#clojure – 2013-11-11

2013-11-11

00:10	arrdem	TimMc: trash-cli is quite nice... thanks!
00:19	`cbp	bitemyapp: give me your steam id sometime ;)
00:20	arrdem	`cbp: starcrafts?
00:20	`cbp	arrdem: only broodwar
00:27	arrdem	`cbp: aw.. bitemyapp and I've been playing when I have access to an sc2 capable box.
00:27	bitemyapp	my main games lately have been starcraft 2 and LoL. I'll play DotA2 if somebody invites me.
00:28	`cbp	arrdem: maybe ill get it once i remember my damn blizzard password :P
00:29	bitemyapp	arrdem: DotA2 is available for Linux.
00:29	bitemyapp	arrdem: if you played that, you wouldn't need the other boxen.
00:29	bitemyapp	arrdem: why not install Steam and see how it goes?
00:29	arrdem	M-x pacman -S steam
00:34	arrdem	oh right. now I remember why I didn't have steam on linux....
00:34	bitemyapp	arrdem: ?
00:34	arrdem	there's a Kerbal Space Program build for linux.
00:35	bitemyapp	arrdem: well install steam, I'm really curious if DotA2 works.
00:35	arrdem	bitemyapp: installed and installing..
00:35	bitemyapp	I like DotA2 quite a bit, better than LoL in most respects and would like to play it if possible.
00:37	arrdem	I know a bunch of people who favor d2 over LoL I've just never gotten into d2
00:37	technomancy	kerbal sounds nuts (in a good way)
00:37	bitemyapp	arrdem: it's like any other MOBA. It's somewhere in between HoN and LoL in design principle.
00:38	bitemyapp	the pacing and design of DotA2 is a lot better though. Much more balanced and interesting game.
00:38	`cbp	dota2 works kinda crappily on my macbook air
00:38	`cbp	not like its a machine meant for gaming
00:38	`cbp	not that*
00:38	bitemyapp	:(
00:38	`cbp	(i play on a pc though)
00:39	bitemyapp	I keep thinking about getting a gaming PC so that I don't have to play on a mac.
00:39	`cbp	it should be pretty affordabe to build one :P
00:39	arrdem	bitemyapp: if you do don't just build a gaming box. build a multihead dev box with enough power to play games.
00:40	bitemyapp	arrdem: I have a Thinkpad for work, what specifically are you angling for with your recommendation?
00:42	arrdem	bitemyapp: I got a 3x 23" monitor desktop with more than reasonable specs out of a research position freshman year
00:42	bitemyapp	I don't want a ton of hardware.
00:42	bitemyapp	just a machine that can game.
00:42	arrdem	shrug. I don't mind the metal, and I don't know enough about portable gaming boxes to comment.
00:43	technomancy	any machine can game; you just have to pick the games right
00:43	arrdem	my perception is that there's a significant price and power primium, but I know some people who are happy with theirs.
00:44	arrdem	technomancy: that's more of an operating system than a game...
00:44	technomancy	don't get me started on nethack
00:46	arrdem	bitemyapp: it's gonna take a bit to get dota2 downloaded.. I'll let you know what happens tomorrow.
00:47	bitemyapp	arrdem: okiedokie.
00:48	swarthy	nethack the Roguelike? I really enjoy Roguelikes but I haven't played nethack. I loved Brogue, I have a hope to one day code a squad based ala Dungeon Siege style rogue-like.
00:49	bitemyapp	swarthy: have you played dungeon crawl stone soup?
00:49	swarthy	bitemyapp: I have that one was fun as well, I need to fire that one up again soon.
00:50	bitemyapp	I like roguelikes a lot partly because of the ephemerality
00:50	bitemyapp	I was really pleased to see the indie game scene learn from the genre and incorporate it into new ideas.
00:50	bitemyapp	I really like 2d games too.
00:51	swarthy	Yeah, making a word out of ascii or bland tilesets is amazing fun.
00:51	swarthy	I played d&d for the first time the other day and realized how advanced modern gaming is. The average gamer copes with a ton of systems.
00:51	swarthy	Other than making up and 'playing a role' d&d was so simple.
00:51	bitemyapp	I like 2d pixel based graphics too.
00:52	swarthy	Bastion is really cool. 3d but locked into isometric.
00:52	swarthy	So it has this cool 2d feel.
00:54	bitemyapp	I really liked that.
00:55	john2x	ooh roguelikes. I love Brogue. It's so pretty, considering it's ASCII "graphics"
00:55	justin_smith	nethack is my jam
00:56	justin_smith	I want to port nethack.el to clojure
00:56	justin_smith	make clojure nethack bots
00:56	swarthy	yeah Brogue did look really good. I'm eager to play the new Dwarf Fortress style game from the makes of Dungeons of Dredmor when it comes out.
00:57	swarthy	makers*
00:58	bitemyapp	swarthy: oh damn, I really liked Dredmor, what game is that?
01:00	bitemyapp	ohhh clockwork empires.
01:00	bitemyapp	yeah that's been long coming
01:00	justin_smith	one of these weekends I will play amnesia: a machine for pigs some more
01:00	justin_smith	works great on my linux system
01:01	justin_smith	(no wine, native support)
01:01	bitemyapp	justin_smith: yeah, indie folk are pretty good about mac and linux support.
01:01	swarthy	I have played games most of my life, and I've recently killed all my old Windows for Gaming Only harddrives. I thought I would rarely play games now. Turns out the Linux story for games is really good now. So that didn't curb my addiction at all.
01:02	bitemyapp	that's partly why I'm struggling with whether or not I want a dedicated gaming machine.
01:02	bitemyapp	I would consider outsourcing that to consoles, if consoles had the kind of games I like.
01:02	arrdem	bitemyapp: wait 12 months see what happens with steambox, re-evaluate.
01:03	justin_smith	I used to have a really good rhythm, where I would play some nethack, and when I started doing real good I knew my wits were sharp, so I would do some coding, then I would get stumped and play some more nethack
01:03	bitemyapp	arrdem: that was one option I was considering.
01:03	justin_smith	but people at work would think I was just slacking
01:03	bitemyapp	justin_smith: that's a good idea.
01:03	bitemyapp	justin_smith: I usually take a walk or make tea to accomplish the same thing.
01:03	justin_smith	heh, maybe healthier :)
01:04	bitemyapp	I suppose so.
01:04	bitemyapp	The real point is to empty my mind.
01:04	justin_smith	I keep my instrument by my desk, so I can stop and play if I am home
01:04	justin_smith	yeah
01:04	justin_smith	playing an instrument is good for that
01:05	swarthy	Yeah I built an HTPC but it turns out it was so cheap and so nice I could just put my gaming GFX card in it was a gaming pc.
01:05	swarthy	Technology is too good to get any real work done.
01:05	justin_smith	alter right next to me on the first level, first room
01:05	justin_smith	*altar
01:06	bitemyapp	arrdem: http://i.imgur.com/tOTISJF.jpg
01:07	arrdem	bitemyapp: nice
01:15	muhoo	wait, did thrown? get removed from the language since these docs? http://clojuredocs.org/clojure_core/clojure.test/is
01:16	muhoo	it isn't in clojure.test ns public vars AFAICT
01:25	coventry	muhoo: I've used thrown? relatively recently. It's a method of the assert-expr multimethod.
01:33	mindbender1	Please can somebody delete the page at http://en.wikipedia.org/wiki/Communicating_sequential_processes?
01:33	justin_smith	why?
01:33	clojurebot	why is startup slow is busy compiling the `for` macroexpansion
01:34	mindbender1	justin_smith: because of http://en.wikipedia.org/wiki/Communicating_sequential_processes#Examples
01:35	mindbender1	How does that represent the objectives of CSP?
01:35	mindbender1	Boring!
01:36	justin_smith	/ignore mindBender1
01:39	bitemyapp	mindbender1: you just got plonked old school style son, you should consider that a Sign (TM)
01:40	justin_smith	bitemyapp: my classic newb mistake, typing a space before the / that made it a command
01:40	muhoo	coventry: CompilerException java.lang.RuntimeException: Unable to resolve symbol: thrown? in this context, compiling:(/tmp/form-init1896962088566288833.clj:1:771)
01:42	muhoo	maybe i don't understand multimethods then. i thought the symbol would resolve. pretty sure i've been able to (doc some-multimethod) before
01:56	arrdem	bitemyapp: dota2 is totally playable
01:56	arrdem	now bed
01:57	bitemyapp	arrdem: awesome, catch you later.
02:18	amalloy	muhoo: clojure.test/assert-expr is a multimethod that dispatches on first
02:18	amalloy	it has a defmethod for the symbol 'thrown?
02:18	amalloy	but there is no multimethod named thrown? anywhere
03:26	muhoo	oh! duh. then thrown? isn't a function, it's the firt arg to (is
03:26	muhoo	(is thrown? (expr....)) not (is (thrown? (expr....)))
03:28	amalloy	no, it's (is (thrown? ...))
03:28	amalloy	but it's still not a function
03:28	muhoo	now i have to go look at the source of assert-expr
03:29	amalloy	'is is a macro that uses the assert-expr multimethod to decide how to expand
03:29	amalloy	it passes assert-expr the first symbol in the list you're expanding
03:30	Jarda	hmm https://github.com/mylesmegyesi/metis seems great but does anyone know of a variant that could be used to export a 'sanitized' version of the passed data?
03:32	Jarda	I mean, if I hava data like {:my-number "123.4"} and I run it through a validator that :my-number actually is numeric and I would then get {:my-number 123.4} as output
03:32	Jarda	or do I have to bake it myself
03:37	muhoo	amalloy: thanks, just looked at it, saw where the dispatch is. neat.
05:33	sm0ke	i just copied my project to a new one a changed namespaces
05:34	sm0ke	now lein repl fails with Unable to resolve symbol create: is in this context
05:34	sm0ke	i doesnt give anything useful in stack trace
05:35	sm0ke	oh wow i misspelled a namespace
05:46	sm0ke	wtf is :ssl :yes!!!11} in
05:46	sm0ke	posta wiki
05:46	sm0ke	postal*
05:46	sm0ke	cant he just use a straight :yes!
05:59	CookedGryphon	sm0ke: you can just use true
05:59	CookedGryphon	that annoyed me too
05:59	CookedGryphon	i think it's just anything truthy
06:01	sm0ke	CookedGryphon: thanks i thought that so
06:01	sm0ke	funny that author would give such contrived example
06:01	CookedGryphon	send a pull request chganging the docs? It genuinely put me off using postal for a while
06:02	sm0ke	hehe yea i will do that
06:02	CookedGryphon	I only used it as a last resort, and functionally it's the best I've tried
06:02	sm0ke	CookedGryphon: do you seem to have used html formatting in your mails?
06:02	sm0ke	any easy way to do that?
06:02	CookedGryphon	yeah
06:02	sm0ke	tell me
06:02	CookedGryphon	one second, i'll find it
06:04	CookedGryphon	Instead of a plain text string, send it [:alternative {:type "text/plain" :content "blah"}{:type "text/html; charset=utf-8" :content (hiccup [:html [:body [:p "blah"]]])}]
06:05	CookedGryphon	that was probably too long to type inline, sorry, should have done a refheap
06:05	sm0ke	awesome sauce
06:05	sm0ke	i am copying that
06:10	sm0ke	CookedGryphon: just wondering whats that :alternative?
06:10	sm0ke	fallback from clients which dont have html
06:10	sm0ke	for*
06:13	CookedGryphon	I can't remember, I think so yes. It's in the README
06:13	CookedGryphon	you can do attachments, multipart and all sorts in that format
06:13	sm0ke	ok thanks
07:00	ambroseb_	only 20 hours left in the Typed Clojure dev campaign. Don't miss out! ;) http://www.indiegogo.com/projects/typed-clojure
07:01	rstandy	hello Clojure community
07:01	ambroseb_	rstandy: hello there
07:01	Jarda	ambroseb_: why should I care about typed clojure
07:02	rstandy	I'm sorry to ask a possible flame war starting question, but, what framework do people use in Clojure when developing webapps?
07:02	justin_smith	rstandy: not a flame-invoking question at all
07:03	rstandy	well, then this is a really neat community :-)
07:03	ambroseb_	Jarda: it's AFAIK the only way to make strong static claims about Clojure code.
07:03	cark	rstandy: i only use ring, but it's not a full featured framework
07:03	justin_smith	liberator, pedastel, Luminous are popular
07:04	CookedGryphon	I tend to avoid frameworks and use the small libraries i need
07:04	justin_smith	there have been flames about the design of Noir in the past, but even the autor of noir depricated it :)
07:04	gwillickers	rstandy: i just started too, i went with compojure
07:04	ambroseb_	Jarda: prevents you from misusing null, really helps using plain maps.
07:04	gwillickers	which sits atop ring
07:04	rstandy	justin_smith: thank, I've seen there are various frameworks, but being new to Clojure I would love to invest some time in the "right" one...
07:04	ambroseb_	Jarda: and it's just a library, sprinkle it where you want.
07:05	cark	rstandy: most frameworks are based on ring, you c&n't go wrong by learning that
07:05	Jarda	I'm using compojure for routing, haven't felt the need for anything else than compojure+ring
07:05	rstandy	justin_smith: "the right one" being that which is popular enough to give me documentation, great community to ask, ecc.
07:05	CookedGryphon	thing is, if you just go with small libraries, building on top of ring and maybe compojure, anything else you use isn't a large investment
07:05	noidi	Jarda, +1
07:05	CookedGryphon	you can swap out smaller components when a better library comes along
07:05	justin_smith	rstandy: I actually am a developer for a framework, that will likely be getting its official "public release" in a day or two, I could say more after that, don't want to overshare before the event
07:05	noidi	we've been happily using ring + compojure + stencil for templating
07:06	justin_smith	CookedGryphon: yeah, start with ring and compojure, most things should modularly combine with those
07:06	justin_smith	inc CookedGryphon
07:06	Jarda	I use hickup templates, haven't really looked into alternatives
07:06	justin_smith	(inc CookedGryphon)
07:06	lazybot	⇒ 1
07:06	justin_smith	there we go
07:06	rstandy	wow so many choices
07:06	gwillickers	i went hiccup too, no complaints
07:06	justin_smith	rstandy: they are made of modular parts, mostly
07:06	Jarda	anything that builds on top of ring is fine as long as they do not try to hide ring
07:07	justin_smith	yeah
07:07	rstandy	any commercial site I can see to have an idea of what I can accomplish with which framework?
07:07	justin_smith	it is not so much "frameworks" you get with clojure as 'sets of libs that work together'
07:07	cark	there are some frameworks
07:07	rstandy	justin_smith: right, I understand it's one of the objectives of Clojure too
07:07	cark	medestal maybe ?
07:07	rstandy	justin_smith: this is so good
07:07	cark	pedestal*
07:08	Jarda	in my philosophy: frameworks that try to do everything _always_ fail at doing every possible use case -> go with the one that helps but doesn't block you when you need to go low-level
07:08	justin_smith	the one my group is releasing in a day or two from now is a similar model to luminous - giving you a decent set of default libs as middleware, and leaving things modular
07:08	rstandy	Jarda: very true! I used Weblocks from CL, which is very nice, but forced me again to work its way
07:09	cark	ah there was hunchentoot in CL ... nothing else was need"ed, and so easy to use !
07:09	justin_smith	pedastel is a more "total solution" top to bottom design (for better and/or worse)
07:10	CookedGryphon	a rule of thumb I have found from experience - avoid anything which requires you to use a defblah macro
07:10	rstandy	ok, so pedestal force you to model in its way
07:10	justin_smith	CookedGryphon: you are trying to get inced by me again aren't you?
07:10	CookedGryphon	usually means they are doing something unidiomatic which isn't going to compose or play nice with other things down the line
07:10	justin_smith	very good advice
07:11	CookedGryphon	:)
07:11	rstandy	looking for compojure
07:11	cark	i think the defXXXX thing isn't too bad as long as there is a functional interface as well
07:13	CookedGryphon	cark: yeah, and as long as it's not doing too much to obscure how it actually works underneath. I've never really liked the magic of compojure's defroutes, but it does make for a good quick-start guide
07:13	rstandy	so for example, if I need a webapp with session management and login/registration, what can I use to start very fast?
07:13	justin_smith	rstandy: there are ring middlewares for session and auth
07:13	justin_smith	friend is great for auth
07:14	justin_smith	I did a kind of roll-my-own but plan to upgrade to friend
07:14	justin_smith	ring-session for session management
07:14	rstandy	justin_smith: thanks! looking for friend right now
07:14	justin_smith	https://github.com/cemerick/friend
07:14	justin_smith	http://clojuredocs.org/ring/ring.middleware.session/wrap-session
07:17	rstandy	thank you very much guys, very helpful!
07:18	justin_smith	CookedGryphon: regarding def-foo ugliness, for the webapp template I've been working on, we ended up dropping compojure because it was so defroutes centric, and we wanted to manage the routes via the cms
07:19	justin_smith	so that a frontend guy who does no clojure could use the cms interface to define the page he needs
07:19	justin_smith	and it is messy integrating routes created as data in a db with a defroutes model
07:20	rstandy	looking for all of this libraries
07:20	justin_smith	rstandy: that is what lein templates are good for
07:21	justin_smith	it basically copy/pastes a sane default app setup, complete with requiring the right deps
07:23	justin_smith	https://clojars.org/search?q=lein-template
07:24	justin_smith	for anything that defines a foo/lein-template you should be able to run lein new foo
07:26	justin_smith	so from that list you have compojure-app, cljs-template, mies, noir, pedestal-service, luminus, each providing some kind of web service / web app functionality
07:27	justin_smith	sadly clojars has not been auto-indexing quite right, because I know of others that exist that are not on that results page
07:37	sm0ke	what a mess i am unable to send mails with postal
07:37	justin_smith	there is a good postmark lib for clojure
07:37	sm0ke	it says {:code 0, :error :SUCCESS, :message message sent} but i dont get no mail
07:38	CookedGryphon	odd, spam?
