#clojure log - Feb 24 2016

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

0:28 Guest73345: I'm having a problem trying to create an uberjar with boot. I've set up a simple repository with a build.boot file containing a task called "build". After running "boot build" I don't find anything in my target directory. I'm following the instructions on several websites (https://lionfacelemonface.wordpress.com/2015/01/17/boot-getting-started-with-clojure-in-10-minutes/) but it doesn't seem to work. Can anyone point out what I'm do

0:28 wrong? Here's a link to the project on github: https://github.com/patterkyle/pietro/blob/master/build.boot

5:29 WickedShell: oO I just noticed that constant math expressions aren't evaluated at compile time... Is there anyway to get them to evaluate? I have some layout math in the hotloop that I did with some const def's as that was much more readable then hardcoding numbers, but since its actually doing the math per loop that's really painful...

5:32 TMA: WickedShell: you could try read-time evaluation

5:34 WickedShell: TMA I'm not quite following can you elaborate? Is there already a way to get this evaluated once instead of per loop?

5:37 TMA: WickedShell: say you have an expression (+ a (* 3 5)) ; you want to evaluate the (* 3 5) to 15 -- either you can (+ a #.(* 3 5)) or you can (defmacro eval-now [x] (eval x)) (+ a (eval-now (* 3 5)))

5:38 WickedShell: the #. is Common Lisp syntax, I would need to check, whether it is the same in clojure or whether it uses another squiggle

5:40 WickedShell: the first thing actually performs the evaluation at read time -- the compiler sees (+ a 15), while the second approach performs the evaluation at compile time

5:41 WickedShell: read time would be better, the first version causes the repl to choke though...

5:41 TMA: WickedShell: it's spelled #= in clojure

5:43 ,'(+ a #=(* 3 5))

5:43 clojurebot: #<RuntimeException java.lang.RuntimeException: EvalReader not allowed when *read-eval* is false.>

5:44 TMA: WickedShell: you might get this error when in restricted environment though :)

5:46 hyPiRion: ,(binding [*read-eval* true] '(+ a #=(* 3 5))) ; ??

5:46 clojurebot: #<RuntimeException java.lang.RuntimeException: EvalReader not allowed when *read-eval* is false.>

5:46 hyPiRion: mean

5:46 WickedShell: TMA it works on a simple expression bnut it appears to crash as soon as I put it in a expression that as ^:const member :/

5:46 TMA: hyPiRion: it has already been read by the time the binding comes in

5:47 WickedShell: oh, that means it's not a constant expression after all

5:48 WickedShell: and it can't evaluate nested expressions which is kinda weird :P #=(+ 3 (- 2 5)) wont work for example

5:48 TMA pretty sure it really is constant, its showing up in the bytecode successfuly replaced anyways

5:49 TMA: WickedShell: try the other approach with the named macro

5:51 WickedShell: there might be an issue with the #= that the thing to be evaluated is not yet in existence (I do not know when forms are evaluated when loading a clojure file, it might as well read the whole file before any of the forms therein gets evaluated)

5:52 WickedShell: TMA that still complies to a runtime eval based on the bytecode

5:52 RT.uncheckedFloatCast(((IFn)const__60.getRawRoot()).invoke(Numbers.unchecked_minus(Numbers.divide(-200L, 2L), 25L)))

5:53 TMA: WickedShell: could you share the source snippet?

5:55 WickedShell: TMA http://pastebin.com/yhBETpCm the actual float is used in an interop call but thats outside the scope

5:55 clojurebot: Alles klar

5:56 hyPiRion: TMA: oh yeah, true that

5:57 TMA: WickedShell: let me experiment a bit

6:00 somehow running lein repl from a cygwin terminal and pasting into that from the clipboard is very slow -- one character per second or so

6:01 would anyone know why that might be the case?

6:02 WickedShell: how did you obtain the RT.uncheckedFloatCast(((IFn)const__60.getRawRoot()).invoke(Numbers.unchecked_minus(Numbers.divide(-200L, 2L), 25L))) translation?

6:04 WickedShell: TMA jd-gui over the AOT'd result

6:05 I'm sure no.disassemble could also do it, but I haven't succeeded in making it work locally so jd-gui was the easiest way to go :P

6:07 TMA: WickedShell: try (defmacro eval-now [x] (eval x)) followed by (eval-now (float (- (/ HUD_WIDTH 2) 25)))

6:09 WickedShell: pretty sure thats what I tried (well the eval-now was inside the (float) will retry now

6:10 yeah thats what I did to get the disassembly I sent you

6:11 TMA: that's weird

6:11 WickedShell: if I dont put that macro in the invoke() call will go away but the math is still there

6:11 TMA: or maybe it is not

6:19 I don't seem to be able to compile ahead of time. My magic is weak

6:39 dazlow: How do I time function like "time" does but rather than print to stdout return time to me?

6:40 TMA: alas, when I figured it out, WickedShell is no more

6:40 andyf: dazlow: You could make your own slightly-modified version of the time macro that does that. It is pretty short and wouldn't be difficult to write a modified version.

6:42 MJB47: TMA: im still interested

6:43 andyf: dazlow: For example: (defmacro time2 [expr] `(let [start# (. System (nanoTime)) ret# ~expr] (/ (double (- (. System (nanoTime)) start#)) 1000000.0)))

6:43 dazlow: andyf: yeah, I'm writing macro now, thanks

6:45 TMA: MJB47: my (eval-now ...) approach does in fact work; if it is wrapped in a body of a function [say (defn do-foo [] (eval-now ...))], the compiled function doesn't do the computation. it has the result precomputed in static {} block and stored in const__1 static final field, which it loads and returns

6:45 MJB47: ty

6:48 TMA: MJB47: probably the difference is when the (eval-now ...) form is a top-level form

6:48 MJB47: i wonder if its actually due to inlining

6:50 TMA: MJB47: it should not be. macros perform transformations on the parse tree, which is then compiled after it is macro-free

6:51 MJB47: the eval-now forces the evaluation at compile time, inserting the reduced result into the tree, so the further stages of the compiler do not even see the original code

6:52 MJB47: in a sense it is manually inlining the result of the evaluation

9:56 rcassidy: TimMc: up() { for n ({1..$1}) do cd ..; done }

9:56 oops, that was waaaay in backscroll, didnt realize iw as reading the past

10:08 sdegutis: Anyone noticing wrap-file-info being highlighted with a red-ish background color by CIDER?

10:09 I just saw that this morning and don't know why it's happening, and the CIDER readme doesn't mention it at all.

10:10 rcassidy: is it because it's marked deprecated?

10:10 sdegutis: That was my first guess, but -- oh yeah, I see it now in the metadata, :deprecated "1.2

10:10 Thanks rcassidy.

10:11 Cool feature!

10:20 rcassidy: yeah, that's super neat.

10:22 sdegutis: Is it possible to replace (if a b c) with a solution that uses (and) and/or (or)?

