#clojure log - Dec 18 2015

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

0:09 owlbird: It's possible to use info->'food' ? 'carrots' to query 'carrots' whether or not a member of info (jsonb) in PostgreSQL 9.4 , but the jdbc library treats '?' as a prepared statement, how do I fix that?

0:28 zereraz: hi, I have a doubt

0:28 I did, lein new test, then cd test, then lein run but it gave me error

0:28 No :main namespace specified in project.clj.

0:29 don't know what to do

0:31 ok I found the problem, I had to create using "lein new app my-app"

0:31 rather than just lein new my-app

2:57 keep_learning: Hello everyone

2:57 aztak: mornin'

2:58 keep_learning: Could some one please tell how to download a web page containing java script

3:08 I got one idea but it's not working with my functional mindset.

3:09 I am trying to use clj-webdriver but it seems to me that it's changing the states of browser by invoking the commands.

4:54 Hello everyone

4:54 I am getting error for this code.

4:54 http://lpaste.net/147323

4:54 IllegalArgumentException Don't know how to create ISeq from: net.cgrand.enlive_html$html_content$fn__4387 clojure.lang.RT.seqFrom (RT.java:528)

4:55 Right now I am trying to find the links of google search

4:55 so I am running a web driver to execute the javascript and enlive to find all the tags [:a]

7:36 noncom: keep_learning: well, first of all the top section of the namespace is weird... are you typing this in the repl?

7:36 keep_learning: noncom, Thank you for reply

7:36 noncom, Yes

7:37 noncom, but now I have moved that code to file

7:37 so I have written the same thing in file

7:38 noncom: keep_learning: next: https://github.com/cgrand/enlive/blob/master/src/net/cgrand/enlive_html.clj#L655

7:38 the html-content function returns a function

7:39 this https://github.com/cgrand/enlive/blob/master/src/net/cgrand/enlive_html.clj#L553 , however idk what it expects, most likely not a function

7:40 keep_learning: are you sure these two functions can work together like in your code?

7:41 keep_learning: noncom, I was trying to figure and changed the code

7:42 noncom, http://lpaste.net/147329

7:42 noncom: does it give the same erorr?

7:43 keep_learning: noncom, So now it's working but I am newbie so I don't know how to idiomatic clojure code

7:43 noncom: keep_learning: oh

7:43 did you read this? https://github.com/swannodette/enlive-tutorial

7:43 gives a good insight into the topic..

7:43 keep_learning: also, one rule of a thumb: never use use :)

7:44 it is deprecated so much

7:45 keep_learning: noncom, Yes, From there I am trying to figure out bits of enliven

7:46 noncom: keep_learning: well, that your second example of code looks ok to me, except that it still uses repl-like requirements section and uses (use)

7:46 :)

7:47 but as you said, you have the regular file-like ns declaration in your actual code, so it is ok

7:47 but still i'd strongly advice you to get rid of use

7:47 keep_learning: noncom, I have knowledge of Haskell so I am comfortable with functional part except how to import library :)

7:48 noncom: keep_learning: well, but i think you already know this thing: (ns my-ns (:require [aaa.a :as b] [ads.asd :as ggg :refer [n]]) (:import some.java.Clazz))) ... ?

7:48 like, if you look into any clojure file from any code, like the enlive lib itself, you'll see that at the top

7:49 keep_learning: doh, enlive uses a thousand :require here https://github.com/cgrand/enlive/blob/master/src/net/cgrand/enlive_html.clj#L11

7:49 keep_learning: noncom, Thank you. Do you have quick link

7:49 noncom: one is enough

7:49 wait a sec

7:50 keep_learning: https://gist.github.com/ghoseb/287710

7:50 keep_learning: look here for a good example: https://github.com/dakrone/cheshire/blob/master/src/cheshire/core.clj

7:50 and if you see :use or (use) - ignore it, it's not welcome these days

7:51 keep_learning: noncom, Thank you

8:11 troydm: let's say I have '((1 2) (3)) how do I convert it into '(1 2 3) ?

8:11 opqdonut: ,(apply concat '((1 2) (3)))

8:11 clojurebot: (1 2 3)

8:12 troydm: opqdonut: thx

8:12 opqdonut: ,(apply concat '((1 2) (3 4 5) (6))) -- more generally

8:12 clojurebot: (1 2 3 4 5 ...)

