#clojure – 2013-03-22

2013-03-22

00:54	clifton	hi, whats the favored pattern for reading/accessing environmental variables from config/(dev\|production).clj
00:54	clifton	sneaking some clojure into our production stack :(
00:54	clifton	:)
01:04	hyPiRion	clifton: oh, I've heard people talking about environ for that
01:04	hyPiRion	https://github.com/weavejester/environ
01:07	clifton	thanks hypirion
01:07	hyPiRion	no problemo
01:26	tomoj	what's the other one?
01:26	tomoj	ah, carica
02:44	tomoj	why design IWatchable as it is instead of having add-watch return a remove-watch function?
02:44	tomoj	s/IWatchable/IRef/ for jvm
02:46	amalloy	tomoj: it wouldn't be a lot of fun if you had to keep track of the remove-watch function elsewhere. the way it is now, you can often remove a watcher without having to save the result of add-watch anywhere
02:47	amalloy	eg, perhaps you have a bunch of files open, and a watcher for each, keyed by the filename
02:47	amalloy	when you close a file, you can remove its watch
02:48	tomoj	that does sound more fun
02:49	tomoj	but should we design our protocols to maximize fun?
02:49	amalloy	just imagine the fun of having to add another ref to manage the state of the watchers on your first ref
02:49	amalloy	i can hardly wait
02:50	tomoj	I see
02:50	tomoj	I was thinking since you can implement one in terms of the other, the protocol should specify the easiest one to implement
02:50	tomoj	but it's not symmetrical
02:51	tomoj	hmm
02:53	tomoj	it does seem strange to put the choice of having another ref right in the protocol
02:55	tomoj	I mean, I'd believe it makes the most sense for the refs in clojure and taking all the other factors I don't know about into account. but in general hiding another ref inside every ref seems strange
02:57	tomoj	oh, of course it makes sense
02:57	tomoj	atoms would be unusable otherwise
03:00	tomoj	ämalloy: thanks :)
03:42	ivan	I see https://github.com/strangeloop/clojurewest2013/tree/master/slides has something now
05:22	muhoo	huh, that's weird, an app just stopped working completely on lein 2.1.1 that worked fine on 2.0.0
05:23	muhoo	lein cljsbuild hangs, lein trampoline repl hangs, just hangs. in production, of course. works fine on my dev system :-/
05:24	muhoo	only difference i can see is producion is java 6, my machine is java 7.
05:30	ivan	muhoo: lein 2.1 changed the default :jvm-opts, could be something else though
06:39	necronian	How do I set up nrepl to allow remote connections? If I do a lein repl and then try to connect to it from a remote machine I get ConnectException Connection refused: connect
06:45	borkdude	necronian might be a firewall issue?
06:46	necronian	borkdude: That was my first thought, so I specifically set the machine to allow any incoming connection from anywhere and it didn't help
06:47	borkdude	necronian are you connecting to the right port number?
06:47	borkdude	necronian lein repl :connect <portnr>
06:48	necronian	It says nrepl server started on port 48333 and I try to connect with lein repl :connect adserver:48333 on the remote machine
06:48	borkdude	necronian no idea why it wouldn't work. try #leiningen also perhaps?
06:49	necronian	Ah, thanks good idea. I'm sure it's something stupid but I have no idea what.
06:54	ironm	necronian, did you check with "namp" listening ports on your server? Further you can trace with wireshark what is going on
06:54	ironm	"namp"
06:54	ironm	nmap
06:54	ironm	sorry
06:55	necronian	I can try to scan it with nmap, I've never used wireshark before
06:58	ironm	necronian, if you run linux it should be easy (to use wireshark)
07:02	pwned	it's even easier to run a tcpdump
07:07	necronian	ironm: Well the port is clearly not open I have no idea why. It'll probably be easier for me to set up a ssh tunnel.
07:08	ironm	necronian, have you tested locally on the server ?
07:08	ironm	in case you have an ssh access
07:09	necronian	ironm: Yes I can connect fine locally, just have an issue remotely.
07:10	ironm	it looks like a "firewall" issue
07:14	jcromartie	I want a lib that lets me go into a namespace and records def.* calls into a corresponding file
07:14	jcromartie	does that exist?
07:14	necronian	ironm: I know, but I can't see how, it's all on the same lan, iptables has no rules and I put ALL: ALL in host.allow
07:15	pepijndevos	Theory: There are more functional programmers than functional jobs.
07:15	jcromartie	so like, when I go into a namespace at the REPL, it replaces all of the def.* bindings with ones that rewrite the file
07:16	jcromartie	pepijndevos: corollary: functional programmers could fight to the death over functional jobs as a form of amusement
07:17	pepijndevos	jcromartie: amusement?
07:17	jcromartie	that is: be made to fight, for the amusement of others
07:17	jcromartie	e.g. gladiators
07:17	pepijndevos	ah
07:19	jcromartie	this is a start
07:19	jcromartie	(defmacro scribe! [code] (spit "scribe-session.clj" (str (prn-str code) "\n") :append true))
07:19	jcromartie	at least I find it handy
07:27	necronian	ironm: Thanks for the help, found the problem. Set the LEIN_REPL_HOST to the ip I wanted to bind to. Binds to localhost by default.
07:46	eckroth	hi, I'm trying to use a java constructor that is overloaded based on the 2nd param, but I need to pass nil to this second param; since nil apparently can't have a type, the wrong constructor is being called
07:46	eckroth	how can I fix this?
07:48	Bronsa	eckroth: have you tried (let [^type x nil] (Ctor. foo x)) ?
07:49	Bronsa	(I don't know if this will work, worth a try though)
07:50	eckroth	Bronsa: wow, nice job! I tried a similar thing that failed, because I didn't immediately call the Ctor
07:50	eckroth	Bronsa: thanks!
07:51	Bronsa	np
07:56	ironm	necronian, yes .. thank you. Such setting is often a source of "strange" issues
08:01	eckroth	In case anyone is interested, I updated the machine learning library clj-ml (Weka wrapper for Clojure) to support string-to-word-vector filters (for text classification, etc.): https://clojars.org/cc.artifice/clj-ml
08:03	tomoj	cool
08:06	necronian	ironm: Yes, I don't think I can be held responsible for not finding it, I grepped the leiningen repo, there are 5 references to the variable none of them can be found by by reading the docs or through lein help.
08:06	tomoj	I'm guessing :words-to-keep is the number of most frequent not-stoplisted words, and that is the dimension of the vectors?
08:07	tomoj	..number of words to keep, obviously
08:14	ironm	necronian, could you document it somehow in the wiki (in case there is one) ? I guess a lot of newcomers will be very thankful for it (inluding me) ;)
08:17	necronian	ironm: I'll try filing an issue for the documentation on the leiningen repo tonight
08:18	eckroth	tomoj: yes, sorry, was away; words to keep is number of final words in vector
08:19	eckroth	tomoj: I forget how weka does it but I presume it's the most frequent, or ones with highest tf-idf or whatever measure you choose
08:19	tomoj	thanks, I found the original javadoc
08:19	eckroth	ok
08:19	tomoj	I never tried weka because their docs scared me away, thanks for clj-ml, maybe I'll try it
08:19	eckroth	tomoj: for what it's worth, I plan to further improve this library so I can fully use it :) stay tuned
08:19	eckroth	tomoj: ha, yeah, and the weka interface is even worse
08:20	ironm	necronian, thanks a lot :)
08:20	eckroth	tomoj: I have a lecture about weka; you can check out this page for an example of using its interface: http://cse3521.artifice.cc/weka.html
08:21	tomoj	hmm
08:21	babilen	tomoj: clj-ml is still maintained?
08:21	tomoj	I think that was part of what scared me away
08:21	eckroth	babilen: I have a fork I'm improving, because I want to use it for a project
08:23	eckroth	babilen: clj-ml may not fit the bill; I don't plan to reimplement any ML algs, but rather just utilize weka and whatever else
08:23	babilen	eckroth: That is nice - I stumbled over clj-ml some time ago, but dismissed it because it did not look maintained at all. Great that you took over and might bring it back to life.
08:23	eckroth	babilen: thanks, stay tuned
08:23	eckroth	got to go
08:24	tomoj	the core.matrix protocols seem very large
08:24	babilen	eckroth: Yeah, that's fine, I never meant that you should. It is just that I observe several things to take shape that would allow for the implementation of something that begins to feel like scipy/numpy/...
08:24	tomoj	do we really need all that?
08:25	tomoj	I'd like for it to be the case that pretty much any matrix/vector operation can be implemented via application of a single simple protocol method :(
08:26	babilen	tomoj: I haven't looked at it in some time, but the aim is to support a number of numeric libraries ... They also reported some impressive speed-ups recently. Let me find the mail.
08:27	tomoj	speedups relative to previous versions of core.matrix?
08:27	tomoj	seems a bit presumptuous to put it in clojure.* :)
08:27	babilen	no, to something not using core.matrix -- but let me try to find it before I make things up on the spot :)
08:28	tomoj	I was thinking the interface seems oriented around strictness
08:28	babilen	http://clojurefun.wordpress.com/2013/03/07/achieving-awesome-numerical-performance-in-clojure/
08:29	tomoj	but I guess there's no reason you couldn't make a crazy matrix type that doesn't perform operations strictly
08:29	babilen	But yeah .. that person is definitely convinced that it is great.
08:30	babilen	+ https://gist.github.com/siscia/5172981
08:30	tomoj	hmm
08:30	tomoj	I think we should be able to get much closer to vectorz
08:30	tomoj	"mutable operation" ?!
08:33	tomoj	I'd like to be able to compose a bunch of matrix operations in the style of clojure.core.reducers, and let the particular implementation how best to physically realize the entire composition
08:34	tomoj	s/how/decide how/
08:37	tomoj	thinking primarily of jcublas
08:37	tomoj	won't core.matrix's protocols be sort of problematic in that environment?
08:40	babilen	tomoj: Maybe you want to raise that on their mailing list (numerical-clojure on google groups) - I can't really comment on that.
08:55	tomoj	it looks like whether you can, for example, mset! on the return value of new-vector depends on the implementation used?
08:59	tomoj	I guess that makes sense
09:45	jcromartie	how can I get it to stop throwing an exception when I try to print my defrecord in the REPL
09:45	jcromartie	i.e. this print-method BS
09:45	jcromartie	I've set up a print-method for the type itself
09:46	jcromartie	never mind… stateful multimethod mell
09:46	jcromartie	*hell
09:56	jcromartie	invitation to critique my first take on this event sourcing design: https://gist.github.com/jcromartie/5221441
09:57	jcromartie	it needs more examples, but the basic idea is that you use multimethods to define how events change your state, and then you can record and replay (in memory or when loading from disk) them on some initial state
09:59	jcromartie	and it doesn't let you write things that can't be read back… although in a rather brute force fashion for now
09:59	jcromartie	this is for small numbers of writes, large numbers of reads
10:00	borkdude	can the chunk size of lazy sequences be changed (for demo purposes)?
10:02	borkdude	can the chunk size of lazy sequences be changed (for demo purposes)?
10:04	pjstadig	borkdude: as far as i understand the chunk size is a decision the data structure makes usually at the level of the java implementation, since it know the most efficient way to batch things
10:04	pjstadig	you might be able to unchunk it and rechunk it at the clojure level
10:04	borkdude	pjstadig I just want to show that side effects only occur once per produced item in some demo coe
10:05	pjstadig	so you want a chunk size of 1?
10:05	borkdude	pjstadig yes
10:05	pjstadig	you can create a lazy seq that is not chunked to demonstrate that using lazy-seq
10:05	pjstadig	or maybe use a list?
10:05	borkdude	pjstadig lists are not lazy
10:06	pjstadig	i think generally vectors are the thing that usually causes chunking
10:06	pjstadig	i mean map a side effecting functionacross a list
10:06	pjstadig	map will only chunk if the seq it is mapping is chunked
10:07	borkdude	pjstadig I could just simplify the output in the demo, this is about lazyseqs which are always chunked I think?
10:07	pjstadig	,(first (map (partial println) [1 2 3 4 5]))
10:07	clojurebot	1\n2\n3\n4\n5\n
10:07	pjstadig	,(first (map (partial println) '(1 2 3 4 5)))
10:07	clojurebot	1\n
10:09	borkdude	pjstadig ah ehm, ok yes :)
10:09	borkdude	pjstadig so it depends on the input whether map will chunk its ouput
10:09	pjstadig	yeah
10:09	pjstadig	~source map
10:09	borkdude	pjstadig I had the example of map over (range)
10:10	pjstadig	yeah range is another one that creates a chunked seq
10:10	borkdude	every day I learn something new :)
10:10	pjstadig	~source range
10:11	Ember-	hi, I have a slight problem with enlive
10:11	Ember-	set-attr seems to encode my string no matter what I do
10:11	Ember-	that's not what I want since I have get parameters in it
10:11	Ember-	and & is transformed into &
10:12	Ember-	I cannot find any options or anything to bypass this behavior
10:12	Ember-	enlive is great but it's documentation is really lacking
10:20	Ember-	so either everyone who would know answer to my question is sleeping or no one actually knows the answer
10:20	Ember-	damnit
10:23	devn	Hm, is org-babel now basically dead due to nrepl?
