far is the airport from the hotel / arcade? I'll head right over there

19:33 dnolen: bbloom: cool, I get there around 7:30

19:34 bbloom: gfredericks: the reason this thought about requesting auto-deref occurs to me is b/c it would enable some lazy evaluation DSLs :-)(

19:35 gfredericks: bbloom: I think I've thought of that sort of thing as well. would definitely make code more potentially confusing though

19:35 bbloom: but i guess it can be done w/ symbol macrolets pretty straightforwardly.... so maybe ill look into that if i ever need to do something extra lazy

19:37 OneFourSeven: Do I have to worry about sql injections if I use sqlkorma?

19:40 lynaghk: bbloom: you can take the max from the airport pretty much directly to the arcade downtown---it'll be about a 40 minute ride or so.

19:40 (the max = light rail)

19:43 tomoj: bbloom: how about deref destructuring instead

19:43 (let [xd (delay 5) @x xd] ..)

19:43 hmm no room for :as

19:46 (let [(:as @x xd) (delay 5)]) ? :(

19:47 er, no..

19:59 amalloy: tomoj: what about (let [^:delay x (some computation)] (if whatever (inc x) 0))? i have that written somewhere

20:00 tomoj: I think I prefer [@x (some computation)]

20:00 but either way, no :as

20:02 what would (if-let [^:delay x y]) mean?

20:03 hyPiRion: tomoj: what are you attempting?

20:03 Seems like you're trying to reinvent dataflow variables.

20:03 tomoj: nothing very closely related to this discussion

20:04 but yes, I am trying to reinvent dataflow variables :)

20:06 wmax: how do i html escape url strings?

20:21 tomoj: are there reducers which can't be CPS'd?

20:22 assume the continuation will only be called once

20:22 (or zero times)

21:26 xeqi: non-clojure, but any recommendations on places to get an ssl cert?

21:34 callahad: xeqi: Check with your registrar? Lots provide them free with registrations (or at least, gandi.net does…)

21:34 bbloom: i have forgotten how anyone gets anything done in the presence of operator overloading....

21:35 this scala library defines <+ and <* to mean subtly different things

21:35 i have had to look up their meaning to differentiate them EVERY SINGLE TIME I ENCOUNTER THEM. which has been dozens of times in the past hour. i simply can not remember them b/c they have no pronounceable name or provide any hint to their meaning

21:38 tomoj: "+" and "*" had better be a hint

21:38 bbloom: <+ means "composition via deterministic choice" and <* means "unconditional linear composition"

21:39 tomoj: if you have a nemonic, i'm all ears

21:39 hiredman: porting python code that mixes math and string concatenation, :/

21:39 egghead: freakin' scala

21:39 hi we always eval left to right except when the operator starts with two colons

21:39 don't ask why

21:40 why does odersky hate programmers

21:40 bbloom: heh

21:40 tomoj: bbloom: + means "or", * means "and", dunno what those descriptions mean but it looks like it lines up OK

21:41 bbloom: <+ is like clojure.core/cond, <* is like clojure.core/comp

21:41 i'm calling them choice and pipe

21:42 egghead: scala, where fn references aren't real references unless you partially apply them to some magical thing named _

21:42 bbloom: ALL HAIL THE GREAT _

21:44 SegFaultAX: egghead: Actually, you only need one colon to change the associativity.

21:44 egghead: oh? didn't know that

21:45 i thought for instance a :: b was a method of a but a ::: b was a method of b

21:45 SegFaultAX: egghead: The folding operators are a good example

21:45 john2x: is there a utility method to create ring requests? can't seem to find it in ring's api page.

21:45 egghead: john2x: ring mock is nice

21:46 bbloom: at least with mathematica you can FullForm[Hold[...]] to debug associativity

21:46 with haskell & scala, you just need to wait for the compiler to give you a type error, or wait until you have a bug in production

21:47 john2x: egghead: cool thanks

21:47 SegFaultAX: bbloom: The #1 #2 and #3 things I dislike about Scala are how fricken complicated the syntax is.

21:47 bbloom: I have a really hard time getting used to it.

21:48 egghead: SegFaultAX: ya scala seems very context sensitive

21:48 stuff like if you call a method with or without parens, might work in some places but not others

21:49 bbloom: i have nothing fundamentally against infix operators, it's just that they add significant notation complexity which is extremely unwelcome while you are learning concepts. all operators should be required to map 1-to-1 with a named prefix operator/function/macro/whatever & then i should be able to OPT IN to operators as i desire the notation to tackle bigger and bigger problems with the semantics i need

21:49 SegFaultAX: egghead: Eh, that's something else. That's really more a matter of "does this class I'm working with observe the uniform access principle"

21:49 bbloom: mathematica gets that half right: you need to use FullForm to opt OUT

21:49 SegFaultAX: egghead: Most do, some don't.