10:22 Disclaimer: this is a homework-style problem and not practical or something I would ever consider using

10:25 rcassidy: ,(defn if-bad [a b c] (or (and a b) c))

10:25 clojurebot: #'sandbox/if-bad

10:25 rcassidy: ,(if-bad true 1 2)

10:25 clojurebot: 1

10:25 rcassidy: ,(if-bad false 1 2)

10:25 clojurebot: 2

10:25 TimMc: ,(if-bad true false 2)

10:26 clojurebot: 2

10:26 rcassidy: fiiiince

10:26 *fine

10:26 that's why I called it if-bad

10:26 xemdetia: if-lucky

10:28 Glenjamin: until python 2.5, that was the only way to do ternary expressions :D

10:28 TimMc: ,(({true (constantly 1)} (= 1 1) (constantly 2)))

10:28 clojurebot: 1

10:28 TimMc: ,(({true (constantly 1)} (not= 1 1) (constantly 2)))

10:28 clojurebot: 2

10:29 TimMc: (the fundamentals of how to do it in swearjure)

10:39 sdegutis: rcassidy: nice!

10:40 Glenjamin: in Go it still is

10:41 wutf: what's a better way to do this? (->> ... (apply concat) (apply concat) (apply concat) (apply concat))

10:41 sdegutis: Glenjamin: also, your map-based solution uses strict evaluation on the clause that won't be run

10:41 wutf: flatten

10:56 TimMc: ~flatten

10:56 clojurebot: flatten is rarely the right answer. Suppose you need to use a list as your "base type", for example. Usually you only want to flatten a single level, and in that case you're better off with concat. Or, better still, use mapcat to produce a sequence that's shaped right to begin with.

10:56 TimMc: wutf: ^

10:57 sdegutis: TimMc: either he has lists 4 levels deep and wants exactly only that gone away with, or he just wants the behavior flatten gives. I betted on the latter. It's a gamble I admit, but I took the chance anyway.

10:57 wutf: please confirm whether TimMc or I am right, so we can settle this dispute once and for good.

10:58 TimMc: Or you could do (for [a ..., b a, c b, d c] d) I think...

10:59 might need another level

11:01 sdegutis: TimMc: I don't think that'll produce the lists back, will it?

11:01 Hmm. I wonder now.

11:01 wutf: i am working on it

11:01 spion_: is there a macro that acts like `use` for namespaces, but only within the passed s-expression?

11:02 TimMc: ,(for [a [[[[4 5]]]], b a, c b, d c] d)

11:02 wutf: the lists are super nested (((((

11:02 clojurebot: (4 5)

11:02 TimMc: wutf: Can you avoid the extreme nesting in the first place?

11:03 spion_: e.g. (using ['ns1 'ns2 'ns3] (expression))

11:03 wutf: it's due to using nested for loops to destructure a nested hash map

11:03 TimMc: wutf: You know that for can take multiple clauses?

11:04 wutf: yeah i guess i knew that. i don't technically need to nest them, it will destructure them right?

11:04 if i just include the appropriate bindings

11:04 Glenjamin: wutf: can you show us an example of the hash-map you're destructuring?

11:05 wutf: it's literally a nested hash map, exported to edn from json in python

11:06 zerokarmaleft: perhaps get-in would get you most of the way

11:06 wutf: i tried using that but couldn't get it to work

11:06 Glenjamin: you can nest destructuring, i'm not clear why you'd need for loops

11:07 wutf: yeah so basically there is almost never a need for nested for loops right

11:07 trying this out

11:07 Glenjamin: there are some needs, but it's not clear whether or not yours is one of them

11:08 sdegutis: spion_: you can make a macro that creates a temporary namespace (and destroy it when done) and calls 'use' inside it.

11:08 zerokarmaleft: ,(let [[_ fruit _] (get-in {:foo {:bar {:baz ["cat" "apple" "car"]}}} [:foo :bar :baz])] fruit)

11:08 clojurebot: "apple"

11:08 sdegutis: spion_: I would advise against this however.

11:08 wutf: can you gist the json file or the hash-map itself?

11:09 TimMc: spion_: https://github.com/timmc/handy/blob/handy-1.7.0/src/org/timmc/handy.clj#L97 but I'd only use it for testing...

11:09 org.timmc.handy -- with-temp-ns

11:09 wutf: {"blah" : {"blerg" : {"bloom": {"blum": ...

11:10 i do have code that destructures it in one go using a doseq

11:10 sdegutis: wutf: there's gist.github.com

11:10 spion_: sdegutis: would be very convenient for things like om, e.g. (using ['om.dom] (dom goes here))

11:10 sdegutis: spion_: Personally I use (:require [some.lib :as lib]) and then just use (lib/foobar).

11:11 spion_: It's much easier to keep track of the symbols being used and where they come from.

11:11 spion_: also, you may want to split out some code into a namespace specific to using that library, where *all* the code in that file will be using that library.

11:11 TimMc: spion_: Yeah, don't use with-temp-ns for your main code. :-)

11:12 spion_: sdegutis: I think `using` would provide a compromise between ease of use and ease of tracking, as its very local so its only worth using if you are using many of those functions

11:12 (in that sense its better than :use)

11:12 Glenjamin: ,(let [{ { { { :keys [blum] } :bloom } :blerg } :blah } {:blah {:blerg {:bloom {:blum 'abc} } } } ] blum) # wutf

11:12 clojurebot: abc

11:13 sdegutis: spion_: except then you can't scan the top of the file to see what other namespaces it requires, since you mix in uses anywhere in a file that could be any number of pages long

11:14 wutf: Glenjamin i'm having a hard time reading that :)

11:14 spion_: sdegutis: it would have to be require'd

11:14 sdegutis: spion_: but if you're firm-set in creating a macro like this, use basically does a combination of clojure.core/require and clojure.core/refer, so you could use those functions yourself in your own macro.

11:14 spion_: I can write that macro for you for $200 right now and be done in 20 minutes.

11:14 wutf: you declared the structure of the data first, then provided the actual data, in the binding?

11:14 Glenjamin: yeah, the left-hand-side is the key bit

11:14 wutf: then had transparent access to blum. i also need blah, blerg, bloom.

11:14 spion_: sdegutis: no need, I was just wondering if it was already there and I'm missing something. Thanks :)

11:14 sdegutis: Surely.

11:15 Glenjamin: ,(let [{ { { { :keys [blum] } :bloom :as bloom } :blerg :as blerg } :blah :as blah } {:blah {:blerg {:bloom {:blum 'abc} } } } ] [blum bloom blerg blah])

11:15 clojurebot: [abc {:bloom {:blum abc}} {:blerg {:bloom {:blum abc}}} {:blah {:blerg {:bloom {:blum abc}}}}]

11:15 wutf: ah

11:15 neat

11:15 i don't really understand the :keys [blum] part

11:16 also not clear that this will actually loop over the entire map giving me access to all blah-blerg-bloom-blum combinations

11:16 Glenjamin: starts to get a bit weird at that point, so a big (let [blarg (get-in data [...]) blah (get-in blarg [...]) might be clearer

11:16 oh, you want to convert the entire map into locals?

11:17 wutf: yeah, i'm looping over the whole shebang and inserting it into datomic

11:17 i've done it like five different ways by now

11:17 trying to avoid side effects at this point

11:26 destructuring is magic :)

11:29 i piped the (((()))) mess to flatten flatten which was at least less verbose than (apply concat) (apply concat)

11:52 justin_smith: TMA: regarding the discussion last night about evaluating things at compile time, anything defined with def is fully evaluated at compile time

11:53 I mean I guess reader macros could help if you want compile time eval inside a lambda somewhere, but defs are already compile time expressions

11:58 sdegutis: I'm back.

