#clojure log - Aug 09 2015

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

0:01 rpaulo: grrr, cat again

2:22 domokato: hm, when i put in a hex value, 0xFFBEBEBE for example (for an argb color), I can't convert it to an int because I get a "value out of range" exception

2:23 ,(int 0FFBEBEBE)

2:23 clojurebot: #<NumberFormatException java.lang.NumberFormatException: Invalid number: 0FFBEBEBE>

2:23 domokato: er

2:23 ,(int 0xFFBEBEBE)

2:23 clojurebot: #error {\n :cause "Value out of range for int: 4290690750"\n :via\n [{:type java.lang.IllegalArgumentException\n :message "Value out of range for int: 4290690750"\n :at [clojure.lang.RT intCast "RT.java" 1198]}]\n :trace\n [[clojure.lang.RT intCast "RT.java" 1198]\n [sandbox$eval47 invokeStatic "NO_SOURCE_FILE" 0]\n [sandbox$eval47 invoke "NO_SOURCE_FILE" -1]\n [clojure.lang.Compiler eval "...

2:24 hiredman: ,(Integer/toHexString Integer/MAX_VALUE)

2:24 clojurebot: "7fffffff"

2:25 hiredman: ,(unchecked-int 0xFFBEBEBE)

2:25 clojurebot: -4276546

2:26 domokato: sweet, thanks!

2:27 man, that's not even in the cheatsheet

2:28 hiredman: if everything was on whatever cheatsheet, the cheatsheet would just be a listing of everying in clojure.core

3:09 irctc: Hello, quick question : I am trying to import core.async in my REPL : (require '[clojure.core.async] :as async) what's wrong with this ? (runtime exception unable to resolve symbol: async)

3:10 this seem to work though at least doesn't throw an error : (require '[clojure.core.async])

3:11 aah nevermind, it's me being stupid : (require '[clojure.core.async :as async])

4:03 Hello, can someone help me debug my protocol ? I can't find the right way to call it https://www.refheap.com/7739d3471f05b3681fde6b0bf

4:04 justin_smith: irctc: the first argument to every protocol method must be the object implementing the protocol

4:04 irctc: traditionally we name it "this"

4:05 irctc: also, the protocol method is a function, it is not an argument to the object implementing the protocol

4:05 irctc: Do you mean I should do (add-subscriber subscription-store id context) instead ?

4:05 hmm thanks just let me 30 secs to process it ^^

4:05 justin_smith: irctc: that, and you also need to change the declaration and definition of add-subscriber

4:06 the declared args must include the this arg

4:06 irctc: ooh I see

4:07 every time, I should have this as first arg ok

4:07 and then how will I call it ? like I pasted here ?

4:07 justin_smith: yes

4:07 irctc: justin_smith: thanks a lot I'm going to try that now

4:07 justin_smith: it's OK not to call the arg "this" - eg. if you don't use the arg, it can be called _

4:09 irctc: Oh I saw that convention thanks again. I shouldn't need it right ? (the memory store will have a redis store counterpart in case it matters-

4:13 justin_smith: Seem to work ! Thanks a ton !

10:58 lodin_: Is it possible to construct a value x such that (let [[& {a :a}] x] ...) makes sense?

11:06 Or as a test, can (is (nil? (let [[& {a :a}] x] a))) fail?

12:17 justin_smith: lodin_: what would an ampersand in a let destructure even mean?

12:19 AimHere: ,(let [[& {a :a}] x] a)

12:19 clojurebot: #error {\n :cause "Unable to resolve symbol: x in this context"\n :via\n [{:type clojure.lang.Compiler$CompilerException\n :message "java.lang.RuntimeException: Unable to resolve symbol: x in this context, compiling:(NO_SOURCE_PATH:0:0)"\n :at [clojure.lang.Compiler analyze "Compiler.java" 6704]}\n {:type java.lang.RuntimeException\n :message "Unable to resolve symbol: x in this context"\n ...