21:49 egghead: uniform access principle, ya

21:50 I don't even want to even know that that is a thing but when writing scala you have to

21:50 SegFaultAX: if that's not the syntax would you say it's the semantics?

21:50 SegFaultAX: egghead: Well, omitting the parens is a nice to have when accessing properties. It is by no means a requirement, though generally idiomatic.

21:50 xeqi: john2x: ring-mock is the base one. If you also need a cookie jar you could look at my peridot or kerodon libraries

21:51 SegFaultAX: egghead: The point is you shouldn't care if the thing you're accessing is a property or a field.

21:51 egghead: SegFaultAX: but having it not be really uniform means you do have to care

21:52 SegFaultAX: egghead: Don't write shitty classes that don't follow the convention, then. ;)

21:52 egghead: I try not to write scala if I can help it, but I inherited some scala last month from another team :(

21:52 :p

21:53 bbloom: the uniform access principle is not-mega-objectionable for reads, but holy crap is it a bad idea for writes via methods named foo_=

21:53 consider a calls representing a rectangle, if you try to do rect.right = 500

21:53 what happens?

21:54 does it move the rectangle or resize it?

21:54 SegFaultAX: bbloom: I /really/ hate how overloaded the character _ is in Scala.

21:54 bbloom: even if you agree on it moving or resizing, then suddenly the order of updates matter! if you change left and right or right and width or whatever

21:54 even if you have an immutible rectangle and left_= returns a new rectangle, these problems exist

21:55 in fact, i'd argue that property setters is one of the dumbest ideas... ever.

21:55 SegFaultAX: bbloom: I concur. I don't know why you'd ever want a setter on a computed value.

21:56 bbloom: SegFaultAX: the only time it makes any sense is when you have a trivial converter, like if you wanted to data bind a number to a text box and needed Integer/parse with white space trimming

21:56 but even then, you're gonna likely want to override that policy, so who says what gets to be theInteger_=

21:56 ugh. setters. dumb idea

21:58 tomoj: interesting

21:58 ivaraasen: setters. it's what plants crave.

21:59 tomoj: you'd need a function with domain and range subsets of XxYxWxH

21:59 er

21:59 the products of subsets of {X,Y,W,H} I mean

22:00 like your multidirectional dataflow graph functions

22:00 bbloom: tomoj: heh, rectangle representation is a soap box i can stand on for days

22:00 rectangles expose so many problems with object oriented programming, it's not even funny :-P

22:01 ivaraasen: bbloom: taking an OOP course at the moment. so much silly stuff

22:01 bbloom: rectangles fundamentally have 4 fields. but what are they? left/right/top/bottom? x/y/w/h ? center-x/y, half-w/h ?

22:01 what about representations as two points? top-left and bottom-right?

22:02 or a center and two half widths in a size vector?

22:02 but let's dumb the problem down: circles! they only have a position and a radius

22:02 still, point or pair of ints?

22:03 let's dumb it down! a Range of integers: minimum and maximum!

22:03 oh wait, is it [min,max) or [min,max]

22:03 this shit is hard.

22:03 or is it min & length ?

22:03 add some getters and setters, paper over the problem

22:03 and suddenly range.length = 5 is a mega confusing arbitrary decision that yields signficant order of statements bugs

22:04 * bbloom will shut up now

22:05 lynaghk: bbloom: have you seen a nice way to handle that with records in Clojure?

22:05 ivaraasen: the correct implementation of a triangle is a class inheriting from Shape, which wraps CartesianSpace<Point>, which in turn is a superior ArrayList, because I wrote it. sigh.

22:06 lynaghk: bbloom: i.e., sometimes it's natural to want to (assoc my-rect :top 1 :left 2 :bottom 5 :right 11), and other times I want to refer to width and height

22:06 bbloom: lynaghk: not really.... afterall, deftype and defrecord are slighly-less-featureful (in a good way) OOP

22:06 lynaghk: bbloom: right, so basically you need to pick one canonical basis and then use functions to get/set the others?