8:15 ridcully_: ,(sequence cat '((1 2) (3) (4 5)))

8:15 clojurebot: (1 2 3 4 5)

8:18 BRODUS: say i have an anonymous function, how can I operate on that function as if it were a list?

8:19 luma: what do you mean "as if it were a list"

8:19 how would a function be a list?

8:20 BRODUS: well i assumed that it was since we write them as such

8:22 luma: what would you do with it "as a list"?

8:23 MJB47: ,(first '(+ 1 1))

8:23 clojurebot: +

8:23 MJB47: i wonder..

8:23 ,((first '(+ 2 3)) 4 5)

8:23 clojurebot: 5

8:26 BRODUS: luma: in my case, i'm trying to convert the function to a string,

8:27 MJB47: ,(name (first '(reduce + (range 3))))

8:27 clojurebot: "reduce"

8:27 BRODUS: i figured the function would be stored in a list data structure, am i wrong?

8:28 luma: when the code is executed, it creates a function object

8:35 hyPiRion: BRODUS: (first '(+ 2 3)) returns the symbol +

8:36 ,[('+ 4 5) ('+ 4)]

8:36 clojurebot: [5 nil]

8:45 oic3ieNg: Hello. Could you recommend me a short and easy tutorial on datomic pull api?

8:55 ridcully_: oic3ieNg: you have seen the docs, right?

9:02 oic3ieNg: ridcully_: I mean something really simple and to the point, without mbrainz, grammars, java, led zeppelin and whatnot

9:18 sdegutis: Hi.

9:18 How are you going?

9:28 oic3ieNg: Slowly.

9:30 noncom: BRODUS: if you read it, then it is a list, however, in runtime the functions get transformed into function objects

9:30 wow, swift! )

9:36 drinking green tea

9:36 beaky: i love greentea

9:36 im drinking dong ding oolong rite now

9:37 noncom: nice one!

9:37 beaky: running out of it tho :(

9:53 sdegutis: Oh.

11:09 benjyzz: anyone in #clojure interested in cryptocurrencies?

11:21 TimMc: Vaguely. I've really cooled off on bitcoin because it seems like a waste of electricity and the consensus protocol is now controlled by just a few miners.

11:22 Stellar seems neat, though. It doesn't use proof-of-work and they're trying to use it to help the unbanked exchange money.

12:18 BRODUS: any ideas on how I can get a reference to the newline character without using \n

12:19 justin_smith: \newline - \n is the letter n

12:19 ,(= (first "\n") \newline)

12:19 clojurebot: true

12:19 justin_smith: ,(= (first "n") \n)

12:19 clojurebot: true

12:21 BRODUS: I don't think I can use that either, that expression will eventually get read as a string and I need it to not break on the \newline

12:21 justin_smith: there is no character literal for newline that will work both inside and outside a string

12:21 ,(str \newline)

12:21 clojurebot: "\n"

12:22 justin_smith: though...

12:23 momerath: benjyzz & TimMc : I work at Ripple - feel free to msg me if you want to talk cryptocurrencies. I hope we'll be hiring at least one more clojurist in the next year too.

12:24 yenda: https://news.ycombinator.com/item?id=10758888 nice article on clojure design patterns

12:26 BRODUS: ,(eval '(char 10))

12:26 clojurebot: \newline

12:31 justin_smith: ,(eval \newline)

12:31 clojurebot: \newline

12:34 TimMc: momerath: Heh, thanks, but I try to stay far, far away from the financial sector. :-P

12:34 beaky: ,(eval "(eval \"(eval '()))\"")

12:34 BRODUS: ,(println "(eval \newline)")

12:34 clojurebot: "(eval \"(eval '()))\""

12:34 (eval \newline)\n

13:30 jdeisenberg: Want to do a websockets server. Am I better off using Clojure w. jetty/ring or ClojureScript w. node.js?

13:31 It may be running on a raspberry pi level CPU; for now, it's a normal fairly reasonable Athlon.

13:32 justin_smith: jdeisenberg: well, I definitely would not use jetty for a clojure websocket server

13:32 jdeisenberg: Ring would be a better choice, then?

