#clojure – 2014-10-31

2014-10-31

00:00	TimMc	I may have found it by searching for "clojure jobs", not really sure.
00:05	cfleming	In my macroexpanded code, I'm occasionally seeing vars like G__1234 - does anyone know when the compiler inserts those?
00:06	cfleming	I've looked at all the uses of RT.nextID() and I can't figure it out.
00:06	justin_smith	,(gensym)
00:06	clojurebot	G__27
00:06	cfleming	Aha, thanks.
00:06	cfleming	Shows how many macros I write.
00:06	justin_smith	cfleming: they are in the reader for ` when you name something binding#
00:07	justin_smith	,`(let [foo# 1] foo#)
00:07	clojurebot	(clojure.core/let [foo__50__auto__ 1] foo__50__auto__)
00:07	cfleming	justin_smith: Those are named <name>__1234__auto__, right?
00:07	justin_smith	ahh, right
00:07	justin_smith	so the others are manual genzyms
00:07	justin_smith	*gensyms
00:07	cfleming	justin_smith: The implicit parameters from a #() are p<num>__1234#
00:08	justin_smith	,'#(%)
00:08	clojurebot	(fn* [p1__75#] (p1__75#))
00:08	justin_smith	yeah
00:08	cfleming	Ok, makes sense - those must be named in clojure.core - I'll take a look.
00:08	justin_smith	,(gensym "foo")
00:08	clojurebot	foo102
00:09	cfleming	Ew
00:09	cfleming	So that one sucks
00:09	justin_smith	I think it is fairly rare...
00:09	cfleming	All this is for detecting the changed region in the macroexpander I'm working on.
00:10	justin_smith	ahh, right, of course
00:11	cfleming	Some of these are stable across expansions, but some of them are generated every time (G__1234 and pn__1234#)
00:11	cfleming	I guess I'll have to check whether two syms start with the same string but are suffixed with some different number for the named gensyms.
00:12	justin_smith	cfleming: what about a placeholder for the number part?
00:12	justin_smith	yeah
00:12	justin_smith	and not strictly a suffic, thanks to # being there sometimes of course
00:13	cfleming	Yeah, I'm planning to shorten them all to things like name<n># for auto-gensyms, p<n># for the parameters and so on, I guess <name># for named gensyms and perhaps G<n># for the plain gensym case.
00:14	cfleming	Fortunately there's only a couple of patterns.
00:16	cfleming	Thanks!
00:40	razum2um	from clojure survey'14 I get to know, that people would like to see "feature experessions" in the language. the idea initially comes from cljx and reminds me on c-ifdef's. i didn't write any cljs yet, but i ask you: is it really needed? I thought, that business logic should be ideally platform independent, why it won't get solved by proper "interface" for functions which deal with host-related things?
00:44	justin_smith	razum2um: in practice you can reduce the complexity of the codebase by allowing certain expressions to be swapped out based on the backend
00:44	justin_smith	razum2um: the advantage over C macros is that they are syntax aware, so it's much harder to get the bizzarre results that a bad C macro could give you
00:45	razum2um	justin_smith: is it really needed only as "syntactic sugar"?
00:45	razum2um	only to reduce file count?
00:46	justin_smith	razum2um: I don't know about calling it sugar. Each ns where you can use a feature expression means at least 2 files you don't need. Also, it's more straightforward to read and understand the code when key parts of it aren't split into separate files for backend rather than conceptual organization reasons
00:48	justin_smith	separation of concerns can simplify things, but when you are splitting things based on impl but sharing others, that can easily be fragmentation of the design rather than simplification.
00:49	mindbender1	is there a function I can use to output the bit representation of a number in 1s and 0s?
00:49	justin_smith	mindbender1: I think there is something in cl-format that can do that
00:49	razum2um	justin_smith: hm, this reminds me on nodejs and meteorjs. people are trying to merge codebases and finally they end up with "is_this_a_server_or_client" functions. I still think, that only some business logic should be shared and it could and must be written platform independent, or am I an idealist?
00:50	justin_smith	,(require '[clojure.pprint :as pprint])
00:50	clojurebot	nil
00:50	justin_smith	,(pprint/cl-format "~1r" 88)
00:50	clojurebot	#<IllegalArgumentException java.lang.IllegalArgumentException: Don't know how to create ISeq from: java.lang.Long>
00:51	justin_smith	,(pprint/cl-format nil "~1r" 88)
00:51	clojurebot	eval service is offline
00:52	justin_smith	razum2um: tangent - when using the term "business logic" - what is the distinction, what part of the code is not business logic?
00:54	mindbender1	justin_smith: I'm trying to read up the docs
00:54	justin_smith	razum2um: what happens is that you can have, for example, a common data structure between clj and cljs, and want to be able to do the same operations, but you have a small number of operations that need to be implemented differently. being able to describe those differences (if they are very small in number) inline rather than through two more namespaces, can make everything much simpler.
00:54	razum2um	hm, say "valid user should have name and email" - (defn valid_user? [u] (every? some? ((juxt :email :name) u))))
00:55	mindbender1	justin_smith: see http://www.cs.cmu.edu/afs/cs.cmu.edu/project/ai-repository/ai/html/cltl/clm/node199.html
00:55	justin_smith	,(pprint/cl-format nil "~b" 88) ; mindbender1
00:55	clojurebot	"1011000"
00:55	mindbender1	ok
00:55	justin_smith	mindbender1: I don't mean the format function in common lisp - the clojure function called cl-format
00:56	razum2um	justin_smith: perhaps in practice you're right and i need to write more cljs %)
00:56	mindbender1	justin_smith: yeah. The docstring points there.
00:56	justin_smith	mindbender1: what you link to is different - it's not about the ascii base 2 rep, it's about outputting raw bytes
00:56	mindbender1	okay
00:56	mindbender1	thanks for clarifying
00:57	justin_smith	razum2um: I think it's something about which reasonable people can disagree, you asked about the rationale so I do my best to present it
00:57	razum2um	justin_smith: but actually appearence of cljx points on the fact, that there is already a huge amount of cljs code, what are companies who do have them?
00:58	justin_smith	razum2um: my previous employer, who I now do some contracting for, has some cljs code
00:59	justin_smith	razum2um: for example here http://www.onenorthpdx.com/
01:10	mindbender1	justin_smith: can't cl-format output hex as well?
01:10	mindbender1	can
01:10	justin_smith	,(pprint/cl-format nil "~x" 99999)
01:10	clojurebot	#<CompilerException java.lang.RuntimeException: No such namespace: pprint, compiling:(NO_SOURCE_PATH:0:0)>
01:11	justin_smith	,(clojure.pprint/cl-format nil "~x" 99999)
01:11	clojurebot	#<ClassNotFoundException java.lang.ClassNotFoundException: clojure.pprint>
01:11	justin_smith	,(require '[clojure.pprint :as pprint])
01:11	clojurebot	nil
01:11	justin_smith	,(pprint/cl-format nil "~x" 99999)
01:11	clojurebot	"1869f"
01:12	justin_smith	,(pprint/cl-format nil "~x" 16rdeadbeef)
01:12	clojurebot	"deadbeef"
01:12	justin_smith	,16rdeadbeef
01:12	clojurebot	3735928559
01:13	mindbender1	okay. thanks
01:43	mindbender1	,(pprint/cl-format nil "~2,8,'0r" x)
01:43	clojurebot	#<CompilerException java.lang.RuntimeException: No such namespace: pprint, compiling:(NO_SOURCE_PATH:0:0)>
01:44	mindbender1	,(require '[clojure.pprint :as pprint])
01:44	clojurebot	nil
01:44	justin_smith	x is still unbound
01:44	mindbender1	,(pprint/cl-format nil "~2,8,'0r" 5)
01:44	clojurebot	"00000101"
01:44	mindbender1	just wanted to say thanks
01:44	justin_smith	np
01:54	marshall_	hey clojure
01:54	justin_smith	hello
01:55	marshall_	i'm almost 100% completely new to clojure and java. I'm trying to run a simple web service that uses a cool clojure library and i'm getting this error: https://gist.github.com/anonymous/3519fa9c8b5714d209fd
01:55	marshall_	is it something simple that i just don't understand? syntax maybe?
01:57	TEttinger	marshall_: just off the top of my head, you have a vector that contains just :post, could it be expecting and converting to a hash-map (which needs an even number of entries, 1 key to 1 value)
01:57	TEttinger	:access-control-allow-methods [:post]
01:57	TEttinger	that part
01:57	TEttinger	oh!
01:58	marshall_	TEttinger: yeah, that's the part that clojure was complaining about
01:59	marshall_	i just pasted from here: https://github.com/r0man/ring-cors
01:59	TEttinger	it apparently converts all the args to wrap-cors to a hash-map, you should double-check the usage of that code (it seems like it's a macro)
01:59	marshall_	i only want to allow the POST method, so I removed the other methods from that part
02:01	TEttinger	oh, gotcha
02:01	TEttinger	it has to do with ->
02:01	justin_smith	marshall_: take out app-routes
02:01	justin_smith	from the wrap-cors call
02:01	justin_smith	it throws off the hash-map construction
02:02	TEttinger	-> is called a threading macro, it takes an initial value and threads it (not concurrency threads, like threading a needle) by inserting it as the first arg to the next function given. then the result of that is inserted as the first arg to the function after that, and so on
02:03	TEttinger	,(-> 100 (+ 10 10) (/ 12))
02:03	clojurebot	10
02:03	TEttinger	so that takes 100, then makes the next piece into (+ 100 10 10) getting 120, then makes the next piece into (/ 120 12) getting 10
02:04	fairuz	,(->> 100 (+ 10 10) (/ 12))
02:04	clojurebot	1/10
02:04	TEttinger	and ->> does the same thing with the thread inserting in the last argument position
02:05	marshall_	ok
02:06	TEttinger	so what justin_smith is saying is, app-routes is already supplied as the first arg by -> , so when you have it again it treats it as part of the hash-map it's constructing and you and up with an odd number of elements it tries to pair into keys and values
02:06	marshall_	ahh I see
02:06	marshall_	that's a handy feature
02:06	TEttinger	quite!
02:07	marshall_	i think they call that a currying function in javascript
02:07	marshall_	or i guess in most other languages
02:09	TEttinger	kinda, -> actually works a bit differently. we have currying in the form of a function that returns functions, called partial
02:09	TEttinger	(doc partial)
02:09	clojurebot	"([f] [f arg1] [f arg1 arg2] [f arg1 arg2 arg3] [f arg1 arg2 arg3 & ...]); Takes a function f and fewer than the normal arguments to f, and returns a fn that takes a variable number of additional args. When called, the returned function calls f with args + additional args."
02:10	TEttinger	,(map (partial + 10) [1 2 3])
02:10	clojurebot	(11 12 13)
02:10	marshall_	hmm
02:10	marshall_	awesome
02:10	TEttinger	map I think is called map in javascript?
02:11	TEttinger	map, apply, filter are some of the most commonly used functions in clojure
02:11	marshall_	map in javascript is when you apply a function to the elements of an array and replace each element with the result of that function
02:11	TEttinger	yep
02:11	TEttinger	here, it returns a new collection, with state possibly shared
02:11	TEttinger	that's just an optimization clojure does
02:12	marshall_	[1,2,3].map(function(num){ return num +1 }) // >> [2,3,4]
02:12	marshall_	i removed the app-routes parameter
02:12	marshall_	but now I get a 404 for my one route
02:13	marshall_	could it be the order of my middleware?
02:13	TEttinger	but you don't have to think about defining some collection, running a map on it several times in an uncertain order, and then the collection was changed in some weird order the next time you use it -- that doesn't happen in clojure
02:13	TEttinger	uh, I am not familiar with ring
02:14	marshall_	hmm
02:14	TEttinger	it looks like wrap-cors returns something that expects one arg?
02:14	TEttinger	https://github.com/r0man/ring-cors/blob/master/src/ring/middleware/cors.clj#L67
02:17	marshall_	what's clojure really good for? what do you use it for?
02:17	marshall_	i'm coming from python and node.js
02:17	marshall_	javascript
02:19	justin_smith	marshall_: it makes concurrency sane, and once you "get it", it helps facilitate a programming style that eliminates incidental complexity
02:19	shane1	It's jvm based, so I use it for doing functional programming with interop with java
02:19	justin_smith	marshall_: part of that is the idioms / supplied libs, but things like immutibility really help with that
02:20	justin_smith	marshall_: I wouldn't use it if I need low ram usage or fast startup, but for a process that can use as much ram as it might need, and stays running a long time, it is awesome
02:23	marshall_	justin_smith: ahh i see
02:23	marshall_	shane1, justin_smith: thanks
02:26	justin_smith	marshall_: for a proper function written to act on immutible data, all you need to understand in order to understand what that function does is what it explicitly acceses in the body. This is a major game changer that effects everything from code reading, to unit testing, to refactoring.
02:27	justin_smith	marshall_: also code reuse, it is easier to write modular functions that can be reused if you know they contain no implicit dependencies or state
02:29	marshall_	interesting
02:30	marshall_	anybody have experience with ring? I'm trying to use CORS middleware
02:30	justin_smith	yes, I use ring all the time
02:33	marshall_	justin_smith: what am I doing wrong here? https://gist.github.com/jeffmarshall/b85a3f5fa6fe0e307112
02:33	marshall_	i only have one route and it gives me a 404 now that I've included the cors middleware
02:34	justin_smith	marshall_: I don't know that particular middleware, but it seems that it only allows post, and only allows requests from that specific host regex?