22:06 bbloom: this is something i've thought a bit about tho. i think that there is some sort of algebra of nouns or something yet to be discovered

22:06 tomoj: isomorphisms are lenses

22:06 lynaghk: bbloom: or go to town with constraints and macro-generated code for an IAssociative implementation

22:07 bbloom: lynaghk: but then you have the setter problem

22:07 lynaghk: bbloom, tomoj: yeah. right. are you aware of any lense implementations for Clojure records

22:07 (Not that I'm sure it's a good idea whatsoever)

22:07 bbloom: i'm anti lenses :-P

22:07 lynaghk: bbloom: that's why I'm thinking you should have to specify everything with constraints

22:08 I mean, core.logic is just SITTING THERE.

22:08 tomoj: no, I've been working on it slowly

22:08 bbloom: i spent some time studying them & decided that symbolic algebra is fundamentally superior

22:08 constraints < algebra

22:08 tomoj: symbolic algebra?

22:08 lynaghk: bbloom: I don't know enough about your termonology to understand; do you have some references?

22:09 bbloom: lynaghk: http://en.wikipedia.org/wiki/Computer_algebra_system

22:10 tomoj: you'd write an equation for the isomorphism between x,y,w,h and (say) sw,ne ?

22:10 bbloom: the problem is that anyone who ever tries lenses or constraints or anything like that, tends to write a paper with some linear algebra & fundamentally ignores business logic domains

22:11 tomoj: yeah, i think you could define a rectangle in terms of some algebraic relations & then the compiler could find a spanning tree of terms. every spanning tree is one possible representation

22:11 however, i can only assume this is undecideable lol

22:12 ohpauleez: lynaghk: I'm going to jump in here

22:12 I have record invariants working

22:13 lynaghk: ohpauleez: gist?

22:13 ohpauleez: a semi-working generator from Alloy, and an implementation of basic constraints from core.contracts

22:13 lynaghk: ohpauleez: that's my mantra for the over-educated inhabitants of #clojure---gists or it's not a real idea =P

22:14 ohpauleez: this is what you'll be talking about at Clojure/West, yeah?

22:14 ohpauleez: in my branch of core.contracts, I have a protocol with (constrain ...)

22:14 lynaghk: Yeah, this is my talk on Monday

22:15 my examples are all about functions, because I think more people can relate to it. But the underlying pieces also work with Records. I haven't fully solidified the API wrt records

22:15 so it's subject to change

22:16 lynaghk: cool, I didn't know you had features for inline usage---I thought it was all out-of-band verification and maybe some runtime assertions

22:16 ohpauleez: it has open handlers, that act as decorators

22:17 lynaghk: ohpauleez: yeah, when I last thought about this kind of thing I wasn't sure what I wanted the API to look like---ideally for some class of relationships everything could be patched into the standard IAssoc and ILookup implementations and Just Work (tm)

22:17 ohpauleez: so it'll generate you n-versions of the thing you're spec'ing

22:17 I did not take that approach

22:17 and I'm with bbloom

22:17 there's more power in an algebra and discrete math

22:18 bbloom: ohpauleez: i want to express my program assertions in first order logic :-P

22:18 lynaghk: ohpauleez: see my comment re: gists.

22:18 ohpauleez: lynaghk: see my response: Monday

22:18 :)

22:18 lynaghk: ohpauleez: =)

22:18 bbloom: that reminds me... i need to write a README before I open source some stuff .... heh

22:19 *shrug* sounds like a good use of my time on the flight

22:19 ohpauleez: bbloom: flying from?

22:19 bbloom: JFK

22:19 ohpauleez: ahh you're already in NYC

22:20 bbloom: yeah, since end of january

22:23 john2x: how do I write tests for my handlers that go through the routes? (instead of just passing mock requests directly to the handlers)

22:24 xeqi: john2x: pass the mock requests directly to the route functions

22:26 john2x: xeqi, hmm, by route functions, you mean those middleware wrappers?

22:29 xeqi: john2x: I was assuming you were using (defroutes some-routes ...). Then you could do (some-routes mock-request)

22:29 if you are wrapping middleware around them then just use the handler that returns

22:30 Raynes: xeqi: You've been using laser, right?

22:30 xeqi: Raynes: I'm using it to build my screencast site

22:31 Raynes: xeqi: Good experience so far?

22:32 lynaghk: xeqi: you're building a screencast site too?

22:33 xeqi: lynaghk: yes, and eric normand did the kickstarter

22:34 lynaghk: are you making one?

22:34 * not that mine and his project are related, just another person I know making screencasts

22:34 lynaghk: xeqi: yeah, though not with Clojure as a topic right now

22:34 xeqi: Raynes: its been pretty good

22:35 Raynes: xeqi: Pretty good sounds horrible.