07:38	justin_smith	sm0ke: the reason people use postmark, is email providers use whitelists
07:38	CookedGryphon	what sort of account are you sending from sm0ke
07:38	sm0ke	nothing in spam too
07:38	justin_smith	roll your own email server is very hard because you need to get on whitelists and shit, easier to use a whitelisted service
07:39	CookedGryphon	depending on your use case though, you might want to just send from a gmail account (that's what i have been doing)
07:39	justin_smith	CookedGryphon: from the looks of it, postal is not imap or such, it is actual smtp, so it would not get far on the open net
07:39	sm0ke	CookedGryphon: yes me too
07:39	sm0ke	i am using a gmail accout
07:39	CookedGryphon	all the appropriate gmail options set?
07:39	sm0ke	why would postal be so cruel to give success message
07:40	CookedGryphon	it's not suspicious about the sending and asking for a captcha on the web interface is it?
07:40	sm0ke	CookedGryphon: i guess so .. everything he has in docs
07:40	CookedGryphon	i mean on the gmail site itself
07:40	CookedGryphon	imap and smtp aren't on by default
07:40	sm0ke	i am using smtp settings only
07:40	CookedGryphon	but presumably you wouldn't then get a success message
07:40	sm0ke	i only need to send
07:41	justin_smith	sm0ke: does mail within localhost work?
07:41	sm0ke	hmm i guess i need only sendmail installed on my machine right?
07:41	justin_smith	postal should be doing what sendmail does already
07:42	sm0ke	i think it uses a mail agent underneath
07:42	CookedGryphon	you don't need anything set up locally to send with a gmail account
07:42	justin_smith	CookedGryphon: the reason I suggested localhost testing, is that eliminates a bunch of weird variables
07:43	CookedGryphon	yeah, introduces a load more though to have a working local mailserver
07:43	sm0ke	hmm weird i jusr installed ssmtp and everything fails now
07:43	justin_smith	sm0ke: what os are you using? a normal linux os will set up local mail (for system messages to administrator)
07:43	sm0ke	i get {:code 1, :error nil, :message nil}
07:43	sm0ke	ubuntu 12.04
07:43	justin_smith	yeah, ubuntu already has mail out of the box
07:44	sm0ke	eee
07:44	justin_smith	that is how root gets messages about system status stuff
07:44	justin_smith	with a normal install
07:44	justin_smith	err... maybe I was wrong
07:45	justin_smith	debian does this, but ubuntu may not
07:45	justin_smith	postfix sets it up though
07:45	sm0ke	i have postfix
07:45	sm0ke	ssmtp stopped it though
07:45	sm0ke	and now i get errors
07:45	sm0ke	rightfully*
07:46	CookedGryphon	sm0ke: so my settings for sending via gmail are (send-message {:username "...@gmail.com" :password "***" :host smtp.gmail.com :port 465 :ssl true} {:to "" :subject "" :body ""})
07:46	CookedGryphon	are you doing exactly the same?
07:47	justin_smith	sm0ke: you can test local mail by using mail or mutt, it should just install and work
07:47	justin_smith	CookedGryphon: so you are able to use the postal lib with gmail?
07:47	CookedGryphon	yeah
07:47	CookedGryphon	wait, sorry, user and pass, not username and password
07:47	sm0ke	hmm weird
07:48	CookedGryphon	i rename them on teh way in so my config is consistent
07:48	guns	I use postal with smtp.gmail.com; works well. Maybe the machine sm0ke is hitting is failing
07:48	sm0ke	i am using ^{:user "... " :pass "**"} ..with that meta sign
07:48	sm0ke	weird thats what the doc mentions
07:48	CookedGryphon	yeah, don't know what on earth that's about
07:48	CookedGryphon	teh docs need editing
07:48	CookedGryphon	makes far more sense to just send settings then message
07:49	sm0ke	hmm let me try your version
07:49	CookedGryphon	i was including a :from in the message too
07:49	CookedGryphon	gmail might need that
07:49	sm0ke	yup i have that already
07:51	CookedGryphon	hmm, there's a pull request to fix the docs from 2 years ago...
07:58	sm0ke	some progress, i get com.sun.mail.smtp.SMTPSendFailedException: 530-5.5.1 Authentication Required. now
07:58	justin_smith	sounds like its time to fork it just to get better docs
07:58	justin_smith	lol
07:58	sm0ke	weird though
08:04	sm0ke	CookedGryphon: hey it worked
08:04	sm0ke	actually it needs :user and :pass
08:04	sm0ke	not :username and :password
08:04	sm0ke	i think main problem was with the meta sign
08:06	sm0ke	bye for now
08:17	tim___	dnolen: ping
08:26	gtuckerkellogg	forcer, with the simple setup, elpy-shell-send-region-or-buffer only seems to work for me with an active region
08:27	justin_smith	gtuckerkellogg: wrong channel?
08:27	gtuckerkellogg	justin_smith, indeed!
08:29	justin_smith	this here city slicker thinks he can use INDENTATION instead of PARENTHESIS as God intended
08:29	justin_smith	get 'im boys
08:29	noidi	:D
08:33	tbaldridge	~guards
08:33	clojurebot	SEIZE HIM!
08:33	mdrogalis	Heh
08:33	gtuckerkellogg	lol
08:53	hyPiRion	~gourds
08:53	clojurebot	SQUEEZE HIM!
09:26	terom	There seems to be a JSON-RPC library for Clojure called clj-json-rpc for the server side. I'm looking for a library to help building a JSON-RPC client, is there a such thing already?
09:28	justin_smith	terom: seems like with cheshire and one of the http client libs it would be pretty simple without a dedicated json-rpc client
09:29	justin_smith	terom: the spec is simple enough http://json-rpc.org/wiki/specification
09:30	justin_smith	oh, I see that version 2 is a bit more complex http://www.jsonrpc.org/specification
09:37	rstandy	just another newbie question: is there some in production web app or site developed in clojure?
09:37	justin_smith	I have deployed a few
09:37	justin_smith	(my company, I did the clojure part on the backend)
09:38	justin_smith	there is an out of date list on clojure.org somewhere
09:38	justin_smith	soundcloud
09:38	rstandy	justin_smith: wow, can I see of them (those which are publicly available)
09:38	justin_smith	http://corp.zinio.com/ the only evidence of clojure is how fast it loads
09:38	justin_smith	I guess
09:39	justin_smith	http://weareinstrument.com/work/featured/the-build this one is pretty awesome actually
09:39	rstandy	justin_smith: souncloud.org is simply wonderful!
09:39	justin_smith	both that blog, and the site it links to, use clojure with the framework I helped write, though I have not coded much for those two projects
09:40	justin_smith	http://thebuildfilm.com/ be sure to see this, the actual site that last blog post was about
09:40	justin_smith	flash-like ui in html5
09:41	justin_smith	with clojure backend, of course
09:42	justin_smith	almost forgot this one: done in clojure + clojurescript http://onenorthpdx.com/
09:42	rstandy	justin_smith: by backend you mean that clojure builds the html pages or not?
09:42	justin_smith	it handles the cms
09:42	justin_smith	it serves all the data
09:42	justin_smith	handles the requests
09:42	justin_smith	renders templates
09:42	rstandy	justin_smith: ok, got it
09:43	justin_smith	resizes images for your screen resolution (caching existing resizes)
09:43	rstandy	justin_smith: thank you very much for the info man
09:43	justin_smith	with that last one (one north) the whole UI is built in clojurescript using core.async to handle events
09:44	rstandy	justin_smith: wow
09:44	rstandy	justin_smith: awesome sites!
09:44	justin_smith	thanks, that's mostly because we have awesome designers
09:44	justin_smith	and very meticulous frontend devs that can realize those designs
09:45	justin_smith	but it helps that I can make sure the pages load fast and get the right data too :)
09:45	rstandy	justin_smith: is your backend donwlodable free software? :-)))
09:45	justin_smith	rstandy: it is open source, we will be announcing and publicising it soon
09:46	rstandy	justin_smith: this is very interesting!!! Will you email the clojure ml?
09:46	justin_smith	basically "watch this space", we have a nice frontend site for intro and docs that will be public in a day or two
09:46	justin_smith	yeah, I think that is the plan
09:47	rstandy	justin_smith: exciting news, I hope to be productive with clojure pretty soon, I want to push it in my workplace
09:47	justin_smith	it is a pretty cool setup, I can take only marginal credit, the main dev is an old friend of mine that pulled me in to help because he knew I liked programming lisp
09:48	justin_smith	rstandy: huge selling points are reduced server resources (thus less deployment cost) massively reduced development time (meaning more agile to accomidate design changes, and your project will never be behind schedule for backend reasons)
09:48	rstandy	justin_smith: nice, what was your language and environment before clojure
09:48	rstandy	justin_smith: huge selling points!
09:48	justin_smith	I got hired to do clojure
09:49	justin_smith	before that they were using ruby
09:49	justin_smith	but server costs and reliability sucked
09:49	rstandy	justin_smith: dream job for me :D
09:49	justin_smith	before that php
09:49	justin_smith	rstandy: it is very nice
09:49	rstandy	justin_smith: yeah, ruby sucks in many ways
09:49	rstandy	justin_smith: are you coming from CL+Emacs
09:49	rstandy	+Slime?
09:49	justin_smith	it is especially since I can be productive enough in clojure that I spend more of my time honing my skills and improving our libs than I spend developing sites
09:50	justin_smith	yeah, common lisp / scheme / ocaml
09:50	justin_smith	definitely emacs no matter what, but that's the tool not the medium :)
09:51	rstandy	justin_smith: yeah, any advice/tip for an CL+Emacs+Slime developer approaching clojure?
09:51	justin_smith	actually I was not even a professional programmer before jumping into this, I was just fascinated by algorithms and functional programming
09:51	justin_smith	hmm
09:51	justin_smith	get cider
09:51	justin_smith	which is the update to nrepl.el
09:52	justin_smith	use paredit, after a few days learning it it is indespensible
09:52	rstandy	justin_smith: yeah! Wonderful mode I already use in CL
09:52	justin_smith	and (source fn) (doc fn) are your friends :)
09:53	justin_smith	oh, and throw (ns-publics ns) in the mix too
09:53	rstandy	justin_smith: I installed cider, and I'm starting to explore clojre through it, but I miss the Slime Inspector
09:53	justin_smith	we are working on a cdt web frontend
09:53	justin_smith	I have not had much luck with ritz
09:54	rstandy	justin_smith: so a Slime-like inspector is still missing in clojure
09:54	justin_smith	cdt assumes you are connecting to a separate program that is being debugged, so we can connect to the process with the server, have an html UI, and click on stack trace lines to evaluate expressions in that context / see locals
09:54	nDuff	...well, there _is_ ritz
09:55	nDuff	...if you're willing to go through the pain to set it up.
09:55	justin_smith	rstandy: theoretically ritz can do that stuff, I have not had luck with it
09:55	rstandy	justin_smith: will check that then
09:55	justin_smith	we looked at ritz as the backend for our web-based debugger, and decided cdt was easier and had all the functionality we needed
09:55	rstandy	justin_smith: ok, sometimes do you miss some CL and Slime features when using clojure (sorry for so many questions)?
09:55	justin_smith	yeah, for sure
09:56	justin_smith	like the step debugging etc.
09:56	rstandy	justin_smith: just trying to understand what I'll miss and what I'll gather with clojure
09:56	justin_smith	but, with immutible data structures and a more strictly functional approach, I find invasive debugging less neccessary
09:56	justin_smith	I can do a lot just by verifying each function produces the right output for the expected input
09:57	justin_smith	also I often miss static checks and currying from ocaml
09:58	rstandy	justin_smith: man your answers are really useful for me, many thanks!
09:58	justin_smith	speaking of debugging, I released a middleware for ring the other day called groundhog, that archives requests so that they can be reconstructed and replayed to verify bugfixes
09:58	justin_smith	https://github.com/noisesmith/groundhog
09:58	rstandy	justin_smith: wow, nice
09:59	justin_smith	I need to update that readme, the latest is 0.4
09:59	justin_smith	*0.0.4
10:00	rstandy	justin_smith: You gave me so much info to starts with that I'm tempted to abandon the idea to learn clojure... :D
10:01	justin_smith	?
10:01	justin_smith	why would you not want to learn clojure?
10:01	rstandy	justin_smith: just joking, it's so much good info that I want to learn everything now!
10:01	justin_smith	hah
10:02	rstandy	justin_smith: very intriguing language and community with fantastic libraries, this is what the clojure world is for me right now
10:02	justin_smith	I concur
10:02	justin_smith	it's great to work with a language that I unreservedly love
10:03	rstandy	justin_smith: hope to be able to become very productive soon. You know, when the boss ask what work has been done, I have to answer something :D
10:03	rstandy	justin_smith: I tried to stick with CL, but lack of libraries is becoming to restrictive
10:03	mdrogalis	rstandy: Just take your time. It's worth learning, but go slowly and thoroughly.
10:03	justin_smith	well, clojure does have some features (and default / recommended coding style) that leads to productivity once you internalize it
10:04	rstandy	justin_smith: very interesting perspective about OOP. I think I more or less agree. It's interesting to see how complexity is managed with the clojure way
10:05	cemerick	dnolen: is continuing to support :export meta carrying an alternative name (rather than truthiness) a real objective? In the face of supporting downstream mutability, it complicates things significantly.
10:06	rstandy	justin_smith: thanks, man, very nice and helpful!
10:06	rstandy	justin_smith: have a nice day and hope to talk with you again
10:06	justin_smith	thanks, you too
10:07	dnolen	cemerick: I don't really understand what you just said
10:07	shoshin	hello
10:07	shoshin	how does one iterate in clojure?
10:08	shoshin	i want to iterate a body twice. how do i achieve this using doseq?
10:08	justin_smith	shoshin: iterate to do an action, or iterate a transformation of a sequential input?
10:08	cemerick	dnolen: right now, you can (defn ^{:export "foo"} bar 5); bar will be emitted and optimized away by GClosure, foo will be an exported "symbol", changes to which will not affect any fns that refer to bar in cljs source.
10:08	Morgawr	,(doc dotimes)
10:08	clojurebot	"([bindings & body]); bindings => name n Repeatedly executes body (presumably for side-effects) with name bound to integers from 0 through n-1."
10:08	justin_smith	,(dotimes [i 2] (print i))
10:08	clojurebot	01
10:09	justin_smith	Morgawr: beat me to it
10:09	Morgawr	:P
10:09	Morgawr	how do you @ clojurebot to somebody?
10:09	Morgawr	(if it's even possible)
10:09	justin_smith	hmm
10:09	llasram	,(str "like this?") ; Morgawr
10:09	clojurebot	"like this?"
10:09	Morgawr	I could do something like
10:10	dnolen	cemerick: yes I know that, but I don't understand what you said above.
10:10	cemerick	dnolen: The fix for CLJS-670 will make such an arrangement impossible, since two properties will need to change in order to maintain consistency between the exported property, and the "internal" one that GClosure has renamed, etc.
10:10	Morgawr	,(str "shoshin:" (doc dotimes))
10:10	clojurebot	"shoshin:([bindings & body]); bindings => name n Repeatedly executes body (presumably for side-effects) with name bound to integers from 0 through n-1."
10:10	Morgawr	:D
10:10	shoshin	Morgawr thanks! will try!
10:10	Morgawr	refer to justin_smith's example (which was probably more useful)
10:10	seangrove	,(println (str "seangrove:" (doc dotimes)))
10:10	clojurebot	seangrove:([bindings & body]); bindings => name n Repeatedly executes body (presumably for side-effects) with name bound to integers from 0 through n-1.\n
10:11	Morgawr	seangrove: that's better
10:11	seangrove	Morgawr: Just wondering if there's another chance for a bot quine... it's almost been one year now
10:11	dnolen	cemerick: I'm pretty sure this has come up before, what's the problem with ^:export ^:expose or something like that?
10:11	Morgawr	a bot quine?
10:12	justin_smith	Morgawr: making the bots trigger one another, infinitely
10:12	rstandy	justin_smith: so, one last question: the most advanced ide, editor, tool for writing and debugging cloujre apps?
10:12	Morgawr	haha
10:12	cemerick	dnolen: you mean expose in the @expose GClosure sense?
10:12	shoshin	Morgawr http://clojuredocs.org/clojure_core/clojure.core/dorun i'm looking for something like example in the first case. repeat some action n number of times.
10:12	Morgawr	rstandy: probably emacs. I personally use both vim and LightTable (LightTable is getting really good)
10:12	seangrove	Morgawr: Yeah, last one was in December 2012...
10:12	justin_smith	rstandy: : we had a "show your editing environment" night for the last clojerks pdx
10:12	Morgawr	shoshin: that's another way, yes
10:13	justin_smith	rstandy: emacs and vim looked the most impressive
10:13	justin_smith	though cursive looked nice if you want a java style IDE experience
10:13	shoshin	Morgawr doesn't seem to work though. :/
10:13	shoshin	i get a null pointer exception.
10:13	Morgawr	dorun is to resolve lazy sequences
10:13	justin_smith	light table was surprisingly disappointing, though it could be because the guy presenting it wasn't adept enough with it
10:13	Morgawr	dotimes is to run N times the same expression
10:13	rstandy	justin_smith: I'm definately on the Emacs side, but what should I use? swank-clojure + swank-cdt? Or cider?
10:14	rstandy	and nrepl?
10:14	justin_smith	rstandy: cider or nrepl
10:14	justin_smith	swank is to be avoided at this point
10:14	dnolen	cemerick: yes, if I recall.
10:14	rstandy	justin_smith: no one is the real winner right now, then
10:14	justin_smith	rstandy: try cider first, since nrepl is transitioning to that
10:14	Morgawr	,(dotimes [_ 5] (println "hi"))
10:14	clojurebot	hi\nhi\nhi\nhi\nhi\n
10:14	Morgawr	^
10:14	rstandy	oh ok, so cdt is not recommended anymore
10:14	Morgawr	or use n instead of _ if you care about the variable bound to the iteration count
10:14	justin_smith	rstandy: cdt is only partially tied to swank
10:15	justin_smith	rstandy: I use it without using swank at all
10:15	rstandy	justin_smith: ah ok
10:15	justin_smith	via the browser
10:15	justin_smith	not even in the editor
10:15	rstandy	justin_smith: sorry for the silly question
10:15	justin_smith	np
10:15	justin_smith	glad to help
10:15	rstandy	justin_smith: :-)
10:16	CoconutCrab	okay, maybe it would be more proper to ask in #emacs instead, but how can one remove nrepl from emacs (via package.el) to install cider?