10:23	devn	org-babel-clojure*
10:23	TimMc	muhoo: Oh, the reflection stuff! I'd basically forgotten about that.
10:24	stuartsierra	devn: I use org-babel with nREPL.
10:25	TimMc	muhoo: Is it in a basically-usable state? If so, I should just push what's there and tweak it later.
10:30	devn	stuartsierra: you have your dotfiles out on github?
10:30	devn	stuartsierra: i found http://blog.jenkster.com/2013/02/org-babel-clojure-and-nrepl.html -- is that sufficient?
10:30	stuartsierra	devn: Yes: github.com/stuartsierra/dotfiles
10:30	Anderkent	Any help with debugging horrid compilation times? (3+ minutes to AOT compile a project)
10:30	devn	stuartsierra: cool, thanks.
10:30	stuartsierra	np
10:47	jcidaho	anyone using midje and getting junit output as to work with a CI server?
10:48	jcidaho	I've seen a google groups post a year old about it, and an issue on midje github been going a year or so
11:05	Anderkent	jcidaho: I was looking at junit output for clojure.test a while ago, but the writer was quite buggy. Pretty much gave up on that. Don't know about midje though, but they'd have to write it themselves
11:05	Anderkent	clojure.test.junit is pretty much useless
11:31	jtoy_	is anyone here using this library? https://github.com/dakrone/cld
11:32	jtoy_	if so, could someone test this line for me: (cld.core/detect "你好吗") when I run it freezes the repl
11:33	jtoy_	which seems like a crapp bug
11:35	wastrel	你好
11:35	jtoy_	??
11:35	lazybot	jtoy_: Definitely not.
11:35	jtoy_	lazybot: it works for you?
11:35	wastrel	?foo
11:35	wastrel	??foo
11:35	wastrel	maybe it's a bot response
11:35	wastrel	is it abot?
11:35	jtoy_	ic
11:36	wastrel	??
11:36	lazybot	wastrel: What are you, crazy? Of course not!
11:36	dakrone	jtoy_: yes, we use it
11:36	wastrel	heh
11:36	dakrone	jtoy_: it may be that your REPL is freezing, not actually cld, can you try it in a file?
11:36	jtoy_	dakrone, yeah, I just ran the examples in your README, I tested it with any random chinese strings and it just freezes the repl
11:36	wastrel	i just saw 你好吗 and thought i was in #mandarin for a sec
11:36	jtoy_	every time
11:37	dakrone	if I recall correctly, jline has issues with some UTF-8 stuff when used for the REPL
11:37	dakrone	but feel free to open an issue on cld
11:37	jtoy_	dakrone: ok, ill assume its the repl then for now
11:37	jtoy_	ill do some more testing first
11:41	jtoy_	this is my first time usine language-detect, but on short sentences its got some funny results: (cld.core/detect "hello")
11:41	jtoy_	["nl" {"nl" "0.5714277977843022", "fi" "0.42856886520061827"}]
11:42	trptcolin	jtoy_ / dakrone: right, non-BMP characters break jline, but those characters are looking fine to me in `lein repl`. so maybe they're in the BMP set?
11:42	jtoy_	(cld.core/detect "bye") => ["no" {"no" "0.9999913661473905"}]
11:42	jtoy_	trptcolin: im on lein1, not sure if its fixed in lein2
11:42	dakrone	jtoy_: right, so cld uses statistical models for determining the languages, the shorter the text, the less likely it is to be correct
11:43	trptcolin	jtoy_: yup, that'll do it
11:43	dakrone	jtoy_: I know lein1 has that issue, I used to run into it all the time
11:43	dakrone	jtoy_: you'll need to pass more than a single word in to get a better answer
11:47	jcromartie	please critique my new lib: https://github.com/jcromartie/deck
11:47	jcromartie	I am new at this. so I don't have a clojar or anything
11:52	Raynes	jcromartie: Looks kinda like https://github.com/flatland/cassette. But there aren't any docs yet so sorry. I'll probably write some today.
11:53	jcromartie	aw snap
11:53	jcromartie	:)
11:53	jtoy_	dakrone: do you know in what situation this error could occur: CompilerException java.lang.RuntimeException: com.cybozu.labs.langdetect.LangDetectException: no features in text, compiling I already remove all nil and empty strings
11:53	jcromartie	well great minds think alike huh
11:53	Raynes	It probably isn't quite the same though.
11:53	dakrone	jtoy_: if the string were entirely numbers I think it might cause that also
11:54	Raynes	jcromartie: Cassette is essentially an implementation of Apache Kafka's file format in Clojure. It writes binary data to disk using gloss sequentially in a simple format and you can iterate through each entry and play it back. It's an append-only format.
11:54	jtoy_	ah, yes
11:54	jcromartie	yeah this is not so slick
11:55	jcromartie	it's literally a single append-only log file
11:55	Raynes	jcromartie: deck is?
11:55	jcromartie	yeah
11:55	Raynes	So is cassette. :D
11:55	jcromartie	oh I see something about subdirs
11:56	jcromartie	but I get it
11:56	Raynes	Except it partitions log files after n bytes.
11:56	Raynes	(like kafka)
11:56	jcromartie	neat
11:56	Raynes	By partition I mean it just starts a new file once the old file runs out of the space pre-allocated.
11:56	jcromartie	maybe I can make mine not as dumb in the future :)
11:57	Raynes	Well, I wasn't saying yours is dumb or anything. I only gave it a cursory glance.
11:57	jcromartie	oh no, mine is dumb
11:57	jcromartie	but for now it's tested stable and fast enough for many n in terms of concurrent writes and numbers of writes
11:57	Raynes	lol
11:57	jcromartie	my use case is for fairly small datasets
11:58	bttf	what does 'foo actually return ?
11:58	Raynes	Our use case for cassette is "Let's save ALL the things!" so that we can play them back later if we want.
11:58	bttf	a symbol named foo ?
11:58	jcromartie	like an administrative type web app, where you're just configuring a few things
11:58	Raynes	bttf: Yes.
11:58	bttf	ij
11:58	bttf	ok*
11:58	Raynes	Darn keyboards.
12:01	mpenet	mapdb is also nice for simple persistance
12:01	mpenet	it's just maps (more or less) from clojure
12:02	jcromartie	I like that but I have problems keeping the data model consistent
12:02	jcromartie	the idea behind my scheme is to only store inputs
12:02	jcromartie	(hello OOTTP)
12:03	jcromartie	so I can change the way the inputs build up the in-memory "current" data model
12:03	devn	Does anyone know the rationale behind choosing SSE over BrowserChannel?
12:03	devn	(in Pedestal)
12:17	pbostrom	devn: I was wondering something similar, but s/BrowserChannel/Webockets/
12:18	ppppaul	is it easy for me to integrate a pedestal service with angularjs?
12:18	technomancy	websockets are a lot harder to support in servers, proxies, routers, etc
12:28	pbostrom	technomancy: interesting, do you think that's a result of lack of adoption in the industry or technical limitations of websockets spec?
12:30	tbaldridge	devn: Mostly it was because it was low hanging fruit. Two co-workers and I implemented SSE in about 6 hours. Websockets would have taken longer.
12:30	tbaldridge	Websockets also allow for duplex communication, something that hasn't been added to pedestal yet.
12:31	tbaldridge	That being said feel free to submit a patch :-)
12:31	technomancy	pbostrom: SSE is just strictly easier technically
12:31	technomancy	lots of HTTP infrastructure assumes half-duplex
12:32	technomancy	websockets gets more buzz and is more flexible, but most people who say they want it would be happy with SSE
12:33	devn	tbaldridge: Gotcha. :)
12:34	pbostrom	I also wonder if an SSE implementation can handle more clients than websockets
12:35	jtoy_	can somone show me an exampe that uses map and inc together, Im trying to do (map #( stuff ) ["foo","bar"]) => ["1foo" "2bar" ]
12:36	tbaldridge	jtoy_: I often use this (map #(str %1 %2) (range) ["foo" "bar"])
12:37	llasram	&(doc map-indexed)
12:37	lazybot	⇒ "([f coll]); Returns a lazy sequence consisting of the result of applying f to 0 and the first item of coll, followed by applying f to 1 and the second item in coll, etc, until coll is exhausted. Thus function f should accept 2 arguments, index and item."
12:37	jtoy_	tbaldridge: ah, range works too
12:37	tbaldridge	there's also maped index
12:37	tbaldridge	*mapped-index
12:37	jtoy_	how does mapped-index implement it? with an atom?
12:37	tbaldridge	https://github.com/clojure/clojure/blob/master/src/clj/clojure/core.clj#L5618
12:38	tbaldridge	bleh
12:38	tbaldridge	that's the link: https://github.com/clojure/clojure/blob/master/src/clj/clojure/core.clj#L5618
12:38	tbaldridge	it's not pretty, because it has to deal with chunked seqs
12:39	jcrossley3	tbaldridge: why io.pedestal.service.log instead of clojure.tools.logging?
12:40	jtoy_	i ddint know you could pass multiple collections like that to map, nice
12:41	devn	tbaldridge: I'm trying out the chat application and am getting some exceptions. Are there some known issues?
12:42	tbaldridge	devn: not to my knowledge, it's been about a week or so since I personally have run the app though.
12:42	tbaldridge	what's the error?
12:42	ppppaul	any benchmarks comparing pedestal to ring?
12:42	tbaldridge	ppppaul: pedestal uses ring....so....
12:42	devn	tbaldridge: https://gist.github.com/devn/5222833
12:43	ppppaul	oh
12:44	tbaldridge	ppppaul: there were some design things that we decided to change in ring, those have then been back ported into mainline ring. So pedestal is now built right on top of ring
12:44	ppppaul	hmmm
12:44	ppppaul	when was this done?
12:45	ppppaul	i'm always looking at the ring docs, haven't seen anything related to pedestal
12:45	tbaldridge	https://github.com/ring-clojure/ring/commit/a8d5aa69b1153af7491bbbf4b70042f7c8c407be
12:45	ppppaul	woah, that's a huge commit
12:46	tbaldridge	So for our pedestal we needed support for stopping/resuming ring middleware. So the interceptors library was born, from there we refactored all the default middleware. So the fact that you never noticed was by design. We wanted to work with ring, but get the "extra features" that we needed.
12:46	ppppaul	pretty cool.
12:47	gfredericks	is "ring" supposed to refer to the interceptors style now then?
12:47	ppppaul	i'm going to read through the whole commit
12:47	devn	interceptors are bigger than basic ring handlers
12:47	tbaldridge	devn: I've contacted a co-worker about that issue. I'm not sure what the answer is, perhaps email the pedestal google group?
12:47	weavejester	gfredericks: No, the standard middleware was just split into two parts to aid async usage.
12:48	gfredericks	weavejester: are there two SPECs now?
12:48	weavejester	gfredericks: Essentially as well as wrap-foo, you also get foo-request and foo-response
12:48	devn	tbaldridge: it's not a big issue for me, but i figure i'd ask someone who might now
12:48	weavejester	gfredericks: No, the SPEC hasn't changed.
12:48	devn	know*
12:48	gfredericks	weavejester: and wrap-foo can be derived from the other two
12:48	tieTYT	i was developing yesterday and I have this issue where I have a function and I want to run map on it, but I'm not sure if I'm running on a collection or a collection of collections. I really want the former. How can I be sure I'm passing in the correct argument?
12:49	weavejester	gfredericks: Correct. Ring 1.2.0 exposes some new functions, but the existing middleware acts the same way.
12:49	tieTYT	i'm coming from haskell and this is very easy. You just write a type signature that takes a list instead of a list of lists
12:49	gfredericks	tieTYT: check what sort of argument it is? knowing what kind of thing a thing is is sort of separate from passing it to map
12:49	tieTYT	gfredericks: how do you do that?
12:50	gfredericks	tieTYT: if you have the thing at the repl you can just look at it
12:50	tieTYT	right, how?
12:50	tieTYT	i've got enlive output
12:50	weavejester	tieTYT: If you want to check, you can use a pre-condition, or if you're particularly bleeding-edge, there is a typed Clojure library.
12:50	gfredericks	type the name for the thing and it'll get printed out?
12:50	tieTYT	and it's so freaking huge I can't really tell what it is by printing it out
12:50	gfredericks	ah
12:50	gfredericks	there's a type function
12:51	gfredericks	so you can (type my-thing) and (type (first my-thing)) to get some idea
12:51	tbaldridge	devn: I just heard back, it's a confirmed bug.
12:51	ppppaul	does the interceptor ring commit make clojure channels redundant?
12:51	tieTYT	gfredericks: ah ok great, thanks
12:51	tieTYT	weavejester: what's that library called?
12:51	weavejester	ppppaul: Clojure channels?
12:51	weavejester	tieTYT: https://github.com/clojure/core.typed
12:51	weavejester	tieTYT: It's rather new, however.