22:35 xeqi: there was some confusion about fragment vs parse-fragment, and which one worked with content, and which with remove

22:35 lynaghk: xeqi: but I'm helping a friend do one for Photoshop (retouching in a production environment) and using the same tech stack for a "personal devops" site

22:35 Raynes: xeqi: Well, parse-fragment and fragment do completely different things.

22:35 parse-fragment just parses html into something you can pass to fragment and do work on.'

22:36 xeqi: lynaghk: I put up a sample video at http://www.clojurewebdevelopment.com/videos/friend-interactive-form to gauge interest, and based on it and in person conversations it seems like there is a small, but willing to pay market for clojure videos

22:36 Raynes: xeqi: Come to LA in April. I'm gonna give a talk at the LA Clojure meetup about it.

22:36 ;)

22:36 xeqi: Raynes: hmm, maybe I meant nodes

22:36 lynaghk: xeqi: yeah, I'm not surprised. I might dip into it later this year; Eric Normand's kickstarter certainly validates that

22:36 Raynes: nodes is basically parse-fragment.

22:37 It just also normalizes things. If you give it a node, it gives you back something you can pass to fragment.

22:37 Bleh. My docs probably suck.

22:37 john2x: xeqi: ah got it. this is cool

22:37 Raynes: Feel free to suggest improvements. You have the benefit of being a guy who didn't write laer.

22:37 alandipert: i would pay for a service that forced others to watch clj videos

22:37 xeqi: Raynes: css selectors :p

22:37 technomancy: alandipert: get y combinator on the phone; I smell a market

22:37 lynaghk: alandipert: your best bet is to pay off a flight crew

22:37 Raynes: alandipert: Pretty sure that's called terrorism.

22:37 :p

22:38 xeqi: I meant doc improvements. :p

22:38 technomancy: Raynes: or advertising. fine line.

22:38 alandipert: ingenius ideas, all of them

22:38 egghead: nice Raynes -- will catch that laser talk :)

22:38 Raynes: But sure, css selectors could be done.

22:38 lynaghk: alandipert: when are you getting into town?

22:38 Raynes: egghead: Oh right, you're in LA aren't you?

22:38 alandipert: lynaghk, tomorrow at 2pm

22:38 egghead: mhm

22:38 visiting pdx this week but la regularly

22:38 Raynes: egghead: It's actually May that I'm talking. May 5th I think, but I might be wrong. Some day in May.

22:39 lynaghk: alandipert: okay, cool. I think I'm just going to head over to downtown and start drinking coffee with clojure friends nonstop starting tomorrow at 10am.

22:40 * technomancy is looking forward to trying Public Domain

22:40 lynaghk: (hear that Clojure friends!?! Find me tomorrow for free coffee and tours of hipsters in their natural habitat)

22:40 technomancy: Public Domain is the only coffee shop I've ever been to that has an "espresso happy hour"

22:41 Raynes: egghead: Okay, May 2nd. This time for sure (I looked at the email).

22:41 If you do come, make sure you introduce yourself to me.

22:41 egghead: cheers Raynes -- at the clj meetup?

22:41 lynaghk: technomancy: just don't order an espresso to go, or they'll get pissy at you.

22:41 dcolish: Public domain is pretty good, you should also check out Spella

22:41 Raynes: Just walk up and say "Yo, I'm Mr. Egghead."

22:41 egghead: haha lynaghk I might take you up on that :)

22:42 alandipert: lynaghk: yes excellent, i look forward to caffeinating with you

22:42 bbloom: lynaghk: i get in at 830 or so. please bring my free expresso to the arcade party, thank you :-)

22:42 * alandipert also already wishing and dreaming about public domain

22:42 lynaghk: bbloom: you forefeited it when you said, "expresso"

22:43 egghead: I've been wine tasting today in mcminnville

22:43 bbloom: lynaghk: i'd say that i'm surprised i made that typo, but i'm not really

22:43 egghead: everyone is so nice over here, i'm used to la where everyone hates everyone

22:43 expresso b/c it makes you go fast vrooom

22:44 lynaghk: egghead: yeah, that weirded me out when I moved here from North Carolina---I distinctly remember wondering why all the cars were stopping and realized it was because I was on the sidewalk *looking* across the road.