10:16	dnolen	cemerick: I'm guess didn't affect set-print-fn because we didn't give it an initial value?
10:16	dnolen	er print-fn
10:16	cemerick	dnolen: Two different topics, really. First, we definitely don't want @expose, it takes a bunch of analysis/optimizations off the table when applied to fns, which inevitably yields piles of (spurious AFAICT based on testing) warnings.
10:17	cemerick	dnolen: the issue with the alternative ^:export name is that we're setting up two different properties when it's used, a natural source of inconsistency if there's mutability about.
10:17	justin_smith	CoconutCrab: in the package interface, d will mark a package for deletion, and x will act on all marks
10:18	CoconutCrab	justin_smith: oh, silly me... sorry for bothering you...
10:18	CoconutCrab	and thank for answering
10:18	justin_smith	CoconutCrab: no problem, it is not intuitive unless you are used to gnus, mh, dired, and all the other little emacs things that use the mark / act paradigm
10:19	justin_smith	the downside with emacs is it is not consistent with "normal", the upside is it is pretty good at being internally consistent
10:19	CoconutCrab	justin_smith: I see now uninstall option and tend to use '-' for marking
10:19	CoconutCrab	(like in debian's aptitude)
10:20	CoconutCrab	then I forgot that emacs has C-h m...
10:20	justin_smith	yeah, never expect emacs to act like another program you are used to...
10:20	augustl	what _does_ harmonikit sound like? :) https://twitter.com/richhickey/status/399918185450704897
10:20	cemerick	dnolen: well, set-print-fn! was never really "exposed", at least w.r.t. mutation, because it was just aliased to the exported symbol.
10:21	justin_smith	augustl: I would say a clojure library for midi controlled compressed air actuated harmonica robots
10:21	napper	does anyone have a recconmendation for image manipulation libs with clojure?
10:21	justin_smith	augustl: and I will be severely disappointed if it is not that
10:21	justin_smith	napper: what kind of manipulations?
10:22	justin_smith	napper: for basic stuff interop with BufferedImage has worked well for me, but there is also imgscalr
10:22	cemerick	dnolen: anyway, my fundamental Q is whether I can sacrifice e.g. ^{:export "some.other.name"} in order to support downstream mutation of exported properties. They're really fundamentally incompatible. Alternatively, we could say that properties exported with alternative names are read-only.
10:23	cemerick	i.e. retain the same semantics they have now, and don't get the same support as described in CLJS-670
10:23	napper	justin_smith: to begin, just plastering images together to make css sprites - but I'm always a fan of graph-based-algorithum-image-manipulation
10:23	karls	napper: i tried to find an image manipulation library that does more than scale/rotate, but couldn't. had to drop down to java.
10:24	napper	karls: it's been a few years for me. what is the current java img-lib?
10:24	napper	karls: ie, in python its PIL
10:25	justin_smith	napper: I use java.awt and javax.imageio
10:25	karls	napper: i re-wrote something that originally used PIL, in clojure. i didn't use a library at all. the interop is so good that i just ended up using BufferedImage related java facilities.
10:26	karls	napper: what justin_smith said :-)
10:26	justin_smith	that's java.awt.image.BufferedImage, if you are looking for the javadoc
10:26	napper	justin_smith, awesome, I'll look into them. BufferedImage + those libs
10:27	justin_smith	napper: they come as part of the jvm, you just need to import them
10:27	napper	justin_smith, karls, I'll see what I can do about building something more then just rotation and scaling. :)
10:27	justin_smith	next part is adapting yourself to learning via javadoc
10:27	justin_smith	napper: please do! though you may just find the built in interop is enough (as we did)
10:28	napper	justin_smith, heh, there was a time when I was a j2ee dev, then I found python o:
10:28	napper	thanks for the leads!
10:28	justin_smith	np
10:34	johnpete	Hi - I'm trying Clojure and Compojure for the first time. The lein template has given me some code that has an application of some function called -> How do I find what this function is? I can't google it!
10:35	seangrove	cemerick dnolen: Any tickets or help you'd like with cljs or cljsbuild?
10:35	justin_smith	wow, it looks like the analemma svg generation library doesn't use java interop, so it may be clojurescript compatible
10:35	karls	,(doc ->)
10:35	clojurebot	"([x] [x form] [x form & more]); Threads the expr through the forms. Inserts x as the second item in the first form, making a list of it if it is not a list already. If there are more forms, inserts the first form as the second item in second form, etc."
10:35	karls	or http://clojuredocs.org/clojure_core/clojure.core/-%3E
10:35	seangrove	,(println (str "johnpete: " (doc ->)))
10:35	clojurebot	johnpete: ([x] [x form] [x form & more]); Threads the expr through the forms. Inserts x as the second item in the first form, making a list of it if it is not a list already. If there are more forms, inserts the first form as the second item in second form, etc.\n
10:36	mdrogalis	johnpete: Yeah it's a weird one. It's called "thread first".
10:36	justin_smith	johnpete: also there is a cool site for searching symbols that other engines don't index http://symbolhound.com/?q=clojure+-%3E
10:36	johnpete	great, thanks everyone
10:38	cemerick	seangrove: cljsbuild is semi-frozen until 1.0.0 is out, then we can start seriously recasting it. If you're in an investigative mood, you could look into what is in certain Chrome extensions that breaks the browser-REPL, e.g. https://github.com/emezeske/lein-cljsbuild/issues/262 and https://github.com/cemerick/austin/issues/17
10:38	seangrove	cemerick: Uhg
10:39	cemerick	heh
10:39	cemerick	The one in the first link is far simpler than the Google Voice extension, so it might be tractable.
10:39	seangrove	cemerick: Will check into it, we make a chrome extension as our primary product right now :)
10:39	cemerick	seangrove: my exact sentiment
10:40	cemerick	seangrove: That's interesting. As the primary business, or is the extension a complement to an online service, etc?
10:40	seangrove	cemerick: http://www.zenboxapp.com/ - clojure and clojurescript
10:41	seangrove	cemerick: As an extension working inside of gmail with other unaware extensions... it's all a nightmare
10:42	cemerick	seangrove: what's the business model, if I may ask?
10:42	seangrove	cemerick: We charge people for premium features, pretty straightforward
10:43	cemerick	seangrove: oh, cool; I guess the upsell is in the extension itself? I didn't see anything obvious about it on the site.
10:43	seangrove	cemerick: Yeah, exactly. Most things are free and it just works, but when someone reaches deeper into the integrations we start a time trial, and then get cc info, charge, etc.
10:44	cemerick	nice
10:44	cemerick	seangrove: the reviews are glowing, congrats. :-)
10:44	justin_smith	or if they want the sidebar to wear a custom hat or get that powerup they need to beat the level...
10:44	seangrove	cemerick: Wsan't always that way, but things have been moving in the right direction :) Getting reified keywords into cljs helped quite a bit
10:45	justin_smith	seangrove: that does look cool, our producers all use gmail, I'll see if they use it already or might find it useful
10:45	seangrove	justin_smith: IAP for level upgrades coming soon, I suppose
10:45	justin_smith	heh
10:45	justin_smith	angry birds does well with holiday editions
10:46	justin_smith	"$1 to add the christmas theme, $100 to remove it again"
10:46	seangrove	hah
10:47	justin_smith	so for a chrome extension do you basically compile to a javascript package?
10:48	seangrove	justin_smith: I believe it's a tar, but we use openforge to take care of the gritty details these days
10:48	justin_smith	cool
10:54	mdrogalis	Anyone here using ZooKeeper for coordination? #zookeeper is a bit quiet.
10:54	justin_smith	it's near the top of my "we should be using this lib" pile
10:54	justin_smith	haven't tried it though
10:54	TheLastProject	,(println ((first '[One Two Three]) (nth '[One Two Three] 2)) (second '[One Two Three]))
10:54	clojurebot	nil Two\n
10:54	TheLastProject	Hmm...
10:54	mdrogalis	Not really a lib, more of a service.
10:55	justin_smith	mdrogalis: I meant the clojure client lib
10:55	mdrogalis	justin_smith: Ah, understood.
10:55	TheLastProject	,(println (join " "((first '[One Two Three]) (nth '[One Two Three] 2))) (second '[One Two Three]))
10:55	clojurebot	#<CompilerException java.lang.RuntimeException: Unable to resolve symbol: join in this context, compiling:(NO_SOURCE_PATH:0:0)>
10:55	justin_smith	TheLastProject: did you mean (println [(first ...
10:55	mdrogalis	It has some annoying semantics around watching for changes.
10:56	TheLastProject	justin_smith: I'm trying to get the dispatch feature to work better, but every attempt I throw at it just turns into chaos
10:56	justin_smith	TheLastProject: ((first '[One Two Three])) will try to execute 'One as a function
10:57	TimMc	mdrogalis: We are, but I can't claim to be familiar with it.
10:57	mdrogalis	TimMc: Hmm, alrighty. Thanks anyway :)
10:57	justin_smith	one of these weekends I plan to make a zookeeper client as a ring middleware
10:57	mdrogalis	It pretty much supports "watching" a piece of state. But after the event triggers, it kills the watch.
10:57	justin_smith	so coordination is ensured with each request
10:58	mdrogalis	So it's more like "watch-once" semantics.
10:58	justin_smith	can't you make your response action create the watch again?
10:58	justin_smith	recursion :)
10:58	TheLastProject	justin_smith: http://pastebin.com/F9rEx7KG (line 29) - I still have two problems with it. It somehow complains about "java.lang.String cannot be cast to clojure.lang.IFn", but I can't figure out why, and I still need to figure out how to have it also send all additional arguments given, ugh...
10:58	mdrogalis	justin_smith: It's a distributed system. Interesting things can happen inbetween that would be missed.
10:59	justin_smith	mdrogalis: true
10:59	justin_smith	TheLastProject: that error means that you put something that returns a string in the calling position
11:00	TheLastProject	justin_smith: I can't see what that would be, though. If I use print to test it it nicely resolves to "list clothing" and the likes, which (dispatch) should resolve to the function before calling...
11:01	justin_smith	TheLastProject: I am going to do a small cleanup of your code to make it a little easier to think about, one moment
11:01	ltw	what's the best approach for creating a library with support for core.typed, but that doesn't require it?
11:01	TheLastProject	justin_smith: I guess it could need that. I can't say it doesn't look like an utter mess. Wonder how much different your version will be from mine, hmm
11:02	ltw	i was thinking multiple versions (i.e. 0.1.0 and 0.1.0-typed)
11:03	justin_smith	TheLastProject: mainly I am using let so you don't have repeated calls to the same function noising things up, I'll paste it in a moment
11:05	justin_smith	TheLastProject: https://www.refheap.com/20719
11:06	justin_smith	that is not yet correct, but I think it is easier to read
11:06	justin_smith	I put it on refheap because refheap is less spammy
11:06	terom	justin_smith: thanks. I've heard of cheshire but not used it. Version 2.0 of JSON-RPC definetely seems a little bit more complex... maybe I'll try to come up with a library of my own.
11:07	justin_smith	terom: cheshire is very easy to use, it turns clojure vectors / maps / strings / numbers into json, and the reverse
11:10	TheLastProject	justin_smith: I should've probably updated the println above, but I'm trying to make it execute (list-clothing upper) if the user types "list upper clothing" and, after I get that working, also take the other parameters so I can get add-clothing to work, but those as just strings
11:11	TheLastProject	justin_smith: I see that your command thing only takes the first parameter :P
11:11	justin_smith	TheLastProject: was that my mistake? I thought I was just following the nesting of the original, one sec
11:12	justin_smith	oh, I did mess it up, but neither of our versions are correct :)
11:12	TheLastProject	Well, I know mine isn't correct
11:12	TheLastProject	At least yours looks more readable
11:13	justin_smith	in your original, if you linebreak and autoindent, you can see that the result of calling dispatch on the third arg is being called as a function
11:13	justin_smith	with (collection ...) as its arg
11:14	TheLastProject	Huh. I tried to very carefully retype it with print and it returned "first third" when I tried to do that. Now I wonder if I'm really so used to how most object-oriented languages use () that I just can't even read it properly
11:15	TheLastProject	s/print/println
11:17	justin_smith	TheLastProject: https://www.refheap.com/20719
11:17	justin_smith	edited, I think closer to what you want
11:17	justin_smith	using amalloy_'s suggestion to use apply
11:18	justin_smith	so we split that input into tokens, first of those looks up the command, rest are args, first arg looks up the collection, and then we just pass in the rest
11:18	TheLastProject	The first and third arg combined is supposed to be the command :x
11:18	justin_smith	ok, but you know how to fix that, right?
11:19	justin_smith	by changing the definition of command in the let statement
11:19	TheLastProject	Well, no, because command ((first tokens) (nth tokens2)) causes Clojure to complain about it being a string
11:19	TheLastProject	Err...
11:20	TheLastProject	Yeah, that's what I'm trying to do, but I can't understand why it complains it is a string that it can't convert to a function
11:20	justin_smith	((first tokens)) tries to execute the return value of (first tokens)
11:20	TheLastProject	s/tokens2/tokens 2
11:20	TheLastProject	Oh...
11:20	TheLastProject	Oh, darn it
11:20	justin_smith	, (:hello {:hello "world"})
11:20	clojurebot	"world"
11:20	justin_smith	, ((:hello {:hello "world"}))
11:20	clojurebot	#<ClassCastException java.lang.ClassCastException: java.lang.String cannot be cast to clojure.lang.IFn>
11:20	justin_smith	that is your error
11:21	justin_smith	() is not for grouping in clojure
11:21	justin_smith	ever
11:21	justin_smith	it can make a literal list, or apply a function
11:21	justin_smith	that's it
11:21	TheLastProject	Okay, I need to get that in my head
11:21	TheLastProject	But now I wonder how to "group" it then because otherwise it will complain about having too many arguments, hmm...
11:22	justin_smith	what you want is to construct the string so it matches your lookup table
11:22	justin_smith	,(str "add" \space "clothing")
11:22	clojurebot	"add clothing"
11:22	justin_smith	,(clojure.string/join " " ["add" "clothing"])
11:22	clojurebot	"add clothing"
11:22	bja	,(+ 1 2 3)
11:22	clojurebot	6
11:22	bja	wow, didn't realize that worked for anybody
11:23	justin_smith	bja: you can also send a private message to clojurebot
11:23	TheLastProject	Ah...
11:24	justin_smith	TheLastProject: the funny thing is you even use clojure.string/join at the top of the file
11:24	TheLastProject	Okay, yeah
11:24	TheLastProject	I get it, [] is for turning it into a list, which is the closest to grouping you get
11:24	justin_smith	there is also the list function
11:24	TimMc	[] is for creating a vector
11:24	TimMc	,[1 2 3]
11:24	clojurebot	[1 2 3]
11:25	TimMc	,(list 1 2 3)
11:25	clojurebot	(1 2 3)
11:25	justin_smith	or () if quoted
11:25	TheLastProject	Ugh, yeah
11:25	TheLastProject	You people call it a vector
11:25	TheLastProject	My bad
11:25	TimMc	Well, we also have lists.
11:25	TheLastProject	Every language uses different names, and a "list" isn't the same thing in all of them :P
11:25	justin_smith	eventually keeping track of which one you are using avoids weird bugs :)
11:25	TheLastProject	Okay, well, thanks
11:26	TheLastProject	Now I just need to do the args stuff, let's see how that goes
11:26	justin_smith	TheLastProject: here is a trick I sometimes use when debugging
11:26	justin_smith	,(let [a 2 _ (println a) b 3] (+ a b))
11:26	clojurebot	2\n5
11:26	justin_smith	so it prints a, then returns 5
11:27	TheLastProject	I'm having trouble following that, heh...
11:27	justin_smith	the _ is a convention that says "I am not using this return value" in a let
11:27	justin_smith	so I assign a with the value 2
11:27	justin_smith	I assign _ with the value nil (returned by print)
11:27	justin_smith	the reason to do it there is I can then look at things as I construct them
11:27	justin_smith	it helps sometimes
11:27	TheLastProject	Hmm
11:28	TimMc	Clojure lists are basically stacks (append and peek at the front); vectors are log-time random-access and append at the end.
11:28	justinjaffray	justin_smith: that is so useful, thanks :D
11:29	justin_smith	justinjaffray: np
11:29	TheLastProject	Is there a way to have something like (rest) but starting at the 4th point in the list, instead of the third?
11:29	TheLastProject	s/third/second/
11:29	TheLastProject	Oh man, such a headache, heh...
11:29	TheLastProject	I'm not used to coding this way
11:30	TheLastProject	It's interesting, but it's... different
11:30	justin_smith	TheLastProject: nthrest
11:30	justin_smith	TheLastProject: http://clojure.org/cheatsheet is helpful
11:30	TheLastProject	That actually makes sense
11:30	justin_smith	TheLastProject: the pain is your brain growing stronger!
11:31	TheLastProject	Yay! It works!
11:31	TheLastProject	Thank you SO much justin_smith
11:31	justin_smith	very nice, can I see the final result?
11:32	justin_smith	no problem, glad I could help
11:32	TheLastProject	Of course
11:32	TheLastProject	justin_smith: https://www.refheap.com/20722
11:32	faust45	is it possible assign metadata to types ?
11:33	faust45	like (deftype User ^{})
11:33	TimMc	What would the behavior be?
11:34	justin_smith	TheLastProject: the order of assigning the values in the let is a little tangled, but yeah, that looks great
11:34	justin_smith	by tangled I don't mean incorrect, just not in the easiest to read order
11:34	TheLastProject	I guess collection-string should go for args
11:34	TheLastProject	Makes more sense
11:35	TheLastProject	But I'm mostly glad it works now, heh...
11:35	justin_smith	heh, yeah, it is an accomplishment
11:36	napper	has anyone putout a repl like ipdb in clojure?
11:36	justin_smith	another suggestion is to put everything after the last defn (and all the printlns) into a -main function
11:36	justin_smith	then lein run will work
11:36	napper	I'd like the opition to break in my code without having to use an ide.