02:35	marshall_	hmm
02:35	marshall_	ok
02:37	justin_smith	marshall_: I am not sure because I don't know that middleware though
02:38	marshall_	justin_smith: have you ever done anything in ring with CORS?
02:41	justin_smith	marshall_: no I have not
02:47	marshall_	justin_smith: dang
02:48	justin_smith	marshall_: but the quantity of things to know about ring that you don't know is probably smaller than the quantity of things about cors that I don't know, and if we discuss it we can probably sort out what is going on
02:48	justin_smith	or is it visa versa? no matter
02:56	justin_smith	marshall_: I am working on a deploy to staging so my client can verify some work, (deploying a ring site even :)) but I can definitely help you sort this out afterward if you like
04:26	sveri80	Hi, I am trying emacs again and started a repl in my project and connected to it with cider, now, then I try to load a buffer with cider-load-current-buffer, nothing happens in the connected repl, what is wrong there?
05:41	vyvup	hi, I am trying to use ring.server.standalone from the REPL iso. using lein ring server-headless. For some reason their auto-reload behavior is slightly different, does anyone know why?
05:42	vyvup	e.g. changing strings inside a defroutes macro do not update using standalone, but do in lein
05:43	vyvup	looking at the sources I couldn't find a real difference between how they handle things
05:55	clgv	vyvup: I haven't used these modes (I guess). but you'll only get updates when the server runs in the same repl you use to update your code
05:55	clgv	errr. function implementation instead of code.
05:56	clgv	vyvup: for routes you usually need to pass the variable (instead of the value) of the routes constant or routes function to the call that starts the server
06:06	vyvup	clgv: see ix.io/eZp for handler code
06:06	vyvup	the repl commands that i use are there inside comments
06:07	vyvup	clgv: I am sorry but don't quite understand what you mean
06:08	clgv	vyvup: change your startup to (def server (ring.server.standalone/serve #'clojure-testapp.core.handler/app))
06:08	vyvup	ok, will try
06:08	clgv	vyvup: this way you have an indirection through the variable and get the latest changes
06:09	clgv	vyvup: you know about refheap.com ?
06:09	vyvup	nope, but now i do :)
06:12	vyvup	hmm, still does not work: what does the #' combination do?
06:12	TEttinger	##(clojure.string/join ", " (range 500))
06:12	lazybot	⇒ "0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, ... https://www.refheap.com/92509
06:13	TEttinger	neat feature of lazybot is the automatic link to refheap for long printed returns
06:13	vyvup	clgv: the namespace is reloaded: the do-bye function to return another string it updates correctly
06:13	TEttinger	vyvup: it's the var-quote , it's documented at the clojure site but there's another guide out there
06:14	TEttinger	http://yobriefca.se/blog/2014/05/19/the-weird-and-wonderful-characters-of-clojure/
06:14	clgv	TEttinger: haha great
06:14	TEttinger	it's a great guide
06:14	vyvup	TEttinger: ah, nice :)
06:15	clgv	vyvup: yeah, it should reload your routes assigned to the variable "app"
06:15	vyvup	yes, but it also did that without the #'
06:15	clgv	vyvup: do you reload the namespace where "app" is defined or just reevaluate the definition of "app"?
06:16	clgv	vyvup: didn't you say your routes were not updated?
06:16	vyvup	I depend on the reload behavior of (serve)
06:17	clgv	if `serve` is not a macro you should pass the variable `app`
06:17	vyvup	well, they are reloaded partly: everything outside the macro definition is reloaded, but e.g. changing the string returned by "/hello" does not propagate
06:18	clgv	vyvup: yes, that is because you do not use the var ;)
06:18	vyvup	hmm, but when I do use the var it also does not work
06:18	clgv	vyvup: you have to pass the var and you need to reevaluate the whole namespace or: first `app-routes` and then `app`
06:20	vyvup	ah, it seems I screwed up something else :)
06:23	vyvup	clgv: yay, works \o/ thanks! I will read up on the macro stuff
06:28	TEttinger	(inc clgv)
06:28	lazybot	⇒ 32
06:47	clgv	thanks
07:01	luxbock	is it just me or has the syntax hilighting for Clojure code at Github got significantly worse?
07:04	jonathanj	is there a solution to the crappy let/with-open nesting you get when you need things to be defined in a certain order?
07:05	jonathanj	(let [path (temp-file)] (with-open [output (output-thing path)] (let [obj (Obj. output)] ...
07:05	jonathanj	i need to return path, so i can't just inline it
07:09	SagiCZ1	i dont understand how could i not care about types of my sequences when conj appends to beginning or end depending on wheter its a vector or not.. thats absolute madness
07:31	raspasov	SagiCZ1: do you have specific example in mind?
07:33	borkdude	SagiCZ1 of course you care, but you usually know what you're dealing with
07:33	SagiCZ1	when i call (last (conj coll)) for example..
07:33	SagiCZ1	borkdude: no i dont.. some clojure functions return random weird types how am i supposed to know what they return
07:34	borkdude	SagiCZ1 they don't return random types. most sequence functions return lazy sequences
07:34	SagiCZ1	for example take-last returns PersistentVector, or PersistentVector$ChunkedSeq, or clojure.lang.Cons depending on the constellation of stars i guess
07:34	clgv	luxbock: it's just you. the colors and fonts change afaik
07:34	borkdude	,(doc take-last)
07:34	clojurebot	"([n coll]); Returns a seq of the last n items in coll. Depending on the type of coll may be no better than linear time. For vectors, see also subvec."
07:34	borkdude	It returns a seq.
07:35	SagiCZ1	,(type (take-last 2 [5 4 3 2 0]))
07:35	clojurebot	clojure.lang.PersistentVector$ChunkedSeq
07:35	SagiCZ1	looks like PersistentVector$ChunkedSeq to me
07:35	borkdude	(conj (take-last 2 [1 2 3]) 0)
07:35	borkdude	,(conj (take-last 2 [1 2 3]) 0)
07:35	clojurebot	(0 2 3)
07:35	borkdude	that's perfectly predicable
07:35	borkdude	SagiCZ1 yes, PersistentVector.....Seq <- seq
07:35	clgv	SagiCZ1: that is a Seq just an implementation detail which class provides the seq methods ;)
07:35	SagiCZ1	but why does it return java.lang.Cons sometimes?
07:36	SagiCZ1	i mean clojure.lang.Cons
07:36	raspasov	SagiCZ1: in most cases you don't care about either of those, if you have a bigger example or gist we might be able to suggest more constructively - usually there's a nice idiomatic way to do things
07:36	borkdude	,(type (list 0))
07:36	clojurebot	clojure.lang.PersistentList
07:36	borkdude	,(type (list))
07:36	clojurebot	clojure.lang.PersistentList$EmptyList
07:36	clgv	&(-> clojure.lang.Cons ancestors)
07:36	lazybot	⇒ #{clojure.lang.ASeq clojure.lang.Sequential clojure.lang.IObj java.util.Collection java.lang.Iterable clojure.lang.Seqable clojure.lang.Obj java.lang.Object java.util.List clojure.lang.IMeta java.io.Serializable clojure.lang.IHashEq clojure.lang.ISeq clojure.lang.IPersistentCollection}
07:36	borkdude	SagiCZ1 I don't even remember what returns a Cons, and I work with clojure pretty much every day
07:36	borkdude	SagiCZ1 not important
07:37	clgv	SagiCZ1: cons ia a seq as well as you can see above ;)
07:37	SagiCZ1	i am acessing only last n itemes from a vector.. if i use it as is, it works fine.. if i call (take-last (* 2 n) coll) on that vector before that, my program no longer works
07:37	clgv	borkdude: not-so-lazy lazy functions do that sometimes, because there impl starts with `cons` instead of `lazy-seq`
07:37	raspasov	as a rule of thumb in Clojure, you try to stick to vectors [1 2 3] or maps {:a 1 :b 2}; and you remember that map, filter, and reduce return a sequence
07:38	borkdude	yes, and if you really want to keep it strict, use mapv, filterv or transducers
07:38	clgv	SagiCZ1: you mean two time take-last?
07:38	clgv	SagiCZ1: the problem is take-last returns a seq and does not preserve the type vector
07:39	raspasov	SagiCZ1: use subvec for vectors
07:39	SagiCZ1	clgv: isnt that what i am talking about?
07:39	clgv	SagiCZ1: you could also implement as special take-last via sub-vec
07:39	borkdude	,(vec (take-last 2 (range 10))
07:39	clojurebot	#<RuntimeException java.lang.RuntimeException: EOF while reading>
07:39	raspasov	https://clojuredocs.org/clojure.core/subvec
07:39	borkdude	,(vec (take-last 2 (range 10)))
07:39	clgv	SagiCZ1: no. because take-last always returns seq and you talk about alternating return types
07:39	clojurebot	[8 9]
07:39	raspasov	use subvec, it's guaranteed O(1)
07:40	clgv	splitting hairs: take-last is O(1) as well - just not the realization part ;)
07:40	borkdude	use sequence functions and don't care about concrete types, until you expose them. premature optimization if you do I think
07:40	SagiCZ1	clgv: but sometimes it returns PersistentVector$chunkedSeq and sometimes clojure.lang.Cons .. andi guess my program depends on it for some reason
07:40	clgv	SagiCZ1: both are seqs just different implementations
07:40	Bronsa	SagiCZ1: those two types behave the same way unless you are explicitely testing for instanceness
07:41	clgv	no it does not. it might depend on the fact that you assume a vector where you get a sequence
07:41	SagiCZ1	ok thanks.. there is some super weird bug then.. i guess im just going crazy
07:41	Bronsa	SagiCZ1: when clojure talks about returning a seq, it means it will return an object whose type implements ISeq
07:41	Bronsa	no concrete type is guaranteed
07:41	raspasov	SagiCZ1: again, feel free to post a larger gist, we might be able to suggest something better
07:41	SagiCZ1	i use the return type in these functions:
07:41	SagiCZ1	apply, count, nth, last
07:41	SagiCZ1	shouldnt all these work fine for both types?
07:43	raspasov	it gets subtle, for example nth on a sequence is O(N) https://clojuredocs.org/clojure.core/nth
07:43	raspasov	for some details I recommend https://www.youtube.com/watch?v=iQwQXVM6oiY
07:43	Bronsa	SagiCZ1: yes, they should. can you paste a gist with the code that you're talking about as the other suggested?
07:44	SagiCZ1	Bronsa: its spread through some files, i will look into it more and then construct some minimal example
07:45	Bronsa	SagiCZ1: early on you said that take-last returns a PersistentVector, that's never the case by the way.
07:49	SagiCZ1	,(type (take-last 3 [2 0])
07:49	clojurebot	#<RuntimeException java.lang.RuntimeException: EOF while reading>
07:49	SagiCZ1	,(type (take-last 3 [2 0]))
07:49	clojurebot	clojure.lang.PersistentVector$ChunkedSeq
07:49	SagiCZ1	Bronsa: ok i guess i get confused.. im trying to solve this bug for about 5 hours now..
07:56	clgv	SagiCZ1: build small composable functions, test these individually
09:14	si14	urgh. is there any reason why Lein can skip local .m2 repo? I have my Datomic installed there, and yet lein deps fails
09:22	hyPiRion	si14: different Datomic versions – I don't think lein would check if the version is already there
09:22	hyPiRion	si14: you don't have to use lein deps anymore though, fyi
09:24	si14	hyPiRion: yeah, lein repl also fails :) versions are same
09:25	si14	ah, I see: ls ~/.m2/repository/com/datomic/datomic-pro/0.9.4956 > datomic-pro-0.9.4956.jar.sha1 datomic-pro-0.9.4956.pom.sha1 _maven.repositories _remote.repositories
09:25	hyPiRion	si14: yes
09:26	si14	hyPiRion: looks like a maven/Datomic bug then, because mvn build finishes without errors
09:29	hyPiRion	si14: Perhaps, but it may be that they store it somewhere else. Why the directory contains the sha but not the jar, I don't know.
09:31	si14	finally solved it. it seems that if you use remote repository and then switch to local one, mvn install will not actully install datomic locally
09:32	si14	you need to remove ~/.m2/repository/com/datomic manually first
09:37	TimMc	hyPiRion: I superstitiously run `lein do clean, deps, compile, javac, midje, junit, uberjar` when I build.
09:39	hyPiRion	TimMc: Let's :prep-task ourself?
09:40	hyPiRion	sounds reasonable.
09:41	fairuz	anyone use Titanium here?
09:42	csd_	anyone around that hosts on heroku?
09:43	daniel__	my laptop has some titanium in it
09:43	fairuz	daniel_ :)
09:46	clgv	~anybody
09:46	clojurebot	Just a heads up, you're more likely to get some help if you ask the question you really want the answer to, instead of "does anyone ..."
09:47	csd_	:)
09:47	csd_	my sql server connection seems to be timing out on heroku. i'm wondering what the best way to fix it is.
09:47	csd_	happens after like 30 or so minutes of inactivity
09:48	TimMc	I don't think Heroku supports persistent connections like that.