22:44 technomancy: lynaghk: I would never order espresso to go

22:44 you have to give it your full attention immediately

22:45 lynaghk: technomancy: yeah, nor would I. Just saying, I've heard you'll get a lot of lip and/or be refused service there if you do that

22:45 technomancy: haha

22:45 dcolish: well, proximity to the hotel is another key factor

22:45 lynaghk: technomancy: Coava Coffee on the east side (directly off the Hawthorne Bridge, basically) is also very good.

22:45 technomancy: I'm more interested in the pour overs anyway

22:46 lynaghk: I think there's a place up here that serves their beans

22:46 egghead: I like my espresso burnt and bitter

22:46 generally I wait around 5 mins after they are pulled to drink them to make sure they don't taste right

22:47 lynaghk: egghead: it's fun to play pretentious now and again, but I wholeheartedly agree that the best way to have anything is to just have it the way you like it

22:47 egghead: which is why I plan on continuing to drink only sub-$10 bottles of wine, preferably in parks.

22:48 alandipert: and why i choose to drink only microwaved day-old mcdonalds coffee that i find on my desk

22:48 lynaghk: see, what's nice about Clojurians is just how damn pragmatic we are =P

22:48 bbloom: i'll be honest: i'm quite enjoying the bland black coffee that every corner deli has in NYC - it's just so much faster than waiting for someone with some skill to prepare your beverage :-P

22:49 it means that when somebody drags my ass down to tribeca or some other rich person neighborhood, i can really enjoy a good coffee as a novelty of sorts

22:49 egghead: why bother drinking it when you can take it via drip

22:49 bbloom: egghead: where do i sign up?

22:49 dcolish: technomancy: Spella is a 1.5 blocks from Public Domain

22:49 i highly recommend getting both

22:49 egghead: :p

22:49 dcolish: stumptown is ok

22:50 barista is good if you get heart or verve beans

22:50 xeqi: Raynes: I was surprised by the difference in https://www.refheap.com/paste/12647

22:50 dcolish: we take coffee a little too serious in pdx

22:50 xeqi: also that error message sucks

22:50 though we had this discussion at one point and I got something working well enough

22:50 besides that its been good

22:52 Raynes: xeqi: You shouldn't be calling fragment on either of those things.

22:53 xeqi: Also, in the latest versions you can actually now just throw strings in there unescaped. You just have to call me.raynes.laser/unescaped on them.

22:53 xeqi: Raynes: its a simplified example, imagine replacements happening inside (laser/fragment ...)

22:53 Raynes: Gotcha.

22:54 Yeah, I should probably look into why that doesn't work. It doesn't look dumb at first glance.

22:54 Not that anything you'd ever do would be dumb.

22:54 <3

22:54 xeqi: Is this happening with the latest laser?

22:54 xeqi: use case: I'm processing a page and turning it into a fragment and send it in a ring response, and then have a middleware do the laser/document with the layout

22:55 [me.raynes/laser "0.1.21"]

22:55 so no

22:55 or.. untested

22:55 Raynes: Yeah, there were lots of changes in content over the last 0.x.x releases. And bugs.

22:56 Make sure you update. The latest version is stable and not going to change from underneath you.

22:56 xeqi: $latest me.raynes/laser

22:56 Raynes: $latest me.raynes/laser

22:56 lazybot: [me.raynes/laser "1.1.1"] -- https://clojars.org/me.raynes/laser

22:56 [me.raynes/laser "1.1.1"] -- https://clojars.org/me.raynes/laser

22:56 Raynes: !!

22:56 o/

22:56 hyPiRion: zing

22:57 Raynes: http://dl.dropbox.com/u/23745600/Screenshots/aOry.png Immortalized.

22:58 technomancy: dcolish: you're talking to someone from seattle =D

22:58 but thanks; hope to try them both

22:58 Raynes: technomancy: You should come visit me. <3

22:58 We can have dinner and you can push lein issues at me.

22:59 dcolish: technomancy: well a big NW highfive

23:01 xeqi: john2x: how is the ring testing going? any problems?

23:18 Raynes: ahh, html-content disappeared

23:26 but looks like the (l/content (l/nodes ...)) works

23:27 Raynes: looks like the example paste works fine in 1.1.1, so no more confusion for me

23:38 Raynes: xeqi: Yes, content serves both purposes now.

23:43 Rich_Morin: FYI - Light Rail notes for Portland - http://pastie.org/6578091

23:44 technomancy: wow, pastie is still around

23:45 I remember being in the channel when that was created in 06

23:45 Rich_Morin: yep, and accepting donations for Josh's brother Nate.

23:47 jeremyheiler: Rich_Morin: Thanks!

Logging service provided by n01se.net