12:01 xitrium: Anyone here use transducers much? to programmatically build up a series of steps, do you typically pass around both the combined steps and the collection until you're ready to realize it?

12:02 justin_smith: xitrium: I've never actually used them that way. I've used them to replace a series of steps with a single combined transformation though.

12:03 xitrium: yeah, that's my aim too - just my steps are generated at runtime

12:03 justin_smith: sounds like a great use of transducers, yeah

12:04 xitrium: sweet

12:10 sdegutis: Hi.

12:10 xitrium: what's the use-case btw? I'm intrigured by such a situation.

12:15 This just in: All of my code sucks.

12:18 Abstractly, how would you implement a loop which runs (f) every 2 seconds until stopped?

12:18 I'm specifically asking you justin_smith and Glenjamin.

12:19 Glenjamin: erm, do you already have core async in the project?

12:19 justin_smith: sdegutis: how accurate does the timing have to be?

12:19 sdegutis: Glenjamin: not yet.

12:19 justin_smith: not super accurate, just so long as it's no less than 2 seconds.

12:19 justin_smith: sdegutis: I'd use a ScheduledExecutorService (or a wrapper of that, like at-at)

12:20 Glenjamin: Probably Thread.new or future with a sleep then, and something to check for thread interrupts

12:20 sdegutis: Also I'm using Clojure 8 and Java 8.

12:20 Glenjamin: depends how throwaway this is i suppose

12:20 sdegutis: Not very throwaway.

12:21 Glenjamin: I'm not opposed to including core.async.

12:21 Glenjamin: probably go with justin's then

12:21 sdegutis: justin_smith: thanks I will look into that Java Class.

12:21 justin_smith: the advantage of ScheduledThreadpoolExecutorService or whatever it's called (I think I had it wrong above) is it gives nice tools for managing tasks and saying "start every X" or "run every X" (different things...)

12:21 sdegutis: justin_smith: it's called ScheduledExecutorService I think.

12:21 "An ExecutorService that can schedule commands to run after a given delay, or to execute periodically."

12:21 justin_smith: sdegutis: https://docs.oracle.com/javase/7/docs/api/java/util/concurrent/ScheduledThreadPoolExecutor.html

12:22 sdegutis: So yeah it sounds precisely as what i need.

12:22 justin_smith: that's a class that implements SES

12:22 so I was right both times, one is the interface, the other is the class

12:23 sdegutis: anyway, the class is easy to use, but depending on how you feel about adding deps on random thin wrappers, at-at is even easier to use

12:24 sdegutis: justin_smith: I think I'm somewhat similar to you philosophically in that I'm okay with using the Java class directly and dealing with slight Java ugliness in order to avoid adding an extra dependency.

12:25 Yes ScheduledThreadPoolExecutor is right.

12:35 This appears to be perfect.

12:36 justin_smith: gratitude

12:41 justin_smith: wow you weren't kidding! https://gist.github.com/sdegutis/d1109109bc56ab7ffb55

12:50 max3: if conversions is a map with functions as values (of the correct types) and vamp-key and value are both strings can someone tell me what this does: ((get conversions vamp-key) value)

13:00 justin_smith: sdegutis: nice, glad I could help - and yeah in my experience it is that simple

13:01 TimMc: hyPiRion: Do you remember how to implement boolean in swearjure?

13:01 justin_smith: max3: that looks up a function in conversions, and calls it with value as an argument

13:01 ,((get {:frob inc} :frob) 0) ; max3 - it is doing this

13:01 clojurebot: 1

13:02 TimMc: nvm, found it, I did the stupid thing: https://github.com/timmc/swearjure/blob/08895fd01e30b5c93225577ecb4165fd9dd3830e/src/org/timmc/swearjure.clj#L29-L31

13:09 max3: thank you justin_smith

13:17 sdegutis: Strange error.

13:17 "No matching method: SECONDS"

13:17 java.util.concurrent.TimeUnit/SECONDS

13:17 justin_smith: sdegutis: I think your doto is a bit off...

13:18 sdegutis: Sounds right.

13:18 justin_smith: sdegutis: SECONDS should be an arg, not something you call in the doto

13:18 so you need some parens somewhere

13:18 sdegutis: Ahh yes. I see. Completely forgot parens.

13:18 justin_smith: how did you know so quickly?

13:18 How did you know it was a doto?

13:18 justin_smith: sdegutis: I was gonna say something but figured it was just a hasty paste with a bad edit

13:18 sdegutis: you shared a paste earlier, also I am psychic so there is that too

13:19 sdegutis: Oh.

13:19 Haha.

13:19 * justin_smith gazes into his coffee grounds and sees your future.

13:19 justin_smith: "There are bugs, so many bugs"

13:19 "things throwing exceptions"

13:19 your future is now known

13:20 sdegutis: LOL

13:21 justin_smith: "the stack traces - ongodly deep stack traces seemingly without end"

13:22 sdegutis: Wow it's like you're reading the code right off my screen!

13:25 Yay now my email queue can be a component!

13:34 TMA: justin_smith: thanks, I did not knew that

13:38 justin_smith: sdegutis: on the psychic topic, I had the idea for an html widget to replace radio buttons and list selectors. It would be a magic eight ball (like the toy) and you click it and one of the options comes into view.

13:38 sdegutis: Haha! That is so impractical but fun.

13:38 Like 2048!

13:38 or http://mienfield.com/

13:40 That was in response to justin_smith.

13:43 justin_smith: TMA: common source of bugs in deployment is the expectation that things inside def are not run when aot compiling (they are, because def is always fully evaluated as part of compilation)

13:45 sdegutis: TMA?

13:45 The More A-you-know?

13:45 Time Matters Asynchronously?

13:45 Hold on don't tell me, I can figure it out.

13:45 justin_smith: sdegutis: too many apples

13:46 https://en.wikipedia.org/wiki/TMA

13:46 TMA: sdegutis: you can figure out something, but I doubt it would be it

13:47 rhg135: I agree with the last one, cuz apples

13:47 sdegutis: Oh! It's a nickname!

13:47 Haha I should pay more attention.

13:48 So, I randomly emailed a random programmer whose blog I found via HN the other day. Actually had a pleasant correspondence! 9/10 would recommend.

13:56 rhg135: I thought it was normal to not be unpleasant to strangers

13:58 sdegutis: rhg135: but it's not normal to email strangers and discuss code

13:58 rhg135: what's your favorite proggin lang?

13:59 rhg135: sdegutis: I'm surprised you have to ask

13:59 sdegutis: rhg135: ah so Lua?

13:59 TMA: rhg135: if you are right, I probably want to move to your country. which country that might be?