09:48	csd_	do you know how to have korma reconnect to the db in that case? i was looking through the source and it wasnt clear to me how to do it
09:49	jonathanj	why is the result of (if-let [{x :x} {:not-x 5}] x "nope") different to (if-let [x (:x {not-x 5})] x "nope")?
09:49	jonathanj	err, :not-x in the second example
09:49	TimMc	csd_: Wait, Korma relies on a persistent connection? What do you mean by timing out, exactly?
09:49	jonathanj	,(if-let [{x :x} {:not-x 5}] x "nope")
09:50	clojurebot	nil
09:50	jonathanj	,(if-let [x (:x {:not-x 5})] x "nope")
09:50	clojurebot	"nope"
09:50	csd_	TimMc: I think so? Because that's what I'm relying on and and it doesn't seem to be reconnecting right away.
09:50	csd_	What happens is that I have to make two http requests. The first fails and returns a blank page, which seems to reconnect to the sql server. And then the second request works.
09:51	fairuz	I can connect OK when I run Titan DB on the same machine as my app, but when I try to connect to a remote Titan DB, it fails with IllegalArgumentException. I'm using Titanium library to connect to this DB. Any advise on this error?
09:55	csd_	I suppose I could have my app do something like SELECT 1 every 5 minutes. I think that might work.
09:58	xeqi	jonathanj: the if portion is checking against {x :x}. This will always be truthy. Destructuring doesn't effect the if portion
09:58	jonathanj	xeqi: ah, okay
09:59	jonathanj	xeqi: thanks
10:00	xeqi	well, checking against {:not-x 5}, which is truthy
10:06	jonathanj	is putting complex if-let forms in let-like bindings usually a good idea or not?
10:06	jonathanj	(i'm guessing if i'm asking this, it's probably not)
10:10	csd_	i wouldn't want to sacrifice clarity for brevity personally
10:12	clgv	jonathanj: depends. you'd have to show examples to judge
10:13	jonathanj	is there a preferred paste-bin for this channel?
10:14	clgv	refheap.com
10:15	clgv	or something else with clojure syntax highlighting and without tons of ads ;)
10:15	clgv	gists on github work as well
10:15	xemdetia	Was refheap ever compromised or something?
10:16	clgv	why?
10:16	clojurebot	clgv: because you can't handle the truth!
10:16	clgv	clojurebot: right
10:16	clojurebot	Gabh mo leithscéal?
10:16	jonathanj	https://gist.github.com/jonathanj/30530c685d44cf45fbb7
10:16	xemdetia	I'm just getting a warning when I tried to look there
10:16	xemdetia	oh well, probably just an oversight
10:16	clgv	humm I dont get any
10:17	borkdude	refheap rules
10:17	clgv	xemdetia: firefox accepts the ssl certificate as valid
10:17	borkdude	because Raynes built it
10:17	clgv	xemdetia: is someone eavesdropping on you?
10:18	clgv	xemdetia: looks fine to me. could be improved with some->
10:18	clgv	jonathanj: ^^
10:18	xemdetia	clgv, yes- employer is and it just gets flagged as possibly malicious
10:18	xemdetia	that's why I was curious
10:18	jonathanj	clgv: some->?
10:19	clgv	jonathanj: ah wait, cond->
10:19	borkdude	conj->
10:20	jonathanj	how would you rewrite it?
10:20	clgv	jonathanj: page (:signature options), input-path (cond-> input-path page (vector (fill-fields page)))
10:21	clgv	jonathanj: you could also pull `page` in a local scope within a `let` surrounding the `cond->`
10:22	clgv	jonathanj: I'd recommend not to do "input-path' (some-expr input-path)" you'll probably for get ' or add it to the wrong symbol which bad since it is visually hard to detect
10:23	clgv	*forget
10:23	jonathanj	can i shadow existing names?
10:23	clgv	jonathanj: yeah
10:24	jonathanj	is the name only bound after evaluating the RHS of the let expression?
10:25	clgv	jonathanj: you could say so. the result of the RHS expression is referred to via the name in the remaining scope
10:26	jonathanj	clgv: your cond-> example, what comes before the "page (:signature" bit?
10:26	jonathanj	oh, i see, it's in the containing let
10:27	clgv	yeah exactly
10:27	clgv	but you can just use (let [page (:signature options)] (cond-> input-path page (vector (fill-fields page)))), so the name `page` is not bound in the remaining scope
10:29	jonathanj	yeah, that code i pasted is already in an existing let (as i mentioned), the let alignment is getting quite... unmanageable
10:29	jonathanj	thanks, that helped
10:30	clgv	jonathanj: maybe you could use smaller functions then
10:32	jonathanj	clgv: yes, that is almost certainly likely
10:32	jonathanj	i'm enhancing what was my first ever clojure application
10:34	jonathanj	if i have a function that takes ±7 arguments, is it preferable to start passing a map and using map destructuring instead of having a huge pile of positional arguments?
10:34	jonathanj	i guess it's more writing from the caller's perspective but it's certainly a lot more readable
10:37	clgv	jonathanj: maybe you should organize the data in a map in the first place?
10:41	jonathanj	clgv: hrm, well it's mostly wrestling command-line options to call some Java code
10:41	jonathanj	but i'll take a look at refactoring once i'm done implementing the enhancement
10:41	jonathanj	so i'm getting this warning: Reflection warning, clj_neon/core.clj:262:11 - call to setField can't be resolved.
10:42	clgv	jonathanj: is clj-neon your implementation=
10:42	jonathanj	yes
10:42	jonathanj	i'm not quite sure how to convey which overload i'm trying to call, i was under the impression there was exactly one 2-argument implementation of setField
10:43	jonathanj	https://gist.github.com/jonathanj/f5a419717c782a827b0d
10:43	clgv	ok so in your java interop, it is either not clear which type the object has or you must specify the types of the arguments to resolve to the correct overload
10:43	jonathanj	http://api.itextpdf.com/itext/com/itextpdf/text/pdf/AcroFields.html#setField%28java.lang.String,%20java.lang.String%29
10:43	jonathanj	oh
10:44	clgv	jonathanj: maybe you have a different version?
10:44	jonathanj	i'm calling the method on the wrong object
10:45	jonathanj	okay, i'll stop spamming the channel with my trivialities, thanks for your help clgv
10:45	clgv	good luck
11:14	sdegutis	Are transducers as wonderful as the memes make it out to be?
11:14	sdegutis	zoom__: uhh
11:14	sdegutis	zoom__: why did you PM me some spam link just now?
11:14	hyPiRion	sdegutis: bot
11:15	borkdude	sdegutis always be decomplecting is the first meme of all
11:16	csd_	what's the easiest way to extend a package that i'm using--a relatively small edit
11:17	csd_	i dont really want to have to fork it
11:18	bbloom	csd_: depends on the edit and the package, but you could 1) forcibly redef a var (please don't) 2) send a patch to the maintainer 3) copy paste some of it's code in to your codebase 4) ....
11:19	bbloom	but really, forking isn't so bad when you've got an application to ship
11:19	csd_	well i don
11:19	csd_	dont
11:19	csd_	this is more just for fun
11:20	csd_	i want to override korma to send a sql command over its jdbc connection
11:20	wink	csd_: why does exec-raw not work?
11:22	csd_	oh that should actually do it. i didnt see that
11:22	csd_	thank you
11:23	wink	the docs are a bit.. yeah :)
11:23	csd_	haha
11:23	wink	I fell into that trap as well. "such a cool library, why wont it let me exec sql?"
11:23	csd_	"such a cool library, why does it have zero foreign key support"
11:24	wink	so you seem to use postgres? .oO( No one uses FKs with mysql )
11:24	csd_	i use mysql
11:24	csd_	i guess i just do things the hard way
11:26	EvanR	i used foreign key constraints with mysql, at some point
11:26	wink	nah, I just find it atypical
11:26	EvanR	it was a real bitch
11:26	wink	most people don't
11:26	wink	maybe historical reasons as it was hard with 3.x or 4.x
11:26	csd_	ive never used postgres. my brother keeps telling me to try it
11:26	EvanR	use postgres if its got to be sql
11:27	csd_	i'd be interested to try something like couchdb at some point but there's definitely value in sticking to one thing and really knowing it
11:27	EvanR	really knowing postgres is really good
11:28	EvanR	and sql sucks
11:28	EvanR	so do key value stores
11:33	csd_	wink: are you pretty familiar with korma?
11:33	wink	csd_: nope, used it in 2 small projects
11:42	jez0990	hey #clojure, why is the ordering of transduce arguments opposite to the ordering of comp - is it is just a stylistic decision?
11:43	justin_s`	jez0990: each arg acts not on the output of the previous, but on the previous funcion
11:43	justin_s`	by wrapping it
11:44	nkoza	jez0990: there is a recent thread on the clojure mailing list about that exact issue
11:44	justin_s`	so it inverts the logical order
11:44	borkdude	jez0990 ah way you can remember: transducers correspond to ->>
11:48	jez0990	justin_smith: nkoza borkdude: thank you :)
11:56	sdegutis	borkdude: Good to know, thanks.
11:59	EvanR	what is the function to convert a string of digits into a number
11:59	justin_smith	EvanR: one for each of the primitive numeric types
11:59	borkdude	or just edn/read-string
11:59	justin_smith	,(Long/parseLong "1")
11:59	clojurebot	1
12:00	EvanR	parseLong, done
12:00	EvanR	,(Long/parseLong "123e")
12:00	clojurebot	#<NumberFormatException java.lang.NumberFormatException: For input string: "123e">
12:00	EvanR	,(Long/parseLong "")
12:00	clojurebot	#<NumberFormatException java.lang.NumberFormatException: For input string: "">
12:00	borkdude	,(read-string "123e")
12:00	clojurebot	#<NumberFormatException java.lang.NumberFormatException: Invalid number: 123e>
12:00	EvanR	,(Long/parseLong "-1")
12:00	clojurebot	-1
12:00	borkdude	is that a thing, 12e3?
12:00	EvanR	i was testing what happens if the prefix is a valid number
12:01	justin_smith	12e3 is scientific notation, dunno about 123e
12:01	EvanR	,(Long/parseLong "12e3")
12:01	clojurebot	#<NumberFormatException java.lang.NumberFormatException: For input string: "12e3">
12:01	borkdude	,(read-string "12e3")
12:01	clojurebot	12000.0
12:01	borkdude	yeah ok
12:01	justin_smith	,(Double/parseDouble "12e3")
12:01	clojurebot	12000.0
12:01	EvanR	i picked a bad letter for that test
12:01	justin_smith	:)
12:01	justin_smith	,(Double/parseDouble "123e")
12:01	clojurebot	#<NumberFormatException java.lang.NumberFormatException: For input string: "123e">
12:01	justin_smith	still fails as a suffix
12:02	clgv	,(Double/parseDouble "123abc")
12:02	clojurebot	#<NumberFormatException java.lang.NumberFormatException: For input string: "123abc">
12:02	clgv	;)
12:02	EvanR	,(Long/parseLong "123z")
12:02	clojurebot	#<NumberFormatException java.lang.NumberFormatException: For input string: "123z">
12:04	justin_smith	EvanR: for more complex sequences of data types in one string, there is of course edn/read, or the more flexible java.util.Scanner. And finally the various parser libs, of course.
12:04	sdegutis	bbloom: I wouldn't say forking isn't so bad. You've got to maintain your fork. At the very least, you've got to regularly merge from the repo you forked from to get their changes. That could get messy if they never merge your change.
12:05	bbloom	sdegutis: assumes you care to upgrade ever
12:05	sdegutis	bbloom: That said, I have done something similar with a really, really small lib, where the code is like 70 lines. But it was more of a wrapper around another lib.
12:06	sdegutis	bbloom: Right. Depends on things like whether the lib has inherent security needs, etc.
12:09	EvanR	,(edn/read "true")
12:09	clojurebot	#<CompilerException java.lang.RuntimeException: No such namespace: edn, compiling:(NO_SOURCE_PATH:0:0)>
12:09	EvanR	,(edn/read "123")
12:09	clojurebot	#<CompilerException java.lang.RuntimeException: No such namespace: edn, compiling:(NO_SOURCE_PATH:0:0)>
12:10	noonian	,(clojure.edn/read-string "true")
12:10	clojurebot	#<ClassNotFoundException java.lang.ClassNotFoundException: clojure.edn>
12:10	sdegutis	EvanR: you gotta require it
12:10	EvanR	required to require
12:10	noonian	,(do (require '[clojure.edn :as edn]) (edn/read-string "true"))
12:10	clojurebot	true
12:10	justin_smith	,(require '[clojure.edn :as edn])
12:10	clojurebot	nil
12:10	justin_smith	heh, beat me to it
12:10	EvanR	,(edn/read "123")
12:10	clojurebot	#<ClassCastException java.lang.ClassCastException: java.lang.String cannot be cast to java.io.PushbackReader>
12:10	EvanR	,(edn/read-string "123")
12:10	clojurebot	123
12:10	sdegutis	,(edn/read-string "(map map)")
12:10	clojurebot	(map map)
12:11	andyf	I worked on a project where I was used to require, but was required to use. refer madness
12:11	noonian	,(edn/read-string "(map map in your hat)")
12:11	clojurebot	(map map in your hat)
12:11	EvanR	,(edn/read-string "'(1 2 3)")
12:11	clojurebot	'
12:12	justin_smith	EvanR: with read, it would also move the stream forward, so the next read would get the next token
12:12	justin_smith	but you need a pushbackreader for read
12:13	clgv	,(read-string "'(1 2 3)")
12:13	clojurebot	(quote (1 2 3))
12:15	justin_smith	,(let [in (java.io.PushbackReader. (clojure.java.io/reader (.getBytes "1 2 :a [1 2 3] \"hello, world\"")))] (take-while identity (repeatedly #(try (edn/read in) (catch Exception _ nil)))))
12:15	clojurebot	justin_smith: I don't understand.