12:52	ppppaul	then channels library (starts with an A... it recently got removed from clojure werkz)
12:52	ppppaul	https://github.com/ztellman/aleph
12:52	tieTYT	weavejester: I see, thanks
12:53	weavejester	ppppaul: Personally, I prefer Alephs use of channels over interceptors.
12:53	weavejester	They're a better abstraction for websockets.
12:53	gfredericks	so the interceptors approach would primarily require new adaptors -- is that correct?
12:53	weavejester	gfredericks: Well, they wouldn't be Ring adapters per se, but yes, it would require some integration with a HTTP server. It wouldn't happen automatically.
12:54	ohpauleez	tieTYT: I haven't scrolled back up to see the full conversation, but you might look at using protocols too
12:54	ohpauleez	which dispatch on type
12:54	gfredericks	and that doesn't exist yet?
12:54	tieTYT	ohpauleez: well I want to reject a certain type of input instead of handle them both
12:54	weavejester	gfredericks: I assume it does, if Pedistal is being used
12:55	gfredericks	weavejester: okay, so it's integrated with pedistal
12:55	gfredericks	or built into rather
12:56	gfredericks	I'm just trying to get a sense for where the clojure webstack generally is headed
12:56	weavejester	gfredericks: Yes, I believe they have something which is loosely equivalent to an adapter, but supports interceptors
12:56	ppppaul	https://github.com/ring-clojure/ring/commit/a8d5aa69b1153af7491bbbf4b70042f7c8c407be i'm looking at this and i'm failing to see the introduction of interceptors
12:56	ppppaul	though, i am only half way through the commit
12:58	tbaldridge	ppppaul: So here's the issue. In ring (before that commit) the request and response of parts of the middleware were combined into a single function. For async work they need to be separate.
12:58	weavejester	gfredericks: I should note that I'm a little skeptical of interceptors. :)
12:58	weavejester	And of HTTP async in general.
12:58	ppppaul	i see those changes
12:58	ppppaul	i didn't realize that was all that was needed to be done
12:58	weavejester	HTTP async seems to me to be essentially a hack to get around a lack of websockets
12:59	ppppaul	so, i guess interceptors make use of those functions separately, while regular ring middleware uses them in the same place/time/thread
12:59	weavejester	ppppaul: Right
12:59	powr-toc	weavejester: is there a video up of the pedestal talk at clojure/west yet? I'm yet to decide on interceptors, but they seem pretty complicated... and don't feel very functional
12:59	tbaldridge	The interceptor library (in pedestal) then uses these functions in a different way. Looks at a vector of fns and executes them in order, or pauses at certain point.
12:59	tbaldridge	ppppaul: bingo.
13:00	ppppaul	interesting
13:00	weavejester	tbaldridge: Do you think there will be any reason to support async HTTP when websocket support becomes universal?
13:00	tomoj	what is "async http"?
13:01	ppppaul	tbaldridge, you guys tested the use of interceptors via simulant?
13:01	weavejester	tomoj: Anything that involves deliberately delaying a HTTP response I guess.
13:02	devn	tbaldridge: another thing you might want to fix in the chat client:
13:02	devn	<IMG style="width:300px; height: 300px" SRC=# onmouseover="document.location='https://www.youtube.com/watch?v=dQw4w9WgXcQ&noredirect=1'">
13:02	tbaldridge	weavejester: I'm not sure. Personally I like async when I have to run of to a DB to do work, or perhaps make calls to other servers. I've done pure TCP work before, and it's a bit of a headache.
13:02	devn	;)
13:02	tomoj	isn't async still sometimes valuable when you have many concurrent requests which don't hang open?
13:02	devn	i know it's just a sample, but i set one up and of course the first thing Brad does is try to inject
13:03	devn	s/one/an instance of the chat app/
13:03	tbaldridge	(disclaimer) I work for relevance, but nothing I say here should be taken as the gospel
13:03	weavejester	tbaldridge: That's an interesting point. I guess you can improve performance of the web server by not blocking.
13:04	ppppaul	is pedestal.io running off of pedestal?
13:04	ppppaul	wondering... cus it's buggy as fuck
13:04	technomancy	async stack traces =(
13:05	devn	ppppaul: why do you say that?
13:05	ppppaul	but, it's not communicating with the back end, so i think there is some weirdness going on with the dom or JS
13:05	ppppaul	devn, i can't scroll properly on the site, and the buttons on the site break
13:05	technomancy	someone found a way to take clojure stack traces and make them worse
13:05	devn	technomancy: haha
13:05	technomancy	devn: they said it couldn't be done
13:05	devn	well, not funny actually
13:05	devn	sort of depressing
13:06	ppppaul	technomancy, need a smart logger
13:06	tdignan	ppppaul: well, what browser are you using
13:06	llasram	technomancy: I really thought the stackholm syndrome would have set in for you by now
13:06	tomoj	it would be nice if you could swap async for sync and get the exact same semantics
13:06	tbaldridge	technomancy: that was Guido's argument against tail calls in Python as well "horrible stack traces"
13:06	ppppaul	using Version 25.0.1364.160 Ubuntu 12.10 (25.0.1364.160-0ubuntu0.12.10.1) chrome
13:06	tomoj	maybe. is that impossible and/or point-defeating?
13:06	technomancy	llasram: nope; I'm just filled with guilt over ignoring clj-stacktrace
13:07	devn	ppppaul: I'm on Version 25.0.1364.172
13:07	tomoj	if you could do that you could replay your bugs in a sync environment for easier debugging?
13:07	tdignan	ppppaul: do you have any chrome add ons?
13:07	devn	OSX -- and no issues
13:07	technomancy	tbaldridge: I'm not saying it's never worth the pain; just pointing out that you should make sure it's worth the trade-off because it's a big one.
13:07	tdignan	I'm using chrome on linux mint, no issues.
13:07	technomancy	(but guido was just using that as an excuse)
13:07	tbaldridge	real men debug with println :-P
13:08	tbaldridge	(just kidding)
13:08	ppppaul	tdignan, the site works a little better in firefox
13:08	devn	speaking of debugging... ritz...
13:08	devn	what's up with ritz, yo?
13:08	tdignan	do you have javascript enabled and everything
13:09	devn	anyone in here use ritz?
13:09	ppppaul	i have extensions and JS enabled. i don't see anything coming up in the console.
13:09	ppppaul	i don't see any network requests
13:10	ppppaul	my extensions don't effect other sites
13:10	devn	ppppaul: if you have user scripts or other user installed extensions try disabling them and see what happens
13:10	ppppaul	cept for flashblock, which usually improves performance
13:10	devn	oh, the pedestal.io site is completely written in flash
13:10	devn	(jk)
13:12	tieTYT	is there any plans to make declare unnecessary in the language?
13:12	`fogus	tieTYT: I doubt it.
13:13	tieTYT	oh
13:13	llasram	tieTYT: Note that `declare` is something you find very rarely in most Clojure code
13:13	nDuff	Multi-pass compilation would be a pretty big change.
13:13	nDuff	...would mean things wouldn't behave the same way in and out of the REPL.
13:14	llasram	Multi-form, really
13:14	tieTYT	llasram: is that because people have written their code to adjust to that?
13:14	technomancy	the talk on racket at clojurewest made me really appreciate that clojure got its compilation model right
13:14	nDuff	tieTYT: People usually write their code to declare things in dependency order, yes.
13:14	technomancy	so much nonsense you just don't have to think about =)
13:14	tieTYT	personally, I like to start at the high level and break it down
13:14	nDuff	tieTYT: If you don't do that, you can't run the same code in a REPL, and that's silly.
13:14	tbaldridge	(inc techomancy)
13:14	lazybot	⇒ 1
13:14	nDuff	(well, can't run it in a REPL starting from a blank namespace)
13:15	ppppaul	tbaldridge, can you help me understand how pedestal scales?
13:15	`fogus	tieTYT: Does "high-level" need to mean "top of file"?
13:16	tieTYT	no, it's just the way I write my code. I code outside in because if I code inside out, I often create code I realize I don't need to use
13:16	tbaldridge	ppppaul: I could try, but you'll probably get a better answer off the google group list. I have more experience with the Pedestal App code.
13:16	tieTYT	I don't think clojure gets in the way with that, I'm just used to languages that let me put the code in any order I want so I'll have to adjust
13:17	`fogus	tieTYT: Me too. :-)
13:17	tieTYT	`fogus: so do you start at the bottom and add upwards?
13:17	tieTYT	add code upwards, I mean
13:18	`fogus	tieTYT: I don't know. It depends on the problem. I usually start with INDATA -> OUTDATA and work inward from there
13:19	tieTYT	i see
13:20	`fogus	I work in Emacs, so top or bottom doesn't matter. What matters is when I hit C-e
13:20	tieTYT	what does that do?
13:20	tieTYT	i've never used emacs
13:20	`fogus	eval form
13:21	`fogus	I meant C-xC-e :p
13:22	tieTYT	i think i'm missing a step. Are you rearranging the code after you do that?
13:22	tieTYT	after=before
13:23	`fogus	tieTYT: No.
13:23	tieTYT	then you're naturally writing your code in a way that avoids declare or you use declare?
13:24	`fogus	tieTYT: You're conflating where the text lies in the file with when it needs to be written
13:24	`fogus	I write the functions when I need them and lay them where the compiler likes them
13:25	tieTYT	ok
13:25	`fogus	Moving up or down is the same number of keystrokes... no productivity lost
13:26	tomoj	until next repl session :)
13:27	`fogus	tomoj: Isn't that a problem in every Lisp?
13:28	`fogus	Any time you run in a REPL you run the risk that what's in memory is not the same as in text.
13:28	`fogus	That was a problem that Smalltalk didn't have... ;-)
13:29	technomancy	we call that "getting slimed"
13:29	`fogus	technomancy: I like that
13:41	pjstadig	but we'll have to change it
13:41	pjstadig	to getting nrepl'ed?
13:41	pjstadig	doesn't have the same ring
13:41	technomancy	let it live on as a piece of history
13:41	pjstadig	hehe
13:42	pjstadig	like car and cdr
13:42	kanwei	anyone here using Pedestal?
13:43	tbaldridge	pjstadig: it's funny to here the miniKanren people "so if we simple cooder down the list..."
13:43	tbaldridge	wwaaatt? oh....cdr-down the list...I see....
13:43	tbaldridge	<-- no scheme background
13:45	TimMc	cdadadddddaaaar
13:45	TimMc	better than juxt
13:45	pjstadig	tbaldridge: it's a lisp problem too
13:46	tbaldridge	skeuomorphs ftw
13:47	amalloy	TimMc: if i am ever trapped in a lisp factory, i will find a reason to write "help i'm trapped in a lisp factory", converted to ascii and encoded with a=0,d=1 into a giant cadr function and hope that someone finds my message in a bottle
13:50	pjstadig	amalloy: don't worry...i got your back
13:53	antoineB	hello, i would like to change how a defrecord are printed to sreen, which prototype should i change?
13:53	gfredericks	is that possible?
13:53	technomancy	print-method
13:54	jcromartie	:) just did it this morning antoineB
13:54	davidchambers	I've spotted a typo at http://clojure.org/sequences, which doesn't appear to be part of the gh-pages branch of clojure/clojure. Is this wiki publicly writable?
13:54	antoineB	is it a multi method?
13:54	jcromartie	just (defmethod print-method YourType [obj ^java.io.Writer writer] ...)
13:54	jcromartie	then (.write writer ...)
13:54	antoineB	ok thanks
13:55	tieTYT	if I have a situation where a multimethod might work, should I attempt to use a protocol instead?
13:56	technomancy	tieTYT: no
13:56	antoineB	i will suspect prototype are a little bit faster
13:56	tbaldridge	tieTYT: I suggest use protocols if you will always dispatch on the object type, or you need speed
13:57	tieTYT	can multimethods dispatch on the object value?
13:57	tbaldridge	yes
13:57	tieTYT	oh ok
13:57	tieTYT	i need to find some better tutorials on it
13:57	tbaldridge	multimethods take a dispatch function, the output of that function is what it dispatches on
13:57	tbaldridge	(defmulti foo (fn [x] (:bar x)))
13:57	technomancy	multimethods have a lot fewer confusing edge cases around reloading and AOT
13:57	technomancy	>_<
13:58	antoineB	tieTYT: multi dispatch on everything you want about the arguments you pass in
13:58	tbaldridge	technomancy: really? because if you reload you can't redef the dispatch fn of multimethods
13:58	tbaldridge	or has that changed?
13:58	technomancy	tbaldridge: right; there is one confusing edge case. which is a lot less than the number of confusing edge cases surrounding protocols.
13:59	technomancy	it's still stupid, but not as stupid.
13:59	technomancy	err--not as annoying
13:59	tbaldridge	what are the edge cases for protocols? I can't think of any I've hit.
14:00	technomancy	if you AOT code that uses a protocol and the protocol's .class files go away, it can't fall back to JIT
14:00	technomancy	if you reload a defrecord it breaks equality; you can have two visually indistinguishable objects that are not equal.