12:19 AimHere: ,(let [[& {a :a}] 10] a)

12:19 clojurebot: #error {\n :cause "Don't know how to create ISeq from: java.lang.Long"\n :via\n [{:type java.lang.IllegalArgumentException\n :message "Don't know how to create ISeq from: java.lang.Long"\n :at [clojure.lang.RT seqFrom "RT.java" 535]}]\n :trace\n [[clojure.lang.RT seqFrom "RT.java" 535]\n [clojure.lang.RT seq "RT.java" 516]\n [clojure.core$seq__4116 invokeStatic "core.clj" 137]\n [clojure.co...

12:20 AimHere: ,(let [[& {a :a}] {:foo 10 :bar 20}] a)

12:20 clojurebot: nil

12:21 justin_smith: ,(let [[& {a :a}] [:a 0]] a) ; wut

12:21 clojurebot: 0

12:21 justin_smith: it makes me kind of unhappy that that works

12:40 lodin_: justin_smith: Hmm, I thought I tried that one.

12:40 Apparently I didn't.

12:45 ,(let [{a :a} (seq [:a :A])] a)

12:45 clojurebot: :A

12:50 lodin_: but,

12:50 ,(let [{a :a} [[:a :A]]] a)

12:50 clojurebot: nil

12:51 lodin_: Presumably because [] is associative, and there's no :a in the vector.

12:51 justin_smith: ,(let [[{a :a}] [[:a :A]]] a)

12:51 clojurebot: nil

12:51 lodin_: ,(let [[{a :a}] [(seq [:a :A])]] a)

12:51 clojurebot: :A

13:34 kwmiebach: Hi. What does the cljr extension stand for? it is mentioned here: http://clojure.org/reader#The Reader--Reader Conditionals

13:37 hyPiRion: clojure clr probably

13:37 http://clojure.org/clojureclr

13:41 kwmiebach: thank you. and my link above was broken, it should be http://clojure.org/reader#The%20Reader--Reader%20Conditionals

13:49 yeah probably :cljr is reserved for the dotnet implementation in reader conditionals

14:02 gfredericks: are there any strong opinions about indenting styles for top-level reader conditionals?

14:08 justin_smith: gfredericks: I lament bitter tears over emacs not being able to indent them like hash-maps

14:10 gfredericks: what is backing the emacs identation stuff currently?

14:22 justin_smith: gfredericks: the magic happens here https://github.com/clojure-emacs/clojure-mode/blob/master/clojure-mode.el#L638

14:45 lentils_: anyone here experienced with enlive?

14:52 lodin_: ,(let [[a :as x b] [1 2]] a)

14:52 clojurebot: 1

14:52 lodin_: ,(let [[a :as x b] [1 2]] x)

14:52 clojurebot: [1 2]

14:52 lodin_: ,(let [[a :as x b] [1 2]] b)

14:52 clojurebot: #error {\n :cause "Unable to resolve symbol: b in this context"\n :via\n [{:type clojure.lang.Compiler$CompilerException\n :message "java.lang.RuntimeException: Unable to resolve symbol: b in this context, compiling:(NO_SOURCE_PATH:0:0)"\n :at [clojure.lang.Compiler analyze "Compiler.java" 6704]}\n {:type java.lang.RuntimeException\n :message "Unable to resolve symbol: b in this context"\n ...

14:52 lodin_: ^^^ Bug or features?

14:52 *feature.

14:52 justin_smith: lodin_: nothing can follow :as

14:53 lodin_: justin_smith: Why not?

14:53 justin_smith: because after ":as x" the destructure ignores all input

14:53 lodin_: justin_smith: But why?

14:53 justin_smith: so anything can follow it I guess, it just won't do anything

14:54 lodin_: because it was written that way. The destructuring code is weird.

14:54 gfredericks: lodin_: it's an impl detail, and is confusing, but doesn't prevent anything

14:54 lodin_: It's not clear from the docs. But when I read "http://clojure.org/special_forms#Special Forms--Binding Forms (Destructuring)-Vector binding destructuring" I see that you could interpret "Finally, ..." as meaning "in the last position".

14:54 gfredericks: i.e., (let [[a b :as x] [1 2]] b) is what you really want, right?

14:55 lodin_: I actually prefer to put :as first.

14:55 I always do that with maps.