13:33 justin_smith: jdeisenberg: ring is a protocol that has adaptors, the jetty ring adaptor does not do websockets

13:33 but aleph or netty or http-kit with ring are all options that do websockets with ring

13:34 jdeisenberg: jvm clojure is not apropriate for doing anything serious on arm, at least not with raspberry pi quantities of RAM available

13:35 clojure will do amazing things on a machine that has sufficient RAM, but it doesn't really do light weight

13:36 jdeisenberg: Yeah, I sort of thought that would be a problem.

13:36 justin_smith: clojurescript and pixie are both definitely usable on the raspberry pi

13:36 pixie is experimental, and clojurescript has an odd dev toolchain

13:37 keep_learning: Hello everyone.

13:37 I am working with clj-webdriver

13:37 and trying to run this code

13:37 (get-url "https://www.tripadvisor.in/")

13:38 (quick-fill-submit (xpath "#GEO_SCOPED_SEARCH_INPUT" city)

13:38 (xpath "#mainSearch"))

13:38 but getting IllegalArgumentException No implementation of method: :find-elements of protocol: #'clj-webdriver.core/IFind found for class: java.lang.String clojure.core/-cache-protocol-fn (core_deftype.clj:554)

13:38 justin_smith: keep_learning: get-url will return a string, right?

13:40 oh, never mind - it navigates the browser

13:40 keep_learning: justin_smith, I think Navigate the browser to `url`.

13:40 yes

13:41 justin_smith: keep_learning: looks like quick-fill-submit wants its args to be hash-maps https://github.com/semperos/clj-webdriver/wiki/Taxi-API-Documentation#quick-fill-submit

13:42 keep_learning: and xpath returns a string https://github.com/semperos/clj-webdriver/wiki/Taxi-API-Documentation#xpath

13:43 keep_learning: seems like that would be the error

13:43 keep_learning: justin_smith, Yeah

13:43 justin_smith, Thank you

13:46 jdeisenberg: justin_smith: thanks for the info.

13:57 ghost_: justin_smith: what do you use for development? I mean, text editor

13:58 keep_learning: justin_smith, Could you please tell me if I my CSS selector is correct or not ?

13:58 (quick-fill-submit {"global_nav_search_form#GEO_SCOPED_SEARCH_INPUT" city}

13:58 {"global_nav_search_form#mainSearch" "Hotel"}

13:58 {"global_nav_search_form.SEARCH_BUTTON" click})

13:59 I am taking the form id global_nav_search_form and then filling both forms

13:59 GEO_SCOPED_SEARCH_INPUT and mainSearch

14:00 https://www.tripadvisor.in/

14:00 I am getting TimeoutException Timed out after 10 seconds waiting for clj_webdriver.wait.proxy$java.lang.Object$ExpectedCondition$66be6a90@23d96cee

14:00 because I have (wait-until false 10000)

14:00 to wait for 10 sec

14:01 justin_smith: yeah, I'm not sure about the selector syntax for that lib, sorry

14:02 keep_learning: justin_smith, No problem. I am just trying to figure out all the way. Thanks for your help

14:24 ghost_: justin_smith: it's kind of important for me, because I want to finally settle on something

14:28 kungi: I am trying to use the BarbaryWatchService (https://github.com/gjoseph/BarbaryWatchService/blob/master/Demo.asciidoc) from Clojure.

14:29 In https://github.com/gjoseph/BarbaryWatchService/blob/master/src/main/java/com/barbarysoftware/watchservice/WatchEvent.java there is a public static interface Kind<T> how can I address this through java interop?

14:30 ridcully_: WatchEvent$Kind ?

14:31 kungi: ridcully_: Sadly no

14:31 I get an "Unable to resolve symbol" error

14:32 ridcully_: I am trying to call this function: https://github.com/gjoseph/BarbaryWatchService/blob/master/src/main/java/com/barbarysoftware/watchservice/WatchableFile.java#L36

14:34 ridcully_: aren't those then StandardWatchEventKind/ENTRY_CREATE ?

14:35 kungi: ridcully_: Yes they are. Let me show you a bit more code

14:35 justin_smith: kungi: should be something like (.register wf (into-array WatchEvent$Kind [event, event, ...]))