11:36	TheLastProject	Ah, I'm running it with "clojure" but I've heard leiningen is really popular, should probably strive for leiningen-compatibility
11:37	justin_smith	TheLastProject: it makes things much easier, eventually
11:37	justin_smith	and even if you are just running the clojure command, you can put (-main) at the end of the file, to explicitly execute it
11:38	TheLastProject	clojure-main?
11:38	justin_smith	napper: cdt can do breaks that you define in a repl, and you can use it without swank pretty easily
11:39	justin_smith	napper: we've been working on a web UI that would let us set breakpoints and investigate values / run code in a clojure project we attach to
11:39	justin_smith	we use cdt for that
11:40	justin_smith	TheLastProject: ?
11:40	TheLastProject	justin_smith: I'll take that question mark as a "no"
11:40	justin_smith	I don't understand the question
11:40	justin_smith	the magic function name defaults to -main
11:41	TheLastProject	justin_smith: (defn -main [] (doseq [line (repeatedly read-line) :while line] (getinput line) (request-command)))
11:41	TheLastProject	So, that stuff?
11:41	napper	justin_smith, interesting, I'm curious as to how you would break on a line with a gui per thread. - How do you trace anom functions?
11:41	TheLastProject	(Put it on one line to not spam, I'm not undoing your improvements)
11:41	justin_smith	yeah, put all the printlns in there too
11:41	TheLastProject	Ah
11:41	justin_smith	napper: that is a UI for stack traces on exception
11:41	TheLastProject	So that they execute when it is first run
11:41	TheLastProject	Got it!
11:42	justin_smith	napper: we haven't done source line breakpoints, but that would be a different UI
11:43	napper	justin_smith: when using ipdb, it would break/stop tho code execution in a single thread allowing you to inspect what is in memory. Much like the gdb interface ( at least what I know of the gdb interface. )
11:43	justin_smith	napper: yes, same idea
11:43	napper	justin_smith: ideally, you'd build a pure function and concat them together with threads in mind.
11:43	justin_smith	but we show all the available stacks at once
11:44	napper	justin_smith: neat, I'm eager to see it - it's the last thing I need to convert. :P
11:44	justin_smith	*available frames
11:44	justin_smith	source line breakpoints is something we may do later - but that gets weird when you are using the repl / debugger to redefine things
11:46	CookedGryphon	justin_smith: it'd be cool if defn attached the source as metadata
11:46	justin_smith	CookedGryphon: it does
11:46	justin_smith	but that doesn't give you line number breaks inside it :)
11:46	CookedGryphon	oh, cool, well in that case I don't care what line it is in a file, as long as i can look at the source.
11:46	justin_smith	CookedGryphon: (source fn)
11:46	napper	justin_smith: redefine? Well, yeah, but if a user does that at run time, it's their prerogative - using in debugging would be ideal, redfined code at runtime, I agree can create issues
11:46	CookedGryphon	or is it that your stack traces no longer relate to anything
11:47	justin_smith	that's in the clojure.repl ns, so you may need to use that first
11:47	justin_smith	napper: this debugger lets you redefine things while debugging
11:47	napper	are we talking about cdt or your webui?
11:48	justin_smith	the webui is just a view of cdt
11:48	justin_smith	all the functionality, including changing remote definitions, is in cdt
11:48	napper	right
11:49	napper	I'm reading over the main-page. it looks like what I want is already accomplishable via the debug-repl
11:49	justin_smith	napper: I think the problem with that is it is swank
11:49	napper	oh
11:50	justin_smith	ritz tries to provide an alternative
11:50	justin_smith	but I couldn't make any sense of ritz
11:53	justin_smith	if your workflow is "I am starting the debugger and the debugged process as one action", then the nrepl ritz module probably does what you want
11:53	justin_smith	in my case, I am running ring, ring creates a repl I can interact with, and I want to fire up an external debugger to connect to it and investigate if something breaks
11:54	muhoo	http://www.youtube.com/watch?v=9m3lu59Cz4g
11:54	justin_smith	muhoo: nice
11:54	napper	eventually I'll be running ring, my next question is how can you pause execution. looking through cdt, it seems that might be pretty straightforward. - I think I have what I need.
11:55	dnolen	cemerick: yeah I don't think we should change the current behavior
11:55	justin_smith	napper: yeah, you can set a break, if you know how to specify the location
11:55	napper	justin_smith, is your webui public or closedsource?
11:55	justin_smith	napper: open source, but not really "there" yet
11:55	napper	I understand
11:55	napper	do let me know when it's there
11:56	justin_smith	someone else is doing the main development, I don't even know how much he has pushed as opposed to just mostly working on his box at this point
11:56	justin_smith	https://github.com/prismofeverything/schmetterling repo
11:56	justin_smith	may be a pile of bugs right now
11:56	cemerick	dnolen: hah, here I just finished a writeup on the ticket :-P
11:57	justin_smith	napper: it is something you would check out and run, standalone
11:57	justin_smith	not a lib you include
11:58	napper	this would be great for pointing it to production servers if anything ever went south
11:58	justin_smith	yeah, though you would need to open a tunnel with ssh -X
11:58	justin_smith	because you sure as hell don't want a dt-thread socket available to the outside world :)
11:59	dnolen	cemerick: also I don't really like the idea of adding any compiler support for this - just export a config function?
11:59	napper	hah
11:59	napper	i take it you've seen that before?
12:00	justin_smith	no, but it's something that could be a disaster
12:00	justin_smith	full control of your app's state
12:00	justin_smith	and full access and introspection
12:00	justin_smith	someone with malice could do a lot with that
12:00	lpvb	how do I use vector-of to create a filled vector of a specified length?
12:01	lpvb	(vector-of :int & elements)
12:01	lpvb	,(vector-of :int)
12:01	clojurebot	[]
12:01	lpvb	,(vector-of :int (repeat 5 0))
12:01	clojurebot	#<ClassCastException java.lang.ClassCastException: clojure.lang.LazySeq cannot be cast to java.lang.Character>
12:02	justin_smith	,(apply vector-of :int (repeat 3 6))
12:02	clojurebot	[6 6 6]
12:02	justin_smith	lml
12:04	cemerick	dnolen: The original use case was being able to change test-time parameters. There are workarounds, including the config function notion, having a fixture that looks up params set from the test runner on window, etc.
12:04	lpvb	oh cool
12:05	justin_smith	lpvb: apply is good when a function expects a bunch of args at the end that you already have in a list
12:05	justin_smith	with how many varargs functions clojure has, this comes up a lot
12:06	cemerick	What ^:export does now is pretty surprising tho. If someone wants some low-hanging contrib fruit, documenting that @ https://github.com/clojure/clojurescript/wiki/Quick-Start#using-clojurescript-on-a-web-page (the only place I know of that mentions ^:export?) would be one such spot.
12:06	justin_smith	,(apply vector-of int "Hello World")
12:06	clojurebot	#<NullPointerException java.lang.NullPointerException>
12:06	justin_smith	erm
12:07	justin_smith	,(apply vector-of :int (seq "hello world"))
12:07	clojurebot	[104 101 108 108 111 ...]
12:07	justin_smith	neato
12:07	dnolen	cemerick: you're one of the few people that have asked for this or been surprised by the behavior. ^:export is mostly used for exporting functions note constants that will be mutated as this isn't considered good CLJS style. The testing case is not even particularly interesting IMO since it's trivially worked around.
12:07	dnolen	s/note/not
12:09	dnolen	cemerick: it has come up once before in a different ticket I believe, but I'm not convinced any of this requires compiler changes. I think just make the documentation more clear is sufficient for now.
12:10	cemerick	dnolen: Fair enough, I'm not lobbying hard or anything. :-) My understanding of ^:export was that it was primarily there to provide entry points and to make providing JS-friendly APIs (where mutable properties are common).
12:11	cemerick	APIs easy*
12:11	dnolen	cemerick: I don't recall anyone suggesting ^:export for mutable properties, nor do I see such documented anywhere (could be wrong)
12:12	cemerick	dnolen: No, I inferred "JS APIs" => "mutable properties".
12:13	cemerick	I'll close the issue and leave a note potential workarounds, etc.
12:13	dnolen	cemerick: cool, thanks
12:14	technomancy	huh; I could have sworn the `hub` CLI tool had gist support in it; am I making that up?
12:14	nDuff	...I've seen a separate ''gist'' tool
12:15	nDuff	perhaps its functionality has been split out from ''hub''.
12:15	Sigma	technomancy: not that I remember. there's a separate "gist" tool though
12:15	technomancy	oh, I was thinking of gist-mode for emacs
12:16	technomancy	which supports fetching but not cloning; oh well
12:16	Sigma	technomancy: patches welcome :)
12:17	deech	Hi all, how do I make a set with a custom comparator?
12:17	technomancy	Sigma: nice =)
12:17	Sigma	btw I'm considering (as of yesterday) implementing a tramp method for accessing gists. guess it might make the whole experience a bit better
12:17	faust45	guys, how i can attache some static info to my deftype ?
12:18	TEttinger	deech, like sorted-set-by but not sorted?
12:18	technomancy	Sigma: huh... I never thought of it that way, but tramp can be considered a FUSE alternative in a way.
12:18	Sigma	technomancy: I already have one for git itself, sounds like a logicial next step :)
12:19	technomancy	that's wild =)
12:19	deech	TEttinger: I thought that function took a set and sorted it. Guess I was wrong.
12:19	justin_smith	tramp access to well designed RESTful apis all over the web would finally catch us all up to plan9 networking
12:19	faust45	any ideas ?
12:20	technomancy	Sigma: lemme know if you ever end up tackling issues; that's the main pain point for me with github currently
12:20	faust45	how attach static info to deftype ?
12:20	Sigma	technomancy: I guess the final step would be to reimplement magit over a tramp method that would use github API instead of local git commands. Then I can die
12:21	technomancy	Sigma: the technical term is "ascend"
12:21	justin_smith	faust45: you could make the data be a method that ignores the value of this and returns the data
12:22	faust45	justin_smith: yes, but sounds strange
12:23	faust45	justin_smith: i don't need behavior i need static data
12:23	justin_smith	(getStatic [this] {:foo "bar"})
12:23	TEttinger	deech, fun stuff: ##(sorted-set-by #(= %1 %2) 1 1 1 2 2 3 3)
12:23	lazybot	⇒ #{1 1 1}
12:23	justin_smith	a behavior can mimic static data :)
12:23	faust45	ok thanks
12:23	justin_smith	faust45: or you could use genclass instead of deftype
12:24	justin_smith	they generate the same kind of thing under the hood
12:24	justin_smith	then you can have a final data member in the class itself
12:24	justin_smith	just think of how java would do it, then figure out the interop that would make that happen, because all this is defined by the java model
12:34	TimMc	faust45: I think you can attach :static to methods in a gen-class :methods form.
12:45	justin_smith	TimMc: is that the only way to make static methods?
12:45	justin_smith	I was looking at it, it seemed like you would only be able to define one :static key?
12:46	justin_smith	sorry
12:46	justin_smith	not :static
12:46	justin_smith	it's :state
12:46	justin_smith	or are these two separate things?
12:46	justin_smith	http://clojuredocs.org/clojure_core/clojure.core/gen-class
12:46	dobry-den	gen-class OO in clojure is fun
12:47	dobry-den	feels like some newfangled OO language
12:47	tbaldridge	for certain definitions of fun
12:47	dobry-den	^
12:48	dobry-den	funny*
12:48	justin_smith	clearly gen-class is missing the essential :laffs key
12:48	dobry-den	every time i use gen-class, i do the same google search to knit together the same 4 blog posts that show how to use it
12:52	lpvb	how do I split a vector into chunks of a specified length like [1 2 3 4 5 6 7 8 9] -> [[1 2 3] [4 5 6] [7 8 9]]
12:52	AimHere	lpvb > partition
12:52	rasmusto	,(partition 3 (range 10))
12:52	clojurebot	((0 1 2) (3 4 5) (6 7 8))
12:53	justin_smith	,(partition-all 3 (range 10))
12:53	clojurebot	((0 1 2) (3 4 5) (6 7 8) (9))
12:53	justin_smith	if you wanted all of it :)
12:53	rasmusto	,(partition 3 1 (range 10))
12:53	clojurebot	((0 1 2) (1 2 3) (2 3 4) (3 4 5) (4 5 6) ...)
12:53	rasmusto	if you want overlap
12:53	lpvb	thanks
13:18	danneu	I've never done something like this before - If I wanted to distribute Datomic along with my app.jar, what would that sort of thing look like?
13:18	llasram	I've come to the opinion that there's ~0 situations where `gen-class` is the right solution
13:19	danneu	Naively, I imagine my app.jar would boot another jvm that runs the transactor and java would handle things like shutting down the child jvm
13:20	mdrogalis	danneu: Can you put a wire between those 2 components?
13:20	mdrogalis	Just avoid all that.
13:20	danneu	What do you mean?
13:20	bitemyapp	yeah don't do that.
13:21	mdrogalis	Well, if I was to tell you that you can avoid the oddness that is multiple JVMs on a machine, doesnt that sound nice?
13:21	danneu	Basically, all I know is that I can't distribute a single jar that contains my app + datomic-db, so I'm looking up solutions
13:21	mdrogalis	danneu: What's your deployment scenario?
13:22	danneu	mdrogalis: I have a bitcoin node that uses datomic to store the blockchain. I'm not sure how to distribute it to other people
13:24	danneu	really, just a project that uses datomic as the db, and im not sure how to package it for end users
13:24	mdrogalis	danneu: I mean, do you have to ship just 1 jar?
13:25	mdrogalis	I dont think there's anything wrong with saying "Here's my app jar. Go stand up Datomic and they'll place nicely."
13:25	nDuff	...folks don't try to ship PostgreSQL built into their applications. :)
13:25	mdrogalis	nDuff: Thanks, was trying to verbalize that.
13:25	bitemyapp	danneu: yeah I'm not buying that should be baked in.
13:25	bitemyapp	danneu: use environment variables so users can point your app at their datomic instance.
13:26	coventry	danneu: Could you pull in the datomic depepndency from your code using something like pomegranate?
13:26	bitemyapp	danneu: This is a good library for that: http://github.com/weavejester/environ/
13:26	bitemyapp	coventry: dude. No.
13:29	danneu	nDuff: yeah, you're right. i'm just trying to arrive at the easiest solution for the end-user
13:31	lpvb	,(replace {1 0} (vector-of :int 0xF0))
13:31	clojurebot	[240]
13:31	lpvb	why is that not working
13:31	coventry	lpvb: What answer are you expecting?
13:31	lpvb	oh wait I get it now
13:32	lpvb	I thought the key was the index to replace
13:33	justin_smith	,(replace {0 0} (vector-of :int 0xF0))
13:33	clojurebot	[240]
13:33	justin_smith	ok, that I did not expect
13:34	justin_smith	,((vector-of :int 0xF0) 0)
13:34	clojurebot	240
13:35	bitemyapp	justin_smith: what are you even doing that these experiments come to mind? seriously, wtf.
13:35	justin_smith	bitemyapp: it was lpvb's question
13:35	Morgawr	I have a static java method that can take 1, 2 or 3 parameters, I want to write a wrapper in clojure, but I can't get it to work with "apply"
13:35	bitemyapp	lpvb: okay, my question, applied to you. What the fuck?
13:35	Morgawr	(apply StaticClass/method args) doesn't work
13:35	lpvb	bitemyapp: ?
13:35	sritchie	Morgawr: you could write a macro that expands out into all the arities
13:35	bitemyapp	Morgawr: make a clojure function
13:35	coventry	"Here's a nickle, kid. Buy yourself your own goddamn repl." :-)
13:35	llasram	lpvb, justin_smith: `replace` replaces by value, not key
13:35	lpvb	I know that now
13:36	llasram	kk
13:36	noncom	what is the fastest and the most idiomatic way to add N same elements to a vector?
13:36	sritchie	Morgawr: but I think you need to recreate each one
13:36	Morgawr	I know I could do that but a one liner feels better, isn't there a way?
13:36	justin_smith	llasram: the doc string says key
13:36	mdrogalis	bitemyapp: Possibly relevant for encryption, by the way
13:36	justin_smith	oh, = key, not at key
13:36	Morgawr	I mean, I can't use "apply"?
13:36	justin_smith	never mind
13:36	sritchie	Morgawr: nope
13:36	danneu	noncom: conj with repeat?
13:36	Morgawr	why not? doesn't apply work with java methods?
13:37	noncom	danneu: is it the fastest? i can think of several ways, but i wonder which is the fastest...
13:37	danneu	,(into [1 2 3] (repeat 5 :a))
13:37	clojurebot	[1 2 3 :a :a ...]
13:37	bitemyapp	Morgawr: why would some random Java method implement IFn?
13:37	Morgawr	bitemyapp: I don't know how apply works, so I don't know why it wouldn't work ^^
13:37	bitemyapp	Morgawr: it's not like java libraries are designed with Clojure integration in mind.
13:37	rasmusto	,(replace {0xF0 1} (vector-of :int 0xF0))
13:37	danneu	noncom: oh. in that case i would just mock up those several ways and bench them
13:37	clojurebot	[1]
13:37	bitemyapp	Morgawr: well...learn how Clojure works :)
13:37	bitemyapp	Morgawr: in particular, Seqable, IFn, etc.
13:38	Morgawr	I'll go check apply's source
13:38	bitemyapp	enables you to anticipate what should and shouldn't work, what makes sense, etc.
13:38	sritchie	to be fair that's what he's trying to do
13:38	noncom	danneu: yeah, looks liek..
13:38	danneu	noncom: oh, you asked for the fastest
13:38	sritchie	Morgawr: a clojure function is an object, a first class object
13:38	noncom	yes, speed is the criterium
13:38	sritchie	that implements IFn
13:39	coventry	Morgawr: The only way I can think of doing it in one line is uglier than using multiple lines.