14:56 gfredericks: yeah I just mean that it means the same thing

14:58 lodin_: gfredericks: Right.

14:58 Do you think a patch to allow :as anywhere, or at least in the beginning, would be accepted?

15:00 justin_smith: I doubt it

15:01 gfredericks: more likely a patch that throws if you put anything after the :as clause

15:01 justin_smith: yeah

15:01 gfredericks: which won't give you what you want of course

15:02 justin_smith: the only reason :as doesn't need to be last in a map is because maps are unordered I guess

15:02 lodin_: I don't really understand the rationale for requiring :as to be the last thing.

15:02 gfredericks: I haven't thought about it till not but

15:03 you could argue that the normal mode of sequential destructuring where each thing is positional gets interrupted in a confusing way if you allow arbitrary clauses in between

15:04 justin_smith: destructure is already insane and hairy as code, it would be worse I bet if it had to accept :as anywhere

15:04 gfredericks: it's visually more confusing because in [a b :as c d e f]

15:04 c and d are *very* different

15:04 lodin_: Yeah. Which is why I also gave the option of "in the beginning".

15:05 gfredericks: same for [:as c a b d e f]

15:05 lodin_: I agree that it would be confusing to have keyword in the middle there, particularly when the subforms are not just symbols.

15:07 Interesting. I find the leading style to be much clearer, because when you have even just slightly non-trivial destructuring the :as clause gets hidden far to the right.

15:07 justin_smith: lodin_: I always put :as on its own line

15:08 lodin_: like [a {foo :bar {this :that}} :as v].

15:08 justin_smith: like that, but with a newline before :as

15:09 lodin_: Yeah, I was giving an example of v being pushed to the left. Maybe I'm bad at reading destructurings, but I need to look back left to see what v actually contains.

15:10 If I get :as v in the beginning, I know that "what I'm reading now is part of v", instead of having to backtrack to update when v started.

15:11 justin_smith: It makes sense to put :as on it's own line. I should try that.

15:11 I've tried to not use multiline destructurings, because I find it disrupting when reading the let. But I guess that's something I can get used to.

15:12 justin_smith: How do you layout the init-expr that follows the binding-form when binding-form is multiline?

15:14 justin_smith: usually on the same line as the :as clause, depending on how the rest of the let is laid out

15:16 lodin_: Ah. Lately, when I have let forms that are too long (usually because it's a vector with long names in it), I have put the init-expr on the next line, but indented.

15:20 gfredericks: are macros just a lot harder to write in cljs and that's just the end of the matter?

15:20 (w.r.t. symbol expansion)

15:46 justin_smith: I've definitely noticed that things that work in clojure macros sometimes don't work in cljs ones

15:47 gfredericks: I'm starting to think that you just can't use the autoexpansion feature of ` anymore

15:47 which I'm sure encourages simpler macros

15:48 justin_smith: gfredericks: time to port hygeinic macros from scheme

15:58 amalloy: gfredericks: you mean instead of syntax-quote you just get backquote?

16:00 gfredericks: amalloy: what's backquote?

16:00 amalloy: like form other lisps

16:00 *from

16:00 gfredericks: oh I'ven't another lisp

16:01 amalloy: just like, `(a b c) returns (a b c) instead of (foo/a foo/b foo/c)

16:02 gfredericks: yeah, you only get the unquoting and the gensym features

16:02 but what this means in real life is that you still have the symbol expansion feature, it just usually does the wrong thing

16:02 e.g. `(* 2 3 7) is wrong I assume

16:02 * gfredericks hasn't tried it

16:10 gfredericks: I think I'm wrong about that, but I have no idea why

16:10 I.e., I'd assume that expands to ##`(* 2 3 7)

16:10 lazybot: ⇒ (clojure.core/* 2 3 7)

16:10 gfredericks: and clojure.core/* doesn't exist in cljs

16:29 aconz2: off topic to discuss how to find jobs as a soon-to-be college grad?