14:35 oh wait, I missed an arg there

14:35 kungi: https://gist.github.com/47635b4ea28dae353be4

14:36 justin_smith: I tried that and is does not work ... :-(

14:36 justin_smith: what symbol is it unable to resolve?

14:36 kungi: WatchEvent$Kind

14:37 justin_smith: kungi: what does your import look like?

14:37 importing WatchEvent won't import WatchEvent$Kind - they each have to be imported separately

14:38 kungi: justin_smith: AHA! Sweet

14:38 justin_smith: thank you I was not aware of this

14:38 justin_smith: in the java view WatchEvent$Kind "belongs to" and is "inside of" WatchEvent, in the jvm, WatchEvent$Kind is just a funny name in the same package WatchEvent is in

14:39 ridcully_: (inc justin_smith)

14:39 justin_smith: well, not *just*, but...

14:39 kungi: (inc justin_smith)

14:39 ridcully_: bah stupid bot

14:39 started weekend already

15:05 spieden: i needed to "patch" datomic entities today by diffing current vs previous attribute values and generating assertions. so amazingly straight forward in clojure -- even handles cardinality many (but not rels) https://gist.github.com/spieden/79255a3a0f27e2c372d8

15:40 felixn: anyone have autotesting that works for them? I'm using midje, but it's not identifying cljc files o_O. also if I remove a test, it keeps the stale output when reloading

15:40 j-pb: we abandoned midje long ago, I'd go with plain clojure test :)

15:41 we had a pretty nasty bug in out production system that we didnt' find because we forgot a => in midje wich made the test always pass

15:55 felixn: j-pb: how do you rerun the test? `test-refresh` seems to be working a lot better, but sometimes I do have to restart

15:56 * spieden is trying midje for the first time.

15:57 spieden: j-pb: yeah, i wish they didn't use magic symbols like that

15:57 j-pb: felixn: intellij actually :)

15:57 cursive has really nice core.test integration

15:57 on the cljs end there is doo

15:57 spieden: yeah, does it work at all with test.check? our codebase would just collapse without it ^^

15:58 spieden: j-pb: hmm not to my knowledge. does cursive have integration with the intellij test runner now or is it still just "send to REPL" style?

15:59 i'd switch back in a heartbeat if it did

16:00 hmm looks like just repl

16:00 felixn: j-pb: nice, I'll give cursive a try. intellij is amazing for python, though atom is very easy on the eyes

16:01 j-pb: spieden: yeah, you do send tests in current repl to ns, I agree that it would be nicer to have continuous checking :)

16:02 spieden: midje seduced me with its code reload and autotest

16:02 felixn: anyone know if it's possible to get the name of an 'clojure.core.logic/lvar? I think it maybe (.getField (lvar "foo" false) "name"), but not sure how to import the datatype functions

16:02 j-pb: spieden: but you don't need midje for that I think https://github.com/jakepearson/quickie

16:07 spieden: cool i'll check it out

16:07 definitely would be nice to use that "Re-run last test action in REPL" action. do you guys use the cursive auto code reloading or tools.namespace for your in-ide test workflow?

16:08 j-pb: ^^

16:08 j-pb: manual reloading

16:08 breaks less stuff and is only one keystroke :)

16:39 lokien_: What do you guys use for editing Clojure? IDE, good old Emacs with Paredit?

16:46 futuro: lokien_: emacs + paredit + cider + clj-refactor

16:47 lokien_: futuro: standard emacs bindings?

16:47 futuro: yeah

16:47 lokien_: Masochist

16:47 * futuro shrugs

16:48 lokien_: futuro: anyway, thanks. I think I'll do the same

16:48 futuro: I hear spacemacs is pretty good if you want modal editing

16:48 lokien_: np

16:48 j-pb: lokien_: intellij and cursive is pretty nice

16:48 but if you're already an emacs head, then you maybe want to stay one ;)

16:49 lokien_: futuro: I'm using it now but emacs bindings for workspaces and vim's for editing are making me cringe

16:49 futuro: and huge tutorials aren't really my thing

16:50 futuro: yeah, pretty much whatever editor you're currently using is the one you should use for clojure, assuming you're not specifically looking for a new one

16:50 lokien_: why not vim+fireplace+other stuff?