13:39	sritchie	a static method isn't first class, it's defined on a particular class -
13:39	Morgawr	I see
13:39	Morgawr	makes sense
13:39	sritchie	so you have to make a first-class clojure function that delegates through in each arity
13:39	Morgawr	alright, thanks
13:39	justin_smith	lpvb: this is what you wanted, I think:
13:39	justin_smith	,(assoc (vector-of :int 0xF0) 0 0)
13:39	clojurebot	[0]
13:40	lpvb	I know how to use assoc, I thought I could use replace to do a list of assocs at a time though
13:40	justin_smith	assoc is varargs
13:40	justin_smith	,(assoc (vector-of :int 0xF0) 0 0 1 6)
13:40	clojurebot	[0 6]
13:40	lpvb	ah I keep forgetting most clojure functions like this have varargs
13:41	llasram	Morgawr: Not necessarily recommended, but you can use reflection to wrap methods as functions w/ correct arities, then `apply` those functions
13:41	justin_smith	,(apply < (range 42))
13:41	clojurebot	true
13:41	Morgawr	llasram: sounds like a lot of work for nothing
13:41	llasram	Morgawr: https://github.com/llasram/method-fn for a proof-of-concept
13:41	llasram	haha
13:41	llasram	Well, depends on what you're trying to do
13:42	Morgawr	llasram: I was just being lazy
13:42	danneu	noncom: is (concat [1 2 3] (repeat n :a)) too slow?
13:42	Morgawr	hoping I could get a simple and elegant one-liner for a java wrapper
13:42	llasram	I see
13:42	coventry	As long as we're proposing awkward ideas, there's (eval `(StaticClass/method ~@args)). :-)
13:42	Morgawr	but it's just a function with 3 different parameter combinations
13:42	Morgawr	so I can just wrote those, no biggie
13:43	Morgawr	coventry: that's actually interesting
13:43	coventry	Morgawr: Yeah, it's interesting, but people will look askance if you do that.
13:43	justin_smith	eval is a big gun for such a small rabbit
13:43	noncom	danneu: umm, looks quite fast.. "Elapsed time: 0.08244 msecs" after warm-up
13:44	noncom	thats faster than with (into)
13:44	Morgawr	I mean, wouldn't it just possible to do (defmacro wrapped-method [& args] `(Static/Methodname ~@args)) ?
13:44	danneu	noncom: right, into uses conj for each element. concat does it all at once
13:44	Morgawr	be possible*
13:44	llasram	noncom, danneu: Er, `concat` is lazy, so just calling `concat` does almost nothing
13:45	coventry	Morgawr: You lose function composition that way.
13:45	Morgawr	mmm.. true
13:46	noncom	llasram: well, i tried it and it printed the output to the console, so it actualle did evaluate it. however, i wonder, if it did before or after (time) :)
13:46	danneu	noncom: you can realize it trivially with `count` or `dorun`
13:46	justin_smith	noncom: if you are wondering what is efficient, use criterium
13:46	llasram	noncom: If you use criterium for benchmarking, it handles many of these issues
13:47	justin_smith	also, criterium is very simple to use, so there is no excuse
13:48	Morgawr	is there any downside in using eval?
13:48	llasram	Especially if you also use hugod's alembic, because then you can slurp it into any REPL :-)
13:48	danneu	sometimes you've just gotta grind it out on that dollar menu with (time (dotimes [_ N] ...))
13:48	llasram	danneu: 100% in disagreement!
13:48	justin_smith	Morgawr: it creates a whole environment context, we were using eval in our templates and when we switched to read our performance shot up more than an order of magnitude
13:49	danneu	;)
13:49	justin_smith	danneu: that would make sense if criterium somehow cost money
13:49	Morgawr	this is just for a simple function (binding to OpenAL) that creates an audio device, it's used just in the initialization of a game engine
13:49	Morgawr	I'm wondering how lazy I can be
13:49	danneu	it's nice to have criterium's `bench` and `quick-bench` imported into your util namespace so theyre always handy
13:49	Morgawr	as in, I know I'm talking about bike shedding atm
13:49	danneu	quick-bench because time is money
13:50	Morgawr	but honestly unless there's any security/consistency/stability problem, if it's only just losing a bit of performance, then I'd say that eval looks more elegant
13:50	Morgawr	unless I'm missing something
13:50	justin_smith	client data sneaking arbitrary code into the string?
13:50	justin_smith	hah
13:50	Morgawr	it's a library, it's not an end-user application
13:51	Morgawr	as in, other developers call that function
13:51	Morgawr	if they want to break their own application, it's not my fault D:
13:51	lpvb	what complexity is concat on two vectors?
13:51	justin_smith	just covering the bases, mentioning other downsides that may apply :)
13:51	Morgawr	true true
13:52	justin_smith	lpvb: they are implemented by narrow hash trees, with something like 7 items on each level iirc
13:52	coventry	Morgawr: If you want to use eval, you probably actually want something more like (eval (list* StaticClass/method args)), so that it works at runtime.
13:52	justin_smith	lpvb: http://hypirion.com/musings/understanding-persistent-vector-pt-1
13:53	Morgawr	coventry: alright, thanks
13:53	TimMc	lpvb: O(1), i think
13:53	coventry	Morgawr: And wash your hands afterwards.
13:53	TimMc	(It's lazy.)
13:53	justin_smith	TimMc: no way, it is log something
13:53	Morgawr	doesn't seem to be working though
13:53	noncom	danneu, llasram: thanks, i'll check the criterium out!
13:54	Morgawr	should it be (list* 'StaticClass/metod args) ?
13:54	Morgawr	with the quote?
13:54	coventry	Morgawr: Right.
13:54	Morgawr	ah ok, at least I understand what it does ;) thanks
13:55	TimMc	justin_smith: concat returns a lazy seq; I'm pretty sure it does nothing involving the elements of input vectors
13:55	coventry	,(eval (list* 'System/getProperty '("java.vm.version")))
13:55	clojurebot	#<Exception java.lang.Exception: SANBOX DENIED>
13:55	Morgawr	haha
13:55	TimMc	lpvb: Concat returns a seq; if you want a vector from a seq you have to call vec on it. If you want to combine two vectors into another vector, the efficient thing is to use clojure.core/into.
13:56	TimMc	&(into [1 2 3] [4 5 6])
13:56	lazybot	⇒ [1 2 3 4 5 6]
13:56	justin_smith	TimMc: ok, now I get it
13:56	TimMc	That should be something like O(n) or O(n log n).
13:56	noncom	conventry: what is the difference of list* and list?
13:56	justin_smith	so since concat only returns a seq, you look at the complexity of realizing the seq (if it needs full realization)
13:57	coventry	noncom: &[(list* 'System/getProperty '("java.vm.version")) (list 'System/getProperty '("java.vm.version"))]
13:57	justin_smith	realizing a seq will be O(n) for the length you need to realize
13:57	danneu	I see that `into` uses transients
13:58	noncom	conventry: sorry i don't get it..
13:58	danneu	justin_smith: yeah, i made that mistake too. i didn't think about how concat is a lazy-seq of `cat`s
13:58	coventry	noncom: list* splices the last argument into the list, like ~@ in a syntax quote.
13:58	noncom	oh!
13:58	coventry	,[(list* 'System/getProperty '("java.vm.version")) (list 'System/getProperty '("java.vm.version"))]
13:58	clojurebot	[(System/getProperty "java.vm.version") (System/getProperty ("java.vm.version"))]
13:59	cYmen	Is there a list of clojure user groups somewhere?
13:59	coventry	Hmm, I could have used juxt, there. :-)
14:00	justin_smith	coventry: so is what differentiates list* from cons the resulting data type?
14:01	Morgawr	talking about clojure user groups... anybody from Amsterdam/Netherlands going to wednesday's clojure meeting in Amsterdam?
14:01	danneu	Speaking of juxt, I often do this: ((juxt :a :b :c) {:a 1, :b 2, :c 3}) -> [1 2 3]. Is there a different way to do that?
14:01	coventry	justin_smith: It's a series cons under the covers.
14:01	coventry	~def list*
14:01	Morgawr	danneu: vals
14:01	Morgawr	if you want all of them, of course
14:02	justin_smith	coventry: ahh, so it is a reverse varargs with the conses in front
14:02	danneu	My trouble with `cons` is that I don't know what to do with the resulting Cons object
14:02	cYmen	danneu: the result is a list not a cons object ;)
14:03	cYmen	well...I guess it could also be a tree
14:03	cYmen	nevermind ;)
14:03	danneu	,(class (cons 1 [2 3]))
14:03	clojurebot	clojure.lang.Cons
14:03	justin_smith	,(map type [(list* 1 [2]) (cons 1 [2])])
14:03	clojurebot	(clojure.lang.Cons clojure.lang.Cons)
14:04	amalloy	danneu: ##(map {:a 1, :b 2, :c 3} [:a :b :c])
14:04	lazybot	⇒ (1 2 3)
14:05	llasram	justin_smith: Yeah, `list*` returning a `Cons` is kind of brutal
14:05	Morgawr	,(vals {:a 1 :b 2 :c 3})
14:05	clojurebot	(1 3 2)
14:05	Morgawr	:V
14:06	justin_smith	llasram: from an implementation perspective, it is perfectly understandable though - a list is just a cons tacked onto another cons
14:07	llasram	justin_smith: That's actually not quite true in the Clojure implementation. Clojure's actual persistent list is `Counted` and supports acts as a (counted) stack with `conj`, `peek`, and `pop`
14:08	justin_smith	oh, interesting
14:09	llasram	If you just need a seq, `Cons`es work fine, but sometimes you really want a PersistentList
14:09	justin_smith	,(pop (into () (list* 1 [2])))
14:09	clojurebot	(1)
14:09	justin_smith	I never though I would need (into ())
14:09	llasram	I think then you might as well just ##(apply list 1 [2])
14:09	lazybot	⇒ (1 2)
14:10	justin_smith	llasram: yeah, probably true
14:10	llasram	Oh, yeah -- `into` will end up reversing the list
14:10	danneu	reduces with conj
14:10	justin_smith	ahh
14:10	justin_smith	fun!
14:11	TimMc	list* is awful
14:11	justin_smith	it seems so!
14:11	amalloy	TimMc: aw, list* is great!
14:12	TimMc	The funcionality is nice; the name is not.
14:12	TimMc	and the doc
14:12	amalloy	it's much nicer than repeated conses, and it's not like i've ever in my life cared that it doesn't return a List
14:12	amalloy	if it did, i'd have to stop using it
14:14	Raynes	TimMc: Don't be so mean to list*.
14:14	amalloy	(defn all-integers [n] (lazy-seq (list* n (- n) (all-integers (inc n))))), for example, suddenly can't use list*
14:14	Raynes	He doesn't mean to be different.
14:15	coventry	What are the drawbacks to list*?
14:15	mischov	He's just looking for a connection.. :(
14:15	llasram	coventry: It doesn't actually return a list
14:15	amalloy	coventry: whenever you use it, you have to listen to someone in irc reminding you it doesn't return a list
14:15	llasram	heh
14:15	TimMc	(inc amalloy)
14:15	lazybot	⇒ 76
14:15	llasram	(inc amalloy)
14:15	lazybot	⇒ 77
14:15	danneu	haha
14:17	technomancy	no, it's the other way around
14:17	amalloy	anyway, it does return a list*, so what's the problem?
14:17	amalloy	* sequence
14:17	technomancy	list? is the rubbish one; list* is entirely reasonable
14:17	TimMc	hrmph
14:17	technomancy	focusing on concrete types is misguided
14:18	TimMc	Actually, the problem is memoize.
14:18	amalloy	yeah, list? is awful. i love to agree with technomancy when he wanders in to say that
14:18	llasram	I'm sure this will never happen to me again, but I once did want the behavior `pop` documents for "a list", and was surprised the result of `list*` didn't work with it
14:18	justin_smith	we should have a database of functions in core clojure and everyone's yes or no opinions on them
14:18	justin_smith	then we could find the winner
14:19	justin_smith	juxt for president!
14:19	llasram	~juxt
14:19	clojurebot	juxt is the bestest option though if it doesn't weird you out
14:19	technomancy	whaaat why is pop a c.l.RT method?
14:19	llasram	Aww, not the one I was hoping for
14:19	justin_smith	~juxt
14:19	clojurebot	juxt is a little hard to grok but it's the best thing ever
14:19	rasmusto	~grok
14:19	clojurebot	grok is a little hard to juxt but it's the best thing ever
14:19	llasram	lol
14:19	coventry	It would be pretty handy to have a table of which clojure.core functions suck and why, actually.
14:19	mdrogalis	Hah
14:20	ToBeReplaced	~complement
14:20	clojurebot	It's greek to me.
14:20	justin_smith	coventry: we could make a funcation that puts a :sucks and a :rocks metadata on each function
14:20	technomancy	am I right in recalling that contains? isn't as bad as it used to be?
14:20	ToBeReplaced	as long as "flatten" is marked with :worst
14:21	justin_smith	yes!
14:21	amalloy	technomancy: i forget if it's contains? or get
14:21	ToBeReplaced	what's wrong with either of those?
14:21	justin_smith	maybe if we could just inc or dec core functions in irc
14:21	coventry	justin_smith: I think we should hack the compiler so that it instruments code with calls to lazybot to (inc) and (dec) the author when :sucks and :rocks functions are used.
14:21	technomancy	ToBeReplaced: contains? is actually contains-key?
14:21	justin_smith	lol
14:21	lpvb	how do I coerce the types of record fields, do I just type hint them?
14:21	technomancy	ToBeReplaced: and it's worse because there's a .contains method that actually does what you'd expect
14:22	ToBeReplaced	yeah, that doesn't smell fishy to me for some reason
14:22	technomancy	I think at least now it throws instead of silently returning false
14:22	ToBeReplaced	throws on what, nil/
14:23	justin_smith	non-int key on a sequential type
14:23	ToBeReplaced	,(contains? [] :foo)
14:23	clojurebot	false
14:23	amalloy	,(contains? 'q :foo)
14:23	clojurebot	#<IllegalArgumentException java.lang.IllegalArgumentException: contains? not supported on type: clojure.lang.Symbol>
14:24	justin_smith	but for that case, I think it always threw
14:24	amalloy	no
14:24	justin_smith	oh
14:24	justin_smith	never mind then
14:24	amalloy	&(contains? 'q :foo)
14:24	lazybot	⇒ false
14:24	ToBeReplaced	amalloy: what's the diff between those two commands?
14:25	ToBeReplaced	the & vs the ,
14:25	amalloy	did you notice anything different happen when i used them?
14:25	coventry	&clojure-version
14:25	lazybot	⇒ {:major 1, :minor 4, :incremental 0, :qualifier nil}
14:25	coventry	,clojure-version
14:25	clojurebot	{:interim true, :major 1, :minor 6, :incremental 0, :qualifier "master"}
14:25	ToBeReplaced	kk sweet
14:30	TimMc	ToBeReplaced: The bots have different Clojure versions in their sandbox evaluators.
14:30	Raynes	Funfact: I have 396 public pastes and 395 private pastes. This was not intentional.w
14:31	Raynes	(on refheap)
14:31	amalloy	Raynes: really? that number seems way too low
14:32	amalloy	or do you mean, just pastes you, personally, made
14:32	Raynes	amalloy: Me of course.
14:32	cYmen	It says here that the vector returned by subvec shares structure with the original one. Isn't that rather odd considering all the immutabiliy everywhere?
14:32	Raynes	If Refheap only had 791 pastes, I should probably have given up on it ages ago :P
14:33	dnolen	cYmen: Clojure data structures generally share structure
14:33	cYmen	dnolen: But that doesn't matter if I never mutate any...
14:33	justin_smith	cYmen: that is one of the advantages of immutability, is being able to safely share structure whenever possible
14:33	justin_smith	in terms of memory usage
14:34	dnolen	cYmen: structure sharing is how/why Clojure data structures perform well
14:34	amalloy	well, not so much memory usage as time, justin_smith
14:34	justin_smith	maybe make that saving graces, making up for the extra mem usage because you don't update in place
14:34	technomancy	cYmen: what you pasted is not odd, but subvec does have some weirdness in its trade-off between efficiency and possible memory leaks
14:34	danneu	cYmen: mutability is opt-in.
14:35	amalloy	usually you just throw away the old version, so the fact that you share structure with it impacts only speed, not space
14:35	amalloy	of course, you don't have to throw it away, which is the point of immutability
14:35	technomancy	amalloy: except it's a reference to the whole original vector, not just part
14:35	amalloy	technomancy: i'm not talking about subvec; that's your hobby horse
14:36	amalloy	just sharing in general
14:36	technomancy	oh, I didn't bring it up
14:36	justin_smith	see, more evidence that we need a :sucks or :rocks metadata on every core function
14:36	technomancy	justin_smith: dunno; I think subvec is more nuanced. needs :perilous.
14:36	cYmen	so if I go (let [foo [1 2 3] bar (subvec foo 2)] ...) I can screw myself because changes to foo or bar also change the other?
14:37	technomancy	cYmen: no
14:37	TimMc	:docs-actually-important
14:37	technomancy	you can screw yourself up because of a memory leak, but you can't change either vector
14:37	amalloy	cYmen: it's impossible to change foo or bar
14:37	amalloy	so that doesn't matter
14:38	cYmen	so how is immutability opt-in?
14:38	justin_smith	cYmen: you can use java arrays
14:38	justin_smith	or java lists
14:38	danneu	mutability*
14:38	TimMc	You can change vectors if you do some very naughty things, but you really can't do it by accident.
14:38	technomancy	cYmen: just try to modify a clojure vector
14:38	TimMc	cYmen: Mutability is opt-in.
14:39	danneu	cYmen: You can also 'mark' clojure datastructures as "mutable" http://clojure.org/transients
14:39	justin_smith	it is opt in, in that you can choose what data structures you use
14:39	amalloy	danneu: that is totally wrong. transients do not make data structures mutable
14:39	danneu	damn
14:39	amalloy	they permit the data structures to use mutability as an optimization; they do not allow you to mutate them without restraint
14:40	technomancy	welllll. not totally wrong.
14:40	technomancy	just misses the point.
14:40	danneu	yeah, it was too haste a simplification
14:40	Raynes	amalloy technomancy fight: go!
14:40	danneu	too dumb*
14:40	danneu	i really just wanted to share the link
14:40	nDuff	...more than that, one can't assume that calculating a mutated version of a transient will change it.
14:40	nDuff	one simply can't assume that doing so _won't_.
14:41	cYmen	Where is #clojure-noobs? :)
14:41	danneu	right hurr
14:41	danneu	right hurr with me, homie
14:43	TimMc	Hmm, what's this TransactionalHashMap thingy?