13:59 * rhg135 joust nods

14:00 rhg135: Just*

14:00 sdegutis: rhg135: meh Lua's ok, but I don't see why it's your favorite, that's a bit excessive

14:00 rhg135: Somewhere in wonderland

14:02 It's a midpoint between clojure and c

14:02 sdegutis: rhg135: true enough

14:03 justin_smith: I think the "politeness to strangers" thing mainly varies with population density. In less crowded places you have more incentive not to piss anyone off, in more crowded places you have less disincentive to be a jerk (both of thse because of probability of ever seeing the person again)

14:04 like, in a town with 100 people making one enemy is a big problem, in a city you can make many enemies and usually never see them again

14:04 sdegutis: Guys help me work on the edges of the hole at http://mienfield.com/948_-957

14:05 rhg135: If I get a random email, I don't reply with 'f*** off' I just won't reply. Maybe that's what you meant

14:06 sdegutis: rhg135: well a more polite version of that, yeah

14:06 rhg135: "sorry, I don't know you and this is weird, please don't ever email me again, good bye"

14:06 rhg135: which I did get once actually a few years ago

14:07 rhg135: Lol

14:08 sdegutis: rhg135: they're actually the more prudent ones of the group, since they know better than to talk to strangers

14:09 rhg135: My approach has been to lure them on irc and convince them it's pointless

14:09 sdegutis: rhg135: in this case I'm not creepy or deranged, but they never know -- I would answer similar to that one person

14:09 rhg135: what's pointless? you mean "pointless" or point-free style in Haskell?

14:09 rhg135: why are you talking about Haskell suddenly? where'd that topic change come from?

14:09 * sdegutis is confused and leaves

14:10 rhg135: Ah, context. I get inquiries about an ancient device and its software

14:11 sdegutis: Oh.

14:13 rhg135: It's pointless in the sense that even if by some miracle you got anything working, nobody knows wtf a gelato is (not the desert)

14:15 justin_smith: rhg135: the gelato desert is a harrowing wasteland of pistachios

14:15 (sorry)

14:17 rhg135: Mmm pistachios

14:17 sdegutis: :)

14:17 fwiw I don't know what a gilato is even as a dessert

14:18 justin_smith: sdegutis: gelato is italian style ice cream - more fat and less water than normal ice cream, the good stuff is like frozen butter with sugar and flavors

14:19 sdegutis: Oh weird.

14:19 justin_smith: it's so heavy that the scoops are much smaller usually (but just as filling)

14:19 sdegutis: Yesterday I learned that I really like Mexican food, which I never thought I would like. My wife made these chicken wraps, and they were delicious. It was chicken, mayo, salsa, cayanne pepper, a little cucumbers, and some honey, wrapped in a tortilla. It was great.

14:20 Also she makes these nachos that are just excellent. Hold on I'll get a photo for you.

14:21 Never mind I can't find it you'll just have to trust me on it.

14:21 justin_smith: it's for the best I am on an immutable data structure diet so it would just be a tease

14:23 rhg135: Food in general is mutable

14:23 justin_smith: rhg135: oh man that makes me think of 2g1c as swap! and uck

14:23 sorry for that image

14:24 rhg135: My brain can't form imagery yet, so it's all good

14:25 justin_smith: lucky

14:26 rhg135: I tried writing C once, I've never been three same

14:38 sdegutis: rhg135: false

14:39 rhg135: food runs through a digest function which hashes it into poop

14:40 My code has a really strange and ugly dependency tree.

14:41 Most of my components do not have any dependence on the app's primary business logic, so the graph strictly flows outward. But two do, and it's weird.

14:41 How do you solve this? Extract it into a function that you give the component which it can then call?

14:43 wouter_: hello, i was wondering if anyone can help me. I have written a macro in Clojure for use in Clojurescript, but I get Uncaught SyntaxError: Unexpected token . . A simplified version of the macro I was writing is: (defmacro test [x] `(defmulti ~x (fn [& _] :default))). can someone help me?

14:44 sdegutis: wouter_: hi

14:44 wouter_: hi sdegutils

14:44 sdegutis: wouter_: Check if ClojureScript has defmulti.

14:44 Maybe it does not?

14:45 wouter_: i have used defmulti and defmethod in clojurescript before

14:45 sdegutis: wouter_: Then I'm wrong.

14:46 wouter_: That was as much assistance as I can be of. Good day.

14:46 wouter_: alright thank you

14:46 sdegutis: wouter_: You're very welcome. Best of luck to you finding out your solution.

14:46 wouter_: thanks

14:47 justin_smith: wouter_: have you tried macroexpanding the macro?

14:47 where does it insert a . ?

14:47 wouter_: yes, that works actually. wait, i will give you the result

14:48 justin_smith: wouter_: I think your [& _] has to be [& _#] because `

14:48 wouter_: so i try the following in clojurescript: cljs.user> (macroexpand-1 '(nedarch.helper/dm bla)) (clojure.core/defmulti bla (clojure.core/fn [& nedarch.helper/_] :default))

14:49 justin_smith: [& nedarch.helper/_] is the problem

14:49 that is not a valid binding

14:49 if you change _ to _# that will be fixed

14:49 wouter_: alright, thanks a lot! i will try to see if it works!

14:52 no, it does not work unfortunately. even if i remove the & and _ for the purpose of testing and replace it with (fn [y] ...), it does not work unfortunately

14:52 justin_smith: wouter_: NO

14:52 you did not do what I said

14:52 replace _ with _#

14:52 or replace your y with y#

14:53 the problem isn't that _ is _, it is that it does not end with #

14:53 ,`(fn [x] x)

14:53 clojurebot: (clojure.core/fn [sandbox/x] sandbox/x)

14:53 justin_smith: ,`(fn [x#] x#)

14:53 wouter_: sorry, it works, i tried that first but somehow it was not compiled. this seems to work!

14:53 clojurebot: (clojure.core/fn [x__49__auto__] x__49__auto__)

14:53 justin_smith: see the difference?

14:54 wouter_: yes, thank you very much justin, i can now continue with my program :)

14:54 justin_smith: cool, sorry for the CAPS

14:54 haha

14:54 wouter_: no worries, thanks a lot! :)

14:56 sdegutis: You know, I've been meaning to write my Clojure-like language for a few years now. But I keep thinking to myself "nobody will use this, in fact I will probably rarely use this," and it's been convincing me to use my time for other things instead.

14:56 But I think it's time to just ignore that fact and start working on it.

14:57 Because even if it's not useful to me or anyone I know, it could eventually be useful to someone. Maybe. Or something.

15:23 ,`(fn [x#] x#)

15:23 clojurebot: (clojure.core/fn [x__25__auto__] x__25__auto__)

15:23 sdegutis: Cool.

15:23 ,`(fn [x#] (str 'x#))

15:23 clojurebot: (clojure.core/fn [x__50__auto__] (clojure.core/str (quote x__50__auto__)))

15:23 sdegutis: ,(`(fn [x#] (str 'x#)) "foo")

15:23 clojurebot: #error {\n :cause "clojure.lang.Cons cannot be cast to clojure.lang.IFn"\n :via\n [{:type java.lang.ClassCastException\n :message "clojure.lang.Cons cannot be cast to clojure.lang.IFn"\n :at [sandbox$eval76 invokeStatic "NO_SOURCE_FILE" 0]}]\n :trace\n [[sandbox$eval76 invokeStatic "NO_SOURCE_FILE" 0]\n [sandbox$eval76 invoke "NO_SOURCE_FILE" -1]\n [clojure.lang.Compiler eval "Compiler.java"...

15:24 sdegutis: ,((eval `(fn [x#] (str 'x#))) "foo")

15:24 clojurebot: "x__100__auto__"

15:24 sdegutis: Hahaha. Silly Clojure, escaping details.