16:50 lokien_: j-pb: I'm not, I just want something simple with autocomplation and repl

16:51 futuro: repl was awful last time, I'll give it a shot tomorrow

16:51 autocompletion*, durr

16:51 futuro: lokien_: word, I'm interested in how that goes

16:52 j-pb: lokien_: cursive has pretty decent refactoring tools, that's why I use it

16:52 lokien_: futuro: I'll let you know

16:52 j-pb: large codebases?

16:53 ridcully_: vim, fireplace, paredit

16:53 futuro: ridcully_: how's it's refactoring support?

16:53 j-pb: lokien_: yeah pretty, I wouldn't know how to navigate them without show usages

16:53 ridcully_: non existing?

16:54 futuro: ridcully_: ah, ok

16:54 lokien_: ridcully_: how do you connect to the repl? is it just sitting in another terminal tab?

16:54 j-pb: futuro: I'm not a fan ob hour long config twiddling, the codebase I have to worry about is enough to keep me up at night :D

16:55 ridcully_: yes. :Connect nrepl-url - or just :Console (via vim-salve)

16:55 lokien_: j-pb: I always end up deleting half of my code doing refractoring in an IDE (or doing other spooky stuff I can't undo)

16:55 j-pb: lokien_: I mean ^^

16:55 lokien_: intellij has a pretty long undo history

16:56 it's not as nice as emacs undo tree, but tbh I never had any problems with cursive refactorings

16:56 lokien_: j-pb: I'm not a smart man

16:57 ridcully_: no tmux or other goofy stuff?

16:58 ridcully_: usually tmux - and tmux things made into my aliases. so your goofy is my normal

16:58 futuro: it looks like there's some work surrounding the refactor-nrepl https://github.com/Deraen/vim-cider

16:59 lokien_: ridcully_: it will eventually become my normal too. big thanks, see goofy-me in a month

17:00 Deraen: Fireplace will also automatically try connecting to nrepl when trying to evaluate something, so :Connect is not usually necessary

17:01 lokien_: futuro: I'll check it tomorrow, thank you too

17:01 ridcully_: for cljs i use a patch - so this works fine for me in both universes

17:01 lokien_: Daraen: really? oh

17:01 ridcully_: yeah, :Connect is usually not needed - but it allows you to connect to lein repl without project.clj e.g.

17:01 Deraen: Yes. Lein and Boot both write a .nrepl-port file which is used by editors to automatically detect nrepl

17:02 lokien_: I've been typing that all the time

17:02 And memorizing repl's port

17:02 Sweet baby jesus

17:03 Deraen: About refactoring stuff, I rarely miss Cursives's features as Vim has great macros and several other features which make it easy to repeat same modifications to multiple places in one file or even to all files in project.

17:04 Not as easy to use, but it

17:04 it's possible to do pretty much anything with those.

17:04 lokien_: Cursive was too much rocket science for me, as every IDE

17:18 felixn: I'm trying to recursively map, but at the same time keep a running total state along side. is there a good way to do this in clojure? I'm having trouble finding an article that explains assoc! / persistent! / transient well

17:20 j-pb: felixn: those are not meant or suitable for mutable state

17:20 just in case that's how you want to use them :)

17:21 felixn: do you want to process a tree, or what do you mean by recursive map?

17:21 felixn: I'm rewriting a nested data structure, but I want to keep track of what I rewrote in a map, and return the rewritten data structure along with the final mapping. trying to figure out how XD

17:22 j-pb: don't use transients for that they are a perfromance optimisation nothing more

17:22 I'd look at zippers maybe, they are perfect for traversing a tree, and accumulating state in a loop

17:27 felixn: it kinda depends, how large is you nested datastructure?

17:27 will a simple update-in suffice?

17:29 felixn: j-pb: it's small, I'm checking out zipper right now. it's really hard to figure out how to use these things from the documentation XD https://clojuredocs.org/clojure.zip/zipper

17:29 j-pb: felixn: could you describe a bit more what you want to do :)

17:29 zippers are great for example for manipulating xml

17:31 felixn: it takes some adjusting, but generaly speaking you never ever need mutable state if you're just processing data

17:44 futuro: reducers was one of those things that took me a while to wrap my head around

17:48 lambda-11235: Would anyone recommend using measjure for unit analysis? Has anyone used it.