14:00	tbaldridge	ah, that's it then. I never AOT code, so that's why I haven't hit it
14:00	tbaldridge	interesting, I'll keep that in mind, thanks.
14:00	technomancy	there are more, but I've literally never used protocols; that's just what I've got by osmosis hanging out in IRC.
14:01	technomancy	I'd be tempted to fix the defmulti reloading bug if it weren't for jira
14:02	antoineB	did you ever need to reloading ?
14:02	`fogus	technomancy: If you fix it I'll deal with Jira. ;-)
14:02	Bronsa	I was just going to propose that too.
14:03	technomancy	`fogus: I'll add it to the seajure project list then =)
14:03	`fogus	But I'll happily defer to Bronsa. ;-)
14:03	`fogus	:p
14:03	`fogus	technomancy: Where is that list?
14:04	technomancy	`fogus: https://github.com/Seajure/seajure/wiki
14:04	technomancy	I really want to do print-method for namespaces
14:05	`fogus	Oh! I've never thought of that. That would be fun
14:05	cap10morgan	ah ha! lein new midje foo-bar is creating the namespace foo_bar.core instead of foo-bar.core. gonna pull request that.
14:05	technomancy	seems like a silly omission
14:10	ToBeReplaced	antoineB: i'm curious in the solution... was the answer just (defmethod print-method ClassName [r writer] (.write writer "my string function")) ?
14:12	antoineB	yes
14:15	antoineB	the question we can ask is why using a defmethod over a protocol for print-method? (i think it is a protocol for cljs)
14:17	technomancy	antoineB: print-method is IO-bound by definition; dispatch overhead is not an issue
14:17	technomancy	also protocols didn't exist back then, but whatever
14:18	sundbp	i'm trying to use gen-class with :impl-ns to have the implementation in another namespace and avoid AOT
14:18	antoineB	ok
14:19	sundbp	i'm strugglign and getting these kinds of errors: com.tulos.lo.examples.ExampleCljAddIn> (com.tulos.lo.examples.ExampleCljAddIn.)
14:19	sundbp	CompilerException java.lang.ClassFormatError: Duplicate method name&signature in class file com/tulos/lo/examples/Exam\
14:19	sundbp	pleCljAddIn, compiling:(NO_SOURCE_PATH:1:24)
14:20	sundbp	that namespace has a :gen-class statement in the (ns ..) part. and uses :impl-ns 'tulos.addin
14:21	sundbp	in tulos-addin are all the functions to implement the gen-class. in the gen-class namespace there's no content apart from the (ns ..) form
14:21	sundbp	does that sound like the right starting point?
14:28	jtoy_	does map-index not happen in order? it seems like after having 10 items, 10 appears first, is there a way to force the order?
14:29	ToBeReplaced	what is map-index?
14:30	mpenet	it does happen in order, it would be a bit useless it that was not the case
14:30	jtoy_	i see what happens, nm
14:30	jtoy_	yes, exactly
14:30	AimHere	ToBeReplaced, there's a function map-index that assigns a numeric value as the first argument of the function you're mapping over
14:31	AimHere	map-indexed, rather. Presumably map-index is the index
14:31	AimHere	&(map-indexed (fn [idx _] idx) [:a :b :c :d])
14:31	lazybot	⇒ (0 1 2 3)
14:32	ppppaul	&(map-indexed (fn [& _] (reverse _) [:a :b :c :d])
14:32	lazybot	java.lang.RuntimeException: EOF while reading, starting at line 1
14:33	ppppaul	&(map-indexed (fn [& _] (reverse _)) [:a :b :c :d])
14:33	lazybot	⇒ ((:a 0) (:b 1) (:c 2) (:d 3))
14:37	devn	Does Michael Klishin hang out in here ever?
14:40	amalloy	devn: you mean antares? yes
14:43	uvtc	How can I iterate over a vector, not only getting its item but also its index? (I think I'm looking for the side-effectey version of map-indexed...) I think I Python it's called `enumerate`.
14:44	tbaldridge	uvtc: fastest way to process vectors is reducers
14:44	tbaldridge	uvtc: unless you use loops, using nth+map won't get you much
14:44	ToBeReplaced	,(map-indexed #(vector %1 %2) [:a :b :c]
14:44	clojurebot	#<RuntimeException java.lang.RuntimeException: EOF while reading>
14:44	gfredericks	at worst (doseq [[i x] (map list (range) xs)] ...do stuff...))
14:44	ToBeReplaced	,(map-indexed #(vector %1 %2) [:a :b :c])
14:44	clojurebot	([0 :a] [1 :b] [2 :c])
14:45	uvtc	ToBeReplaced: Ah, yes. I think that's what I want. Then I can put that in my `doseq [...]`.
14:46	uvtc	tbaldridge: thanks. Though, haven't studied reducers yet. Would use a loop, but I think what ToBeReplaced suggested is what I was looking for.
14:46	gfredericks	uvtc: ToBeReplaced or (map-indexed vector [:a :b :c]) is shorter
14:46	ToBeReplaced	gredericks is better... map list (range) xs
14:46	tbaldridge	uvtc: reducers also have the nice side-effect of being parallel (using multiple CPUs)
14:46	ToBeReplaced	,(map-indexed vector [:a :b :c])
14:46	clojurebot	([0 :a] [1 :b] [2 :c])
14:47	Raynes	cemerick: Asciidoc is excellent.
14:47	cemerick	Raynes: for sure
14:47	uvtc	gfredericks: thanks.
14:47	cemerick	or, it sucks, but it's maximally capable
14:47	gfredericks	uvtc: yerp
14:49	amalloy	gfredericks: i hear your swearjure talk was well received
14:50	borkdude	are there any talks online yet from clojurewest?
14:50	amalloy	borkdude: if the talks went online the next day, i bet attendance would drop quite a bit
14:51	uvtc	The PyCon ones went up quick (though not next-day).
14:51	technomancy	pycon has way more $$$
14:56	gfredericks	amalloy: yeah it went better than I was expecting
14:56	gfredericks	thankfully rich gave me a great opening line
14:56	tbaldridge	gfredericks: that was an awesome talk
14:57	gfredericks	tbaldridge: thx
14:57	amalloy	gfredericks: not having heard this line, i'm now imagining rich introducing swearjure as "A #$*&ing joke of a language"
14:58	jtoy_	if i have a lazyseq and turn it into a set, the order seems to get list, is there a way i can force the order?
14:58	gfredericks	amalloy: he spent half his keynote harping on self-imposed constraints
14:58	rabbit_airstrike	gfredericks: that was a hilarious talk.
14:58	gfredericks	jtoy_: sets don't have order
14:58	ToBeReplaced	if it's ordered, it's not a set... if you just want to strip duplicates, try distinct
14:59	amalloy	ToBeReplaced: not strictly true. if it's ordered it's not a hashset, but you can certainly design sets which are ordered
14:59	jtoy_	hos is this able top keep an order: (def features (set (list (feature "gender" 0 :categorical) (feature "age" 1 :categorical))))
14:59	rabbit_airstrike	sorted-set
14:59	amalloy	jtoy_: coincidental implementation details
14:59	jtoy_	i tested it with many items, it always keeps the order of which the items were added
15:00	jtoy_	i want to do the same thing as that but programitically generate the data
15:00	amalloy	hash sets are not ordered, and any sets which appear to be ordered are just luck
15:00	amalloy	(not actually luck, but you should pretend it's luck, since it's not any kind of guarantee)
15:00	ToBeReplaced	amalloy: what's an example for me to google?
15:00	jtoy_	hmm, i wouldnt be surprised if the library has a bug
15:01	amalloy	&(apply sorted-set (range 40))
15:01	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}
15:01	amalloy	$google flatland ordered set
15:01	lazybot	[flatland/ordered · GitHub] https://github.com/flatland/ordered
15:01	amalloy	are two reasonable ways to have sets which give ordering guarantees
15:01	jtoy_	ok, ill just use sorted set
15:01	amalloy	jtoy_: we'll still be here, when you discover that's not what you want
15:01	jtoy_	i think the library is implemented not very well, i need to give it data like that
15:02	jtoy_	amalloy: haha, ok
15:03	ToBeReplaced	hm... i've never had a use before... when do you end up using something like that?
15:03	gfredericks	when writing legacy code
15:04	jtoy_	amalloy: hmm, that makes it work, yeah! the required format doesnt make sense
15:06	jtoy_	why would one use the flatland stuff vs the the core ones? for clojurescript?
15:06	gfredericks	I think flatland's are insertion-order
15:06	gfredericks	whereas clojure are value-order
15:08	gfredericks	&(apply sorted-set-by #(rem % 3) (range 40))
15:08	lazybot	clojure.lang.ArityException: Wrong number of args (2) passed to: sandbox9396$eval46487$fn
15:09	jtoy_	amalloy: yeah, you are right, that is not what I want
15:09	amalloy	(inc amalloy) :(
15:10	jtoy_	i want the insertion-order sets, without adding more libraries :(
15:10	amalloy	gfredericks: sorted-set-by takes a comparator, unlike sort-by which takes a keyfn
15:11	gfredericks	yeah
15:11	gfredericks	I didn't care enough to redo it
15:11	amalloy	jtoy_: easy. remove the terrible library you're already using, if it really requires sets in a particular order (which, by the way, i very much doubt); that frees up room to add flatland/ordered
15:13	jtoy_	amalloy: this is code from the library and i did tryy with other 'orderings' but they dont work: (def features (set (list (feature "gender" 0 :categorical) (feature "age" 1 :categorical))))
15:14	thm_prover	I have a bunch of jars, like http://hpaste.org/84505 ... I want to use them all from Clojure. Is there a nice way to automate mvn-install them without having to manually mvn-install each one of them?
15:14	thm_prover	[especialy since the file names seems to be encoded in a way that encodes all information]
15:18	sundbp	using gen-class this is the generated class when decompiling it - looks broken: http://i.imgur.com/df9RkuM.png http://i.imgur.com/vQYrBUl.png http://i.imgur.com/0xyMUfe.png
15:19	sundbp	i get the correct Var, but then another one referring some some odd functino. then a double definition of thewrapper, one for each Var. hence the error message..
15:23	devn	what do people use for socket libraries?
15:23	zackzackzack	Is there a way to get more information about why a call is forcing the JVM to do reflection?
15:23	devn	I want to grab a websocket stream over 443
15:24	zackzackzack	I've set warn-on-reflection to be true and tried everything I can think of to type hint in a macro correctly, but to no avail
15:24	amalloy	devn: aleph has a good websocket server, if you're comfortable with aleph and lamina
15:24	pbostrom	2nd aleph
15:25	ppppaul	what is lamina for?
15:25	ppppaul	amalloy,
15:25	TimMc	technomancy: I installed Debian Squeeze. It's so beautiful I could cry...
15:25	TimMc	A thousand papercuts, suddenly removed.
15:26	gfredericks	beautiful compared to what?
15:26	TimMc	Latest Ubuntu.
15:26	pbostrom	ppppaul: lamina provides a "channels" abstraction around the websocket interface
15:27	amalloy	well, around a lot more than websockets
15:27	TimMc	gfredericks: beautiful == usable
15:28	ppppaul	oh. cool.
15:32	sundbp	worked it out. note to others: when using gen-class, don't explicitly list the interface methods you implement!
15:33	arrdem	TimMc: welcome to the Ubuntu expats club XP
15:39	devn	aww, wish ztellman was around
15:40	amalloy	devn: he's very responsive on the lamina/aleph/gloss MLs
15:41	devn	amalloy: cool, thanks -- looks like 0.5.0-SNAPSHOT (channel) is broken
15:41	devn	or the docs are out of date...
15:42	amalloy	i don't really believe that; i've been riding the latest snapshot for ages, and just updated and re-ran my app an hour ago
15:42	devn	hm, weird
15:42	amalloy	of course, X is broken is un-fixable
15:42	devn	amalloy: 0.5.0-SNAPSHOT?
15:42	devn	amalloy: i got a wrong number of arguments when I use (channel)
15:43	devn	amalloy: ArityException Wrong number of args (1) passed to: node$eval4985$queue clojure.lang.AFn.throwArity (AFn.java:437)
15:43	amalloy	lol. yep, looks like. i guess i just don't create empty channels myself
15:44	devn	amalloy: it fails with args as well i think
15:44	devn	like (channel 1 2 3)
15:44	amalloy	devn: bet you a dollar it's fixed within 10 minutes
15:44	devn	haha yeah, i bet you're right, i saw he committed like an hour ago
15:50	tbaldridge	technomancy: can you point me to some resource that'll help with publishing .so/.dynlib files to clojars? There's native-deps, but that looks a little old.
15:56	jtoy_	amalloy: could you help me to fix this set issue? https://www.refheap.com/paste/12838 im really close
15:57	jtoy_	its returned in a list? that is not a set?