12:16	justin_smith	:P
12:18	justin_smith	,(let [in (java.io.PushbackReader. (clojure.java.io/reader (.getBytes "1 2 :a [1 2 3] \"hello, world\"")))] (take-while identity (repeatedly #(and (.ready in) (edn/read in)))))
12:18	clojurebot	(1 2 :a [1 2 3] "hello, world")
12:18	justin_smith	there we go!
12:19	EvanR	push back reader
12:19	justin_smith	yeah, useful when you want to speculatively consume bytes, and later say "oops, I can't use that" and put them back, so another part of the code can consume them
12:19	EvanR	juke box hero
12:20	justin_smith	?
12:20	EvanR	immediately what came to mind based only on the arity of syllables
12:20	clgv	just define a custom lazy-seq it'll end on nil ;)
12:20	EvanR	yeah "ungetc"
12:21	justin_smith	clgv: oh yeah, I could have done it with an fn making a lazy seq with self calls, yeah
12:21	EvanR	self calling a fn, does this bust the stuck potentially?
12:21	EvanR	stack
12:22	clgv	not in a lazy-seq ;)
12:22	justin_smith	EvanR: no, because lazy-seq transforms it into a thunk
12:22	EvanR	fun times
12:23	noonian	recursion is the funnest
12:24	justin_smith	noonian: also, the funnest is recursion
12:25	cbryan	justin_smith: but don't forget that recursion is the funnest
12:25	justin_smith	I propose we rename iterate to turtles
12:25	noonian	GNU has the best recursive acronyms
12:26	clgv	justin_smith: no then it would be mutual recursion ;)
12:27	justin_smith	clgv: that's recursion
12:27	clgv	justin_smith: recursion is almost recursion ;)
12:28	clgv	justin_smith: "to understand recursion you have to know someone who understands recursion" ;)
12:29	silasdavis	is there a way to get leiningen to add some arbitrary files to your jar?
12:29	clgv	silasdavis: put them under the resource path
12:29	justin_smith	clgv: I hear Heinz Von Foerster was fond of having you provide a random word, and see how long it took, starting from that word, to demonstrate a closed system of references in the dictionary. He called this the dictionary game.
12:29	EvanR	to understand recursion you need a base case to avoid an infinite regression
12:29	technomancy	silasdavis: easiest way is to plop them in resources/ put you can use :filespecs too
12:30	clgv	justin_smith: haha :D
12:30	clgv	EvanR: no, base cases take away all the fun! :P
12:30	justin_smith	clgv: his claim was that all human knowledge was recursive
12:31	noonian	infinite recursion works, its just usually not what you want unless you are making an interpreter or something
12:31	silasdavis	technomancy: thanks I thought of using resources. I didn't know about filespecs, these are vendored dependency executable binaries
12:31	justin_smith	clgv: though he preferred the term "circular"
12:31	clgv	justin_smith: there is some wikipedia game as well, to count the number of times to following the first link in an article until you end up at "Philosophy" starting at a random article
12:32	technomancy	silasdavis: resources/ is fine for that; there's no requirement that they be text
12:32	justin_smith	clgv: yeah, I have played that one too :)
12:32	SagiCZ1	,(-> clojure.lang.PersistentQueue/EMPTY (conj 5) (conj 42))
12:32	clojurebot	#<PersistentQueue clojure.lang.PersistentQueue@486>
12:32	SagiCZ1	,(seq -> clojure.lang.PersistentQueue/EMPTY (conj 5) (conj 42)))
12:32	clojurebot	#<CompilerException java.lang.RuntimeException: Can't take value of a macro: #'clojure.core/->, compiling:(NO_SOURCE_PATH:0:0)>
12:32	justin_smith	,(seq (conj clojure.lang.PersistentQueue 5 42))
12:32	clojurebot	#<ClassCastException java.lang.ClassCastException: java.lang.Class cannot be cast to clojure.lang.IPersistentCollection>
12:33	SagiCZ1	,(seq (-> clojure.lang.PersistentQueue/EMPTY (conj 5) (conj 42)))
12:33	clojurebot	(5 42)
12:33	SagiCZ1	isnt it backwards?
12:33	justin_smith	,(seq (conj clojure.lang.PersistentQueue/EMPTY 5 42))
12:33	clojurebot	(5 42)
12:33	clgv	justin_smith: damn! tapped in "Greek" -> "Modern Greek" -> "Greek"
12:33	justin_smith	SagiCZ1: that's what a queue does, it adds to the end, removes from the front
12:33	SagiCZ1	justin_smith: ok i though it was the other way around, np
12:34	justin_smith	,(seq (pop (conj clojure.lang.PersistentQueue/EMPTY 5 42)))
12:34	clojurebot	(42)
12:34	justin_smith	see, the 5 is gone
12:34	justin_smith	FIFO
12:34	SagiCZ1	yeah.. FIFO
12:34	SagiCZ1	5 came first so it gets removed first
12:35	SagiCZ1	queues usually have static size, but i guess its dynamic in clojure, right?
12:35	justin_smith	yeah
12:35	justin_smith	sometimes that is not what you want
12:35	justin_smith	java of course has bounded ones
12:36	SagiCZ1	and if i wanted to remove more than one element by popping, can i just call pop repeatedly?
12:37	clgv	O_o
12:37	justin_smith	SagiCZ1: remember that this is a persistent data structure
12:38	SagiCZ1	i understand it will not mutate
12:38	justin_smith	SagiCZ1: in practice you want to bind the value being removed with pop before doing that, but yeah, you can do multiple pops if you want to take multiple values off
12:38	justin_smith	SagiCZ1: maybe there is a more convenient method for this...
12:39	SagiCZ1	i am still trying to come up with the best data structure for my problem.. cant be that hard
12:39	SagiCZ1	vectors kinda work
12:40	hyPiRion	SagiCZ1: you can use core.async as well, if that fits
12:40	justin_smith	SagiCZ1: if I recall correctly you were taking items off on end, and adding to the other, and because you were using seq operations to do this you were losing the vector format
12:41	SagiCZ1	justin_smith: yes
12:41	justin_smith	SagiCZ1: a queue is actually designed for this FIFO behavior that is inherent to your algorithm
12:41	justin_smith	also, as hyPiRion mentions, core.async is, among other things, an interesting DSL over queues
12:42	SagiCZ1	i was removing the elements from the other end because i no longer needed them and because the vector was growing too much.. this would solve either constant size queue, or just using the regular peristent queue, and popping every time i conj new element once it reaches some predetermined size
12:43	SagiCZ1	does that make any sense?
12:43	justin_smith	I think so, yeah
12:43	clgv	SagiCZ1: you can just do that, by writing a custom `queue-conj`
12:44	SagiCZ1	clgv: true.. while experimenting i often forget that when i put some code in functions it makes it more readable but also easier to reason about.
12:44	justin_smith	SagiCZ1: why not remove items from the queue if they have been processed?
12:45	SagiCZ1	justin_smith: because i always need last n items, not just the very last item
12:45	SagiCZ1	n is constant
12:45	clgv	SagiCZ1: sliding window? moving averages?
12:45	justin_smith	SagiCZ1: OK, then hold onto exactly n items :)
12:45	SagiCZ1	clgv: exactly
12:47	SagiCZ1	so core.async has things that are handy even though i am not dealing with concurrency?
12:48	justin_smith	SagiCZ1: it can be useful for code that can be expressed in terms of reading from and writing to queues
12:49	clgv	SagiCZ1: is that code performance critical?
12:50	justin_smith	the basic concept is that all the code is connected by (chan) instances, and all the logic happens by filling and processing queues
12:51	SagiCZ1	clgv: yes, but i dont think i should optimize it at this point too much
12:53	SagiCZ1	justin_smith: very interesting
12:53	clgv	SagiCZ1: no not really.
12:56	SagiCZ1	clgv: you asked if the code is performance critical, would you have any performance suggestions?
12:57	cbryan	i am not used to paredit yet. just sat there for about 5 minutes trying to delete a bracket
12:58	sdegutis	cbryan: The point of paredit is that you don't have to delete a bracket.
12:58	justin_smith	cbryan: haha - C-space / C-w still works :)
12:58	sdegutis	cbryan: You can delete an s-expression though. Or it's contents. And it'll still be balanced.
12:58	cbryan	exactly ;)
12:58	sdegutis	I've been trying to learn paredit.vim lately.
12:58	justin_smith	sdegutis: you can get in a messy spot if you copy / paste from a non paredit source though
12:58	cbryan	yep ^ that's what happened
12:59	justin_smith	cbryan: so yeah, C-space / C-w
12:59	sdegutis	justin_smith: Good point. In those cases, I edit parens within a comment, and uncomment then when it's ready.
12:59	justin_smith	or just temporarily turn off paredit
12:59	cbryan	justin_smith: thanks :p
12:59	justin_smith	sdegutis: interesting approach
12:59	sdegutis	I find it to be the easiest/quickest way.
12:59	sdegutis	And most flexible.
12:59	cbryan	sdegutis: interesting
12:59	justin_smith	sdegutis: cbryan you can also use C-q to insert the matching paren
12:59	justin_smith	so that it's balanced, and thus deletable
12:59	sdegutis	I haven't found a time when it hasn't worked.
13:13	cbryan	i named my webcrawler project "Parker" and my robots.txt project (to keep Parker in line) "Jameson". it's the little things
13:13	sdegutis	ha!
13:13	sdegutis	(inc cbryan)
13:13	lazybot	⇒ 1
13:14	cbryan	if i'm being honest i was fishing for an inc ;)
13:14	sdegutis	(inc cbryan)
13:14	lazybot	⇒ 2
13:15	clgv	seems not that difficult with sdegutis ;)
13:15	sdegutis	(inc clgv)
13:15	lazybot	⇒ 33
13:15	cbryan	now you're just devaluing it! ;)
13:15	sdegutis	(dec cbryan)
13:15	lazybot	⇒ 1
13:15	EvanR	(dec EvanR)
13:15	lazybot	You can't adjust your own karma.
13:15	sdegutis	(inc EvanR)
13:15	lazybot	⇒ 1
13:16	cbryan	no halloween-themed clojure(docs).org? damn
13:17	EvanR	reminds me of the old pay-per-month MUDs where you got a fixed number of role playing points to give people who role played
13:17	sdegutis	(dec cbryan)
13:17	lazybot	Do I smell abuse? Wait a while before modifying that person's karma again.
13:17	cbryan	sdegutis is out of control!
13:17	sdegutis	No lazybot I'm just being an active participant in the community.
13:17	cbryan	sdegutis: stop playing with my emotions
13:17	sdegutis	+1
13:17	EvanR	(inc lazybot)
13:17	lazybot	⇒ 32
13:18	sdegutis	$karma sdegutis
13:18	lazybot	sdegutis has karma 6.
13:18	sdegutis	Oh nice.
13:19	cbryan	starting my new internship monday.. they're a java shop, maybe i can inject some clojure? :D
13:20	clgv	better not in the first week ;)
13:20	zachncst22	Never hurts to show initiative.
13:20	clgv	I doubt the empirical truth of that statement
13:20	cbryan	haha
13:20	sdegutis	zachncst22: In this case it may :P
13:21	sdegutis	cbryan: Why not Scala instead?
13:21	sdegutis	I hear it's much faster than Clojure, if you're going for speed.
13:21	sdegutis	What's the fastest JVM language?
13:21	hfaafb	Java
13:21	clgv	sdegutis: just hack everything together in C and just call per JNI :P
13:21	sdegutis	clgv: I'd rather write in Lua tbh.
13:22	hfaafb	Just write it in Go
13:22	zachncst22	Ugh, JNI
13:22	clgv	sdegutis: I'll let you know when the next serious post is to be expected from me ;)
13:22	SagiCZ1	i would use Pascal
13:23	zachncst22	Go Fortran, it's tons of fun.
13:23	EvanR	Groovy
13:23	zachncst22	Clojure is love, Clojure is life.
13:24	clgv	guess I never use pascal again
13:25	SagiCZ1	zachncst22: the only thing i hate about clojure is how little i know about it
13:25	clgv	SagiCZ1: change it ;)
13:26	cbryan	4clojure taught me the most/fastest
13:26	mdrogalis	Same
13:26	zachncst22	SagiCZ1: I agree completely. Which is why it's kicked out Scala as my 'play' language
13:26	SagiCZ1	clgv: i usually learn by coding..
13:26	clgv	cbryan: but it leaves knowledge holes you should fill with one of the books^^
13:26	bbloom	(inc 4clojure) ; looking at other people's answers helped me lots
13:26	lazybot	⇒ 3
13:28	clgv	you can't learn the big picture with 4clojure alone...
13:28	mdrogalis	Just saying, it's a nice start.
13:29	mdrogalis	A Datomic equivalent would be fun... :D
13:29	clgv	yeah, for fast practical experience it is awesome.