14:43	justin_smith	technomancy: http://memegenerator.net/instance/42894672
14:43	TimMc	public TransactionalHashMap() { this(421); } Nothing like a magic number to spice things up.
14:44	TimMc	https://github.com/clojure/clojure/blob/clojure-1.5.1/src/jvm/clojure/lang/TransactionalHashMap.java
14:44	technomancy	justin_smith: it's like you read my mind
14:44	justin_smith	or maybe your words on irc
14:45	technomancy	it's another possibility
14:45	danneu	Question: Some contributors on my project have started bikeshedding about my preference to refer to coordinates as {:x _, :y _} instead of a tuple [x y]. [x y] is only nice if you're always destructuring it. but otherwise, (coords 0) and (coords 1) are heinous to my eyes.
14:46	justin_smith	danneu: I agree, I even do :x :y :x' :y' for quads
14:46	danneu	thanks. bro-support. i'll remain adamant
14:46	justin_smith	then you can just do a {:keys [x y x' y']} and be done with it
14:46	justin_smith	destructuring ftw
14:47	TimMc	https://groups.google.com/forum/#!msg/clojure/A9Y9wml49_s/EUF4E43PJRIJ
14:53	cYmen	Could somebody please explain this to me: Also, in some cases keys can be used as functions of maps. For example, keyword keys can, but string and integer keys cannot.
14:54	`cbp	,(:a {:a 1})
14:54	clojurebot	1
14:54	cYmen	I apparently have no useful notion of what a keyword is internally.
14:54	cYmen	Because I am very surprised that it can be looked up in a way that allows using it as a function.
14:54	cYmen	Wouldn't that require knowing that it is in a map?
14:55	justin_smith	cYmen: it is an object, that is callable
14:55	justin_smith	it has the right method to try and look itself up if the arg is a map or struct
14:56	tbaldridge	cYmen: https://github.com/clojure/clojure/blob/master/src/jvm/clojure/lang/Keyword.java it's pretty simple stuff
14:56	`cbp	cYmen: here https://github.com/clojure/clojure/blob/master/src/jvm/clojure/lang/Keyword.java#L24
14:56	`cbp	:-(
14:57	poppingtonic	Hello
14:57	`cbp	hi
14:58	poppingtonic	Is ClojureScript One still relevant? I'm working through it, but the instructions in the Getting Started guide aren't working out...
14:59	poppingtonic	I'm not sure what needs to be fixed
14:59	tbaldridge	poppingtonic: it was replaced by Pedestal App
14:59	tbaldridge	poppingtonic: http://pedestal.io/
15:02	cYmen	tbaldridge, `cbp: thanks
15:07	danneu	tbaldridge: it seems like https://github.com/magomimmo/modern-cljs is more of a Clojurescript One surrogate (a ground-up tutorial)
15:08	tbaldridge	danneu: true, I was just stating that the author of Pedestal App is the same as ClojureScript One.
15:09	TimMc	cYmen: It's more that keyword-as-function and map-as-function are convenient ways to call get.
15:09	tbaldridge	Brenton Ashworth dropped ClojureScript one when he started on the Pedestal App code, so one could be considered the spiritual successor of the other.
15:10	cYmen	TimMc: I totally get map as function but the benefit of allowing to turn the function and the argument around elude me.
15:11	ToxicFrog	I've actually gotten into the habit of using (:key map) and I'm not sure why.
15:11	`cbp	cYmen: so you can get the same attribute from many maps using the function `map` and other such things
15:12	danneu	cYmen: (map :username users)
15:12	justin_smith	also filtering by key (filter :key [{:key false} {} {:key true} {:key 1}])
15:12	cYmen	hm... as opposed to #(% :foo) ?well I mean...why not ..whatever I won't be confused by it any longer! harr!
15:12	justin_smith	or even something like (filter (comp zero? :key) coll)
15:13	justin_smith	that gets all items with :key that is 0
15:13	justin_smith	very handy
15:13	danneu	cYmen: there are some other good cases that are just hard to think of on the spot
15:14	`cbp	like avoiding null pointer exceptions
15:14	lpvb	is there a table somewhere of clojure data structure complexities
15:14	danneu	`cbp: that's a good pt i often take for granted
15:15	danneu	,(:a nil)
15:15	clojurebot	nil
15:20	bja	apparently we can now set! print-fn in cljs. Whenever that support landed was a good day.
15:21	danneu	Is there a good place to manipulate the jvm state like inserting a provider (Security/insertProviderAt (BouncyCastleProvider.) 1)
15:21	danneu	A central place*
15:21	lpvb	,(range \a \z)
15:21	clojurebot	#<ClassCastException java.lang.ClassCastException: java.lang.Character cannot be cast to java.lang.Number>
15:23	lpvb	,(map char (range (int \a) (inc (int \z))))
15:23	clojurebot	(\a \b \c \d \e ...)
15:23	poppingtonic	tbaldridge: ok, pedestal is a great starter
15:24	danneu	lpvb: frankly i like (seq "abcdef...xyz") better
15:27	TimMc	&(map (comp char (partial + 16r61)) (range 26))
15:27	lazybot	⇒ (\a \b \c \d \e \f \g \h \i \j \k \l \m \n \o \p \q \r \s \t \u \v \w \x \y \z)
15:27	TimMc	^ magic numbers like the palindromic 16r61 are just better :-P
15:28	lpvb	,16r61
15:28	clojurebot	97
15:28	TimMc	61 is 97 in base (radix) 16
15:29	lpvb	I know
15:29	lpvb	why not just
15:29	lpvb	,0x61
15:29	clojurebot	97
15:30	TimMc	Because palindromes!
15:30	lpvb	lol
15:30	TimMc	Exactly.
15:30	danneu	isn't it so you can use decimal 26 in the range?
15:30	danneu	oh
15:31	danneu	hope a convo starts so thatll get pushed above the fold.
15:32	TimMc	&(map read-string (for [n (range 10)] (str 1 n \r n 1)))
15:32	lazybot	⇒ (1 12 25 40 57 76 97 120 145 172)
15:32	lpvb	how do I define namespace privte constants?
15:32	lpvb	(def- x 5) doesn't work
15:33	lpvb	(defn- x [] 5) is tedious
15:33	llasram	(def ^:private foo ...)
15:33	TimMc	defn- is just sugar for defn ^:private
15:34	TimMc	(Macro sugar, not reader sugar.)
15:35	TimMc	lpvb: Private vars are overrated; people will argue about this, but you should know that it's still easy to get ahold of private vars from outside of the namespace.
15:35	lpvb	seems like a def- macro should be included by default
15:35	lpvb	unless that isn't idiomatic
15:35	tbaldridge	(inc TimMc)
15:35	lazybot	⇒ 52
15:35	justin_smith	lpvb: information hiding is not very popular with clojure
15:35	llasram	I think people decided that having separate private variations for everything was a bad idea
15:35	tbaldridge	private vars are just wrong, imo
15:35	llasram	I just use ^:private even for defn
15:35	danneu	yeah me too
15:35	llasram	tbaldridge: Er. Why so?
15:36	mdrogalis	There was a good talk by Stu Halloway about why. Can't remember which one.
15:36	tbaldridge	llasram: they make code harder to test, and manage imo. It's much better to put your public API in a different ns. Put all the "private" stuff somewhere else.
15:36	mdrogalis	It was just a slide or two, but he nailed it..
15:36	danneu	:private is mostly helpful for me as the developer since a namespace is otherwise an interleave of "public" functions and helper functions
15:36	TimMc	,@#'loaded-libs
15:36	clojurebot	#<CompilerException java.lang.RuntimeException: Unable to resolve var: loaded-libs in this context, compiling:(NO_SOURCE_PATH:0:0)>
15:37	danneu	Not :private, but the idea of private
15:37	`cbp	(ns cant.touch.this)
15:37	TimMc	,@#'clojure.core/loaded-libs
15:37	clojurebot	#<Ref@1071e40: #{clojure.core.protocols clojure.instant clojure.java.io clojure.repl clojure.string ...}>
15:37	TimMc	,clojure.core/loaded-libs
15:37	clojurebot	#<CompilerException java.lang.IllegalStateException: var: clojure.core/loaded-libs is not public, compiling:(NO_SOURCE_PATH:0:0)>
15:37	danneu	in fact, it's hard to scan a clojure namespace for the intended entry-points.
15:38	lpvb	mdrogalis: can you show me the slides?
15:38	mikerod	mdrogalis: I'd be interested in these slides
15:38	TimMc	lpvb: All it does is control what is exported from the namespace when it is refer'd by another.
15:38	mdrogalis	lpvb: Looking.. Not successful yet.
15:39	mikerod	I've struggled over the idea of no :private vars
15:39	mikerod	hard to say it is frowned upon when it is done in clojure.core though right
15:39	llasram	tbaldridge: I could see that if you're testing private functions. Usually functions I mark as :private I see no value in testing independently
15:39	logic_prog	we have number?
15:40	logic_prog	is there a bool? or boolean?
15:40	`cbp	there is no bool?
15:40	`cbp	ahem, there is no `bool?`
15:40	tbaldridge	llasram: I just don't see a reason to restrict code for no apparent reason. Putting private vars into foo.impl.bar is a much better I idea.
15:40	tbaldridge	llasram: then it's just normal code, why does it have to be special?
15:41	TimMc	mikerod: clojure.core is not a good example -- there's some weird-ass code in there
15:41	mikerod	TimMc: fair enough
15:42	TimMc	It's also supposed to be fairly stable, so minorly bad decisions do not get corrected.
15:42	logic_prog	wtf
15:42	logic_prog	there is no bool?
15:42	mikerod	I have started questioning myself when I want to do a defn-
15:42	logic_prog	how do I check if a varaible is a bool in java?
15:43	logic_prog	maybe insanity like (== x (not (not x)) ?
15:43	`cbp	logic_prog: if by that you mean clojure: (or (false? x) (true? x))
15:43	TimMc	&(== true (not (not 5)))
15:43	lazybot	java.lang.ClassCastException: java.lang.Boolean cannot be cast to java.lang.Number
15:43	mikerod	I start thinking maybe I shouldn't need so many one-off-specific-case private implementation functions
15:43	tbaldridge	logic_prog: perhaps instance?
15:43	TimMc	&(= true (not (not 5)))
15:43	lazybot	⇒ true
15:43	llasram	tbaldridge: I see it mostly as a hint that a function is internal. Putting in a separate namespace gets the same result, but having a separate ns & file seems like a heavier solution programmer-wise than just a :private annotation
15:44	mikerod	private function needed? `letfn` local binding
15:44	tbaldridge	llasram: I hear that, I just hate it whenever I decide to tap into someone's library and find out something is private, because they thought they knew more about it than I did. :-P
15:44	logic_prog	`cbp: nice, thanks
15:44	mdrogalis	Dang, I can't find it. He was making a point that if you only emit X amount of information, you can only perceive, at most X information
15:44	tbaldridge	,(instance? Boolean true)
15:44	TimMc	I tend to use :private if most of my namespace is intended to be public and people should be able to use ns-publics to browse it.
15:44	clojurebot	true
15:44	mdrogalis	And private minimizes X
15:44	tbaldridge	,(instance? Boolean false)
15:44	clojurebot	true
15:44	llasram	tbaldridge: I feel the same way in languages like Java, but I've never had a problem accessing a :private Clojure var when I needed it :-)
15:44	TimMc	I also make dangerous fns private -- ones that back safer interfaces.
15:45	mikerod	private functions are sort of a "code smell" of non-reusable code aren't they? :)
15:45	TimMc	If I have only a couple of API fns in a namespace, I just list them in the namespace's docstring and default to leaving things public.
15:45	mikerod	I guess not if you are using it all over in a lib of some sort, but you don't wnat others to get involved with it
15:46	mdrogalis	There are probably more valuable problems you can spend your brain cycles on that something small like this :)
15:46	mdrogalis	Than*
15:46	noonian	yeah i think they are mainly used to hide things that aren't part of the public api
15:46	TimMc	The problem is that namespaces conflate namespacing and APIs.
15:46	TimMc	APIs/modules
15:48	mikerod	true
15:50	dnolen	tbaldridge: it seems it would be possible to improve core.async source map support w/o waiting on that ticket no? but I guess you would need to change core.async macros to pass around line/col info
15:51	tbaldridge	dnolen: yeah I could use some help on that. I had a patch that passed around line/col info but it doesn't go far enough.
15:51	dnolen	tbaldridge: ok cool is this in a branch somewhere?
15:52	tbaldridge	dnolen: from what I can tell, there's a general issue with how metadata is handled in Clojure. I haven't figured it out yet.
15:52	tbaldridge	dnolen: not yet, I'll try to get it up soon
15:53	dnolen	tbaldridge: I guess I don't see how this could be true since macros seems to ok w/ source maps for the most part?
15:53	dnolen	"seem to be ok"
15:55	bja	has anyone seen: "Uncaught Error: URI file:/robots.txt is invalid for field ppu " out of the browser repl lately?
15:55	bja	I get it when I attempt to clojure.browser.repl/connect
15:55	tbaldridge	dnolen: it's been over a month since I tried out this code, so I could have a bug in it. I'll dig into it again and see what I find.
15:57	dnolen	tbaldridge: I'm happy to contribute here, would give me an excuse to understand the macro stuff
15:58	mdrogalis	tbaldridge: Do you have a general stance on serialization formats like Thrift and Protocol Buffers? e.g. when is it appropriate?
15:59	tbaldridge	mdrogalis: not really. I used Fressian the other day and really liked it.
15:59	tbaldridge	mdrogalis: https://github.com/Datomic/fressian
15:59	mdrogalis	Fressian doesn't really fall into the same category as those two, does it?
16:00	mdrogalis	Its self-describing IIRC?
16:03	tbaldridge	mdrogalis: yeah it is, that's what I default to I guess. I don't have much of a use for contract based serialization.
16:03	mdrogalis	Hm, alright. Thanks.
16:07	llasram	mdrogalis: If you're evaluating thrift, protobufs, etc, I've got Clojure integration for Avro I'm quite happy with: https://github.com/damballa/abracad </pitch>
16:55	danneu	Does anyone have a workflow for frequently running typedclojure type checks?
17:01	rasmusto	danneu: I've heard this is good: https://github.com/typedclojure/lein-typed
17:01	rasmusto	haven't used anything but in-line checks personally, so can't tell you more
17:03	rasmusto	too soon
17:08	danneu	rasmusto: yeah, i'm trying to roll my own emacs hotkey to check the current ns
17:09	jtoy	is there a library that finds the difference between a schema and a db adds those changes to the db?
17:09	jtoy	its not a migrations im looking for, its more auto migration
17:11	justin_smith	jtoy: http://clojuredocs.org/clojure_core/clojure.data/diff would be a start, if you have to make it yourself
17:11	justin_smith	just make schema->map and map->schema functions, and use data/diff to find the differences
17:13	jtoy	justin_smith: do you know of a library that will deal with the lowl levle db stuff ? i also want to write my migration in clojure and not in sql so i can make it db agnotistic
17:13	justin_smith	hmm, if you don't mind a macro-heavy solution there is korma
17:14	jtoy	justin_smith: i use it anyway for writing sql, i'll see if i can continue to use it for this part
17:24	lpvb	,(vector-of :bool true false)
17:24	clojurebot	#<NullPointerException java.lang.NullPointerException>
17:24	lpvb	why is that
17:25	lpvb	oh
17:25	lpvb	nvm needs :boolean
17:33	jtoy	hmm, a lot of work for now
17:41	justin_smith	jtoy: https://github.com/budu/lobos this actually may help you
17:45	justin_smith	jtoy: note it has a db analyzer - you could run that on two dbs, then use data.diff on their tables
17:45	justin_smith	*table-definitions
18:03	tbaldrid_	dnolen: my WIP for source map fixes is in meta-fixes (in the core.async repo)
18:03	clojurebot	You don't have to tell me twice.
18:04	tbaldrid_	dnolen: it should work, but it doesn't for some reason. If you run cljsbuild adv it'll spit out a bunch of info at compile time about the metadata it finds. It looks correct but doesn't seem to make it into the source map
18:08	bitemyapp	(is (= madness?-this-is-sparta blah also)))))
18:08	bitemyapp	descriptive nomenclature? kinda not really.
18:18	akurilin	bitemyapp, Korma defaults to RETURNING for every insert/update, correct?
18:21	bitemyapp	akurilin: I remember changing something about that at some point. might be keys only.
18:22	bitemyapp	akurilin: I dunno, you tell me.
18:22	bitemyapp	I've forgotten most of what I've changed in Korma ^_^
18:22	bitemyapp	specifically I made it so that bulk inserts would "work okay" by not RETURNING
18:22	bitemyapp	what I made the default is not within my memory.
18:24	akurilin	bitemyapp, I remember talking to the PG folks about always using RETURNING even if you don't need the results, and they were revolted by the idea. In practice most of the time you actually benefit from returning that one row you insert/update
18:24	akurilin	for large updates Korma right now returns either the first or the last one, forgot
18:24	akurilin	but essentially only 1 row
18:27	bitemyapp	akurilin: the choices, if I recall were none, keys, and values
18:28	bitemyapp	how those map to what data actually gets returned is a little obtuse.
18:29	bitemyapp	akurilin: you'd benefit most from reading the code, by my estimation.
18:29	bitemyapp	Better than reading tea leaves and guessing.
18:29	akurilin	bitemyapp, I am actively doing that in parallel to this.
18:31	bitemyapp	akurilin: you use korma heavily enough that the familiarity should pay dividends.
18:31	bitemyapp	akurilin: I'm pretty sure you use Korma about 99x more than I do these days.
18:33	bitemyapp	akurilin: in fact, I'm not really sure when I'll use a SQL database at all again in the future.
18:33	akurilin	bitemyapp, I am actively doing that in parallel to this.
18:33	akurilin	oops
18:33	bitemyapp	akurilin: most of my future database use is slated for Datomic, with some falling over to RethinkDB/Cassandra/HBase/etc
18:33	akurilin	I was going to say: it's interesting when maintainers stop being consumers.
18:34	akurilin	Guess it's our job to step up and start tossing some pull requests at you guys.
18:34	bitemyapp	akurilin: I don't think baranosky would be opposed to you contributing to Korma at all if you decided to learn the codebase.