17:50 justin_smith: lol, this page is the only google hit for measjure: https://gist.github.com/ddillinger/5154002#file-gistfile1-txt-L46

17:52 lambda-11235: Its spelled without the s, and I get the github account when I google it (https://github.com/arrdem/meajure).

17:52 Are there any other unit analysis that do algebraic analysis?

17:52 justin_smith: ahh, I googled measjure and only found that amusing gist :)

17:53 ridcully_: you googled ass

17:53 justin_smith: lambda-11235: not that I know of, and arrdem is a smart cookie, that's probably a good lib

17:53 ridcully_: no, I just chose to highlight that line :P

17:53 ridcully_: :)

17:54 amalloy: the only successful clojure project: nonfailjure

17:54 justin_smith: haha

18:03 rhg135: that's amusing

18:26 anoctopus: Anyone here using squiggly-clojure? I can't get it to show anything

18:46 rhg135: hmm, where can I get the latest jar of clojure?

18:47 justin_smith: rhg135: maven should have it

18:47 rhg135: I think they were on github, might have been cljs

18:47 justin_smith: rhg135: http://clojure.org/downloads

18:47 rhg135: I'll do that then. thx

18:47 * rhg135 facepalms

18:48 justin_smith: rhg135: just yesterday I downloaded clojure.jar via that link, because I wanted to do something in a repl on a weak vps instance I was logged into via my phone :P

18:49 rhg135: I'm working on a script to aid in writing acrostic poems

18:51 as an aside, how much ram could you get away with? I have a weak arm board running this irc client

18:51 justin_smith: rhg135: cool

18:52 rhg135: 368M total ram D:

18:52 justin_smith: I haven't really tested, my clojure-1.8.jar repl is taking under 20 megs right now

18:53 but I also haven't asked it to do much :)

18:54 rhg135: This sucks wrt ram. I couldn't even compile lxml; gcc kept dying

18:54 kept getting killed

18:55 justin_smith: rhg135: hmm - did you try something like -O0 ?

18:56 rhg135: justin_smith: I said some choice words, installed the binary from the distro, and copied it to my venv

18:56 justin_smith: :)

18:57 rhg135: fortunately arch had the exact version :D

18:58 the alternative was run lazybot instead; which seems less likely

19:03 felixn: https://gist.github.com/anonymous/0207a6b8e9ad905bd24c <-- j-pb: woo! thanks for recommending zipper, works really well for this

19:04 feel free to criticize the code, second day messing around

19:05 with clojure that is

19:08 cfleming: spieden: No test runner integration yet, although I'm hoping to do that very soon.

19:08 spieden: Unfortunately it appears to be tricky to make that work with auto-runners, but that might just need some investigation.

19:08 spieden: But a one-off "Run all tests" type thing is very easy.

19:09 My ultimate goal is a Wallaby-JS type thing.

19:45 tolstoy: I'm getting a 'Forbidden - redeploying non-snapshots is not allowed' on clojars, but I don't see my lib up there. Has something changed in the last couple of years?

19:46 justin_smith: tolstoy: just because you can't see the lib available doesn't always mean it isn't up there - you might want to see if any clojars maintainer is available on #leiningen

19:47 tolstoy: Ok.

19:48 justin_smith: Well, I can pull it down as a dep, so good enough. Oy. Signing.

19:54 justin_smith: tolstoy: a common one is that the version you pushed up doesn't show up in search (yet?) but if you go to the project/artifact page itself the latest version will be visible

19:55 tolstoy: Hm. Not there. Not even in the org.clojars.<username> part.

19:56 Perhaps the regenerate once an hour or so?

19:56 justin_smith: tolstoy: what about org.clojars.username/project

19:56 does that show the latest jar?

19:56 tolstoy: Nope.

19:58 The site is going down "this weekend".