15:57	ohpauleez	tbaldridge: To my knowledge, there isn't a good solution. I package them in a separate dir and add them to my resources path. The JVM seems to pick them up there fine
15:58	tbaldridge	ohpauleez: yeah, in this case I'm trying to package specific .so for my lib. So users can simply add the clojars reference, and depending on the platform they get the right dlls.
15:59	tbaldridge	Or is that just crazy talk? Does the whole Java package system break down for native libs?
15:59	ztellman	devn: sorry about the issue with (channel), it was in the toString method, which is only used in the repl and didn't have an associated test
15:59	ohpauleez	tbaldridge: Nah, that's definitely reasonable. clojure-leap does this with Windows and OSX binaries.
16:00	ohpauleez	but the Java code does a conditionally lookup and load on launch
16:00	ohpauleez	and since they're all in the resource path, all works out ok
16:00	tbaldridge	ah, so do you simply include all libs in a single package?
16:00	ohpauleez	but I don't know if that's a good solution
16:00	ztellman	devn: a fix is pushed to clojars/github
16:01	ohpauleez	yeah
16:01	ohpauleez	you get the kitchen sink :)
16:01	ohpauleez	but you get portability
16:01	Raynes	amalloy: ^ Just a bit over 10 minutes.
16:01	tbaldridge	yeah. Well that could work. I'll look into it. Thanks.
16:01	ohpauleez	np
16:02	ztellman	Raynes: I was away from my computer when amalloy told me
16:04	Raynes	cemerick: CompilerException java.lang.RuntimeException: Unable to resolve symbol: original-ns in this context, compiling:(NO_SOURCE_PATH:1:12355) WHAT DOES IT ALL MEAN
16:04	cemerick	Raynes: what is that from?
16:05	Raynes	cemerick: This was the result of running (start-server :port port) in a project and then connecting with `lein :connect 4008`. It connected, spewed that exception, and now I'm in reply.eval-modes.nrepl=>
16:05	Raynes	Incompatible versions of reply and nrepl, perhaps?
16:06	ohpauleez	Raynes: You've got to calibrate the qubits, before nREPL stabilizes.
16:07	Raynes	cemerick: I can require my namespaces and stuff, so it at least is kind of working.
16:07	Raynes	It's nrepl 0.2.2 and lein 2.1.1
16:07	cemerick	Raynes: whew. Sounds like a reply bug.
16:08	Raynes	trptcolin: ^
16:08	Raynes	bitchbitchwhine
16:08	Raynes	(it isn't so broken that it is unusable though, so what evs)
16:08	cemerick	Raynes: that doesn't happen with a regular `lein repl` with the same deps though?
16:09	Raynes	No.
16:09	Raynes	Well, I didn't actually try.
16:09	cemerick	right
16:09	cemerick	well, at least you said it with conviction
16:09	cemerick	the profile stuff around e.g. :connect is a bit wonky IIRC
16:09	Raynes	I'm a very insincere person.
16:09	Raynes	Yeah, lein repl works fine
16:11	cemerick	oh, you're starting the server manually, so the profile stuff isn't relevant
16:14	devn	ztellman: Thanks! (amalloy I owe you a dollar)
16:14	ztellman	devn: if you have any other questions, let me know
16:14	jtoy_	amalloy: completely fixed it,thx forthe help
16:14	trptcolin	Raynes: yeah, this is basically a symptom of reply trying to do init code where it shouldn't. completion is probably going to be broken too.
16:15	Raynes	trptcolin: Completion works.
16:15	jtoy_	does clojure have a respond to?
16:15	devn	ztellman: I'm trying to do something quick and dirty: I want to connect to a websocket server as outlined here: https://en.bitcoin.it/wiki/MtGox/API/Streaming
16:15	trptcolin	Raynes: ah, cool. i see, was grepping and mistook a local for a var
16:15	Raynes	devn: Stop stealing my bitcoins.
16:16	devn	hahaha
16:16	devn	This random dude is visiting our office today and he does automated trading stuff. They're trading at 70$ today. He was trying to debug some python script so I told him I'd rewrite what he has for a bitcoin.
16:16	ztellman	(lamina.core/receive-all @(aleph.http/websocket-client {:url "http…"}) prn)
16:16	devn	Mission accomplished.
16:16	ztellman	or something like that
16:17	trptcolin	Raynes: GH ticket me?
16:17	ztellman	I'm not 100% sure the websocket client and https play nicely together
16:17	ztellman	but that's an easy fix, if you have an issue with that
16:17	Raynes	trptcolin: Sure.
16:17	trptcolin	cool, thanks - got some jline non-BMP character debugging i'm in the middle of
16:17	Raynes	trptcolin: Have no clue what the problem is so I'm just going to slap the keyboard for the title.
16:18	trptcolin	just stick the error msg in there, that's good enough
16:25	ztellman	devn: if you're having issues with the SSL certs, add ':ignore-ssl-certs? true' to the websocket-client option map
16:27	technomancy	TimMc: pure and uncut
16:28	ztellman	or just use the non-https endpoint, which I just noticed, and seems to work just fine
16:28	technomancy	tbaldridge: no docs unfortunately; all I can point to is the implementation.
16:28	technomancy	every time someone asks my reply is "it would be great if someone would contribute some docs" since I've never actually used native deps.
16:28	Raynes	cemerick: The server example in the nrepl readme should apparently use :bind "localhost", because an nrepl server listening on anything other than localhost is 'bad practice'.
16:29	cemerick	Raynes: depends on your env though, right?
16:29	technomancy	tbaldridge: all native bits in one jar is simplest; separate jars would reduce download time at the expense of more hassle for the user. recommend the former unless it's huge.
16:30	technomancy	lein can DTRT but if you need it to run as an uberjar you've got to do all the heavy lifting yourself
16:31	Raynes	cemerick: Ask amalloy, he is the one who is all horrified of the world wide web. :p
16:32	technomancy	binding to non-localhost is like using with-local-vars; if you're one of the few people qualified to do it, you'll know without needing docs.
16:32	cemerick	Raynes: oh, so this is unrelated to the stack you hit? Just so that people aren't accepting REPL connections on all network adapters, etc?
16:32	amalloy	i didn't recommend any change to cemerick's readme, Raynes. that was you. i can understand doing something insecure in a getting-started sort of page
16:32	amalloy	though i would like it to change, i suppose
16:32	cemerick	it's a fair point
16:34	jtoy_	how does one typically deal with this kind of issue: WARNING: resolve already refers to: #'clojure.core/resolve in namespace: interests.enrichments, being replaced by: #'clojurewerkz.urly.core/resolve
16:34	amalloy	personally i'd rather leave the readme alone, and have the default bind address be localhost rather than 0.0.0.0
16:35	technomancy	someone yelled at be for making swank bind to 0.0.0.0 by default
16:35	Raynes	Yeah, I brought it up because amalloy just yelled at me for binding to 0.0.0.0.
16:35	gfredericks	jtoy_: use (:refer-clojure :exclude [resolve]) in your ns declaration
16:35	squidz	does anybody know why compojure doesnt automatically load changes made to html/code
16:36	gfredericks	jtoy_: assuming of course you're bringing in the other one on purpose and want to use it and not the clojure one
16:36	technomancy	you could make the argument it should be up to the dev tool to specify localhost, but since the consequences of getting that wrong are fairly disastrous it's really not worth being pedantic
16:36	Raynes	technomancy: nrepl isn't just for dev tools though.
16:36	Raynes	Users embed it in their applications.
16:36	Raynes	And they have to set the port and address in those cases.
16:37	amalloy	Raynes: you would only embed nrepl in an app as a dev tool
16:37	TimMc	localhost-by-default would be very reasonable
16:37	technomancy	Raynes: yep, and forcing them to pay attention or get owned is sad times
16:37	Raynes	amalloy: Whatever.
16:37	Raynes	I think my point was clear.
16:37	jtoy_	gfredericks: thx, was just reading about that
16:38	squidz	i thought clojure was supposed to automatically load code while running
16:39	technomancy	squidz: that would make for some pretty rotten performance
16:39	squidz	but while in some kind of development mode?
16:39	technomancy	lein-ring does it in development mode, yeah
16:40	squidz	because if I have to start my ring server every time I want to see changes, it's not much better than java
16:40	technomancy	but it's much simpler just to connect a repl to it and reload that way
16:40	squidz	okay so maybe im not in development mode
16:41	squidz	so if I start my server in my terminal with lein ring server , how do I send my evaluations from emacs to it
16:41	squidz	I guess when I jack-in it starts a separate repl?
16:41	borkdude	squidz don't start it via lein ring, but via lein repl
16:42	technomancy	squidz: right; so you can just launch a web server via jetty from your repl
16:42	technomancy	https://github.com/technomancy/syme/blob/master/src/syme/web.clj#L161
16:42	squidz	technomancy: from the repl i jacked-in to with emacs nrepl?
16:42	technomancy	aye
16:42	squidz	okay thanks this was bugging me
16:42	technomancy	you can launch a repl from lein ring too, but it's a bit more complicated
16:43	technomancy	indeed; running without a repl is like running blind
16:43	squidz	yeah I thought there has to be something im missing
16:47	drorbemet	Which is the current way of writing the out and printed output of a function evaluated in the repl to a file?
16:47	drorbemet	I read clojure.contrib.io doc and clojure.java.io doc ... which functions to use is still unclear to me ..
16:48	Raynes	cemerick: So yeah, in summary, I think most people appear to be in favor of changing the default bind address to localhost, with the less nuclear option being to edit the README example to bind to localhost or at least mention that it isn't binding to localhost by default.
16:48	nDuff	drorbemet: rebind out to your file
16:48	Raynes	I'd happily provide either of these modifications, but I'm not sure I can bear the Jira patch nonsense for something this simple.
16:49	drorbemet	nDuff: that's to short for me at the moment, I don't even have a file handle, buffer or writer ...
16:50	drorbemet	nDuff: I deleted a file though with delete-file
16:54	edw	OK, I'm playing with the new LightTable and…how exactly do you get a Clojure REPL in a project? Am I crazy in saying that nrepl in Emacs is like a million times easier to grok?
16:55	technomancy	there is no easy, only familiar
16:55	tbaldridge	edw: I've had very bad experiences with LT. Perhaps that's simply because it's still "pre-alpha" though
16:56	aaelony	difficult equates to unfamiliar
16:57	nDuff	edw: There is no conventional REPL in LT, just the instarepl. Which is fine / all one needs, IME.
16:58	edw	So, if I have a project, how do I get LT to start a JVM in the project's directory, with the appropriate deps on the classpath?
16:58	arrdem	nDuff: define instarepl?
16:59	tbaldridge	like a repl but more instant
16:59	tbaldridge	:-P
16:59	nDuff	arrdem: Watch any Light Table demo video. Type code in editor -> evaluation result shown in editor alongside code, updated as you edit.
17:00	nDuff	arrdem: ...includes showing values of locals in other functions called during execution, etc etc. Generally pretty shiny.
17:00	edw	technomancy: Perhaps I have spent too many years (~25) in Emacs.
17:00	technomancy	edw: no such thing
17:01	edw	Yes, but I'm trying not to put my bigotry on display.
17:01	technomancy	"too many years to understand how normal people use computers" is possible though
17:01	borkdude	edw I hope I have mastered emacs in 25 years, but for now I will use it's basic features
17:01	borkdude	its
17:02	edw	But seriously, I am open-minded about LT: I wouldn't mind Squeak + CLJS + Emacs. But I'm utterly lost.
17:03	borkdude	didn't rich say something about emacs on clojurewest? saw smth on twitter
17:04	Raynes	Oh man, if Rich mentioned Emacs it'll validate the past 37 years of work on the editor!
17:04	edw	OK, I give up. Back to Emacs + nrepl + autocomplete.
17:05	ieure	borkdude, Rich said he uses Aquamacs, which, ewww.
17:05	technomancy	Raynes: he said if you spend more than 15m a day writing elisp you'll go blind
17:05	squidz	technomancy: wow thanks for the tip, I started my server from the repl in emacs and now everything works like magic :D
17:05	nDuff	edw: personally, btw, I find that LT is great for quick experimentation -- say, inspecting the behavior of code people are asking about here in IRC -- but for real work, there's no replacement for emacs yet.
17:05	edw	technomancy: More than thirty minutes a day? Hairy palms. Forty-five? Neckbeard.
17:06	Raynes	Rich has never been particularly cutting edge on the environment front.
17:06	edw	So Rich has a 115 baud acoustic coupled modem connected to a daisywheel TTY?
17:06	arrdem	Raynes: we have The One True Editor.. what's to be cutting edge about?
17:07	yogthos	nDuff: I've been getting by with ccw myself :)
17:07	nDuff	Ewww.
17:07	yogthos	nDuff: it does all I need, got repl, paredit works acceptably
17:08	yogthos	nDuff: and I do need to work with java once in a while ;(
17:08	yogthos	nDuff: but purity aside, it's definitely usable for real work is my point
17:09	nDuff	yogthos: I fire up IDLE for working with Java, but find CCW's paredit incompatible enough with my finger memory (need slurp and barf, damnit) that it's totally worth running multiple environments.