15:24 * sdegutis turns that into a legit macro and uses it in production immediately

15:33 Kamuela: I think I'm looking for some pain tonight. Ok. So is it Emacs with Cider that's the ultimate setup?

15:33 TimMc: that'll get you some pain, sure

15:33 more the cider part of things

15:34 amalloy: well, if you're not familiar with emacs there's some pain learning that stuff

15:35 also if you do start learning emacs, my advice is to do something to make the mouse and arrow keys both un-usable and then press ctrl-h, followed by t, to start the tutorial

15:37 TimMc: Good point. Kamuela, how familiar are you with Emacs to begin with?

15:38 amalloy: I should go through the tutorial again at some point. :-P

15:44 sdegutis: Kamuela: in my opinion it's the best interactive experience, but it also needs paredit

15:44 Kamuela: TimMc: pretty much zero

15:44 sdegutis: Kamuela: ok then you're in for about 2 weeks of nothing but pain

15:44 amalloy: nah

15:45 Kamuela: sdegutis: is paredit not part of cider?

15:45 sdegutis: Kamuela: I'm super productive in Emacs with Paredit and Cider but I have absolutely no idea how to use it anymore, only my fingers remember how.

15:45 Kamuela: it is not

15:46 amalloy: sdegutis tries to be helpful sometimes but right now i think he is just trying to scare you

15:47 sdegutis: Kamuela: yah I was exaggerating just a bit.. expect to have some bumps in the road and frustration while teaching your fingers how to use emacs, but in about 1 or 2 weeks you should be pretty comfortable using it.. and in another 2 weeks you should have a decent ~/.emacs.d/

15:47 Kamuela: Is this a decent guide? http://www.braveclojure.com/basic-emacs/

15:47 * sdegutis shrugs

15:48 sdegutis: I forgot what guides I used for learning Emacs like 4 years ago.

15:48 justin_smith: Kamuela: the C-h t thing amalloy mentions (the built in tutorial) has the most important basics

15:49 Kamuela: I'll follow this guide then and make sure to do the built in tutorial

15:49 sdegutis: yeah definitely, although it's written terribly and extremely verbose and time-consuming

15:49 (the builtin guide)

15:50 It's almost as if it were written by RMS.

15:50 justin_smith: sdegutis: it needs big blocks of text so you can practice all the motion commands

15:50 sdegutis: I think it might have been written by RMS

15:50 sdegutis: justin_smith: that explains it then

15:50 justin_smith: if #2 is true then #1 is merely coincidence

15:50 Although I admit #1 is a real benefit.

15:51 rcassidy: i'm doing pretty OK with vim, paredit, and fireplace, if you're already somewhere in that camp, but i bet emacs/cider is awesome too. sounds like it's more integrated.

15:52 sdegutis: rcassidy: lately cider shows you the evaluated result in-line in your text file (highlighted as a comment) and it disappears as soon as you do anything, which is convenient

15:52 justin_smith: I've been using clojure-mode and evil, and might have to try the vim/fireplace thing one of these days. The new lady on my team uses vim/fireplace and it looks decent.

15:52 rcassidy: sdegutis: in-line in the file? as a comment? woah.

15:52 sdegutis: rcassidy: in the buffer, yeah

15:52 rcassidy: woah

15:52 sdegutis: it's super neat

15:53 justin_smith: kind of like light-table?

15:53 rcassidy: i'm not sure if i want that, but it sounds flashy

15:53 sdegutis: justin_smith: probably, I dunno

15:53 rcassidy: it's been very convenient, I have to admit

15:53 It's also got var-aware, namespace-aware auto completion.

15:54 It's pretty much a sweet IDE at this point.

15:55 It's also got a nice function to replace the result of your function call in-line, so if you type (vec (range 5)) and then press C-c C-w, it replaces it with [0 1 2 3 4] in your file.

15:56 That's not super useful but it's neat anyway for the few times you want to put literal data in your file for whatever reason.

15:56 TimMc: Kamuela: This is an emacs pre-guide that I wrote when I was first learning it: https://www.brainonfire.net/blog/emacs-n00b-start/ The idea is to read that first, then go to some other guide. This one just teaches you the basics like how to cancel, how to quit, and some terminology.

15:57 ystael: didn't emacs used to have C-x C-c, C-g, and C-h t mentioned on the startup screen?

15:57 sdegutis: Hmm. I wrote a starter kit slash cheat sheet a few years ago, but there was 0 interest in it so I deleted it.

15:57 justin_smith: ystael: it still does, but who actually reads that stuff :)

15:58 sdegutis: justin_smith, ystael: especially when you can just do (setq initial-scratch-message "" inhibit-startup-screen t inhibit-splash-screen t) to avoid seeing it

15:58 ystael: ooh, I didn't know about `initial-scratch-message`

15:59 sdegutis: You'd be surprised how many variables Emacs comes with, especially ones that seem a bit redundant.

15:59 justin_smith: sdegutis: if you know how to do that you probably don't need C-h t, but maybe someone made a bad choice when designing a starter kit you use, who knows

15:59 sdegutis: :D

16:00 justin_smith: meh I was more or less trying to make a bad joke

16:00 speaking of bad jokes

16:00 hold on im looking for it

16:02 i made this dad-joke this weekend http://i.imgur.com/oeLDBhB.jpg

16:12 Kamuela: i notice that i find ctrl on mac really hard to deal with

16:12 is emacs basically built around ctrl and alt commands?

16:13 justin_smith: Kamuela: yeah, it helps to swap control and caps-lock

16:14 postpunkjustin: Kamuela: how often do you use caps lock?

16:14 dammit double justin again

16:14 Kamuela: hmm

16:14 postpunkjustin: I'll get you one of these days, justin_smith

16:14 justin_smith: postpunkjustin: I'll rate limit and give you a chance

16:14 Kamuela: It's possible that I can just work through my disability for a bit because I do tend to use caps lock without thinking

16:15 postpunkjustin: Kamuela: that sounds... bad?

16:15 justin_smith: clearly a professional spammer or troll

16:15 Kamuela: haha

16:15 justin_smith: writes sql or basic or fortran?

16:16 Kamuela: sql and TOO_MANY_CONSTANTS

16:19 spuz: anyone here use Lighttable?

16:19 I am trying to run a test but it does not seem to load the definitions in the source file I am testing

16:19 how do i get it to include definitions from another file when running the tests?

16:20 justin_smith: spuz: you use require to require the namespace you are testing

16:21 spuz: justin_smith, i've done that

16:22 the test file works if I run it with 'lein test'

16:22 postpunkjustin: spuz: do you have to evaluate the ns declaration maybe?

16:23 spuz: postpunkjustin, maybe but it seems the state of one evaluation does not carry over to the next

16:23 when i evaluate the ns, I see the result is 'nil', but it doesn't seem to help resolve the function call when i evaluate the test definition

16:24 postpunkjustin: oh, that's really odd

16:24 I'm afraid I don't have much else to offer; I don't actually use LightTable

16:25 spuz: I will try something else

16:39 Kamuela: lol it said when you open emacs you may see a lot of activity while it downloads packages. so if I don't see it, here's my first hurlde

16:42 spuz: Kamuela, are you trying to use emacs for the first time?

16:42 I am investigating it right now

16:42 sdegutis: Kamuela: I have rebound caps lock to be control within Mac's system preferences -> keyboartd

16:43 spuz: do you have any guides to point to?