16:34 gfredericks: geography might (or not?) be relevant to that question

16:37 aconz2: currently in Maryland, 90% sure I'll be in Tennessee in 6 months, but ideal is remote

17:10 gfredericks: does anybody understand maven and/or clojure's maven setup well enough to figure out why this commit broke `mvn test`? https://github.com/clojure/test.check/commit/adb0883101e9bb3216539f1f597f6c3075c251f6

17:14 I can't tell if it fails while trying to run some kind of cljs thing, or fails because there are now extra files under src/main/clojure/

17:18 this must have something to do with the pom.contrib thing

17:20 hiredman: well, that is a project.clj for lein, likely what is in the maven pom is out of sync with the changes to project.clj

17:20 gfredericks: hiredman: it's also a file renaming

17:21 looks like it's the extra cljs files under /src/test/clojure that are doing it

17:22 I'm guessing this is controlled by the clojure-maven plugin

18:18 tmtwd: would there be a way to print the value at the key of y in this example?

18:53 justin_smith: tmtwd: which?

18:54 tmtwd: http://pastebin.com/Brw6mEZA

18:55 nvm mind though, I think its impossible/unimportant to do what I wanted to do

18:55 zardoz2: if 3.7V means about 50% of the charge is still there (does it?) , can I rely on it always meaning roughly that, regardless of the age of the battery and the number of cycles?

18:55 uh wrong channel

19:30 tmtwd: so, assoc-in is for vectors, and assoc is for hashes?

19:33 justin_smith: tmtwd: assoc and assoc-in are both for associative collections, associative collections include vectors and hash-maps

19:33 ,(assoc [] 0 :a)

19:33 clojurebot: [:a]

19:33 tmtwd: but they are not interchangeable, i'm trying to see the difference

19:33 justin_smith: ,(assoc-in [] [0 0 0] :a)

19:33 clojurebot: [{0 {0 :a}}]

19:34 justin_smith: tmtwd: assoc-in does a series of nested assoc

19:34 tmtwd: both default to create a hash-map with nil, but are able to set a numeric index on a vector

19:34 ,(assoc [] 0 :a)

19:34 clojurebot: [:a]

19:35 justin_smith: ,(assoc-in [[[]]] [0 0 0] :a)

19:35 clojurebot: [[[:a]]]

19:35 justin_smith: ,(assoc nil :a 0)

19:35 clojurebot: {:a 0}

19:36 tmtwd: but, if I have this: (def users [{:name "James" :age 26} {:name "John" :age 43}]), this fails : (assoc users [1 :age] 44)

19:36 i need to use assoc-in in that case

19:36 justin_smith: tmtwd: right, that needs to be assoc-in

19:36 tmtwd: oh I see

19:36 justin_smith: but that isn't because of vectors versus hash-maps

19:36 tmtwd: assoc-in is more "recursive"

19:36 justin_smith: right

19:36 it digs in

19:36 tmtwd: or its like a "deep" version of assoc

19:37 justin_smith: exactly

19:37 tmtwd: cool thanks

19:37 justin_smith: tmtwd: this is incredibly valuable when you have nested immutable data structures

19:39 tmtwd: i see

19:41 justin_smith: because otherwise you end up writing a lot of boilerplate to get the structures out and put the other stuff in and rebuild it...

20:47 jefelante: Is it a bad idea to use a transducer to write an EDN file for data that gets put on the channel and put it back as-is? https://gist.github.com/anonymous/623164e2ffbbec86432f

20:51 justin_smith: jefelante: you could use mult/tap instead

20:51 jefelante: so tap1 = file writer and tap2 = processor?

20:51 justin_smith: right

20:52 also, the fact that you only have one channel means you can't retrieve the data in parallel

20:52 maybe you actually want that

20:52 jefelante: i dont really understand

20:53 justin_smith: jefelante: you have a single data processing pipeline, where every task is independent, and they could otherwise be done in parallel

20:53 jefelante: oh right

20:53 justin_smith: so you've artificially slowed down your processing (unless there is a pragmatic reason to only make one request at a time)

20:53 jefelante: so with mult/tap, I could have the file writer and the processor happening concurrently

20:54 justin_smith: yes, but with a different design you could be doing all the downloads concurrently

20:54 whether that's an improvement is another can of worms of course

20:54 jefelante: right, i see what you mean

20:55 is there another benefit to mult/tap besides just working concurrently? the thing i'm wondering is if maybe its foolish to introduce a transducer that has a side effect

20:55 justin_smith: jefelante: an idiom I sometimes use is (dotimes [i parallelism] (go ...)) so that the number "parallelism" decides how many parallel data pipelines are created

20:56 jefelante: transducers and side effects are fine, but with mult/tap the two tasks can go at their own speed

20:56 which may or may not be a good thing...