17:09	yogthos	nDuff: ah yeah that's fair :)
17:09	squidz	technomancy: do you know by chance if there is a way to reload when html is changed?
17:09	yogthos	nDuff: at this point I'm holding out for LT to actually mature
17:10	technomancy	squidz: depends on the templating lib
17:10	squidz	it is just plain html with enlive
17:10	technomancy	I think you just recompile the deftemplate, but I haven't used enlive in a while
17:10	squidz	so if I eval the enlive it loads but I wanted to know if there was a way to just edit the html then save
17:10	squidz	yeah thats right
17:11	squidz	but I guess im getting greedy and wanted to just see changes on html save
17:11	technomancy	squidz: you could do it in an emacs after-save-hook
17:11	technomancy	but emacs has no way of knowing which .html file corresponds with which .clj file
17:11	squidz	okay ill keep it simple and just evalueate the deftemplate
17:12	squidz	its already a lot better than java
17:13	squidz	i used hiccup first because it was easiest to grasp, then I switched to mustache, and now I landed at enlive
17:13	borkdude	so, what is the -canonical- templating library nowadays :P
17:13	technomancy	there's no one right answer
17:14	squidz	I decided to switch from mustache to enlive because mustache just wasnt flexible enough, and clojurescript has a library enfocus which is almost the same as enlive with goodies
17:16	borkdude	technomancy I thought there was some self declared canonical new web thing recently ;P
17:17	squidz	yeah I still have to look into pedestal
17:17	squidz	ive glanced over it, but haven't grokked it yet
17:17	technomancy	borkdude: last I checked Canonical is still using Python. =P
17:18	borkdude	lol
17:18	squidz	lol
17:19	technomancy	2-lol combo; not bad but we can do better
17:19	borkdude	plol
17:21	tyler	whats like conj but puts on the end of list?
17:21	borkdude	tyler conj puts at the end if it gets a vector
17:22	borkdude	tyler conj is a polymorfic function
17:22	tyler	ah not sure what class its passing in, i assumed it was list but might be a vector
17:22	tyler	didn't write this code
17:22	dansalmo	would anyone here be interested in trying out a web based interpreter for lambda calculus that uses indices instead of traditional lambda notation?
17:23	dansalmo	It is not practical so much as educational.
17:23	tyler	are you saying education isn't practical? ;)
17:23	yogthos	squidz: I really like clabango myself :) https://github.com/danlarkin/clabango
17:23	borkdude	$findfn '(1 2 3) 4 '(1 2 3 4)
17:23	dansalmo	http://pure-fn.appspot.com/about
17:23	lazybot	[]
17:23	dansalmo	I guess you got me. :)
17:24	borkdude	$findfn 4 '(1 2 3) '(1 2 3 4)
17:24	lazybot	[]
17:24	borkdude	tyler I think you'd have to go with concat then
17:24	borkdude	,(concat '(1 2 3) '(4))
17:24	tyler	borkdude: thnx
17:24	SegFaultAX	technomancy: If you haven't seen it yet, I think you might enjoy the first part of this post: http://www.flyingmachinestudios.com/programming/how-clojure-babies-are-made-lein-run/
17:24	clojurebot	(1 2 3 4)
17:25	technomancy	SegFaultAX: heh; I copyedited that. copious lulz.
17:25	squidz	yogthos: yeah i took a look at clabango, but it seemed pretty similar to clostache, and after my experience with clostace, I wanted something more different
17:25	SegFaultAX	technomancy: :D
17:26	SegFaultAX	For the record, Fullmetal Alchemist is amazing.
17:26	yogthos	squidz: I found it more flexible than clostache myself, it's based off django templates which are pretty nice
17:26	uvtc	tomato, tomahto, clamato, clabango
17:26	yogthos	:)
17:26	amalloy	dansalmo: you might enjoy http://www.ioccc.org/2012/tromp/hint.html if you're into indexed lambda calculus
17:26	yogthos	that's the nice thing about the stack being pluggable
17:27	squidz	I liked the idea of having normal html and being able to transform that xml to your hearts content. Seemed like an clean solution
17:30	dansalmo	amalloy, thanks. It looks interesting. Thanks for your help before, the web interpreter is was I was working on.
17:31	Raynes	SegFaultAX: This is true.
17:31	drorbemet	nDuff: got it, asked a bit to early ..
17:32	jtoy_	is this considered the correct way to write a collection to a file? (with-open [wrtr (writer "fns.txt")]
17:32	jtoy_	(map #(.write wrtr % ) fns))
17:33	yogthos	nDuff: well that can happen, but in practice I haven't noticed it to be a problem
17:33	borkdude	jtoy_ it's pointless to use map for only side effects
17:34	jtoy_	borkdude: it doesnt seem to work , what is the right way to do this then?
17:34	borkdude	jtoy_ also, what is fns? you probably need doseq
17:34	borkdude	jtoy_ give us the complete code via a refheap paste
17:34	jtoy_	borkdude: fns is just my collection
17:35	jtoy_	borkdude: https://www.refheap.com/paste/12839
17:37	jtoy_	I see : (with-open [wrtr (writer "fns.txt")] (doall (map #(.write wrtr (str % "\n") ) [1 2 3 4 5 6] )) )
17:37	jtoy_	im still considered a clojure newb, but so far everysingle time i work with read files, it seems like always doing doall
17:38	jtoy_	or writing files, any file stuff
17:39	yogthos	you probably should be using doseq instead
17:39	yogthos	since you'd doing side effects
17:39	yogthos	(doseq [item [1 2 3 4 5 6]] (.write wrt item))
17:40	borkdude	jtoy_ https://www.refheap.com/paste/12840
17:40	borkdude	yogthos somehow I had to str i(tem) first
17:40	yogthos	yeah that makes sense
17:41	jtoy_	so doseq is like for and doall is like map? is tht the correct way to think about this?
17:41	yogthos	the write method expects a string
17:41	yogthos	joys of dealing with java api
17:41	borkdude	jtoy_ map creates a lazy sequence of which the results are only produced on demand
17:42	borkdude	jtoy_ doall forces the production of all those elements
17:42	borkdude	jtoy_ but since you're only interested in the side effects and not in some resulting collecting, you should use doseq in this case
17:42	jtoy_	borkdude: right, but is doseq to for like doall is to map? or is that analogy totally wrong?
17:43	borkdude	,(doc doall)
17:43	yogthos	not quite
17:43	clojurebot	"([coll] [n coll]); When lazy sequences are produced via functions that have side effects, any effects other than those needed to produce the first element in the seq do not occur until the seq is consumed. doall can be used to force any effects. Walks through the successive nexts of the seq, retains the head and returns it, thus causing the entire seq to reside in memory at one time."
17:43	pbostrom	personally, I wish there was a side-effect equivalent to 'map'
17:43	yogthos	conceptually doseq is specifically for side effects
17:43	borkdude	,(doc doseq)
17:43	clojurebot	"([seq-exprs & body]); Repeatedly executes body (presumably for side-effects) with bindings and filtering as provided by \"for\". Does not retain the head of the sequence. Returns nil."
17:43	technomancy	first person to say mapv gets a kick in the shins
17:43	yogthos	lol
17:43	technomancy	I'm watching you
17:44	jtoy_	i see what doall does, but i dont really understand doseq even after reading that
17:44	borkdude	I don't get why (.write wrtr 1) doesn't return some error, it just silently fails?
17:44	jtoy_	borkdude: without the str, it worked for me, it just outputted garbage
17:45	borkdude	,(= nil (doseq [i (range 5)] (print i)))
17:45	clojurebot	01234true
17:45	SegFaultAX	jtoy_: It's somewhat annoying to remember all the do* forms, but you'll get used to it.
17:45	yogthos	jtoy_: basically doseq will always return nil, and it doesn't retain the sequence in memory
17:45	borkdude	,(= nil (doall (map print (range 5))))
17:45	clojurebot	01234false
17:45	amalloy	.write probably has an overload that accepts ints
17:45	yogthos	jtoy_: but you can do something with each element as you run through them
17:46	yogthos	jtoy_: presumably side effects like printing it out or writing to a file, etc
17:46	jtoy_	yogthos: ah, i see
17:47	jtoy_	yogthos: so in my case i often read in files, and i want to just collect a part of each line, i use doall and that works now, is that considered the best way to do it?
17:47	borkdude	amalloy clojure uses longs right?
17:47	borkdude	amalloy ah, it can deal with ints, but those are interpreted as chars
17:47	borkdude	DOH
17:48	SegFaultAX	jtoy_: doall is for eagerly evaluating lazy things. That's probably not the right tool for what you're doing.
17:48	yogthos	jtoy_: if you're collecting then doall is fine
17:48	tieTYT	newbie here. So ~@ takes a & parameter and evaluates it as-is?
17:48	yogthos	jtoy_: the reason you need doall is because when you leave the (with-open ...) expression the file gets closed
17:49	yogthos	jtoy_: so you want to force evaluation before that happens
17:49	yogthos	tieTYT: in a macro ~@ will splice a sequence into the parent
17:49	tieTYT	what does that mean?
17:49	tieTYT	what's splice?
17:49	jtoy_	SegFaultAX: I use it bc I use with-open
17:49	technomancy	try it and see
17:50	jtoy_	does that mean maybe it makes sense to have a with-open-eager or something like that
17:50	yogthos	tieTYT: eg: (defmacro foo [items] `(println ~@items))
17:50	technomancy	,`(1 2 3 ~@[4 5 [6 7]])
17:50	clojurebot	(1 2 3 4 5 ...)
17:50	technomancy	dang it clojurebot
17:50	technomancy	,`(1 ~@[4 5 [6 7]])
17:50	clojurebot	(1 4 5 [6 7])
17:50	yogthos	what technomancy said :)
17:51	yogthos	whatever items were in the collection you used ~@ get appended into the outer expression
17:52	tieTYT	hm but you used ` there. I thought that forces evaluation
17:52	technomancy	~@ only works inside `
17:52	clojurebot	@ , {a b c]
17:52	tieTYT	how can you evaluate (1 2 3)
17:52	tieTYT	oh
17:52	hyPiRion	,`[[~@[1 2 3]] [~@"wat the"] [~@{:a 1 :b 2}] [~@#{:also :sets}]]
17:52	clojurebot	[[1 2 3] [\w \a \t \space \t ...] [[:a 1] [:b 2]] [:also :sets]]
17:52	technomancy	also, quoting delays evaluation
17:52	yogthos	which can be very handy :)
17:52	tieTYT	i didn't see any quoting there
17:52	tieTYT	only unquoting
17:53	tieTYT	errr
17:53	hyPiRion	backquote -> `
17:53	tieTYT	ohh
17:53	amalloy	clojurebot: tias is <reply> Try it and see! You'll get results faster than asking someone in #clojure to evaluate it for you, and you'll get that warm, fuzzy feeling of self-reliance.
17:53	clojurebot	In Ordnung
17:53	tieTYT	what's that vs '?
17:53	tieTYT	i got mixed up. I swapped ` with ~ in my head
17:53	hyPiRion	,'[~@[:oh :no]]
17:53	clojurebot	[(clojure.core/unquote-splicing [:oh :no])]
17:54	SegFaultAX	I don't usually link to clojure.org, but: http://clojure.org/reader
17:54	jtoy_	clojurebot: help
17:54	clojurebot	Nobody can help with "X doesn't work". Please provide context: what you did, what you hoped would happen, and what happened instead. A stack trace is especially helpful, if applicable.
17:54	SegFaultAX	tieTYT: ^
17:54	tieTYT	ok
17:55	yogthos	personally I'd recommend not focusing on macros too much when you're starting out :)
17:56	tieTYT	yogthos: i'm running into library code that uses them
17:57	yogthos	ah
17:57	tieTYT	are you just recommending I don't use them for now?
17:58	yogthos	yeah, it's rare that you actually need to use a macro legitimately
17:59	tieTYT	cool, I had that intuition :)
17:59	yogthos	:)
17:59	hyPiRion	Whenever I find a new language construct (such as macros) I tend to overuse them because I feel so powerful using it.
17:59	hyPiRion	Then I look at it 1 month later and cry because it's so horrible.
17:59	hyPiRion	(It's a good way of learning that specific construct though).
17:59	borkdude	hyPiRion haha, so true
18:01	SegFaultAX	But when you do need macros, they're amazing.
18:01	ivaraasen	hyPiRion: same thing in journalism. learned a new word or a new idiom, then used it in every single article
18:01	technomancy	drunk on the power of defmacro
18:01	technomancy	"you will cower before me, mortal!"
18:02	SegFaultAX	Obligatory: http://xkcd.com/224/
18:04	borkdude	SegFaultAX I suspect he did it in perl8.org
18:04	hyPiRion	ivaraasen: oh yeah, it's a bit painful there too. Or in reports for that matter. When I found a nice way to cite things, they were suddenly everywhere.
18:06	ivaraasen	hyPiRion: in Trondheim this Easter BTW?
18:07	jtoy_	is it perfectly fine to use with-open to open a file to write and read in 1 binding?