13:29	the_danko	fellows, i am working through the lispcast web development video and part of it involves a postgresql database. when i try to connect, i get Exception in thread "main" org.postgresql.util.PSQLException: The server requested password-based authentication, but no password was provided., compiling:(/private/var/folders/df/0z098cs56496gxlycxf2688c0000gn/T/form-init3442633375570322721.clj:1:124)
13:29	SagiCZ1	i read through Clojure Programming, while i consider it a great book I didn't learn all that much, i need to practice it a lot
13:30	justin_smith	the_danko: did you specify any password in your database config?
13:30	the_danko	anyone familiar with this? i dont know sh*t about psql, and i do need to enter a password when i type in "psql webdev"
13:30	sdegutis	hfaafb: phshaw
13:30	the_danko	justin_smith: no, seems like we are onto something!
13:30	the_danko	justin_smith: (def db "jdbc:postgresql://localhost/webdev")
13:31	justin_smith	{:classname com.postgresql.jdbc.Driver :subprotocol "postgresql" :host "localhost" :database "webdev" :user "youruser" :password "yourpassword"}
13:31	cbryan	the_danko: how are you actually connecting?
13:31	cbryan	or that :p
13:32	the_danko	cbryan: execute in clojure jdbc
13:32	the_danko	justin_smith: thanks! trying now!
13:32	justin_smith	the_danko: or, alternatively, construct the URL with all that info as a string, but easier to use the map form as above
13:33	EvanR	lispcast = eric normand?
13:33	the_danko	evanr: correcto
13:33	EvanR	ok
13:34	the_danko	justin_smith: Caused by: java.lang.ClassNotFoundException: com.postgresql.jdbc.Driver
13:34	justin_smith	the_danko: correction :classname "org.postgresql.Driver"
13:34	justin_smith	the rest should work
13:35	justin_smith	I was lazy, and I was looking at my mysql config on my most recent project :P did a poor on-the-fly translation :)
13:37	the_danko	justin_smith: hmm, missing a required parameter
13:37	justin_smith	does it say which one?
13:38	justin_smith	:classname :subprotocol :host :database :user :password
13:38	justin_smith	I specified all of those right?
13:38	the_danko	Caused by: java.lang.IllegalArgumentException: db-spec {:password "pass", :classname org.postgresql.Driver, :subprotocol "postgresql", :host "localhost", :database "webdev", :user "austin"} is missing a required parameter
13:38	the_danko	justin_smith: correct
13:39	epichero	is there a book on debugging clojure yet?
13:39	justin_smith	I think classname should be a string
13:39	justin_smith	(it is in my code for postgres at least)
13:39	epichero	i always parse name as meaning a string(in my head)
13:43	the_danko	hmm, maybesubname? i'll get googling
13:43	the_danko	maybe :subname i mean
13:44	the_danko	keep googling
13:44	the_danko	thanks again justin_smith
13:46	justin_smith	the_danko: here is an obfuscated version of a config that actually works: {:classname "org.postgresql.Driver" :subprotocol "postgresql" :host "127.0.0.1" :database "favoritedogs" :user "user1" :password "xxxx"}
13:47	the_danko	cfleming: automatic line breaks would really be awesome!
13:47	the_danko	justin_smith: thanks, trying it out now!
13:50	the_danko	justin_smith: (def db
13:50	the_danko	{:classname "org.postgresql.Driver"
13:50	the_danko	:subprotocol "postgresql"
13:50	the_danko	:user "user"
13:50	the_danko	:password "pass"
13:50	the_danko	:subname "//localhost:5432/webdev"
13:50	the_danko	:unsafe true}) worked!
13:50	the_danko	justin_smith: thanks a lot!
13:50	justin_smith	np
13:50	justin_smith	the_danko: ahh, I htink in this project I had a middle layer constructing the subname out of the host and database, sorry for that bit of noise
13:51	justin_smith	I forgot that was in there
13:51	the_danko	justin_smith: ha, you led me in the right direction. i wasn't aware of the map form.
13:52	justin_smith	the_danko: for extra credit you can throw a :datasource in there with a connection pooling class :)
13:52	cbryan	yay! now i can scroll irssi with my mouse. the future is now
13:52	EvanR	the_danko: jealous of the formatting of that map
13:52	the_danko	EvanR: ha, i stole it from here https://github.com/budu/lobos/blob/master/test/lobos/test.clj
13:52	justin_smith	I am sure there is a dodgy emacs library that does it automatically
13:53	the_danko	justin_smith: with hope, cfleming will be rocking this shit soon in cursive!
13:54	the_danko	the more times you say something, the truer it becomes
13:54	Bronsa	align-cljlet can indent map literals that way
13:54	Bronsa	https://github.com/gstamp/align-cljlet
13:54	justin_smith	Bronsa: I knew it! and there is a small chance it isn't even dodgy
13:54	Bronsa	justin_smith: it works really well
13:55	the_danko	cfleming: good code formatting will be huge for making this thing really catch on. it's hard for the intellij/eclipse java guys to live without. ok, that's the last tiem for today!
13:56	mmeix	Light Table has very good indenting, like it
13:58	epichero	I have spent so much time trying out different solutions on vim and emacs for clojure workflow
14:19	mdrogalis	I'm trying to do some stream analysis with Clojure. Can anyone think of something interesting to do with a tweet stream? Creativity is running dry.
14:20	tbaldridge	mdrogalis: pick a hot topic, then a set of words that express emotion, somehow figure out what the public's average view of a given topic is.
14:21	tbaldridge	mdrogalis: heard about some group that does that for motivational speakers. They generate live feeds of what the public thinks of the speaker via the tweets of the people watching the talk.
14:21	amalloy	tbaldridge: (constantly "shallow and trite")
14:21	tbaldridge	(inc amalloy)
14:21	lazybot	⇒ 184
14:22	cbryan	tbaldridge: tweetwall
14:22	amalloy	justin_smith: i'm not a fan of auto-aligned let/cond/map/whatever. it makes it hard for the next guy to edit your code without making it look stupid
14:22	mdrogalis	tbaldridge: Hah, very cool.
14:23	mdrogalis	I think I'll do that! Thanks!
14:23	justin_smith	mdrogalis: another possibility: try to predict who will end up following / favoriting / retweeting
14:23	cbryan	mdrogalis: http://tweetwall.com/ inspiration :p
14:24	cbryan	would it be better style to have each defrecord of a protocol in its own namespace, or in the same namespace as the protocol? or does it really depend..
14:24	justin_smith	amalloy: clearly the solution is for everyone to use identical, policy approved, emacs config
14:24	justin_smith	(nor am I a fan of that kind of indenting btw)
14:24	amalloy	for reasons that i don't recall, i forked align-let myself instead of using that clojure one, for the brief period that ninjudd insisted i align things: https://github.com/amalloy/align-let/commit/a6ece33751735d83b9de08377fd3919f84fe8641
14:24	mdrogalis	justin_smith cbryan: Thanks. :D
14:25	dbasch	mdrogalis: I did a word cloud generator a while back, just ran it for the word “giants” http://images.diegobasch.com/giants.png
14:25	EvanR	is this the right place to ask about datomic advice
14:25	pandeiro	what is the rationale behind namespacing keywords and is it A Bad Idea to introduce namespaced keywords in an API that use a shortened ns portion (eg :foo/thing vs. :foo.core/thing) ?
14:25	justin_smith	mdrogalis: you could also look at what facebook / okcupid have done with their social data (facebook being able to guess who would end up in a relationship, okcupid predicting who would respond to messages, etc.)
14:25	dbasch	mdrogalis: https://github.com/dbasch/twittercloud
14:25	amalloy	EvanR: i'm pretty sure #datomic is a thing
14:26	arrdem	mdrogalis: #gamergate what could go wrong
14:26	mdrogalis	I still dont understand what gamergate is. I tried multiple times :(
14:26	mdrogalis	But, not relevant for #clojure I guess.
14:26	mdrogalis	Thanks guys ^ Ill check those links.
14:27	mmeix	Is there a guide line concerning the order of a function's arguments?
14:27	mmeix	for example:
14:27	mmeix	(defn myfn [x coll] (remove #{x} coll))
14:27	mmeix	or
14:27	mmeix	(defn myfn [coll x] (remove #{x} coll))
14:27	arrdem	mmeix: cols come last typically
14:27	cbryan	makes it easier to partial-ize!
14:27	amalloy	arrdem, mmeix: sequences go last; other kinds of collections usually go first
14:27	dbasch	there are examples of both
14:28	dbasch	strings usually go first
14:28	amalloy	arrdem: see, for example, update-in, assoc, etc
14:29	mmeix	ok, is there some rationale behind this, besides partializing, or is this just a stylistic thing?
14:29	the_danko	amalloy justin_smith not sure whether just was joking or not but forcing a code formatter on all developers in the company works great!
14:30	arrdem	style and allowing ->> as an idiom on sequence last functions
14:30	mmeix	ah, I see
14:30	amalloy	mmeix: ->> for sequences, -> and the unified update model for other stuff
14:30	amalloy	eg, it lets you write (swap! x update-in [:people] conj mmeix)
14:30	arrdem	the_danko: pre-commits that lint, test and reindent are awesome :D
14:31	mmeix	(notes term "unified update model")
14:31	amalloy	which wouldn't work unless update-in and conj both took their collection argument first
14:32	justin_smith	the_danko: forcing everyone to use emacs usually not so great
14:32	amalloy	mmeix: i'm probably the only person who still says it. it was a name that stuart halloway (i think) popularized years ago to describe how updating functions work together in the way i just showed
14:32	mmeix	aha! thanks
14:32	the_danko	arrdem: interesting! would love to hear more. justin_smith: ha, agreed. i made the guys to do it with eclipse and intellij (can use the same format files).
14:32	the_danko	arrdem: we did not do pre-commit checks though. also i have never heard of lint until now.
14:33	the_danko	arrdem: care to elaborate?
14:33	arrdem	the_danko: the idea is just that if you have code climate requirements (whitespace/formatting/tests/test coverage) you write git hooks that prevent code from reaching staging without satisfying all these requirements. I personally choose to inflict them on myself almost all the time.
14:34	tuft	hello clojureplex
14:34	the_danko	yo tuft!
14:34	the_danko	arrdem so you just reject commits instead of autoreformatting?
14:34	arrdem	the_danko: sure
14:35	Bronsa	the_danko: it's more useful to reject commits if the linter returns warnings or if the test fail though
14:35	mmeix	amalloy BTW: reading flatland.org/useful/ source is a great learning experience for me :-)
14:35	the_danko	Bronsa: who is this linter fellow everyone ist alking about?
14:35	arrdem	$google c lint
14:35	lazybot	[Lint (software) - Wikipedia, the free encyclopedia] http://en.wikipedia.org/wiki/Lint_(software)
14:36	Bronsa	the_danko: http://en.wikipedia.org/wiki/Lint_(software)
14:36	arrdem	$google clojure eastwood
14:36	lazybot	[jonase/eastwood · GitHub] https://github.com/jonase/eastwood
14:36	the_danko	arrdem Bronsa ha i got that far but thanks!
14:36	the_danko	arrdem ill check out
14:36	the_danko	arrdem check it otu
14:36	arrdem	$google clojure cloverage
14:36	lazybot	[lshift/cloverage · GitHub] https://github.com/lshift/cloverage
14:37	arrdem	the_danko: ^ test coverage tool for Clojure
14:38	the_danko	arrdem and what did you mean by reindent? reformat?
14:38	amalloy	mmeix: great
14:39	arrdem	the_danko: other languages have tools like gofmt. Clojure doesn't have a good equivalent, although Emacs's clojure-mode indent is good and bbloom's fipp has a "code" mode now.
14:39	arrdem	the_danko: astyle for Java/C++/C
14:40	arrdem	bbloom: did you take the code format patches?
14:40	arrdem	looks like he did
14:43	sdegutis	Gonna tackle Transducers today!
14:43	joobus	sdegutis: don't forget to wear protection
14:46	dwmm	has anybody written a comparison of buddy and friend?
14:46	dwmm	I'm looking to use one of them with liberator and biddi
14:47	amalloy	having never heard of buddy, i'm gonna go out on a limb and say it's an attempt to rewrite friend to be less hairy, but which ultimately falls short by missing several key features
14:48	amalloy	so, take that for what it's worth
14:50	bbloom	arrdem: yeah, i merged it and tweaked it a bit
14:51	bbloom	arrdem: however, i should be clear that gofmt style code convention enforcement is a non-goal
14:51	bbloom	i'm almost exclusively interested in debugging macros
14:51	bbloom	although i guess it also helps as a starting point for hand tuned re-formatting
14:53	dwmm	thanks I'll look into it, but I'm hesitant to try it
14:55	the_danko	cfeming: how about the rainbow parantheses also? to be clear, i have a lot less experience with and confidence in this request as opposed to the formatting as a 1.5 week lisp programmer vs. a 4 year java guy, but it does look helpful.
14:55	the_danko	cfleming: how about the rainbow parantheses also? to be clear, i have a lot less experience with and confidence in this request as opposed to the formatting as a 1.5 week lisp programmer vs. a 4 year java guy, but it does look helpful.