20:56 but usually if one can finish much faster, you'd like it to do so

20:56 jefelante: ok cool

20:57 justin_smith: jefelante: also, your task (storing the data on disk as you do the transactions) is something I am doing in my project at the day job right now

20:58 we are using kafka which stores all messages for a configurable time period on disk

20:58 and then using transit to encode the body of the messages

20:59 jefelante: ive read about kafka, but i have this issue where i have trouble understanding something complicated (like kafka) unless i make the mistakes that necessitated building that thing

20:59 first

20:59 justin_smith: heh, you have a lot of mistakes (or learning one way or another) before you could use kafka, most likely

20:59 distributed systems are very hard

20:59 I'm still very bad at them...

21:00 jefelante: yea, speed isn't an issue for me since i'm rewriting a python script that is plenty fast and runs sequentially, but i'm doing it to try and learn how a larger/async system might work

21:00 justin_smith: me trying to do hammock driven development on my latest project: http://i.imgur.com/8ryWyNJ.gifv

21:01 jefelante: it makes one request every ~15 minutes, extracts some information, and occasionally makes additional follow up requests based on that information

21:01 justin_smith: jefelante: yeah, in that case do the linear version first, and be aware of your options if you want options for speeding it up in the future

21:02 jefelante: also, though mongo has a bunch of gotchas, it is very easy to use it to store a map of vanilla clojure data as a document for a system like yours

21:02 with a bit of the fiddly stuff about using files taken care of

21:03 jefelante: ahh that's an interesting point

21:07 justin_smith: thanks

21:16 tmtwd: I have an atom http://pastebin.com/HFjCJNZF and I want to assoc such that a new item is appended with a new value

21:16 whats the best way to achieve this?

21:16 justin_smith: update-in / conj

21:16 tmtwd: ah

21:18 justin_smith: ,(update-in {:items [{:item "program in clojure"}]} [:items] conj {:item "yolo"})

21:18 clojurebot: {:items [{:item "program in clojure"} {:item "yolo"}]}

21:18 justin_smith: or with clojure 1.7, there is also update

21:18 ,(update {:items [{:item "program in clojure"}]} :items conj {:item "yolo"})

21:18 clojurebot: {:items [{:item "program in clojure"} {:item "yolo"}]}

21:19 tmtwd: cool

21:19 that worked :)

21:19 thanks

21:20 axsk: hey, i need a macro which expands (mymacro a b c) to "a b c" (i need a macro here, as i dont want to eval the symbols, right?) but cant get there since hours. anyone has an idea?

21:21 namra: ,'(a b c)

21:21 clojurebot: (a b c)

21:22 justin_smith: ,(defmacro symstring [& args] (clojure.string/join \space args))

21:22 clojurebot: #'sandbox/symstring

21:22 justin_smith: ,(symstring a b c foo)

21:22 clojurebot: "a b c foo"

21:23 justin_smith: axsk: the thing to remember is that the macro gets the literal form as its input, and should return the form that should be evaluated.

21:28 axsk: wow im surprised by the ease of the answer, (though i still dont understand :>) will come back if i cant work it out, thanks a lot

21:28 justin_smith: axsk: it's not a hard macro to figure out :)

21:28 I mean in terms of seeing what it does

21:28 of course macros are hard, period

21:30 axsk: the args to the macro are a sequence of symbols (thanks to the & destructuring), and the macro simply puts those symbols into a string

21:30 ,(symstring (+ 1 1) a b (inc 0))

21:30 clojurebot: #error {\n :cause "Unable to resolve symbol: symstring in this context"\n :via\n [{:type clojure.lang.Compiler$CompilerException\n :message "java.lang.RuntimeException: Unable to resolve symbol: symstring in this context, compiling:(NO_SOURCE_PATH:0:0)"\n :at [clojure.lang.Compiler analyze "Compiler.java" 6704]}\n {:type java.lang.RuntimeException\n :message "Unable to resolve symbol: symst...