16:43 Kamuela: yeah I'm using this spuz http://www.braveclojure.com

16:43 but no one necessarily says it's the best or anything

16:43 sdegutis: haha

16:43 surely amalloy knows of some guide that he recommends or something

16:45 spuz: does spacemacs offer any advantage over vanilla emacs other than vim like keybindings?

16:45 postpunkjustin: spuz: most definitely

16:46 it also has better discoverability, sensible configurations for a lot of languages, basic package management

16:46 and some other stuff

16:46 spuz: postpunkjustin, would you recommend it to someone who is totally unfamiliar with emacs and only a little familiar with vim?

16:47 my primary aim is to have an intelligent clojure editor

16:47 postpunkjustin: I guess? It's hard for me to say since I was pretty familiar with both when I started using it

16:47 Kamuela: Does this look like an emacs config to anyone here? https://github.com/flyingmachine/emacs-for-clojure

16:47 j-pb: spuz: use intellij + cursi e

16:47 spuz: i.e. with all the shiny things like inline docs, syntax highlighting, repl integration etc

16:47 j-pb: cursive, it's not open source but gets you started in no time without having to follow down any config rabit hole

16:48 spuz: j-pb, cursive is great but I want to see what the free options are like before settling on the $120 price tag

16:49 I've used it for 30 days an am reasonably happy with it but it is a little clunky in some respects

16:49 Kamuela: spuz: are you using it commercially?

16:50 spuz: Kamuela, no

16:50 Kamuela: if you're honestly just dabbling or open source, apply for a license and you'll get one $0

16:50 spuz: hmm

16:50 what is the difference between a personal license and a non-commercial license?

16:51 Kamuela: personal is it's just for you but you sell software

16:51 non-commercial means open source or practice

16:51 personal projects

16:51 <- not a lawyer

16:52 spuz: great

16:52 well I can start coding again then :)

16:54 Kamuela: Should ~/.emacs.d be a directory full of themes?

16:58 justin_smith: Kamuela: I'd say .emacs.d is a decent place to put a theme, but it should also have your init.el (your config file) and other elisp code

16:58 Kamuela: justin_smith: I'm wondering if I set this up wrong by copying a directory into .emacs.d rather than the contents of that directory

16:59 justin_smith: Kamuela: that could be, maybe it wants themes to be in a subdirectory - that seems cleaner at least

17:00 you can probably explicitly tell the theme user where to find the themes, regardless

17:00 Kamuela: justin_smith: well that init.el thing is currently in .emacs.d/subdirectory-of-thing/init.el

17:00 justin_smith: subdirectory of what thing?

17:01 Kamuela: emacs looks for init.el in specific places, the lib might have its own init, but you should have an init.el as well, and that should be directly in .emacs.d

17:01 Kamuela: i will see what explodes by moving it up one

17:02 sdegutis: Ahh, I remember the first time I configured ~/.emacs.d/

17:02 I still have tons of "WTF" code in there that I just don't care about and it gets the job done so whatever.

17:02 justin_smith: hell I remember having to chance from ~/.emacs to ~/.emacs.d/init.el

17:02 sdegutis: mines all refactored into separate namespaces and stuff

17:03 sdegutis: justin_smith: yeah mine too, kinda sorta

17:03 justin_smith et al: This is my init.el https://gist.github.com/sdegutis/30b881ad500660c5abea

17:03 Terribly stupid way to load "modules" but meh it gets the job done.

17:04 justin_smith: right, pretty much how I do it, but I also have some conditionals for emacs versions and whatever

17:04 sdegutis: Plus things are not really in the right files, like necessities.el and startup.el and os.el are all conflated

17:04 justin_smith: Ahh I just use the latest and be done with that.

17:05 Also I have all of ~/.emacs.d/elpa under version control.

17:06 Honestly I did give that bundler thing a try, but I dunno it just didn't settle well.

17:07 Forgot what it was called

17:08 Kamuela: ok moving it to the root did the trick

17:08 I'm at the very least at this guy's setup

17:09 sdegutis: I think it was Cask?

17:10 Kamuela: well I think this setup includes paredit because that bit is happening

17:10 Not sure if the repl or cider bit is happening

17:11 sdegutis: Oh yeah! Cask used to be called Carton. That's when I was trying it out.

17:19 rhg135: sdegutis: I was perfectly fine without that image

17:19 sdegutis: rhg135: sorry what image again?

17:20 rhg135: "<sdegutis> rhg135: food runs through a digest function which hashes it into poop"

17:20 sdegutis: rhg135: oh, right

17:20 rhg135: hey dont blame me i didnt invent digestion

17:21 paul_muadib666: can somone help me with a cider issue? I have ring setup to start nrepl when i run 'lein ring server', when I later try to use 'cider-connect' to connect to the nrepl port, it claims cider-nrepl isn't installed

17:21 though I have cider-nrepl declared in my profiles.clj file

17:23 TristeFigure: Hi. I'm having a silly problem. I'm trying to (use 'some-name-space) I placed in the test folder, but I get FileNotFoundException Could not locate... The file is test/entomologist/gentests.clj and the ns statement reads like (ns entomologist.gentests ....) What am I doing wrong ?

17:24 Besides, there is nothing wrong with underscores in my file names

17:24 rhg135: sdegutis: to be pedantic digestion is hardly of type Food->Poop

17:26 amalloy: Food -> (Calories, Poop)

17:26 sdegutis: rhg135: Yeah amalloy's correct it's more of a monad

17:26 rhg135: and Gas

17:26 amalloy: that's not really a monad at all

17:26 yeah, (a,b) doesn't admit a monad instance

17:26 sdegutis: Oh, well fine then.

17:26 Then DONT implement digestion in Haskell for all I care.

17:27 rhg135: and calories would be an input

17:27 the canonical implementation is in DNA

17:30 sdegutis: amalloy: technically all this is in the BodyProcess Monad which abstracts away the details of how the body finances it

17:30 sdegutis: Um, did str/split change behavior in Clojure 1.8? I'm not seeing it in the release notes.

17:31 It's throwing an exception when I didn't think it should.

17:31 rhg135: It shouldn't have

17:32 sdegutis: Oh no wait I'm just stupid ignore me.

17:32 rhg135: typo?

17:32 sdegutis: Haha! Nope. Someone created invalid data in production.

17:32 Oh man. I should foolproof this stuff a little more.

17:33 rhg135: oh

17:33 * sdegutis goes into data-clean-up mode.

17:34 rhg135: this is why db's have schemas, too bad they don't work well enough

17:35 sdegutis: The more I look into it the more I think it's not actually an inconsistency in the database, it's just retarded code I wrote that assumed things which I guess are no longer true.

17:35 Also, Datomic is just plain awesome.

17:38 rhg135: it'd be rather cool to have a db enforce invariants at transaction time

17:38 TristeFigure: Resolved my own problem : I wasn't in my project folder :facepalm:

17:40 sdegutis: rhg135: what's an invariant btw

17:41 cortexman: why is it called figwheel?