14:56	andyf	Bronsa: Ciao. I?m refining Eastwood? wrong-tag linter, and got a t.a.j wrong-tag question for you. In a project (meltdown) there is one namespace with (defn ^Reactor create ...), no warning. This is ok, because CLJ-1232 only rears its ugly head if ^Reactor were tagged on the arg vector.
14:57	andyf	Then in another namespace, there is a call (mr/create) to that function, and there t.a.j gives a wrong-tag callback. Clojure itself gives no error here like for CLJ-1232 case.
14:57	the_danko	cfleming: the highlight matching when the cursor is near a paranthesis is great, but the always on rainbow thing could be cool.
14:57	andyf	Does t.a.j warn here because the tag is useless/ignored by Clojure?
15:03	sdegutis	Tranducers are still in a beta version of Clojure and therefore not safe to use in production yet right?
15:04	justin_smith	sdegutis: probably safer than a stable version of mongodb
15:04	sdegutis	!!!
15:04	TimMc	oh snap
15:05	andyf	sdegutis: It?s really up to your judgement. I know Sean Corfield said they are using alpha versions of Clojure 1.7.0 in production, but I don?t know if they are actually using transducers.
15:05	mdrogalis	World Singles seems to run edge everything.
15:06	justin_smith	in all seriousness, I would trust a not-yet-official version of a known reliable project over an official version of a known unreliable one. But yeah, they are not in any stable version of clojure yet.
15:06	TimMc	"MongoDB: A ship in the harbour is safe, but that's not what ships are for"
15:07	TimMc	"(ships are for moving fast and breaking things)"
15:07	justin_smith	TimMc: therefor we invite you to climb on our raft made of twine and truck tires
15:08	hiredman	but ships are but boards, sailors but men.
15:14	the_danko	so why do i need parantheses around the anonymous function here?
15:14	the_danko	(def app (-> routes wrap-params wrap-db (#(wrap-resource % "static")) wrap-server ))
15:15	justin_smith	,(macroexpand '(def app (-> routes wrap-params wrap-db (#(wrap-resource %
15:15	justin_smith	"static")) wrap-server )))
15:15	clojurebot	#<RuntimeException java.lang.RuntimeException: EOF while reading>
15:15	justin_smith	err
15:15	justin_smith	,(macroexpand '(def app (-> routes wrap-params wrap-db (#(wrap-resource % "static")) wrap-server )))
15:15	clojurebot	(def app (-> routes wrap-params wrap-db ((fn* [p1__48#] (wrap-resource p1__48# "static"))) ...))
15:15	hiredman	,(macroexpand '(-> a (fn [])))
15:15	clojurebot	(fn* a ([]))
15:15	hiredman	,(macroexpand '(-> a ((fn []))))
15:15	clojurebot	((fn []) a)
15:15	arrdem	the_danko: you don't (def app (-> routes wrap-params wrap-db (wrap-resource "static") wrap-server))
15:15	TimMc	the_danko: Because it's actually (fn [...] ...), and what happens to parenthesized forms when you put them in -> ?
15:16	the_danko	(macroexpand '(-> a f1 (f2))
15:16	justin_smith	the_danko: you need a , before that
15:17	hiredman	,'#(+ 1 %)
15:17	clojurebot	(fn* [p1__249#] (+ 1 p1__249#))
15:17	the_danko	,(macroexpand '(-> a f1 (f2))
15:17	clojurebot	#<RuntimeException java.lang.RuntimeException: EOF while reading>
15:17	TimMc	,(-> a #((((((a)))))))
15:17	clojurebot	#<sandbox$eval322$a__323 sandbox$eval322$a__323@1bd831f>
15:19	the_danko	TimMc: so it looks like (macroexpand `(-> a f1 f2))
15:19	the_danko	is equivalent to (macroexpand `(-> a f1 (f2)))
15:19	the_danko	=> (-dev-main/f2 (-dev-main/f1 -dev-main/a))
15:19	the_danko	(macroexpand `(-> a f1 (f2)))
15:19	the_danko	=> (-dev-main/f2 (-dev-main/f1 -dev-main/a))
15:19	the_danko	(macroexpand `(-> a f1 ((f2))))
15:19	the_danko	=> ((-dev-main/f2) (-dev-main/f1 -dev-main/a))
15:20	the_danko	TimMc: hmm, i don't know why (f2) is equivalent to f2 but not ((f2))
15:20	justin_smith	the_danko: that's just a weird thing about ->
15:21	the_danko	justin_smith: ha, ok, well now i know! thanks.
15:21	justin_smith	the_danko: a pattern you will see is (#(f % b)) in ->> or (#(f b %)) inside ->, beause this switches arg position
15:21	TimMc	the_danko: For convenience, -> allows you to say foo instead of (foo).
15:22	justin_smith	but it is better to use as->, or an embedded cal to -> / ->>
15:23	TimMc	,(macroexpand '(-> a b c))
15:23	clojurebot	(c (b a))
15:23	TimMc	,(macroexpand '(-> a (b) (c)))
15:23	clojurebot	(c (b a))
15:23	the_danko	justin_smith TimMc thanks !
15:23	TimMc	,(macroexpand '(-> a (b 1 2) (c 3 4)))
15:23	clojurebot	(c (b a 1 2) 3 4)
15:23	TimMc	the_danko: ^ This illustrates it.
15:24	TimMc	Then see hiredman's message above for why #() interacts weirdly with this.
15:24	TimMc	Finally, see if you can understand why my example above works the way it does: (-> a #((((((a)))))))
15:24	TimMc	It's a pathological example, but illustrative.
15:25	the_danko	TimMc: ok for the a b c thing, it looks like i should do this (def app (-> routes wrap-params wrap-db (wrap-resource "static") wrap-server ))
15:25	the_danko	TimMc: instead. which works.
15:26	sdegutis	How does this work? &(let [dict {'dup (fn [[x & s] _] (conj s x x)) '* (fn [[x y & s] _] (conj s (* x y)))}] (reduce #((dict %2 conj) %1 %2) () '(5 dup *)))
15:26	csd_	Magit is pretty neat
15:26	sdegutis	(inc csd_)
15:26	lazybot	⇒ 1
15:27	the_danko	TimMc: cool, so i think i have the abc part. now onto hiredman.
15:28	the_danko	TimMc I love the lessons plan by the way!
15:28	justin_smith	sdegutis: looks like a forth interpreter
15:28	sdegutis	bbloom: You wrote a forth interpreter in <=140 chars!?
15:28	justin_smith	sdegutis: actually yeah, it's a small subset of forth, but it uses forth semantics with a list as a stack
15:29	justin_smith	sdegutis: where reduce is used to run the program (the words of the program being in the input)
15:29	sdegutis	That is so fricken cool.
15:29	sdegutis	I wonder how else you could do it, without reduce? There's another core function that I think could do it that's just on the tip of my tongue.
15:30	justin_smith	sdegutis: the idea with forth is words take N words off the stack, do some operation, then put N words back on
15:30	bbloom	it's more fun to see it with reductions:
15:30	justin_smith	sdegutis: iterate maybe
15:30	bbloom	(let [dict {'dup (fn [[x & s] _] (conj s x x)) '* (fn [[x y & s] _] (conj s (* x y)))}] (reductions #((dict %2 conj) %1 %2) () '(5 dup *)))
15:30	bbloom	,(let [dict {'dup (fn [[x & s] _] (conj s x x)) '* (fn [[x y & s] _] (conj s (* x y)))}] (reductions #((dict %2 conj) %1 %2) () '(5 dup *)))
15:30	clojurebot	(() (5) (5 5) (25))
15:30	justin_smith	bbloom: yup, and that does exactly what the forth program "5 dup *" would do
15:30	sdegutis	Could this be easier with transducers maybe?
15:30	cbryan	ugh, "contact us about pricing." my four least favorite words
15:30	sdegutis	bbloom: That is just so cool.
15:31	justin_smith	haha, transducers are more about perf than ease of programming
15:31	sdegutis	Oh I didn't know that.
15:31	sdegutis	I am now tempted to embed a forth interpreter somewhere in our website.
15:31	Bronsa	justin_smith: perf & reusability
15:31	sdegutis	Yeah I thought reusability was the big one.
15:31	the_danko	TimMc: I don't feel like it is clicking with the (-> a #((((((a)))))))
15:32	the_danko	TimMc: if you're bored. otherwise i'll figure it out. thanks again.
15:32	TimMc	First see what the reader does.
15:32	TimMc	,'(-> a #((((((a)))))))
15:32	clojurebot	(-> a (fn* [] ((((((a))))))))
15:32	justin_smith	Bronsa: sdegutis: well, perf unless you are implementing something like core.async where you end up replicating things like map and filter on a new data source, then they do make things easier
15:32	TimMc	Then see what macroexpansion does:
15:32	melipone	hello! I would like to use the marmaduke common-math-3 library on clojars but I can't seem to find its github location. Do u know it?
15:32	TimMc	,(macroexpand '(-> a #((((((a))))))))
15:32	clojurebot	(fn* a [] ((((((a)))))))
15:33	TimMc	,(macroexpand '(-> a #(a))) ;; a simpler version
15:33	clojurebot	(fn* a [] (a))
15:33	justin_smith	melipone: you can get the source out of the jar if nothing else
15:33	clojurebot	Huh?
15:33	justin_smith	melipone: but you don't need the github location in order to use it if that's what you mean
15:34	TimMc	,(-> a #(= (a) a))
15:34	clojurebot	#<sandbox$eval133$a__134 sandbox$eval133$a__134@83c488>
15:34	TimMc	,((-> a #(= (a) a)))
15:34	clojurebot	#<StackOverflowError java.lang.StackOverflowError>
15:34	TimMc	the_danko: Sorry, my example was bad.
15:35	TimMc	,((-> a #(a)))
15:35	clojurebot	#<StackOverflowError java.lang.StackOverflowError>
15:35	the_danko	TimMc why are you using a for the function and the argument?
15:35	justin_smith	melipone: I think the actual source is linked here http://commons.apache.org/proper/commons-math/download_math.cgi
15:35	the_danko	TimMc is that part of the lesson? i assume so. i just don't get why.
15:36	TimMc	There's no argument, actually -- I'm just demonstrating that -> is capable of giving a function literal a name. A terrible idea, but interesting.
15:37	the_danko	TimMc interesting. i'll have to come back to this after a few more days of study it looks like. thanks for your help once more.
15:37	TimMc	(-> a (fn* [] ...)) becomes (fn* a [] ...); "a" doesn't end up in the body of the fn, because that's not the outermost collection.
15:44	melipone	justin_smith: But how do I know what is the clojure interface?
15:44	justin_smith	melipone: I don't know that there is on - but you can open the jar you get from clojars, it's a zip file
15:44	justin_smith	s/on/one
15:48	melipone	justin_smith: If there is no clojure wrapper, why is there a clojar package for it?
15:49	justin_smith	melipone: I am wondering what is up with that package, checking it out now
15:51	justin_smith	melipone: weird, when I specify the dep, it doesn't even seem to download anything....
15:53	melipone	justin_smith: it did download something. I'm checking it out now
15:53	justin_smith	weird, seems like my emacs is misbehaving
15:54	melipone	justin_smith: yeah, it's just the apache jar and nothing else.
15:54	justin_smith	melipone: yeah, that was my first guess (I am seeing that now too)
15:54	justin_smith	my emacs dired for some reason showed me an empty directory the first time... problem on my end clearly
16:20	txgruppi	Hi. I am new to Clojure and ClojureScript. I can't find my answer on Google. Is it possible to sue ClojureScript without Clojure? I have a PHP REST API and I would like to use swannodette/om with it. Is it possible?
16:21	postpunkjustin	IANAL, but yes
16:22	postpunkjustin	ClojureScript just compiles to JavaScript, so it's not dependent on Clojure in any particular way
16:22	technomancy	the compiler is written in clojure, but that's all
16:23	justin_smith	txgruppi: a previous employer was able to put clojurescript in production without deploying any clojure
16:23	txgruppi	postpunkjustin, technomancy, justin_smith: So, I just need to create a clojure project because of the compiler and I can use the js output in my app, right?
16:23	justin_smith	right
16:23	postpunkjustin	Yeah, your output will be HTML+JS
16:24	txgruppi	postpunkjustin, technomancy, justin_smith: Thank you all
16:24	technomancy	txgruppi: my gut feeling is that practically speaking, you have to know clojure because the tools and docs for using cljs all assume you know it.
16:24	technomancy	even if it's theoretically not required.
16:24	justin_smith	technomancy: I assumed it was more about being able to use cljs in production without deploying clojure itself, I could have misinterpreted
16:24	txgruppi	technomancy: the problem is that I can't just change the backend to clojure. Learning clojure will not be a problem
16:25	technomancy	ok, cool
16:26	justin_smith	"Learning clojure will not be a problem." that's the spirit!
16:27	txgruppi	hauhuahuahua
16:27	teslanick	Do you have a chant or mantra that I can pass along to my coworkers?
16:27	amalloy	$timer 30 0 0 hauhuahuahua
16:27	lazybot	Timer added.
16:28	hiredman	~suddenly
16:28	clojurebot	CLABANGO!
16:28	txgruppi	I am tired of PHP, I don't like ruby/python for web development, I don't use M$ stuff. Solution: Slowly migrate to Go, NodeJS, Clojure
16:29	justin_smith	txgruppi: a buddy of mine deploys cljs to backend node
16:30	txgruppi	justin_smith: nice! does he have any github (or some other public place) project with it?