18:07	jtoy_	i mean multiple open files
18:07	hyPiRion	ivaraasen: yeah, I'll be here for the whole Easter
18:08	borkdude	jtoy_ you need a binding for either a writer or reader
18:08	borkdude	jtoy_ and with-open supports multiple bindings afaik
18:09	ivaraasen	hyPiRion: same thing. feel free to call me up if you want to drink strong coffee and talk about swearjure
18:09	hyPiRion	ivaraasen: oh, sweet
18:10	hyPiRion	I should bring my spare Clojure stickers and buttons
18:10	ivaraasen	hyPiRion: damn, I need some Clojure stickers ... for my employees, of course.
18:12	hyPiRion	exactly
18:13	borkdude	any favorite plot libraries for quick function plotting and using ouputted pics in a .tex or .org document? gnuplot?
18:13	tieTYT	ooh #(...) looks useful
18:14	borkdude	tieTYT until you try to nest them :P
18:14	tieTYT	oh that could get ugly
18:14	tieTYT	so should I avoid them or just not nest them?
18:14	hyPiRion	not really, it's not allowed.
18:15	borkdude	tieTYT I tend to avoid them and use only (fn [x] ….), but this is just my personal preference
18:15	hyPiRion	use them where it makes stuff more readable
18:15	hyPiRion	e.g. #(Double/parseDouble %)
18:16	ivaraasen	hyPiRion: I tend to use GNUPlot. somebody needs to make an open source version of the service mentioned in the a grammar of statistical graphics talk
18:16	tieTYT	so that's an anonymous function that takes one argument and passes it to Double/parseDouble
18:17	hyPiRion	tieTYT: yep
18:17	tieTYT	ok
18:17	borkdude	,(#(-> 1))
18:17	clojurebot	1
18:17	hyPiRion	ivaraasen: I think you meant to send that one to borkdude? Regardless, I agree. GNUPlot is nice, but I would like an api for it in Clojure
18:17	hyPiRion	My shell script hacks can only do that much
18:18	ivaraasen	hyPiRion: err, yeah
18:22	hyPiRion	borkdude: gnuplot with epslatex is my favourite for plotting, but it's not quick if you're not familiar with it. I know Incanter has some way of spitting out PDFs though.
18:22	borkdude	hyPiRion ivaraasen ok tnx, I'll stick with gnuplot
18:23	yacin	are there R bindings or interop for clojure?
18:23	yacin	i can't use anything for plotting except R/ggplot2 these days
18:23	tieTYT	ok I think i get ~@ now
18:24	tieTYT	it inlines the structure into the outter form
18:24	ivaraasen	borkdude: another option is just using org-plot on tables created from running print-table on your Clojure data structures
18:24	borkdude	ivaraasen I really need a function plot
18:24	borkdude	ivaraasen but it's good to know that exists
18:24	borkdude	it
18:25	ivaraasen	borkdude: yeah, I see where you're coming from. I tend to use Incanter for stuff like this mostly
18:25	ivaraasen	or Mathematica :P
18:25	llasram	yacin: There's http://www.rforge.net/rJava/ ; but I've only dealt w/ it via researcher code :-)
18:25	ravster	Where can I find an explanation of :available-media-types in liberator?
18:25	borkdude	I used Maple but it's long since I had a license
18:26	llasram	yacin: Lately I've been using org-babel to glue together doing data-processing in Clojure then stats and plotting in R. It works not half-badly
18:26	yacin	ah yeah, rjava. i've used that a bit
18:26	ravster	a "hello world" isn't even working without that, and I can't seem to find docs for as essential a function as that.
18:26	yacin	haha howdy llasram
18:26	llasram	yacin: Yep :-)
18:26	yacin	didn't check the nick right away :P
18:27	llasram	heh
18:27	ivaraasen	yacin: yeah, org-bable and print-table with some R in-between is pretty nice
18:28	ivaraasen	err, wrong nick again.
18:28	yacin	nice, i'll have to give it a whirl
18:29	ivaraasen	I'm considering writing some output stuff so I can write my articles in org-mode and export them as XML to InDesign, ready formatted. reproducible journalism is an interesting concept.
18:30	hiredman	you know org-mode can export to pdf and html?
18:31	yacin	blah, i really need to get on the org-mode bus
18:31	hiredman	the pdf export is via latex I think
18:31	ivaraasen	hiredman: yeah, but we have some in-house quirks and use our own XML dialect that gets put into InDesign
18:31	hiredman	sounds gross :)
18:31	ivaraasen	it is
18:32	ivaraasen	most of the code is from the early 2000's
18:32	hiredman	I haven't used babel yet, but it is intriguing
18:33	ivaraasen	crashes all the time as well. usually half an hour before the paper has to reach the printing company
18:34	hyPiRion	haha
18:35	ivaraasen	which is like a $15k loss in revenue. pretty scary stuff
18:47	zakwilson	clj-time is saying this string is invalid for this formatter, but it looks ok to me according to the jodatime docs: (parse (formatter "d M y") "22 March 2013"). It works with a numeric month though.
18:49	hiredman	zakwilson: if I recall the docs say something like single letter for numbers and multiple letters for longer forms for parsing
18:49	hiredman	oh
18:49	borkdude	goodnight
18:49	hiredman	I am thinking of simpledateformatter, dunno about clj-time
18:52	zakwilson	Actually, hiredman, that does seem to be it; MMM parses "March". That is one possible interpretation of the docs, which say "Month: 3 or over, use text, otherwise use number.", but since they give "July; Jul; 07" as examples, I assumed it meant that 3 or more characters in the string being parsed would be treated as a textual representation of the month, not three or more Ms in the pattern.
19:02	tyler	hugod: are you hugo duncan?
19:06	OneFourSeven	Can somebody explain to me how to use roles in cemerick friend?
19:07	OneFourSeven	I'm just not sure about the syntax behind the #{::user} and how I would store that stuff in a database
19:08	tyler	anyone know if you can simulate tty in clj-ssh?
19:08	hiredman	,::user
19:08	clojurebot	:sandbox/user
19:08	hiredman	it is a set of namespace qualified keywords
19:09	hiredman	the reason he used namespace qualified keywords for roles is because you can put them in hierarchies
19:10	hiredman	hierarchies in clojure are a thing you should google about if you want to know about them
19:12	OneFourSeven	hiredman, what should I store in the db? A string with the role user?
19:14	Glenjamin	hi guys, is anyone aware of a lib for mocks/fakes/spies/stubs ?
19:14	technomancy	Glenjamin: clojure.core/with-redefs
19:14	Glenjamin	ideally i'd want to declare a stub, and then assert on how it was called afterwards
19:15	technomancy	,(let [vals (atom [])] (with-redefs [mocked-fn (partial swap! vals)] (mocked-fn :expected) (= [:expected] @vals)))
19:15	clojurebot	#<CompilerException java.lang.RuntimeException: Unable to resolve var: mocked-fn in this context, compiling:(NO_SOURCE_PATH:0:0)>
19:15	technomancy	oh yeah; can't do that in a sandbox
19:15	technomancy	whatever
19:15	technomancy	don't use a library just to redef a function in tests
19:16	Glenjamin	it's not the redef that i'm after, as you said - that's easy
19:16	Glenjamin	it's just wrapping up that atom + constantly combo
19:17	technomancy	well, you can add one line to project.clj and one line to your ns form to refer it, or you can just add two lines for the above snippet =)
19:18	Glenjamin	per-test :)
19:18	Glenjamin	but, yeah i take your point
19:20	Glenjamin	cluttering up test cases with that every time would distract from the actual meat of the test, so i'd end up pulling it out into a macro, which at that point makes sense to spin off as a separate project
19:20	Glenjamin	since the form is general
19:21	technomancy	most nontrivial projects have a "test helper" kind of namespace
19:21	Glenjamin	exactly, thats the sort of thing we should be sharing
19:22	Glenjamin	anyway, i'll get some stuff working with that for now - cheers
19:25	chessguy	so i hear people talk about the reader in clojure. isn't it just the same as a parser in another language?
19:26	amalloy	chessguy: sorta. the main difference is it returns a data structure that's not complete rubbish
19:28	nDuff	chessguy: ...one of the big things about LISPs is isomorphism -- your parse tree and your code match.
19:28	chessguy	well, parsers return ASTs, right?
19:28	nDuff	chessguy: That makes macros far more powerful and usable than they are otherwise.
19:29	headshot	"homoiconicity" ?
19:29	drewc	reader is different from a parser ... first lex. then parse, then read and circle.
19:29	powr-toc	Can anyone tell me why this core.logic snippet is so slow?
19:29	powr-toc	http://pastebin.com/DrGkZyCW
19:30	powr-toc	distincto seems to perform quite badly
19:30	drewc	nDuff: like we used to have for paste.lisp.org before the spammers took care of it?
19:31	powr-toc	nDuff: Sorry, didn't see the ads... I have adblock
19:31	dnolen	powr-toc: if you dealing with integers you should probably use fd/distinct
19:31	chessguy	why not just use gists?
19:31	dnolen	powr-toc: I'm not sure how to make distincto efficient
19:31	hyPiRion	refheap :)
19:31	powr-toc	dnolen: I'm not dealing with ints... just in that example I wanted to generate some
19:31	powr-toc	values
19:33	powr-toc	dnolen: would I be better using integers and mapping them to my values somehow?
19:33	dnolen	powr-toc: there's probably a way to to do an efficient distincto constraint when you don't have complex terms ... would take a patch of course.
19:33	dnolen	powr-toc: maybe
19:34	dnolen	powr-toc: also CLP(Set) might do what you want when we get there.
19:34	dnolen	powr-toc: if you look at the implementation of disincto you can see that it's pretty naive.
19:35	powr-toc	dnolen: Naive, you mean like I would write ;-)
19:36	powr-toc	dnolen: How else could it be written?
19:37	dnolen	powr-toc: it could probably be done almost exactly the same as fd/distinct
19:39	powr-toc	I don't really grok much of the fd stuff... it's mostly blackmagic to me
19:40	powr-toc	dnolen: where can I find out more about CLP(Set)?
19:40	dnolen	powr-toc: hehe, it not as complicated as it seems though there's a lot of details that you need to be aware of to write a constraint - hopefully that will improve with time.
19:41	dnolen	powr-toc: https://github.com/clojure/core.logic/wiki/CLP%28Set%29, it's like CLP(FD) but trickier
19:42	powr-toc	dnolen: lol, and my head was already hurting!
19:43	dnolen	powr-toc: one day these silly details will be hidden from the user and you can just use the damn thing. kind of how people just use persistent data structures in Clojure - obviously we're not there yet.
19:44	powr-toc	dnolen: I didn't seem to struggle quite as much with the small amount of prolog I've done...
19:45	dnolen	powr-toc: well I'm sure you used a fairly mature Prolog
19:45	dnolen	lots of optimizations in place
19:45	powr-toc	dnolen: yeah, SWI
19:45	dnolen	powr-toc: yeah I think they probably have a decent distinct constraint that does the right thing.
19:46	dnolen	powr-toc: writing a distinct like distincto in Prolog is sure to be bad too.
19:47	arohner	I have an app that works fine, until I AOT it. Now I get: https://gist.github.com/arohner/d1a2d9c0928b0ede769c
19:47	powr-toc	dnolen: from what I remember it was pretty decent
19:48	powr-toc	dnolen: but it seems core.logic is getting better and better all the time
19:48	hiredman	arohner: have you checked to see if that class exists in classes/ ?
19:49	arohner	hiredman: I don't think it does. That's part of log4j, so shouldn't I expect it not to be there?
19:49	arohner	sorry, nvm
19:49	dnolen	powr-toc: slowly but surely. I would look into this, but I'm trying to improve the CLP(FD) stuff at the moment, adding minimize, maximize, etc
19:50	arohner	hiredman: I don't see it there, what should I look at?
19:51	arohner	https://groups.google.com/forum/?fromgroups=#!topic/datomic/6xWGFB-Dx68
19:51	hiredman	arohner: do you have lein setup to delete non-project class files?
19:51	arohner	probably not. I'll look into that
19:51	hiredman	arohner: ugh, get a real build process
19:52	hiredman	and deploy a built artifact
19:52	arohner	huh? I triggered this trying to make 'lein uberjar' work
19:52	hiredman	don't `lein run` in a git checkout
19:52	hiredman	oh, the message there talks about "on various deployments"
19:53	arohner	that's not me posting :-)
19:53	powr-toc	dnolen: One thing I've not seen documented, is what the following does: (distribute q ::l/ff)
19:53	powr-toc	(as seen in sudokufd)
19:54	dnolen	powr-toc: actually I'm a looking a bit closer at your paste
19:54	dnolen	powr-toc: ::l/ff is a custom strategy for how to enumerate the domains
19:55	dnolen	powr-toc: there'll be more, very useful for optimization w/o really changing program
19:55	hiredman	arohner: does any clojure code not in your project end up as class files?