18:34	bitemyapp	akurilin: yeah, exactly.
18:35	bitemyapp	I will say this though, I still think a stripped down "micro-Korma" that wasn't complected with things like connection pooling would be really cool.
18:35	bitemyapp	but that's not really my "problem" right now, so I'm choosing not to be opinionated about it specifically because I'm not a consumer.
18:35	bitemyapp	Alex and I are keeping the lights on and accepting PRs.
18:36	akurilin	bitemyapp, and I appreciate that :)
18:37	akurilin	Would be curious to find out how many people are activeling using this stuff in production
18:37	bitemyapp	quite a few, if I had to guess.
18:37	bitemyapp	not as many as cjj, but they're out there.
18:37	akurilin	Ok, just making sure I'm not survivor man there.
18:37	bitemyapp	not that I've taken a census.
18:38	bitemyapp	HoneySQL is pretty neat, but I'm not convinced it's the right level of abstraction.
18:38	bitemyapp	concatenating predicates is kind of tedious.
18:39	akurilin	Sort of related topic: korma exception handler loves to println stuff. I have my own sql exception handling and logging layer, and extra crap in the terminal is confusing in test runs. Is rebinding out around each korma db call the way to go to /dev/null it?
18:40	bitemyapp	akurilin: the logging in Korma is log4j yo.
18:40	bitemyapp	just configure it.
18:40	bitemyapp	unless I've missed something.
18:40	bitemyapp	I mean, you could do that too, but you don't need to.
18:41	Somelauw	kinda hacky, but i think you can do (define out (StringIO.)) to disable println globally
18:41	akurilin	I'm completely unfamiliar with log4j. Can I specify that I want logging exceptions only for selects for example?
18:45	akurilin	Warnock's dilemma.
18:49	bitemyapp	please don't do what Somelauw just said :P
18:50	bitemyapp	akurilin: I have no idea @ log4j, obviously it would be ideal if you could just dependency inject/AOP the log-fn and handle it yourself.
18:50	bitemyapp	whiiiiich is why I usually just have a log-fn dilly that people can override...
19:01	bitemyapp	another day, another error arising from an unexpected nil creating a nonsensical error when it blows up in my face somewhere unexpected. Delightful.
19:01	bitemyapp	technomancy: ^^ grumble grumble.
19:02	technomancy	bitemyapp: imma write a fuse adapter for github issues in ocaml
19:02	technomancy	I've Decided.
19:02	bitemyapp	technomancy: that's pretty brilliant.
19:03	technomancy	well I haven't done it yet
19:03	bitemyapp	technomancy: grom is still a thing, now that I've written the error handling for Revise grom and Simonides are my next projects.
19:03	technomancy	cools
19:03	bitemyapp	technomancy: I really like the concept in general...you should write it in Haskell instead though :)
19:03	bitemyapp	I need to shanghai noprompt into helping me with Grom since he wanted it too.
19:04	technomancy	maybe if I didn't want it to get written quickly I could use haskell
19:04	bitemyapp	hrm. good point.
19:04	technomancy	but the web UI is buggin' me and I don't want any yaks in the way
19:04	technomancy	context: I just started using noscript for conkeror, and its whitelisting doesn't work on github.
19:05	technomancy	which is mostly a win, but it turns all the icons into stupid unicode-not-found glyphs
19:05	zeebrah	I would like to have an alias λ for core/fn
19:05	zeebrah	is there a way to achieve that?
19:06	technomancy	zeebrah: do it locally in your editor, not on disk
19:06	zeebrah	technomancy: good idea!
19:07	bitemyapp	zeebrah: Haskellers have a pretty well established history of doing stuff like that with the glyph replacement stuff Emacs provides.
19:07	bitemyapp	I forget the proper name for it, #emacs surely knows.
19:08	technomancy	http://marmalade-repo.org/packages/pretty-mode-plus this one maybe
19:09	zeebrah	i like how it is assumed i'm using emacs lol :) Light Table but i'll try to figure it out
19:12	`cbp	emacs live has that by default
19:13	technomancy	it has issues in a few cases
19:13	akurilin	Those of you who worked with PG and JDBC, have you ever noticed the driver throwing a SQLException wrapping a PSQLException rather than directly the latter?
19:13	technomancy	it can affect line-length calculation and in some cases cause wrong indentation
19:13	akurilin	I think the latter case is there most of the time, but I swear I've seen the former once or twice in certain situations
19:16	bitemyapp	awwww yissss. Breaking test case that reproduces my problem.
19:16	akurilin	So the # and the end of a symbol in a macro is there to make sure there isn't accidentally a naming collision once the macro is expanded, avoiding interesting debugging sessions?
19:16	bitemyapp	akurilin: it's a shortcut for gensym.
19:17	akurilin	Oh ok, I didn't know its actual name.
19:17	bitemyapp	,(gensym 'blah-this-fucking-rocks)
19:17	clojurebot	blah-this-fucking-rocks31
19:17	bitemyapp	,(gensym 'blah-this-fucking-rocks)
19:17	clojurebot	blah-this-fucking-rocks60
19:17	bitemyapp	,(gensym 'blah-this-fucking-rocks)
19:17	clojurebot	blah-this-fucking-rocks89
19:17	bitemyapp	,(gensym 'blah-this-fucking-rocks)
19:17	clojurebot	blah-this-fucking-rocks118
19:17	`cbp	no!
19:17	bitemyapp	`cbp: wut
19:17	`cbp	stahp
19:17	bitemyapp	`cbp: sorry :(
19:18	bitemyapp	akurilin: it's to avoid the tedium of managing your gensym aliases yourself.
19:18	bitemyapp	akurilin: which...you had to do in Common Lisp.
19:18	bitemyapp	it sucked.
19:18	bitemyapp	a lot of things sucked. This was one of those things.
19:18	akurilin	bitemyapp, ok but the point was collisions?
19:18	`cbp	unintentional symbol capture
19:18	bitemyapp	you dun wanna shadow the users code.
19:18	bitemyapp	yeah
19:18	akurilin	There you go.
19:19	coventry	It doesn't explicitly check for collisions, though, just updates an internal counter each time you call it.
19:20	akurilin	coventry, I get it, thanks.
19:20	akurilin	Weird, the docs say it should append _number
19:20	akurilin	but there's no _ either here or in my repl
19:21	bitemyapp	akurilin: # appends _ + gensym.
19:21	akurilin	bitemyapp, ah, got it now.
19:24	bitemyapp	`cbp: DotA2 tonight?
19:25	bitemyapp	arrdem: ^^
19:29	TEttinger	Defense of the Ancients 3: The Ancients are in a nursing home for senior citizens. Defend them by giving them their applesauce and meds 3 times a day.
19:29	swarthy	ahhaha
19:32	bitemyapp	TEttinger: hahahaha, that's pretty good :)
19:34	akurilin	bitemyapp, I saw the exec-mode dynamic var approach Korma does to determine how to execute the queries. Is that a pretty good pattern to follow as opposed to passing options to functions to customize their behavior?
19:34	bitemyapp	akurilin: no no no no no
19:34	bitemyapp	akurilin: Korma isn't a good example of many things.
19:34	bitemyapp	The thing it got right was passing around query maps, but even that was done in a brittle fashion.
19:34	`cbp	bitemyapp: sure
19:35	akurilin	(how to learn Clojure: get intimately familiar with other people's code, then forget everything they did)
19:35	bitemyapp	akurilin: if you want to see good code, Revise is a good place to look.
19:35	bitemyapp	there's some mess due to young the codebase is, but `cbp did a damn good job.
19:36	bitemyapp	the patterns and design are solid, there's no global/dynamic bullshit.
19:36	akurilin	bitemyapp, I think the one advantage that approach has is that you can group an unlimited amount of queries under the binding of that dynamic var, without having to do the pure approach for every call.
19:36	bitemyapp	you pass a connection to execute a query. no global singletons.
19:36	bitemyapp	akurilin: yuo can solve that problem just with closures...don't make my eyes roll that hard, it hurts :(
19:37	bitemyapp	there are times when dynamic vars are called for, that wasn't one of them.
19:37	akurilin	Actually the same pattern was there in noir validation with a dynamic var accumulator of validation errors.
19:37	swarthy	bitemyapp: I'm about to grab a smoking jacket and some other people's code for reading. What do you think are good clojure libs to learn from?
19:37	bitemyapp	akurilin: that was a bad fucking idea too.
19:37	akurilin	bitemyapp, I think it's just the author's style.
19:37	bitemyapp	akurilin: well it's the same author.
19:38	bitemyapp	Regardless, it's a bad fuckin' idea.
19:38	akurilin	bitemyapp, yes, my point.
19:38	akurilin	bitemyapp, fair enough, point taken.
19:39	bitemyapp	swarthy: https://github.com/weavejester/hiccup https://github.com/weavejester/environ github.com/bitemyapp/revise/ github.com/bitemyapp/bulwark/ github.com/bitemyapp/blackwater/
19:39	bitemyapp	swarthy: if you want scary code: https://github.com/bitemyapp/brambling
19:39	bitemyapp	swarthy: https://github.com/dakrone/clj-http/
19:39	akurilin	bitemyapp, can you expand a bit on the closure pattern?
19:39	bitemyapp	swarthy: https://github.com/dakrone/cheshire
19:40	swarthy	bitemyapp: Thanks a bunch! I've used a few of those while learning! Can't wait to dig much deeper.
19:40	bitemyapp	akurilin: https://github.com/bitemyapp/bulwark/blob/master/src/bulwark/core.clj#L88-L91
19:41	bitemyapp	swarthy: heavily optimized and kin da razy but still interesting: https://github.com/aphyr/riemann
19:41	bitemyapp	riemann is NOT representative of typical Clojure code (this is true for a lot of aphyr and ztellman's stuff)
19:41	akurilin	bitemyapp, so you're basically being pretty pure there, passing config values around downstream?
19:41	bitemyapp	questionable libraries are the order of the day for those blokes.
19:41	bitemyapp	akurilin: pure'ish, there's an atom, but the closure takes precedence.
19:41	bitemyapp	also the atom isn't dynamic, it's global.
19:42	akurilin	bitemyapp, to avoid magic?
19:42	swarthy	bitemyapp: questionable libraries are the order of the day for those blokes. Care to elaborate?
19:43	bitemyapp	akurilin: to keep it async safe.
19:43	bitemyapp	akurilin: long story.
19:43	coventry	Is there ever any reason to use dorun as opposed to doseq?
19:43	bitemyapp	swarthy: long story, and I'm about to go eat with a friend.
19:43	swarthy	np I'll be idling here plenty I'll ask again later
19:43	bitemyapp	swarthy: suffice it to say, Mr. Tellman and Mr. Kingsbury have very different needs from the rest of us mere mortals.
19:43	akurilin	bitemyapp, mind if I dump that list of libraries you recommended to my tumblr for future reference?
19:43	swarthy	ah cool!
19:43	bitemyapp	swarthy: big distributed systems high performance stuff blah blah bvalh
19:43	bitemyapp	akurilin: go for it.
19:43	amalloy	coventry: of course. if you already have a lazy sequence that contains side effects, and you want to run them now, rather than run new side effects based on an existing sequence
19:43	swarthy	that says enough thanks!
19:44	amalloy	or it can be easier to write (dorun (map foo! xs)) than (doseq [x xs] (foo! x))
19:44	coventry	amalloy: Oh, ofcourse. Thanks.
19:49	coventry	ztellman's sleight and riddley libraries touch clojure's guts in slightly icky ways, but for totally clear and sensible reasons. Reading them with those reasons in mind won't leave any harmful impressions.
19:59	seangrove	I have a string: "2012-10-07T20:23:56.000+0000", and I would like a timestamp. Can't quite figure out the straightforward way to do this
20:00	AimHere	If there's nothing inbuilt, you could split it up into bits with re-match and then look at the individual fields
20:00	amalloy	AimHere: that way lie madness and shattered dreams
20:00	AimHere	Oh ye of little faith
20:01	seangrove	AimHere: Going to have to side with amalloy there
20:01	swarthy	seancorfield: I had to mess with dates recently and I ended up using this: https://github.com/seancorfield/clj-time
20:01	amalloy	seangrove: http://docs.oracle.com/javase/7/docs/api/java/text/SimpleDateFormat.html
20:01	amalloy	there may be a wrapper in clj-time, but SDF isn't that complicated to use
20:01	swarthy	swarthy: I also did what amalloy mentions if you won't need to mess with it much.
20:02	seangrove	I'll check out SDF
20:02	akurilin	clj time or underlying joda should be able to take care of all of those concerns
20:02	akurilin	unless it's some odd edge case
20:02	seangrove	Thanks amalloy and swarthy, and of course AimHere
20:02	seangrove	akurilin: No, just interop with SFDC, nothing too bad
20:02	AimHere	So much for adding on the regex problem to all your other problems!
20:04	swarthy	seangrove: if you haven't already seen it I use clojure-toolbox.com to find libraries for various things
20:04	akurilin	The full list of Joda's out-of-the-box formats is: http://joda-time.sourceforge.net/apidocs/org/joda/time/format/ISODateTimeFormat.html
20:04	akurilin	Wrapped by this guy: https://github.com/clj-time/clj-time/blob/master/src/clj_time/format.clj#L82
20:08	jtoy	if I have 2 arrays like [1 2] and [:foo :bar] jow can I join them so i get {:foo 1 :bar 2} ?
20:08	justin_smith	jtoy: zipmap
20:08	justin_smith	,(zipmap [:foo :bar] [1 2])
20:08	clojurebot	{:bar 2, :foo 1}
20:09	jtoy	sweet
20:11	brehaut	jtoy: to generalize into and map are also useful. zipmap ~== (fn [a b] (into {} (map vector a b)))
20:12	brehaut	jtoy: which is to say, map is a general purpose zipping function for stitching n sequences together, and into lets you pour a sequence into a collection type
20:12	jtoy	brehaut: I should use that more but that confuses me more often
20:13	brehaut	whats confusing about it?
20:13	justin_smith	,(apply assoc {} (interleave [:foo :bar] [1 2]))
20:13	clojurebot	{:bar 2, :foo 1}
20:13	justin_smith	another option :)
20:15	jtoy	brehaut: i dont know why i would do map vector
20:15	jtoy	justin_smith: ye, even more confusing
20:16	brehaut	jtoy: ok so map takes a function of N arguments and N sequences
20:16	brehaut	jtoy: it takes items at the same index from all the sequences and passes tehm to the function
20:16	amalloy	justin_smith: any time you apply assoc, you've probably made a mistake. into is just better for that kind of thing
20:16	brehaut	jtoy: vector takes N arguments and creates a new vector from it.
20:17	brehaut	,(map vector [:a :b] [1 2]) ;; jtoy
20:17	clojurebot	([:a 1] [:b 2])
20:17	brehaut	jtoy: thats the interstitial datastructure we need to use into to fill a map
20:17	brehaut	jtoy: conj on a map takes a key value pair as a vector
20:17	brehaut	jtoy: and into uses conj from memory
20:19	brehaut	,(map + [1 2 3] [4 5 6]) ;; jtoy this zips together two sequences by adding the values at each index, as another example of map as zip
20:19	clojurebot	(5 7 9)
20:25	jtoy	brehaut: ah I see, I havent used multiple lists with map before
20:35	zeebrah	When i enable clojure-mode in a cider repl, i can no longer evaluate any expression in the repl. Any way to solve this?
20:36	coventry	zeebrah: Use nrepl instead. The cider package is not stable yet.
20:36	justin_smith	don't turn on clojure mode?
20:36	zeebrah	justicefries: I don't get the font locking then
20:37	justin_smith	clojure mode and the repl mode are both major modes, the one replaces the other
20:37	coventry	zeebrah: Oh, you were turning on clojure mode in the repl buffer? It's not meant to be used that way.
20:37	zeebrah	coventry: yep
20:49	amalloy	brehaut: into actually uses conj! and transients, by the way
20:49	amalloy	not observably different, of course
20:49	brehaut	amalloy: yeah i figured, but that can be brushed under the carpet
20:51	brehaut	not observable without a timer anyway
20:59	ddellacosta	if you were going to re-write some Clojure code which uses refs in CLJS, generally speaking, are atoms the most appropriate (only) answer?
21:00	ddellacosta	I guess the big distinction is between "coordinated" vs. "independent," huh
21:02	amalloy	ddellacosta: since there are no threads in cljs, atoms are just as good as refs: no swap! ever retries
21:03	amalloy	you have to make sure you don't split up a dosync block onto different event handlers, or whatever js calls them, but otherwise it's pretty hard to go wrong
21:04	bitemyapp	ddellacosta: what did you need the coordination refs offer originally?
21:05	ddellacosta	amalloy, thanks, that's helpful.
21:06	ddellacosta	bitemyapp: I'm just playing around now playing with FSMs
21:06	ddellacosta	bitemyapp: I was converting this code to work in CLJS: http://nakkaya.com/2010/06/22/finite-state-machine-implementation-in-clojure/
21:06	ddellacosta	bitemyapp: now I'm reviewing how core.async works, since I understand that uses state-machines under the covers
21:07	bitemyapp	ddellacosta: FSMs are very useful when they fit your problem.
21:08	ddellacosta	bitemyapp: yep. Specifically I'm managing a validation system which validates as you move from field-to-field, so it seemed like a good approach. And we have a help system we will need to update later on, so I thought I could kill two birds with one stone.
21:08	ddellacosta	*managing -> re-factoring
21:08	bitemyapp	ddellacosta: I'm not convinced that code needed a ref though. They're just doing flat ref-sets.
21:08	bitemyapp	unless you're THAT worried about stale reads.
21:08	bitemyapp	that's pretty paranoid though.
21:09	ddellacosta	bitemyapp: yeah, I wasn't quite sure why that code did what it did, although I suppose if it gets accessed in a multi-threaded context it could be a problem, no?
21:09	bitemyapp	ddellacosta: for contrast, 99% of SQL database access is generating writes that use stale reads that are milliseconds, if not whole seconds, old. A stale read with an atom will be microseconds old and involve only one process whereas my database example might involve hundreds of readers and writers.
21:09	amalloy	bitemyapp: well, it's over three years old - atoms might not have existed back then?
21:09	bitemyapp	ddellacosta: if the SQL database scenario isn't biting 99% of users, it's extremely unlikely an atom would.