21:30 justin_smith: ,(defmacro symstring [& args] (clojure.string/join \space args))

21:30 clojurebot: #'sandbox/symstring

21:30 justin_smith: ,(symstring (+ 1 1) a b (inc 0))

21:30 clojurebot: "(+ 1 1) a b (inc 0)"

21:30 justin_smith: it's literally using the input form

21:32 axsk: (clojure.string/join \space (a b))

21:32 ,(clojure.string/join \space (a b))

21:32 clojurebot: #error {\n :cause "Unable to resolve symbol: a in this context"\n :via\n [{:type clojure.lang.Compiler$CompilerException\n :message "java.lang.RuntimeException: Unable to resolve symbol: a in this context, compiling:(NO_SOURCE_PATH:0:0)"\n :at [clojure.lang.Compiler analyze "Compiler.java" 6704]}\n {:type java.lang.RuntimeException\n :message "Unable to resolve symbol: a in this context"\n ...

21:32 axsk: ,(clojure.string/join \space ('a 'b))

21:32 clojurebot: ""

21:32 justin_smith: axsk: those are function calls

21:32 axsk: ,(clojure.string/join \space ['a 'b])

21:32 clojurebot: "a b"

21:32 axsk: ah ok

21:32 justin_smith: there you go

21:32 ,(clojure.string/join \space '(a b))

21:32 clojurebot: "a b"

21:33 axsk: guess i got it. read so much about syntax quoting and unquoting, that i forgot that args is a seq of SYMBOLS :)

21:33 justin_smith: axsk: yeah, sometimes you have to think outside the "syntax quote" box. If you are not evaluating any args, you don't need syntax-quote

21:34 axsk: imho a good macro tutorial would start with list operations, and only introduce ` after the list operations get tedious, so that it's clear that ` is not magic (and why we actually use it)

21:35 TEttinger: justin_smith: I believe brave clojure is that way

21:35 gfredericks: my macros talk did that

21:35 justin_smith: gfredericks: awesome

21:35 gfredericks: ~` is not magic

21:35 clojurebot: Ack. Ack.

21:35 gfredericks: ~`

21:35 clojurebot: ` is not magic

21:35 TEttinger: ~~

21:35 clojurebot: ~ is how you get factoids from me

21:37 perif: hi all

21:37 TEttinger: hey perif

21:39 justin_smith: TEttinger: today I ended a nethack game that had like 20 hours of total play time - I was in the astral plane, had made it to one altar, and was about to try a second altar

21:40 TEttinger: neutral monk, first two wishes were the eye of the aethiopica and the orb of fate

21:41 TEttinger: I am not very good at nethack

21:41 I would recommend brogue if you aren't totally bound to a true terminal

21:42 justin_smith: TEttinger: oh, what all that means is I was within 10 minutes of winning the game, after about 20 hours playing, and made a stupid mistake and game over

21:42 TEttinger: haha

21:42 well next time

21:42 justin_smith: by the end, I had 300+ hitpoints plus a magic item that cuts all damage in half

21:42 so effectively over 600 hp

21:49 tmtwd: http://pastebin.com/YbhmTiT6 how come the js/alert isn't executed when I call this functions?

21:50 justin_smith: tmtwd: this doesn't answer your question, but you never need a do inside defn like that

21:50 tmtwd: hm okay when do I use do then?

21:51 justin_smith: tmtwd: inside if or for

21:51 or other contexts that only take one expression

21:51 tmtwd: or cond?

21:51 justin_smith: sure, yeah

21:52 tmtwd: that code is weird because line 3 doesn't do anything at all

21:52 update creates a new collection, but you don't do anything with it

21:53 tmtwd: oh

21:53 justin_smith: so it's as if it wasn't done at all

21:53 tmtwd: I thought since atoms were mutable we could change them

21:53 justin_smith: you can, but that's not how

21:54 tmtwd: would I use assoc-in then?