17:41 rhg135: sdegutis: an INvariant

17:41 sdegutis: rhg135: sorry whats an INvariant

17:41 wNNNNNbc

17:41 csd_: Can someone explain what exactly is happening in clojure.main on the first line REQUIRE.invoke(CLOJURE_MAIN)? When I try to reason out what's happening I end up with invoke() being applied to an an instance of Unbound, which is an exception. I'm sure I'm missing some assignment somewhere but not clear exactly what I'm missing

17:43 rhg135: sdegutis: "a function of the

17:43 │ │ coefficients of one or more forms, which remains unaltered, when these

17:43 │ │ undergo suitable linear transformations

17:43 eh

17:43 sdegutis: rhg135: I'm scared

17:43 rhg135: sorry

17:43 sdegutis: rhg135: water under the bridge

17:44 hiredman: why do you think REQUIRE is Unbound?

17:44 rhg135: I've since learned to get a good dictionary

17:44 csd_: hiredman: a newly created var sets root = Unbound

17:45 hiredman: sure, but using RT causes RT's static initializers to run

17:46 csd_: oh maybe now i'm hitting the edge of my java knowledge. what exactly happens w/r/t what you're referring to?

17:47 hiredman: loading the RT class causes clojure.core to load

17:48 sdegutis: Fixed. Phew.

17:49 csd_: hiredman: oh i see that now. but still how does REQUIRE's root not get set to Unbound?

17:52 rhg135: csd_: you have to do something like REQUIRE = RT.var(...) not just instatiate a var

17:52 csd_: rhg135: this is me trying to understand the the opening line of main.java's main()

17:53 rhg135: ah, then nvm

17:54 csd_: it seems that RT.var creates var REQUIRE with root = Unbound but that some static magic in RT changes the value of root, or something

17:57 sdegutis: bye

18:08 Trioxin: why hasn't Clojure and Scala overtaken the popularity of Java? It only makes sense. Academia?

18:09 Java is pretty much #1 and the other two are nowhere on the map

18:10 csd_: i think being effective in clojure requires some knowledge of java.. i imagine scala might be similar

18:10 rhg135: csd_: https://github.com/clojure/clojure/blob/master/src/jvm/clojure/main.java#L20

18:11 csd_: rhg135: yeah that's the code i'm looking at

18:11 rhg135: that's what sets REQUIRE

18:12 csd_: rhg135: my question is about line 36. i cant find where REQUIRE is bound to anything prior to invoke()

18:12 rhg135: lines 19-23 are executed at class load

18:13 csd_: yeah i know

18:13 Trioxin: hell i would just stick with javascript since we have things like nwjs and cordova if concurrency weren't so damn annoying

18:13 csd_: but line 20 AFAICT only creates the var. i dont see where it binds it

18:15 rhg135: it doesn't. it looks it up in the environment and as hiredman said RT loads loads core

18:15 then at line 36 clojure.main is required through it

18:15 csd_: ok but where does it do that lookup

18:16 rhg135: RT#var gets a var by name

18:16 csd_: it interns the var

18:17 doesnt that only mean declaring it

18:17 RT.var = return Var.intern(...

18:18 rhg135: but if it is bound then it will be a bound

18:19 csd_: but where does that happen

18:19 rhg135: and RT loads clojure.core

18:20 thus #'clojure.core/require already exists and RT#var return the var already bound.

18:20 csd_: oh i see what you're saying

18:21 rhg135: and vars are mutable hence those vars like on line 21 are unbound until REQUIRE loads clojure.main

18:22 csd_: and so when load('clojure/core') is run, which files exactly are being loaded? i dont see any reference to the compiler so i dont think it can be the clj files yet

18:24 rhg135: I'm not sure, the clojure java code is a special kind of beautiful

18:25 csd_: haha

18:25 rhg135: the clojure clojure code though...

18:26 csd_: tell me about it... i was looking at the code for (destructure) earlier

18:26 rhg135: it's a mixed bag

18:26 csd_: if someone accustomed to haskell wrote clojure

18:26 but without the helpfulness of type signatures

18:27 rhg135: some of it is so minimal it's beautiful

18:33 csd_: thanks for your help

18:35 rhg135: np

18:57 Guest50097: hi, having this (defmacro mac [a] (list a 2 3 4)) I call it with (mac 1) and it returns (nil 2 3 4). Why?

18:58 hiredman: it doesn't do that, you are mistaken

18:59 Guest50097: hiredman: you mean it returns (1 2 3 4)

18:59 ?

18:59 hiredman: ,(defmacro mac [a] (list a 2 3 4))

18:59 clojurebot: #'sandbox/mac

18:59 hiredman: ,(mac 1)

18:59 clojurebot: #error {\n :cause "java.lang.Long cannot be cast to clojure.lang.IFn"\n :via\n [{:type java.lang.ClassCastException\n :message "java.lang.Long cannot be cast to clojure.lang.IFn"\n :at [sandbox$eval52 invokeStatic "NO_SOURCE_FILE" 0]}]\n :trace\n [[sandbox$eval52 invokeStatic "NO_SOURCE_FILE" 0]\n [sandbox$eval52 invoke "NO_SOURCE_FILE" -1]\n [clojure.lang.Compiler eval "Compiler.java" 6927]...

19:02 AimHere: ,(mac +)

19:02 clojurebot: 9

19:02 AimHere: ,(mac list)

19:02 clojurebot: (2 3 4)

19:05 Guest50097: :]

19:43 Quezion: Is there a way to work-around namespace collisions from protocols in another namespace?

19:43 E.g. Namespace A and B both protocol X. I want to use both versions of the protocol in the current namespace

19:45 If I require namespace A :as A, I still can't access the protocol. A/ProtoX/method naturally throws an error.

19:57 Is there a way to import/use two protocols with the same name into the current namespace?

20:07 justin_smith: Quezion: it's not A/ProtoX/method it's just A/method

20:07 or (.method thing)

20:07 Quezion: just like in Java you wouldn't type the name of the interface into a call on an interface method

20:11 Quezion: @justin_smith: That fixed it, thanks! I'm guessing you can't have a protocol method and a function of the same name in the same namespace, then?

22:16 srruby: How do I accomplish (.indexOf "dog" ["cat" "dog"]) ? answer should be 1

22:17 Java interop here

22:17 TEttinger: well you're calling that on the String "dog"

22:17 (.indexOf ["cat" "dog"] "dog" )

22:17 ,(.indexOf ["cat" "dog"] "dog" )

22:17 clojurebot: 1

22:18 TEttinger: (.indexOf "dog" \g)

22:18 ,(.indexOf "dog" \g)

22:18 clojurebot: #error {\n :cause "No matching method found: indexOf for class java.lang.String"\n :via\n [{:type java.lang.IllegalArgumentException\n :message "No matching method found: indexOf for class java.lang.String"\n :at [clojure.lang.Reflector invokeMatchingMethod "Reflector.java" 80]}]\n :trace\n [[clojure.lang.Reflector invokeMatchingMethod "Reflector.java" 80]\n [clojure.lang.Reflector invokeInst...

22:18 TEttinger: hm

22:18 well the main takeaway is .method takes the object with that method as its first parameter

22:19 srruby: ok parameter order was wrong. (.indexOf [:a :b :c :d] :c) works

22:20 TEttinger: Thanks

22:21 TEttinger: glad to help

22:34 paul_muadib666: two questions. what does :- mean? how can i search that pattern in google?