19:58 justin_smith: tolstoy: oh, wait, I mean clojars.org/org-name/project

19:59 tolstoy: Either one. Try finding com.zentrope/quiescent

20:29 oracle123: I am using clj-webdriver & core.aysnc, in the -main function, I call a function will use firefox1, and then in the go block,I used firefox2. however I found the main thread used the wrong firefox instance.

20:29 please check the sample code: http://paste.lisp.org/display/231639

20:30 In each function I called set-driver! to bind a firefox instance to the thread running that function

20:31 But how could the function uses a wrong firefox instance?

21:46 tolstoy: Weird. I can "git pull" from github just fine, unless I ssh into this box. Then it prompts me for a passphrase (which I don't have).

21:48 Interesting. If I enter the passphrase for the account I'm logging in FROM, it works. Oy.

21:50 amalloy: tolstoy: my guess: the box you're SSHing to doesn't have your github ssh credentials on it, and you don't have your ssh client set up to forward the agent that's running locally

21:53 eriktjacobsen: Hey guys, little bit of an emergency: as of right now, my production uberjar is giving at runtime missing classes errors for my own classes… ie java.lang.NoClassDefFoundError: pipeline/youtube/api$get_bulk_channel_info$fn__354 where get-bulk-channel-info is definitely a function inside of that namespace. have been using latest versions of lein for awhile without issue. Also (and this may or may not be related), the size of the uberjar has

21:53 changed from like 110mb down to 62mb…. without any noticable changes in dependencies…. is lein leaving out my internal classes?

21:55 amalloy: i mean, a jar is just a .zip file. open it up yourself and see if that's what's happening

21:55 personally i'd try lein do clean, uberjar and see what happens - god knows what happens to stale classflies from AOT compiling

21:55 eriktjacobsen: mostly wondering if anyone has seen similar issues latetly with lein, we are looking into whats going on specifically

22:09 domgetter: I'm trying to come to an understanding of transducers. I understand that if I run (map inc) I will get back a transducer, and that it has three arities, and I know that one of them takes a "reduction function", so my question is: How would I construct a reduction function that reduces into a set?

22:14 turbofail: ,(reduce ((map inc) conj) #{} [0 1 2])

22:14 clojurebot: #{1 3 2}

22:15 turbofail: ,(transduce (map inc) conj [] [0 1 2])

22:15 clojurebot: [1 2 3]

22:16 turbofail: there's all sorts of fun ways to do it

22:17 ,(into #{} (map inc) [0 1 2])

22:17 clojurebot: #{1 3 2}

22:19 turbofail: i guess to answer your specific question, the answer is just ((map inc) conj)

22:19 domgetter: Okay let me play with that for a bit, thank you

22:20 oracle123: possible to get a var value for ALL threads?

22:20 since different thread could have different value, can we println all of hthem?

22:29 turbofail: hm... is there a reason the persistent set types don't implement IReduce?

22:38 domgetter: So what is the difference between

22:38 ,(reduce + 1 (->> [3 4 5] (map inc) (filter even?)))

22:39 clojurebot: 11

22:39 domgetter: ,(transduce (comp (map inc) (filter even?)) + 1 [3 4 5])

22:39 amalloy: you shouldn't really use IReduce

22:39 clojurebot: 11

22:39 amalloy: domgetter: the one with ->> builds a lot of cons cells and then throws them away

22:40 domgetter: ah okay so when the transducer grabs the 3, it works on it as a 3 instead of trying to work in seq-land?

22:50 amalloy: "you shouldn't really use IReduce" why is that?

22:51 amalloy: it's not part of clojure's "public" api, or wouldn't be if anyone ever bothered to declare parts of it non-public

22:51 you should just use clojure.core/reduce, which does the right thing (including sometimes dispatching to IReduce) based on the collection you give it

23:01 turbofail: amalloy: i wasn't intending to use IReduce directly, i was just wondering why sets don't do it

23:02 amalloy: it probably just didn't turn out to be necessary for them to implement it, in the bootstrapping code or whatever

23:02 turbofail: i don't think it's ever strictly necessary is it?

23:03 tolstoy: amalloy: Both of computers I'm using have ssh keys registered with github. It's weird. I've never seen this before. *shrug*

23:04 turbofail: IReduce is strictly a performance optimization as far as i can tell

Logging service provided by n01se.net