21:54 justin_smith: tmtwd: you would use swap

21:54 tmtwd: oh of course

21:54 justin_smith: and then you would need to figure out what to do in the swapping function

21:55 I guess you want to take list-of-items and put it inside a hash-map under the key :items?

21:55 tmtwd: no I want to take the value of @upcoming-item and append it to @list-of-items (with a key of :item)

21:56 (def list-of-items (atom {:items [{:item "program in clojure"} {:item "finish web app"} {:item "solve euler problems"}]}))

21:57 justin_smith: so something like (swap! list-of-items conj {:item @upcoming-item})

21:57 nope, not quite

21:57 (swap! list-of-items update :items conj {:item @upcoming-item})

21:58 ,(def list-of-items (atom {:items [{:item "program in clojure"} {:item "finish web app"} {:item "solve euler problems"}]})))

21:58 clojurebot: #'sandbox/list-of-items

21:58 justin_smith: ,(def upcoming-item (atom "yolo"))

21:58 clojurebot: #'sandbox/upcoming-item

21:58 justin_smith: ,(swap! list-of-items update :items conj {:item @upcoming-item})

21:58 clojurebot: {:items [{:item "program in clojure"} {:item "finish web app"} {:item "solve euler problems"} {:item "yolo"}]}

22:01 tmtwd: almost works

22:01 justin_smith: what's missing?

22:03 tmtwd: i get this instead: (def list-of-items (atom {:items [{:item "program in clojure"} {:item "finish web app"} {:item "solve euler problems"} {:item {:item "yolo"}}]}))

22:03 an extra :item

22:03 justin_smith: tmtwd: well chang {:item @upcoming-item} to @upcoming-item

22:04 I didn't know what was in upcoming-item and I just made a random assumption (that was wrong)

22:05 tmtwd: oh, ok

22:05 I'll just get my repl out

22:05 justin_smith: tmtwd: if you get that result, just change {:item @upcoming-item} in my code to @upcoming-item and it will work

22:06 but yes, always have a repl open if you are coding clojure

22:06 tmtwd: justin_smith, yup tried that, the compiler did not like that

22:07 justin_smith, okay got it :)

22:08 justin_smith, you were right

22:44 gfredericks: dnolen: something about the closure-library upgrade in the recent cljs release broke test.check; is there an easy way for me to try out various commits from closure-library?

22:44 (i.e., to git-bisect the closure-library project by running test.check)

22:45 dnolen: gfredericks: I can't think of an easy way to do that.

22:45 gfredericks: what is the problem you're seeing?

22:46 the only thing I can think of is something about the PRNG we use

22:47 gfredericks: dnolen: well this is the new code with the hand-rolled PRNG

22:47 dnolen: something's undefined at a particular point, which feels like it could be distant from the root cause

22:47 dnolen: gfredericks: still then would be goog.math.Long thing

22:48 gfredericks: right; I didn't see any recent changes to goog.math.Long itself; I'll start adding asserts to the code to track down the root cause location

22:49 dnolen: gfredericks: and you think this is a GCL bug because the error originates from there?

22:49 gfredericks: dnolen: because explicitly using the older GCL version with new CLJS works fine

22:49 dnolen: ah interesting

22:52 gfredericks: in ClojureScript repo, in script/closure-library-release there's a script for local Maven install of GCL

22:53 gfredericks: dnolen: okay cool, I'll use that if it comes to it

22:58 dnolen: looks like this returns nil https://github.com/clojure/test.check/blob/master/src/main/clojure/clojure/test/check/random/longs.cljs#L56

22:59 seems to be an easy workaround if I replace with (long/fromNumber 1)

23:01 yep everything's passing now

23:07 I will probably just leave it at that; should I file a ticket?

23:07 I've never been confident that I was actually doing interop the right way

23:07 GCL-interop, rather

23:38 devn: Does anyone here run their production apps with `lein trampoline run`?

23:39 How are people feeling about AOT these days? I've continued to wonder over the years: to AOT, or not to AOT.

23:39 (when you have plenty of room for a choice)

Logging service provided by n01se.net