19:55	dnolen	powr-toc: oh, the issue here is not distincto :P
19:56	arohner	hiredman: yes, it looks like all 3rd party libs are in target/
19:56	dnolen	powr-toc: it's how you're calling value* w/ everyg, you're generating a ton of possiblities whene you do that.
19:56	powr-toc	dnolen: I was wondering if that might be it...
19:56	powr-toc	dnolen: what's the best way to distribute that fact distinctly over the vector of lvars?
19:58	dnolen	powr-toc: if I'm not mistaken that's probably 15! or something?
19:58	powr-toc	dnolen: yeah... the performance sure seemed to suck like a factorial :-)
19:59	dnolen	powr-toc: yeah, you're probably better off encoding the problem in CLP(FD) if you can and writing helpers to move between the representations
19:59	hiredman	arohner: everything runs fine with `lein run`?
19:59	arohner	hiredman: yes
19:59	dnolen	powr-toc: it's annoying, and I've had to do this - it's likely we'll fast track at least a version of CLP(Set) that works when everything is ground as this is very common.
19:59	arohner	hiredman: lein run works as long as target/ is empty
20:00	hiredman	arohner: how weird is your code base? are you doing conditional code loading or anything like that?
20:00	arohner	after `lein uberjar`, everything breaks with that exception until rm -rf target/
20:00	arohner	I'll try undoing the transitive AOT manually, and see what happens
20:01	powr-toc	dnolen: Am I right in thinking that I saw an example recently where someone (you?) bodged in some CLP(Set) stuff, did what you describe?
20:01	dnolen	powr-toc: http://swannodette.github.com/2013/03/09/logic-programming-is-underrated/
20:01	hiredman	arohner: I would try the reverse, force compilation of clojure.tools.logging upfront
20:02	powr-toc	dnolen: yes, that'd be the one :-)
20:02	arohner	hiredman: how would I do that? AOT tools.logging from it's source tree, and copy them in?
20:02	dnolen	powr-toc: yep, an fd encoding of your example takes <30ms, vs. 5000ms on my air
20:02	hiredman	like stick (compile 'clojure.tools.logging) before the (ns ...) of your main namespace
20:03	arohner	interesting :-)
20:03	powr-toc	dnolen: can you pastie?
20:05	dnolen	powr-toc: https://gist.github.com/swannodette/5225702
20:07	dnolen	powr-toc: updated the gist just now
20:09	powr-toc	dnolen: wow... the jit eats it for breakfast too! Some of the benchmarks are coming in at under 5ms on my crappy laptop
20:09	dnolen	powr-toc: yep
20:10	powr-toc	impressive!
20:10	dnolen	powr-toc: like I said, slowly getting there
20:10	arohner	hiredman: progress!
20:10	arohner	I had to compile clojure.tools.logging, and tools.logging.impl
20:10	arohner	now I'm getting similar problems in other libs
20:11	arohner	thanks
20:12	hiredman	arohner: so it may just be an ordering thing, if you move ordering around so libs compile in a specific order it might work out
20:13	hiredman	move the order of require's and use's around
20:14	arohner	hiredman: I'm pretty sure I never use tools.logging without requiring it in that ns. How could my code cause the prob?
20:17	hiredman	arohner: I don't know, clojure.tools.logging actually does some pretty gross stuff under the hood to pick the logging library to use
20:18	hiredman	could be something like something depends on logging, so logging is loaded, logging looks that the available classes and compiles itself to use backend X, but then some other code loads latter that invalidates that so logging dies or something
20:19	Glenjamin	what is it with java logging? pretty much every other language/framework i've used just has the application bootstrap set up the logger
20:19	technomancy	Glenjamin: it is. the worst.
20:20	arohner	and I'd happily call (logger/init :log4j) or whatever
20:25	arohner	success!
20:25	arohner	startup time down from 1:30 to 0:21s
20:26	Glenjamin	21s is good?
20:26	arohner	4.5x improvement is good
20:27	Glenjamin	mm
20:30	arohner	Glenjamin: oh, that also includes DB & web server init
20:30	hiredman	arohner: did you just (compile ...) namespaces one by one until it worked?
20:31	arohner	hiredman: yes, I had to compile clojure.tools.logging, clojure.tools.loggin.impl, and explicitly require ring.adapter.jetty
20:31	clojurebot	clojure is like life: you make trade-offs
20:31	arohner	noir tries to load r.a.jetty at runtime
20:31	hiredman	interesting
20:32	hiredman	and loading jetty would bring in jetty's whatever logging
20:32	arohner	hrm, now immediately after lein uberjar, I get 'Error: Invalid or corrupt jarfile target/circle-0.1.0-SNAPSHOT-standalone.jar"
20:32	hiredman	now that you have an explicit load of jetty, I would try removing the explicit compiles
20:32	hiredman	arohner: always `lein clean`
20:34	arohner	hiredman: same result
20:35	hiredman	does the jar file unzip?
20:37	Glenjamin	can you move the noir require out of the ns declaration
20:37	arohner	hiredman: kind of: https://gist.github.com/arohner/02d57509bc1fe05c20ef
20:37	arohner	other than that, it completes fine
20:38	hiredman	yeah, unzip is most likely more lenient about files then java's jar file reading
20:38	hiredman	if lein is generating invalid jars that sounds like a lein problem :)
20:38	hiredman	what happens if you re-jar the contents?
20:39	arohner	using the jar tool manually?
20:39	hiredman	sure
20:42	arohner	oh wait, I haven't generated a java class
20:42	arohner	for my main ns
20:42	Glenjamin	is there a way to get a function to accept either [a b &c] or [a &c] ?
20:42	arohner	Glenjamin: manually parse the args. otherwise it's ambiguous
20:43	Glenjamin	cheers
20:43	Glenjamin	actually, i think my desired functions is ambiguous
20:43	Glenjamin	heh, rethink
20:45	arohner	what's the syntax to start clojure.main from a jar again?
20:46	tieTYT	in lein is there a way to reload the repl so that it downloads new dependencies you've added to the project.clj? I always ctrl+c and reenter
20:46	arohner	java -jar my-jar.jar clojure.main?
20:46	tieTYT	but then i lose all the requires I've built up
20:47	arohner	java -cp target/my-jar-standalone.jar clojure.main
20:47	Glenjamin	tieTYT: i think if you drop out and run lein deps, then require should work?
20:48	Glenjamin	ctrl+z / fg or separate shell
20:48	tieTYT	hm
20:48	tieTYT	ok i'll try that next time
20:48	arohner	tieTYT: there's also https://github.com/cemerick/pomegranate
20:48	arohner	though that doesn't reload project.clj for you
20:49	Glenjamin	i suspect i'm wrong, having read that
20:50	tieTYT	aroemers: thanks
20:50	tieTYT	arohner: i meant thanks
20:51	amalloy	Glenjamin: yeah, no way that works
21:00	tieTYT	this article should be at the top of google when you search for a multimethod tutorial: http://www.fatvat.co.uk/2009/01/multi-methods-in-clojure.html
21:02	tieTYT	"Reading Programming Clojure shows that multi-methods are very rarely used in most of the current Clojure code (only about 1 occurrence per thousand lines of code). There's some guidelines in the book about choosing whether to use multi-methods, but it's summed up best with "try writing the function in both styles, and pick the one that seems more reliable"."
21:02	tieTYT	is that accurate?
21:02	tieTYT	i mean do you guys agree with that?
21:10	Raynes	Sure.
21:10	Raynes	I think that it is worded strangely.
21:11	Raynes	People don't sit around going "HMMMM, I don't know if I should use a multimethod or not…". A multimethod is either a good solution or it isn't. It solves a specific purpose, and if it isn't used very often it is merely because that purpose doesn't come up all that often.
21:11	drewc	I do not agree, but do not disagree, fwiw.
21:13	tieTYT	well my situation is I'm writing a program that's kind of like a web spider. It's going to traverse different sites but I'm going to program in the logic of how to traverse each site given a starting url. So the logic will differ depending on if you go to a.com/... vs b.com/...
21:13	tieTYT	so I was thinking a multimethod that dispatches on the domain of the url would be a good fit
21:27	lovemuffin	is there a way to keep the jvm 'warm' ? for instance running `lein new` to see the help then running `lein new app project` slow
21:30	mabes	lovemuffin: nailgun does that but it comes with its own set of problems... check out drip https://github.com/flatland/drip
21:32	lovemuffin	ah so its like jvm pools
21:33	lovemuffin	also for vim users, is the official unofficial vim plugin for clojure? https://github.com/vim-scripts/VimClojure
21:45	xeqi	lovemuffin: the common way is to use https://github.com/sattvik/lein-tarsier, though I know there is alot of work being done on https://github.com/tpope/vim-fireplace
21:46	xeqi	not sure if vim-fireplace is the new standard yet
21:51	Glenjamin	"espionajure" good/bad name for a clojure stub/spy lib?
21:51	arohner	Glenjamin: https://github.com/circleci/bond
21:51	arohner	oh, that needs a readme
21:51	Glenjamin	bah, where were you two hours ago >.<
21:52	hyPiRion	Glenjamin: well, you have to set LEIN_IRONIC_JURE for that
21:52	Glenjamin	oo, metadata
21:53	Glenjamin	thats much smarter than dynamic binding a map of function => vector
21:53	lovemuffin	i've been reading programming clojure (pragprog) and eh its ok, does a lot of talking not enough doing
21:53	Glenjamin	this is what i had: https://gist.github.com/glenjamin/7edfefcbf1eaa02d6a58
21:54	lovemuffin	so in a clojure project, the best debugger is the repl ?
21:54	jerryzhou	hi
21:55	Glenjamin	arohner: that's exactly what I was about to write, only done better :D
21:55	lovemuffin	or if it isn't what is the standard way to debug clojure?
21:55	arohner	Glenjamin: thanks. I'll write a readme then [ANN] it
21:55	Glenjamin	[ANN] ?
21:55	arohner	announce, on the google group
21:56	Glenjamin	oh right
21:56	arohner	message titles are typically prefixed with [ANN]
21:56	Glenjamin	i should subscribe to that
21:56	Glenjamin	oh, actually
21:57	Glenjamin	(with-stub [fn retval, fn2 retval2] body) would be cool
21:57	arohner	(with-redefs [f1 (constantly retval) f2 (constantly retval2)] body)
21:58	Glenjamin	yeah, but then i have to do with-redefs & with-spy :)
22:00	Glenjamin	i can knock up a patch for that if you'd accept it?
22:00	xeqi	lovemuffin: if you can use the repl, hugod has done a lot of work on ritz to make a real debugger https://github.com/pallet/ritz
22:00	Glenjamin	i'd say sugar for constant retvals and fake implementations would be handy
22:01	xeqi	... the repl is step one, when a real debugger is needed see ritz
22:02	lovemuffin	ok, so thats the standard way of doing it? most people don't depend on some IDE debugger
22:02	danielglauser	lovemuffin: most folks use emacs as their "IDE"
22:03	arohner	Glenjamin: how would you distinguish between (with-stub [f1 retval f2 retval]...) and (with-stub [f1 f2]) ?
22:03	Glenjamin	arohner: more macros :)
22:04	Glenjamin	or maybe (with-stub {f1 retval f2 retval}
22:05	Glenjamin	and then (with-fake {f1 #(…) f2 #(…)} ...)
22:07	arohner	hrm, vector vs. map isn't bad
22:07	arohner	I'm kind of proud of the fact that the entire lib is only 44 lines :-)
22:08	Glenjamin	yeah, i clearly need to work on my vector manipulation in macros :D
22:09	Glenjamin	anyway, uk time so i'm heading off
22:09	Glenjamin	i'll definitely be using bond though
22:10	Glenjamin	nice work on circle too, used it a couple of times and liked it :)
22:10	john2x	what's the difference between pedestal-app and pedestal-service?
22:12	powr-toc	john2x: apps are client-side (clojurescript) programs... service's are server side (clojure) programs
22:13	arohner	Glenjamin: thanks!
22:14	Glenjamin	only found out it was clojure fairly recently, just seemed like any other app from a consumer pov
22:14	john2x	powr-toc, oh ok. so if I don't want to work in clojurescript (yet), I can just create a service and build the client whichever way?
22:14	Glenjamin	which i guess is a good thing?
22:16	powr-toc	john2x: yup
22:17	john2x	powr-toc: cool, thanks!
22:42	chessguy	'evening, everyone
22:44	jerryzhou	hi
22:45	chessguy	anybody have a good recommendation for some good clojure code to read? i've got plenty of programming experience, including functional programming, and ruby
22:49	dcolish	clojurewerks does good stuff
22:50	ivan	chessguy: core.clj
22:50	ivan	actually, clojurescript core.cljs
22:52	chessguy	ivan\: you mean this? https://github.com/clojure/clojurescript/blob/master/src/cljs/cljs/core.cljs
22:53	ivan	yes
22:53	chessguy	thanks
22:55	ivan	I also remember enjoying trying to understand the cljs emitter
23:47	muhoo	TimMc: yes, it's not only in a usable state, i'm using it constantly now
23:58	technomancy	I think I finished up syme: https://syme.herokuapp.com/