21:10	ddellacosta	amalloy: good point
21:10	bitemyapp	amalloy: aha, good call. I hadn't even noticed.
21:10	bitemyapp	I wasn't critiquing, just explaining why I don't think stale reads are a big deal most of the time.
21:10	amalloy	i think atoms were around then, but i'm not really sure
21:11	bitemyapp	Narullah wrote a lot of good stuff about Clojure.
21:11	ddellacosta	yeah, really love his blog
21:11	bitemyapp	I need to update my blog with an updated projects list.
21:11	bitemyapp	I've added a lot since I last updated it.
21:11	logic_prog	is there any fringe implementation of clojure on go rather than java?
21:11	nightfly	javaSCRIPT
21:12	nightfly	cljs
21:12	bitemyapp	logic_prog: you'd be better served by explaining your needs.
21:12	bitemyapp	nightfly: cljs isn't fringe.
21:12	logic_prog	purely intellectual curiosity
21:12	nightfly	and feels ashamed for a word earlier
21:12	logic_prog	to see if someone has a implmentation of clojure that runs on Go (the langauge) rather than the JVM
21:12	bitemyapp	logic_prog: no, and it's unlikely to happen.
21:12	logic_prog	nightfly: lol, was wondering how cljs / javascript fit in
21:13	ddellacosta	oh it's almost the conj isn't it
21:14	bitemyapp	ddellacosta: hey I'm stuck in SF, so we can suffer together (apart)
21:17	lpvb	(defn destructuring [{x :x y :y}] ...)
21:17	lpvb	the problem with ^ is that I lost the name of what's being passed in
21:18	lpvb	which is a reminder for what should be passed in
21:19	lpvb	or am I being pedantic?
21:19	technomancy	logic_prog: the go runtime has no dynamic loading capabilities at all; it's completely unsuitable for interactive development
21:20	technomancy	(which by extension makes it unsuitable for basically any good development, because good development is interactive)
21:20	xeqi	lpvb: you can name the arg with :as. {:keys [x y] :as point}
21:21	lpvb	xeqi: thanks :)
21:25	coventry	Are there any projects I should look at for examples of extending nrepl.el? Aside from nrepl-discover?
21:25	technomancy	coventry: lein.el maybe; but that's kinda stalled
21:26	technomancy	probably ritz
21:26	brehaut	ddellacosta: what distant clime are you in?
21:26	coventry	technomancy: Thanks.
21:27	technomancy	coventry: what's your plan?
21:30	coventry	technomancy: Well, I was trying to use nrepl-discover, but I need more logic on the elisp side, so I cargo-culted the output of nrepl-discover-command-for and hacked it up a bit (https://www.refheap.com/20753). I'm getting complaints that (".../tst.clj" 1 198) is not a string, which is the result of the (cons fn defun-region) form in that snippet.
21:32	coventry	technomancy: What I'm really trying to do is write an emacs function which will take the current region, get all the top-level forms which intersect with it, pass that to clojure, and instrument all forms in the region with clojure.tools.trace/trace.
21:32	coventry	technomancy: I have the clojure side of this written. Just need the emacs integration now.
21:33	coventry	technomancy: Currently, I'm looking for examples or instructions on how to pass a non-string argument through nrepl-send-op.
21:34	arrdem	bitemyapp, `cbp : eeeeeh not really Q_Q
21:36	technomancy	coventry: huh; that should be doable without any elisp if nrepl-discover is working properly
21:37	ddellacosta	brehaut: I am in Tokyo
21:37	technomancy	but I haven't tested some of the arg types yet
21:37	ddellacosta	wishing I could get to conferences more easily. :-(
21:38	brehaut	ddellacosta: yup totally. its mighty expensive to fly to the US
21:39	ddellacosta	brehaut: I'm thinking I'll try for Clojure West next year though
21:39	coventry	technomancy: nrepl-discover is working to the extent that nrepl-discover-command-for is being run with the nrepl/op metadata I put on the clojure fn, but I don't see a way to get the buffer-substring through to the clojure side within the nrepl-discover framework, at least not without saving the buffer to disk first.
21:40	technomancy	coventry: oh... do you have to send the substring? can't you send the file/positions and do reading server-side
21:40	technomancy	as long as you force a save before it's invoked
21:40	technomancy	(you should compile anyway, so that's probably fine)
21:42	coventry	technomancy: Yeah, I'd rather not have to force a save. The idea is to go in and just change the vars affected by the given region, like C-M-x rather than C-c C-k.
21:43	coventry	If I have to force a save, I have to ask the user whether it's OK, which slows things down in my experience.
21:44	coventry	But my more immediate problem is that sending a list as an argument in nrepl-send-op doesn't seem to work. Normally I'd assume it's my fault, but I'm a little more agnostic here since I'm ape-ing untested code.
21:44	technomancy	hm; maybe adding region-contents as an arg type is a thing we should do
21:45	technomancy	I gotta take off though; feel free to open an issue
21:45	coventry	technomancy: region-contents and nrepl-region-for-expression-at-point contents would work here, yeah.
21:46	coventry	Yeah, I'll open an issue, but may take a bit for me to get to it.
22:11	lpvb	,(-> 1 (fn [x] (+ 3 x)))
22:11	clojurebot	#<IllegalArgumentException java.lang.IllegalArgumentException: Parameter declaration 1 should be a vector>
22:11	lpvb	can someone explain why this doesn't work? I don't think I'm understanding ->
22:12	amalloy	,(macroexpand-1 '(-> 1 (fn [x] (+ 3 x))))
22:12	clojurebot	#<IllegalArgumentException java.lang.IllegalArgumentException: Parameter declaration 1 should be a vector>
22:12	amalloy	&(macroexpand-1 '(-> 1 (fn [x] (+ 3 x))))
22:12	lazybot	⇒ (fn 1 [x] (+ 3 x))
22:12	beppu	,(-> [1] (fn [x] (+ 3 x)))
22:12	clojurebot	#<Exception java.lang.Exception: Unsupported binding form: 1>
22:13	amalloy	that's a rather unfortunate fallout of the change to fix how -> nests, gfredericks (or was it TimMc who noticed the problem?)
22:14	lpvb	so how do I accomplish what I intended?
22:14	amalloy	(+ 3 1)
22:14	clojurebot	4
22:15	beppu	,(-> 1 (+ 3))
22:15	clojurebot	4
22:15	amalloy	actually i don't understand at all why clojurebot fails to macroexpand that
22:15	bitemyapp	,(-> [1] first (+ 3))
22:15	clojurebot	4
22:15	bitemyapp	arrdem: DotA2?
22:15	lpvb	,(-> {:x 1} (assoc :x 5))
22:15	clojurebot	{:x 5}
22:16	bitemyapp	lpvb: macros operate on code structure, not data.
22:17	bitemyapp	lpvb: (-> [1] (fn [x] (+ 3 x))) turns into (fn [1] [x] (+ 3 x))
22:17	lpvb	I defined a function (defn setm [m k f v] (assoc m k (f (get m v))))
22:18	lpvb	and (-> {:x 1 :y 2} (setm :x (fn [x] (+3 x)) :y))
22:18	lpvb	does the same error
22:18	lpvb	shouldn't the macro inject that map after setm
22:20	arrdem	bitemyapp: neg. homework.
22:20	lpvb	oh wait it is working
22:21	lpvb	thanks guys
22:22	amalloy	,(macroexpand-1 '(-> 1 (fn [x] (+ 3 x)))) ;; seriously i'm baffled as to why this no longer returns (fn 1 [x] (+ 3 x)). any ideas, guys? a clojurebot issue somehow?
22:22	clojurebot	#<IllegalArgumentException java.lang.IllegalArgumentException: Parameter declaration 1 should be a vector>
22:24	`cbp	bitemyapp: wanna play?
22:24	`cbp	amalloy: it does return that?
22:24	amalloy	yeah, it must be clojurebot. hiredman, any idea why clojurebot now expands macros even if they're quoted? eg, ,'(with-open x) throws an exception
22:25	`cbp	oh but it shouldnt evaluate
22:26	`cbp	nevermind me :)
22:26	amalloy	right, `cbp. i'm guessing clojurebot does a dumb clojure.walk/macroexpand-all or something
22:27	coventry	amalloy: There's no macroexpand-all in clojurebot master.
22:28	amalloy	coventry: looks like https://github.com/hiredman/clojurebot/blob/master/clojurebot-eval/src/clojurebot/sandbox.clj#L147 to me, a hand-rollwed reimplementation of macroexpand-all
22:29	coventry	Yep.
22:32	akurilin	Does anybody know of any neat tools out there that can tell you what the "state" was when an exception happened? As in, I'm thinking of a mix of clojure.tools.trace and an exception handler
22:33	akurilin	I can write my own, but I'm suspecting this must have been done before.
22:33	akurilin	So something like: function f blows up with exception e, I want to know what the input was.
22:43	amalloy	akurilin: slingshot's throw+ automatically includes the in-scope locals into the exception
22:44	akurilin	amalloy, oh interesting, never heard of slingshot.
22:44	akurilin	amalloy, strongly recommended?
22:44	amalloy	meh
22:44	akurilin	lol...
22:45	amalloy	although, now i think of it, shouldn't that cause head retention? like, (defn last* [xs] (try (last xs) (catch Exception e (throw+ "broken")))) ought to prevent xs from being GCed, since the throw+ clause will attempt to include xs in the ex-info, right? that sounds bad
22:49	akurilin	Can you actually list locals in Clojure without having to wrap anything?
22:53	justin_smith	akurilin: here is code that displays locals in another clojure process: https://github.com/GeorgeJahad/cdt/blob/master/src/cdt/reval.clj
22:54	justin_smith	as you could imagine, it is a very low level thing
22:54	akurilin	justin_smith, yeah, definitely looks like serious business.
22:54	akurilin	thx
22:55	akurilin	Man, the stuff you can do.
22:55	justin_smith	akurilin: here is a webapp that connects to another clojure process, and when the remote gets an exception, it lets you examine the stack frame and eval in context: https://github.com/prismofeverything/schmetterling
22:56	justin_smith	uses clojurescript and websockets
23:00	echo-area	`defmacro' doesn't work inside (binding [warn-on-reflection true] ...).
23:01	amalloy	echo-area: (binding [warn-on-reflection true] ...) doesn't do anything, so that's not a super-compelling complaint
23:01	echo-area	amalloy: What I wanted to say is that, is it because the context is changed?
23:01	amalloy	anyway, defmacro needs to be a top-level form, since each top-level-form is compiled as a whole; warn-on-reflection is a red herring
23:03	bitemyapp	`cbp: grabbing a drink with a buddy, then we can.
23:03	bitemyapp	`cbp: I'll be back soon.
23:04	akurilin	justin_smith, wow that's awesome, thanks for sharing. What is the advantage of attaching a process to it rather than just dumping the entire context somewhere in a log?
23:04	akurilin	justin_smith, is there some kind of killer app of evaling within the context?
23:05	echo-area	amalloy: Why does that binding form not do anything?
23:05	amalloy	because it sets warn-on-reflection at runtime, and warn-on-reflection is only meaningful at compile time
23:08	echo-area	So is it not idiomatic to enclose def's inside such a form? I thought that way the enclosed forms are then compiled with warn-on-reflection being true
23:10	amalloy	that is simply not the case. the entire form is compiled, and then *warn-on-reflection is bound to true, and then the entire form is eval'd
23:11	echo-area	Oh right, it should be set before loading
23:12	justin_smith	akurilin: seeing exactly what is going on in each context
23:12	justin_smith	at the exact moment that the exception occured
23:12	justin_smith	ie. not only having a log of the things you intended to record, but being able to interactively query status of things you had not thought of beforehand
23:12	akurilin	justin_smith, any way you can just dump that to disk and reload it later?
23:13	justin_smith	I think there is a way to dump and reload a vm
23:13	justin_smith	that would be a java thing I guess
23:13	echo-area	amalloy: Is it correct that during loading .clj files the forms are compiled one by one, rather than entirely?
23:13	justin_smith	I think the problem would be that a bunch of things in the dump would refer to stale resources
23:14	akurilin	justin_smith, my issues are pretty simple for now, but I'm sure that with size I'll need to look into more adult debugging practices
23:14	justin_smith	ie. some values on a bytecode level refer to open files or streams, or process ids of other processes
23:14	amalloy	yes
23:14	justin_smith	akurilin: our use case is sometimes you just get that one bug you cannot reproduce and you just wish you could have had a repl in the context where the answer happened
23:14	echo-area	I see. Thanks
23:15	justin_smith	akurilin: complementary to that app I also made a request recorder replayer ring middleware
23:15	akurilin	justin_smith, yeah, I hear you. Right now my solution is: dump the prod db, load on my machine, repro, but that doesn't scale very well with # of external services and size of db :(
23:16	justin_smith	which you may find useful if you are using ring, and want to be able to test updated code with a request that is known to have triggered a bug
23:16	justin_smith	akurilin: also the repro can be odd
23:17	justin_smith	one particular bug I remember: we were using the default name for the cookie "ring-session", and one of our frontend devs had tried some friend's buggy webapp that created a broken localhost ring-session cookie
23:17	justin_smith	so the bug was only reproducible on his machine
23:17	justin_smith	and it wasn't even us, it was a corrupted cookie causing an exception
23:17	justin_smith	well I guess we learned to handle that better :)
23:17	akurilin	justin_smith, ah damn, must have been a pain
23:18	akurilin	I'm also a ring-session cookie kind of guy heh :)
23:18	justin_smith	classic "error only happens for the one guy who can't read the stack trace"
23:21	akurilin	justin_smith, a lot of things get much easier once you can see the full stack trace, huh?
23:21	justin_smith	well, its a first step
23:22	akurilin	On an unrelated note: how do I replace a fn in a library I don't own? Bitemyapp was saying something about AOT
23:22	akurilin	but I haven't found anything useful on Google \
23:23	`cbp	akurilin: alter-var-root
23:23	justin_smith	I don't know how aot plays into that
23:24	justin_smith	I have seen someone actually run their whole app inside a with-redefs foorm
23:24	justin_smith	lol
23:25	akurilin	`cbp, the idea is that I do that once at boot and it'll be set for the rest of the app's lifetime?
23:25	`cbp	akurilin: yeah
23:25	`cbp	unless theres an evil alter-var-root somewhere in that lib i guess
23:25	akurilin	so do people end up with a "do stuff at boot" function ?
23:26	justin_smith	akurilin: yeah, with ring you can register an :init action
23:26	justin_smith	that way you can make sure you have no stateful events other than var bindings at the top level of any of your code
23:26	justin_smith	just def the init in each ns, and call all of them from the main ring init
23:27	justin_smith	that way reloading source doesn't have weird side effects
23:27	justin_smith	of course without ring you have -main
23:27	akurilin	justin_smith, what does it mean to def the init in each ns?
23:28	justin_smith	it is a convention that is useful
23:28	justin_smith	in each ns you make, if the code needs any initialization, you (defn init [] (initialize this namespace... ))
23:29	justin_smith	then you call each of these functions from the function you register with ring as your :init
23:30	akurilin	justin_smith, what if it's some initialization that multiple namespaces need, you still put it up there?
23:30	akurilin	As long as it's idempotent I guess.
23:30	justin_smith	you define it in each ns
23:30	justin_smith	then call them from the init function
23:31	akurilin	justin_smith, so the initialized thing is defined once per namespace this way?
23:31	justin_smith	well for the ones that need it, each gets their own
23:31	justin_smith	really most namespaces don't need an init function
23:32	justin_smith	just the ones that do things like load config files at boot
23:32	justin_smith	or set up a db connection pool
23:32	justin_smith	or create a pool of worker tasks
23:32	justin_smith	that kind of stateful startup stuff
23:32	akurilin	Yeah I have stuff like that and logging levels
23:34	akurilin	justin_smith, well thanks for sharing, that's interesting
23:35	justin_smith	no problem, hope it proves helpful
23:35	akurilin	Can I even alter-var-root private vars?
23:36	akurilin	Guess I'm about to fin dout
23:36	Raynes	akurilin: Yes.
23:37	Raynes	Privacy is more of a "Yo, this person thought you should probably not use this." instead of a "GET ON THE GROUND HANDS BEHIND YOUR BACK HANDS BEHIND YOUR BACK DROP THE WEAPON."
23:40	akurilin	Raynes, well it does seem to be making things a bit trickier. For example, this fn I'm replacing calls another private inside of it, so it's throwing an exception when I try to alter var root it
23:41	justin_smith	akurilin: (in-ns target-ns)
23:42	akurilin	justin_smith, that seems to be stateful, can I do the same with rebinding ns?
23:42	akurilin	or at least I only ever used in-ns from a repl
23:43	justin_smith	hmm
23:43	justin_smith	yeah, that could be tricky I guess
23:44	justin_smith	akurilin: http://clojure.org/namespaces
23:44	justin_smith	"The current namespace, ns can and should be set only with a call to in-ns or the ns macro"
23:44	Raynes	If it throws another private inside it then call the var and not the function, I guess.
23:45	Raynes	&(#'+ 3 3)
23:45	lazybot	⇒ 6
23:45	justin_smith	akurilin: you could always just fork the library instead of monkey patching it
23:45	justin_smith	it may be easier
23:46	akurilin	justin_smith, that's fair, I need to get used to doing that anyway.
23:46	akurilin	supposedly pushing stuff to clojars is pretty braindead these days
23:46	justin_smith	very easy
23:48	justin_smith	the nice thing to do is if you actually fixed a bug or extended functionality you can issue a pull request
23:48	akurilin	justin_smith, that's fair. In my case it's more of a flavor thing which I don't want to impose on other folks who might have their own.
23:49	justin_smith	in that case, you may be yak shaving ;)
23:50	akurilin	well, korma by convention defaults to FKs being named %table-name%_id, which doesn't work if your naming convention is plural for tables and singular for FKs
23:50	akurilin	so I'm overriding their fn to compute that to support that case
23:51	akurilin	and folks name their stuff however they want
23:51	akurilin	I guess the better option would be to let people define their own table name to fk function
23:51	akurilin	with a default
23:52	akurilin	so maybe that'd be a PR
23:53	akurilin	Welp, there's an inflections library, there you go :D