16:30	justin_smith	txgruppi: I'll ask, and let him speak up if he would like to break anonymity :)
16:31	txgruppi	justin_smith: thx =D
16:36	postpunkjustin	I'm kind of puzzled by cljs+node on the backend. I mean, you're basically writing Clojure at that point, but you don't have real threads.
16:36	justin_smith	postpunkjustin: a client very picky about implementation, and not very wise about the tech, I think
16:37	arrdem	postpunkjustin: and you sacrifice access to all the JVM libraries and tooling. No, I agree.
16:37	postpunkjustin	Yeah, I can see how that would happen.
16:37	hiredman	it is a silly thing to do
16:37	technomancy	people do silly things all the time though
16:38	technomancy	https://en.wikipedia.org/wiki/The_Ministry_of_Silly_Walks
16:38	postpunkjustin	I actually thought about doing it for something that would run on a Raspberry Pi, since Clojure is a bit much for the hardware.
16:39	hiredman	don't get me started
16:39	justin_smith	postpunkjustin: I have a beablebone that runs a node webserver by default
16:40	justin_smith	for when you want to use a browser / html5 to push buttons on a screen in order to make an led on a device turn off and on
16:40	justin_smith	*beaglebone
16:42	postpunkjustin	Yeah, Node is way better than the JVM for serving something like that on "cute" hardware
16:42	postpunkjustin	esp. when the JVM has to load all of clojure.core
16:42	seangrove	postpunkjustin: Gotat go ocaml if that's where you're headed
16:43	hiredman	the jvm was designed for embedded set top boxes
16:43	technomancy	just use rackjure if you're low on ram =)
16:44	postpunkjustin	Woah, hadn't heard of rackjure, but I should have guessed it would happen sooner or later
16:44	postpunkjustin	I kind of like Hy
16:44	postpunkjustin	but I used to write a fair bit of Python
16:44	technomancy	hy... http://thisotplife.tumblr.com/post/63360807823/mutable-data-structures
16:45	postpunkjustin	Yeah, Clojure's immutable data structures are part of my brain now.
16:48	sdegutis	bbloom: I'd love to see a full-blown library expanded on your Forth interpreter tweets. That'd rock.
16:49	sdegutis	postpunkjustin: Yeah totally; whenever I write in Ruby it gets really awkward because I try to return stuff all the time, and just use functions, but Ruby doesn't make it easy.
16:49	postpunkjustin	Ugh Ruby.
16:50	sdegutis	postpunkjustin: You prefer Python?
16:53	postpunkjustin	Yeah, but only marginally at this point
17:08	toxmeister	hey guys, anyone know if clojar deploys are currently broken? am getting "403 forbidden Reason:passphrase" but nothing has changed on my GPG or SSH keys (just double checked) and I successfully deployed few weeks ago...
17:10	justin_smith	toxmeister: ssh upload is turned off, gpg only
17:10	hiredman	the scp deploy method was disabled due to, I think shellshock is what it is being called
17:10	hiredman	you have to use `lein deploy` which deploys via https
17:11	toxmeister	yes, using `lein deploy clojars` as usual
17:11	technomancy	403 indicates that it's an http failure though
17:11	hiredman	Oh
17:12	technomancy	toxmeister: I'm getitng the same thing
17:12	technomancy	looking into it
17:12	toxmeister	weird, just tried once more the same... and it worked this time!!
17:13	toxmeister	technomancy: thanks, but use your time more wisely :)
17:13	technomancy	but it's still failing for me
17:15	justin_smith	toxmeister: technomancy: http://i.imgur.com/Y6hqNQP.png
17:17	toxmeister	technomancy: yeah, happening again for me too (there i was hoping to have a release friday...)
17:17	technomancy	xeqi: any changes to clojars recently?
17:18	toxmeister	technomancy: is there a way to get more verbose debug info for these deploy requests?
17:18	toxmeister	bit like `curl -v`
17:19	technomancy	toxmeister: you can try setting LEIN_DEBUG=y
17:19	technomancy	or maybe just DEBUG=y
17:20	technomancy	nothing all that helpful though
17:23	toxmeister	hmmm... looking at https://clojars.org/repo/thi/ng/geom-core/0.3.0-SNAPSHOT/ it seems to have copied the files though (last version 20141031.212027)
17:24	toxmeister	maybe that error can be ignored?
17:24	cfleming	the_danko: You can already align map items like that in Cursive: Settings->Code Style->Clojure->Align map items
17:24	xeqi	technomancy: nothing
17:24	xeqi	technomancy: theres some minor restructuring refactorings on master, but nothing has been deployed in months
17:25	cfleming	the_danko: Also, Settings->IDE->Clojure->Rainbow parens - the default colours are pretty subtle, you can configure them though. I'll be adding something more psychodelic at some point.
17:25	technomancy	logs are rather unhelpful for this
17:26	justin_smith	cfleming: double-rainbow-parans-what-does-it-mean
17:26	erikcw	I have a ClojureScript repl running in my terminal. I’ve connected to the repl process from emacs using cider-connect — however, I’m conntect in a Clojure context. How do I enter the ClojureScript repl from here?
17:27	the_danko	cfleming cfleming: awesome, thanks!
17:27	justin_smith	erikcw: which clojurescript repl do you have running?
17:27	xeqi	looks like maven-metadata.xml didn't get updated, wonder if url that being split wrong
17:27	the_danko	cfleming maybe the rainbow hsould be default? also i agree on the brigher colors. very helpful still. thanks. just turned them on
17:27	erikcw	justin_smith: with figwheel (using the chestnut lein template)
17:28	xeqi	not sure why that would have changed since aug tho
17:28	erikcw	justin_smith: weasle with figwheel — sorry
17:29	justin_smith	erikcw: you should be able to run M-x cider, and then pick the port that weasel is running on
17:29	justin_smith	cider-jack-in starts a process, cider just finds one to connect to
17:29	justin_smith	or maybe it is called cider-connect now
17:30	lambdahands	Is there a way to create a core.async timeout function that accepts microseconds instead of milliseconds?
17:33	amalloy	lambdahands: (fn [t] (timeout (quot t 1000)))
17:33	toxmeister	xeqi: you're right, the last update to maven-metadata.xml for my project was on aug 17, even though there're newer jars uploaded
17:33	amalloy	not a useful answer, i'm sure, but the best one available
17:33	erikcw	justin_smith: seems to just timeout if I connect to the weasel web socket port (9001). I am able connect to the main repl process that is running weasle though. Problem is, I have no idea how to get into the clojurescript prompt from there. I’m stuck in Clojure “user>” instead of Clojurescript “cljs.user>”
17:33	dbasch	amalloy: you could build your own with the high resolution timer
17:34	dbasch	who knows how accurate it would be
17:35	justin_smith	erikcw: have you tried the code suggested on the weasel readme? https://github.com/tomjakubowski/weasel
17:35	justin_smith	the whole (cemerick.piggieback/cljs-repl ...) thing
17:36	dbasch	,(let [[a _ b] [(System/nanoTime) (Thread/sleep 1) (System/nanoTime)]] (- b a))
17:36	clojurebot	1144357
17:36	dbasch	&(let [[a _ b] [(System/nanoTime) (Thread/sleep 1) (System/nanoTime)]] (- b a))
17:36	lazybot	⇒ 14949110
17:36	erikcw	justin_smith: yeah — I get an “Address already in use” error
17:36	justin_smith	erikcw: when specify a different address?
17:37	justin_smith	*then
17:37	erikcw	justin_smith: I guess I’m not being clear. I’m trying to connect to the already running session. Perhaps that just isn’t suppored?
17:38	jcsims	erikcw: aside from the lack of a 'cljs.user' prompt, are you sure you're not in a cljs context when you connect?
17:38	justin_smith	erikcw: I am sorry, I don't know.
17:39	justin_smith	erikcw: so you opened a cljs context in a repl, and then when you connect to the repl in emacs you get clj?
17:40	badyl	hi, I would like to define a comparator for a set of keywords (or in other words define an order of a set of keywords)
17:40	erikcw	jcsims: Looks like you’re right — I am in a cljs context, I just didn’t know it! OOPS!!
17:41	arrdem	so.. an enum with keywords as values?
17:41	badyl	and I am wondering how it can be done in a most clear way
17:41	jcsims	erikcw: good! if that wasn't the case, I was out of suggestions ;)
17:41	justin_smith	badyl: you can use sorted-set-by, and any function that predictably returns -1 0 or 1 when provided with two keywords
17:41	erikcw	jcsims: I assumed they were running under different threads or something
17:41	erikcw	jcsims: thank you
17:41	erikcw	justin_smith: thank you!
17:42	badyl	arrdem: yes, I have maps where there is an entry with keyword values as enums
17:42	jcsims	erikcw: no problem!
17:43	arrdem	amalloy: so why are you not a fan of your own defenum macro?
17:43	justin_smith	,(sorted-set-by (fn [a b] (compare (mod (int (first (name a))) 8) (mod (int (first (name b))) 8))) :a :b :c :d :e :f :g :h :i :j)
17:43	clojurebot	#{:h :a :b :c :d ...}
17:43	dbasch	keywords are comparable
17:44	dbasch	what’s wrong with the default (lexicographic) sort?
17:44	amalloy	arrdem: the one that's like the first project i ever pushed to github?
17:44	arrdem	amalloy: back in 2010 yeah that one
17:45	dbasch	,(compare :a :b)
17:45	clojurebot	-1
17:45	amalloy	arrdem: you want me to look further than https://github.com/amalloy/enum/blob/master/src/enum/core.clj#L3 ?
17:45	arrdem	maybe this isn't the _right_ defenum, but I think in principle a defenum is reasonable.
17:45	arrdem	amalloy: no need I'm wiping blood away too
17:46	badyl	let's say I have maps like {.... :type :a}, {..... :type :b} {.... :type :c} etc and I want them to be compared by :type key, but in my specific order, e.g. :c :a :b
17:46	justin_smith	dbasch: I am aware they are comparable, but I assumed he wanted some non-default sorting
17:46	badyl	I know I can use just (compare :a :b) but I need a specific order, not lexicographics one
17:46	justin_smith	badyl: then use sorted-set-by as I showed above, with a function that provides the right comparison result
17:46	dbasch	never mind, it wasn’t clear to me from the question
17:46	dbasch	but in that case yeah, just define your own compare function
17:47	lambdahands	@amalloy Thanks! So timeout takes a quoted list with two elements?
17:47	justin_smith	badyl: .indexOf on an array may help (subtracting two indexes to get pos / 0 / neg
17:47	justin_smith
17:47	justin_smith	)
17:47	badyl	dbasch: ok, I just thought there is a function that will create one when I provide a seq of my keywords in my specified order :)
17:47	amalloy	lambdahands: that quot is "quotient". it's a facetious suggestion that you divide microseconds by a thousand to get back to milliseconds
17:47	badyl	but it seems I have to write one
17:48	dbasch	lambdahands: do you need sub-millisecond resolution? I guess that’s the question
17:48	amalloy	because there's nothing i know of that takes microseconds, nor any compelling reason to need it
17:48	amalloy	badyl: https://github.com/flatland/useful/blob/develop/src/flatland/useful/map.clj#L243 is kinda close
17:49	amalloy	you can't use it directly but you can perhaps adapt the idea
17:49	dbasch	&(let [[a _ b] [(System/nanoTime) (Thread/sleep 1) (System/nanoTime)]] (- b a))
17:49	lazybot	⇒ 18028774
17:49	dbasch	that’s 18 ms
17:49	lambdahands	@amalloy: Ah, of course.
17:50	lambdahands	dbasch: Yes, that's what I'm looking for.
17:50	amalloy	&(let [[a _ b] [(System/nanoTime) (Thread/yield 1) (System/nanoTime)]] (- b a))
17:50	lazybot	java.lang.IllegalArgumentException: No matching method: yield
17:50	amalloy	&(let [[a _ b] [(System/nanoTime) (Thread/yield) (System/nanoTime)]] (- b a))
17:50	lazybot	⇒ 15650228
17:50	lambdahands	It's for a live performance using overtone's Open Sound Control library.
17:50	amalloy	&(let [[a _ b] [(System/nanoTime) [] (System/nanoTime)]] (- b a))
17:50	lazybot	⇒ 8380435
17:51	badyl	amalloy: thanks it is really close
17:51	badyl	justin_smith: your solutions is also similar
17:51	dbasch	,(let [[a b] [(System/nanoTime) (System/nanoTime)]] (- b a))
17:51	clojurebot	12407
17:52	dbasch	&(let [[a b] [(System/nanoTime) (System/nanoTime)]] (- b a))
17:52	lazybot	⇒ 7417257
17:52	badyl	thanks guys for the ideas. It's my first interaction with clojurians on IRC and you were very helpful
17:52	dbasch	does lazybot run on an abacus?
17:52	justin_smith	badyl: thanks
17:53	dbasch	, (- (System/nanoTime) (System/nanoTime))
17:53	clojurebot	-818
17:53	dbasch	& (- (System/nanoTime) (System/nanoTime))
17:53	lazybot	⇒ -7569558
17:53	lambdahands	Hmm, so System/nanoTime is the answer then?