22:36 srruby: How come this works (update-in ["1" 2] [0] #(Integer/parseInt %))

22:37 but this doesn't (update-in ["1" 2] [0] Integer/parseInt)

22:38 TimMc: paul_muadib666: You probably found that in... let's see, something that uses schema?

22:39 paul_muadib666: TimMc: I believe so I'm playing around with compojure, yet me check the deps tree and see if anything uses that

22:39 TimMc: http://symbolhound.com/ sometimes helps but not in this case

22:39 TEttinger: srruby: because oddly, java interop method calls don't act like normal clojure fns. there is memfn to turn something like that into a fn

22:40 paul_muadib666: yes compojure-api has a dependency on prismatic/schema

22:40 TEttinger: ,(doc memfn)

22:40 clojurebot: "([name & args]); Expands into code that creates a fn that expects to be passed an object and any args and calls the named instance method on the object passing the args. Use when you want to treat a Java method as a first-class fn. name may be type-hinted with the method receiver's type in order to avoid reflective calls."

22:40 TEttinger: hm, not static

22:41 ,(update-in ["1" 2] [0] read-string)

22:41 clojurebot: [1 2]

22:41 srruby: TEttinger: Thank you.

22:42 TimMc: srruby, TEttinger: read-string is a bit over-powered here...

22:42 TEttinger: it looks like you're getting into the trickier parts of learning clojure, srruby

22:42 yes

22:42 TimMc: I would use Integer/parseInt for sure.

22:42 TEttinger: I just needed a clojure fn for that that could be passed as an example

22:42 TimMc: ah, ok

22:42 srruby: No, I just want to turn a string into an int.

22:42 TimMc: read-string has security implications, for one :-)

22:43 srruby: Right

22:43 TEttinger: since Integer/parseInt isn't quite the same as a fn... what is the recommended way for static members like that?

22:44 paul_muadib666: TimMc: I'm not having any luck finding that symbol on symbolhound :(

22:44 TimMc: TEttinger: #(Integer/parseInt %) :-P

22:44 or Long/parseLong, really

22:44 TEttinger: paul_muadib666: it is a keyword, and it might be used in a map

22:44 yeah

22:45 ,(:- {:- 1 :D 2})

22:45 clojurebot: 1

22:45 TimMc: paul_muadib666: What's the context?

22:45 paul_muadib666: TEttinger: oh. that makes sense. I assumed a reader macro or something

22:45 TEttinger: no, that would be # something

22:46 the only cases where reader macros don't start with # are probably things you use all the time

22:47 paul_muadib666: TimMc: going through this tutorial http://nblumoe.github.io/mundane-clojure-for-mortals/org/api-for-mortals.html#sec-7-1 , the symbol is in the last code block

22:47 TEttinger: https://yobriefca.se/blog/2014/05/19/the-weird-and-wonderful-characters-of-clojure/

22:48 paul_muadib666: nice, yall are full of choice links today :)

22:48 TimMc: paul_muadib666: http://clojure.org/reference/reader is a good thing to look through now and then -- that's effectively the page on syntax

22:48 TEttinger: ah! https://github.com/metosin/compojure-api

22:49 that uses :- and is require'd in that link

22:49 TimMc: paul_muadib666: Oh, looks like it's a keyword that has special meaning for that GET* macro.

22:52 paul_muadib666: TEttinger: what do you mean it's require'd ?

22:53 TEttinger: oh, either :require or :use , didn't checj

22:54 I just noticed that that API was being used from the code

22:54 paul_muadib666: ok, so if i check ns-publics in that namespace it should be there

22:55 welp library's closing, thanks for hte help TimMc and TEttinger!

22:55 TEttinger: probably. it looks like they might have not written out namespace GET* is in, which is the one that has the funny meaning for :-

22:55 later!

23:21 kenrestivo: what's the most clojure-ish way to deal with configuration management in a large clustered app?

23:22 rhg135: Environ I think

23:23 kenrestivo: for large clusters? i've used environ. i found it limiting, now i use large edn conf files (with schema). but i'm not sure how this would scale

23:23 tolstoy: Are you talking about multiple services that consult a central service for configuration?

23:23 kenrestivo: something like that might be necessary. or is it?

23:24 tolstoy: IMHO, as soon as you start asking that question, you're into "domain specific" solutions that depend on so many things outside the language of the app. But maybe I'm misunderstanding?

23:24 rhg135: Oh for clusters, hmm

23:27 tolstoy: Hm. Something like puppet/chef plops an EDN file in a known-location on every filesystem, and you slurp it in? ;)

23:30 A long time ago, someone made a Zookeeper thing that had atom semantics. I bet that's fallen by the way side.

23:30 kenrestivo: i mean, there's etcd in coreos, for example. there's also chef/puppet/ansible. there's also just putting conf files in git and scp'ing them up, which is what i've been doing.

23:30 my main curiosity is in clojure specifically

23:30 how do people do it? what's the best practices?

23:30 in other languages, there's a whole gaggle of different things all competing with each other. in clojure-land, what's the answer? i'll bet it's somethign simple and elegant

23:30 then again, maybe the fact that this hasn't been a problem for me means that it just isn't a problem in clojure-land.

23:32 tolstoy: What's an example in another language?

23:32 kenrestivo: etcd, chef, puppet, ansible

23:33 some people also use redis for that stuff

23:33 tolstoy: Ah, so you're wanting a version of one of those, but written in Clojure?

23:33 kenrestivo: no, and i'm having a tough time making the question clear

23:33 it's a devops question.

23:35 rhg135: I'd think synchronizing some data would be rather common actually

23:35 tolstoy: Yeah. Sorry. I assume configuration management is independent of the language a thing that needs to be configured is written in, hence my confusion.

23:36 kenrestivo: if you're rolling out a system with a bunch of microservices, dependent on databases, ip addresses, load balancers, etc, how do i get all that configuration to the clojure artifacts running the services? right now i just do "java -jar someuber.jar config.edn" and i have the config pushed out via scp.

23:36 it's been fine for my small apps. i wonder how the Big Guys do it with large, hairy clojure apps.

23:36 tolstoy: Yeah, that's what I do (now that I'm small time).

23:37 In the past, I've had an RPM what had the central config with a key for individual apps to pick out, then just make apps depend on that.

23:37 But, datomic would be an intereting choice. Maybe.

23:38 kenrestivo: thanks. maybe the only people who'd really know are buried deep inside cognitecht or somewhere, and they're not going to share their secrets.

23:39 who builds massive clojure apps with dozens or hundreds of instances?

23:43 tolstoy: Here's this thing, but you can see it's bitrotted: https://liebke.github.io/avout/

23:47 kenrestivo: thanks. i think *shrug* is a perfectly valid-- and in this case, probably the most accurate-- answer.

23:57 solatis: kenrestivo: it's fairly simple, actually

23:57 i'm not sure if we qualify as "big guy", but we have around 100 servers / services

23:57 we use AWS CloudFormation

23:58 a *huge* template

23:58 and it automatically writes a config.edn

23:58 that is indeed imported in the way you describe

23:58 and then we use luminus/config to read out the config files

23:59 if you want to get inspiration how the Really Fucking Big Guys do it, look at Erlang / Elixir and how they do configuration management

23:59 but it's hairy, and full of "service lookup services", etc

23:59 i like the simplicity of working with CloudFormation

Logging service provided by n01se.net