17:53	justin_smith	,(let [order [:c :b :a]] (- (.indexOf order :b) (.indexOf order :a)))
17:53	clojurebot	-1
17:54	justin_smith	badyl: the above could be easily made a comparison function, with your order of keys specified in a vector
17:54	EvanR	is there a function to convert keyword to string without the prefixed :
17:54	justin_smith	EvanR: name
17:54	justin_smith	,(name :a)
17:54	clojurebot	"a"
17:55	justin_smith	,(map name ['a "a" :a])
17:55	clojurebot	("a" "a" "a")
17:55	justin_smith	I like the above :)
18:05	amalloy	dbasch: lazybot has some weird stuff in its interop-sandboxing code
18:06	amalloy	i don't remember what the deal is anymore, but it makes interop really slow
18:06	dbasch	amalloy: is it sending everything to the NSA? :P
18:06	arrdem	dbasch: this is IRC they get everything anyway
18:07	dbasch	arrdem: you can be lazy and paranoid :P
18:33	amalloy	is there some version of max that uses compare instead of >?
18:34	hiredman	,(doc max-key)
18:34	clojurebot	"([k x] [k x y] [k x y & more]); Returns the x for which (k x), a number, is greatest."
18:34	hiredman	sort of
18:35	amalloy	right, which doesn't actually do that. it's fine if you can map your stuff onto integers
18:35	hiredman	,(doc compare)
18:35	clojurebot	"([x y]); Comparator. Returns a negative number, zero, or a positive number when x is logically 'less than', 'equal to', or 'greater than' y. Same as Java x.compareTo(y) except it also works for nil, and compares numbers and collections in a type-independent manner. x must implement Comparable"
18:35	hiredman	isn't that would compare does?
18:35	hiredman	what
18:36	hiredman	Oh
18:36	hiredman	right
18:37	hiredman	well, wait, I think that should be fine
18:37	TEttinger	(doc sort)
18:37	clojurebot	"([coll] [comp coll]); Returns a sorted sequence of the items in coll. If no comparator is supplied, uses compare. comparator must implement java.util.Comparator. If coll is a Java array, it will be modified. To avoid this, sort a copy of the array."
18:38	TEttinger	so sort and get the last?
18:38	hiredman	max works are two arguments, so you don't need everything mapped to some known fixed set of numbers, it should work
18:38	hiredman	(unless max-key just calls the key once)
18:38	hiredman	on two
18:38	hiredman	you just need to be able to come with numbers for every two things, not for everthing
18:39	amalloy	hiredman: i don't understand. max-key calls (f x) and (f y), then compares them
18:39	amalloy	it doesn't call (f x y), which would be the compare interface, and exactly what i wanted
18:39	hiredman	foo, right
18:40	hiredman	I was not thinking about compare correctly of course
18:41	amalloy	i'm pretty sure that at some point in the past i told someone they were silly for wanting this max-by thing i'm asking for
18:42	TEttinger	amalloy, because you can sort and get the first or last?
18:42	amalloy	TEttinger: no, that is outrageous
18:42	amalloy	performing an O(n*log(n)) operation for no reason when O(n) will do is not something you should do
18:42	TEttinger	fine
18:44	amalloy	i probably argued the opposite reason: that it's not very fast to keep doing this expensive compare operation on pairs of elements, so clojure doesn't encourage it, or something like that
18:44	amalloy	that sounds like me
18:44	dbasch	amalloy: one more thing you can stick in useful
18:45	TEttinger	(doc compare)
18:45	clojurebot	"([x y]); Comparator. Returns a negative number, zero, or a positive number when x is logically 'less than', 'equal to', or 'greater than' y. Same as Java x.compareTo(y) except it also works for nil, and compares numbers and collections in a type-independent manner. x must implement Comparable"
18:46	dbasch	something like
18:46	dbasch	(defn max-by [f a b] (if (pos? (f a b)) a b))
18:47	dbasch	and reduce by that, or whatever
18:48	TEttinger	,(defn max-compare [coll] (reduce #(if (pos? (compare %1 %2)) %1 %2) coll))
18:48	clojurebot	#'sandbox/max-compare
18:49	TEttinger	,(max-compare [2 3 [1 2 3 4]])
18:49	clojurebot	#<ClassCastException java.lang.ClassCastException: clojure.lang.PersistentVector cannot be cast to java.lang.Number>
18:49	TEttinger	well I dunno how that works
18:51	hyPiRion	,(compare 3 [1 2 3 4])
18:51	clojurebot	#<ClassCastException java.lang.ClassCastException: clojure.lang.PersistentVector cannot be cast to java.lang.Number>
18:51	TEttinger	type-independent manner my ass
18:53	hyPiRion	compares numbers and collections in a type-independent manner => ints and doubles can be compared, lists and vectors can be compared
18:54	hyPiRion	It's not like comparing a number to a collection suddenly made sense.
18:54	technomancy	eeeeeh
18:54	technomancy	I can see value in a universal comparator
18:55	technomancy	it doesn't have to make sense across types so much as be consistent across types
18:58	andonilsson	Weird thing. Was using core.async channels for dealing with cards (you know, poker etc). So I tried: (onto-chan (chan) deck-of-cards)
18:58	andonilsson	...but the channel returned from onto-chan never contained anything.
18:59	andonilsson	Instead I had to do something like:
18:59	andonilsson	(def c (chan))
18:59	andonilsson	(onto-chan c deck-of-cards)
18:59	andonilsson	...and then I could pull my cards from channel c.
19:04	nathan7	andonilsson: (doto (chan) (onto-chan deck-of-cards))?
19:04	nathan7	,(macroexpand `(doto (chan) (onto-chan deck-of-cards)))
19:04	clojurebot	(let* [G__27 (sandbox/chan)] (sandbox/onto-chan G__27 sandbox/deck-of-cards) G__27)
19:04	nathan7	shit, I gotta go refactor a whole lot of code using that trick
19:04	andonilsson	or, even better in my case: (to-chan deck-of-cards)
19:05	andonilsson	Found it while browsing the source for onto-chan
19:06	justin_smith	yeah, doto is kind of handy sometimes
19:07	JohnJJ	hi guys I want to add the JDBC jar to a local repository, I've tried Archivia without success, what do u suggest?
19:07	andonilsson	still, I find it weird that the channel returned from onto-chan is not of any use. Instead I have to use the 'original' channel (the one passed into onto-chan)
19:07	justin_smith	JohnJJ: how are you managing your deps? if it is not lein you should probably switch.
19:08	technomancy	JohnJJ: depends on why you want a local repository, really. lots of ways to do it.
19:08	JohnJJ	justin_smith, technomancy: I'm using lein, and I'm using Nightcode as an IDE
19:09	justin_smith	JohnJJ: and using the usual .m2/ cache doesn't suffice for local repository?
19:09	justin_smith	are you serving a dep for multiple users or machines?
19:10	JohnJJ	justin_smith: I've just tried that but couldn't get how the folder structure should be, can I just put the .jar there?
19:10	technomancy	JohnJJ: what are you trying to accomplish?
19:10	technomancy	usually archiva is used to cache dependencies to protect against network failures and stuff
19:10	justin_smith	JohnJJ: the normal thing is to declare the dep, such that lein finds it automatically
19:11	technomancy	or for private stuff, not jdbc drivers
19:11	JohnJJ	technomancy: just trying to get my clojure app to connect to SQL Server
19:11	justin_smith	JohnJJ: add [mysql/mysql-connector-java "5.1.6"] to your deps in project.clj
19:12	justin_smith	and you probably want org.clojure/java.jdbc too
19:12	justin_smith	oh way you didn't say mysql
19:12	JohnJJ	justin_smith: it is SQL Server from Microsoft, not Mysql
19:13	justin_smith	http://claude.betancourt.us/add-microsoft-sql-jdbc-driver-to-maven/ install it locally, then declare a dep
19:13	JohnJJ	justin_smith: I've lein installed but not maven, should I install maven then?
19:14	justin_smith	JohnJJ: technomancy: there is a lein command for that, right?
19:14	technomancy	yeah, check out the second arity of deploy
19:14	JohnJJ	justin_smith: that's my confusion, how can lein run without maven
19:15	justin_smith	JohnJJ: it uses a bunch of the same libs
19:17	justin_smith	technomancy: I'm looking at lein help deploy now - what is the magic repo string for "into .m2"
19:17	technomancy	justin_smith: file:///home/me/.m2/repository
19:17	justin_smith	technomancy: cool
19:18	justin_smith	JohnJJ: so that's the simplest way (other than installing maven and copy/paste from that web site I guess)
19:20	JohnJJ	justin_smith, technomancy: I'm trying to install maven right now
19:26	technomancy	wow, so the MS DB honestly does not publish any official jdbc deps?
19:27	justin_smith	technomancy: all I could find were workarounds via local install
19:27	technomancy	inconcievable
19:28	mlb-	In SKI combinator calculus, is the Clojure `doto` macro analagous to the K combinator?
19:28	technomancy	mlb-: not really, SKI doesn't have side-effects
19:29	technomancy	I would say K is more like constantly
19:29	mlb-	ah, that's fair
19:29	technomancy	but neither functions nor macros map strictly to combinators I guess
19:30	justin_smith	technomancy: JohnJJ: https://social.msdn.microsoft.com/Forums/sqlserver/en-US/db43206a-756c-4291-9c67-e813d18bdbf5/can-microsoft-please-publish-the-jdbc-driver-for-sql-server-on-maven-central?forum=sqldataaccess
19:30	mlb-	I ask because I've recently been abusing `doto` inside of thread first/last macros
19:30	amalloy	mlb-: how is that abuse? doto and -> go together like bacon and eggs
19:30	justin_smith	mlb-: doto inside thread-last seems a bad plan
19:30	amalloy	it's a lot harder in ->>
19:31	justin_smith	now if we had doneby
19:31	justin_smith	(which would be like doto but taking the thing as the last arg)
19:31	JohnJJ	justin_smith: sad MS is sad, I'm progressing with Maven right now
19:31	justin_smith	JohnJJ: cool
19:32	mlb-	I've been doing things like (->> stuff (map transform) (filter predicate?) (#(doto % (->> count (spy :info "Count of stuff")))))
19:32	justin_smith	mlb-: aha - as-> may help there
19:33	mlb-	and then continuing the chain, unaffected by the doto's transformation on the collection
19:33	technomancy	(doto prn) in -> is the best
19:33	justin_smith	yes it is
19:33	Bronsa	true
19:34	mlb-	I'm using a slightly more verbose syntax with timbre's spy to help my coworkers who aren't used to thread-last chains across collections
19:36	mlb-	thanks for the reassurance, everyone =]
20:11	whodidthis	whats an easy way to forward all messages from one channel to another
20:13	justin_smith	whodidthis: looks like that is what pipe does
20:13	whodidthis	oh cool
21:18	wei	i have a long sequence of numbers, and i want the sum at the end as well as the max value that the sum ever took. what’s the best way to get these two values without having to go through the sequence twice?
21:21	justin_smith	,(reduce (fn [[mx sum] v] (let [update (+ v sum)] [(max update mx) update])) [0 0] [1 2 -30 8 12])
21:21	clojurebot	[3 -7]
21:22	justin_smith	,(reductions (fn [[mx sum] v] (let [update (+ v sum)] [(max update mx) update])) [0 0] [1 2 -30 8 12]) ; for good measure
21:22	clojurebot	([0 0] [1 1] [3 3] [3 -27] [3 -19] ...)
21:22	andrewhr	that was fast :O
21:22	andrewhr	I was just tinkering on repl…
21:23	justin_smith	andrewhr: it took me a few tries
21:24	andrewhr	hahaha ok. I will keep up practing my clojure-fu
21:24	muraiki_	you could also use http://clojuredocs.org/clojure.core/reductions
21:24	justin_smith	muraiki_: see my second example
21:24	muraiki_	oh
21:24	muraiki_	haha
21:24	muraiki_	didn't read that far yet T_T
21:25	justin_smith	well I didn't use reductions to find the max - simpler to just update in one pass
21:25	justin_smith	though I could have used reductions of sum, and taken the max of that
21:25	muraiki_	yeah, that's the approach I was thinking of
21:25	muraiki_	my clojure-fu is weak though
21:25	andrewhr	same approach here
21:26	justin_smith	,((juxt #(apply max %) last) (reductions + 0 [1 2 -30 8 12]))
21:26	clojurebot	[3 -7]
21:26	justin_smith	actually the reductions / max version is elegant
21:26	justin_smith	but does an extra walk, which is against the first criteria :)
21:26	justin_smith	actually it walks the full length 3 times
21:29	hyPiRion	speaking of, is there anyone else besides me finding a function à la this useful?
21:29	hyPiRion	,(defn aside [& fs] (fn [vs x] (mapv #(%1 %2 x) fs vs)))
21:29	clojurebot	#'sandbox/aside
21:29	hyPiRion	,(reduce (aside + - max) [0 0 0] (range 10))
21:29	clojurebot	[45 -45 9]
21:30	justin_smith	oh, very nice
21:31	justin_smith	it captures a pattern I often do, where I want to do more than one reduction on the same sequence
21:31	hyPiRion	yeah, I tend to do that a lot too – It's like juxt on reductions
21:32	wei	thanks guys. so my sequence of numbers is the result of a transduce
21:32	wei	is there any benefit to writing the reduction as the final transduce step?
21:33	wei	and is that even possible?
21:33	justin_smith	in that case, you should be able to compose the other ops with the transduce