#clojure – 2015-04-10

2015-04-10

03:12	magnars	Turning on source-maps in my clojurescript profile, makes it basically just hang there for several minutes now - could it be taking so long? (I am including core.async) Any idea what could make it hang otherwise?
03:34	Guest72028	is the excuse , we have too many change requests valid for not using unit tests and version control?
03:35	magnars	there is no excuse to not use version control
03:35	Guest72028	:-), ok and the unit tests?
03:35	Guest72028	the project is very big btw
03:35	Guest72028	a payroll application
03:38	magnars	automated testing is a huge advantage when done right, but it might be prohibitively expensive to shoehorn into an existing huge codebase
03:39	magnars	but too many change requests sounds to me like a rather weak argument
03:39	Guest72028	oh, ok, i see
03:39	Guest72028	what if they implement unit tests gradually?
03:41	magnars	one good way of introducing unit tests is to write tests when reproducing bugs, and them keeping them around to avoid those known bugs to creep up again - that is probably easiest done with some sort of feature tests tho.
03:43	Guest72028	thanks, that seems like a clever idea
03:43	sveri	And adding tests for new features obv.
03:45	TEttinger	gfredericks made a pretty amazing regex tester for use in clojure tests
03:46	TEttinger	https://github.com/gfredericks/test.chuck/blob/master/src/com/gfredericks/test/chuck/regexes.clj#L501
03:46	Guest72028	part of the frustration i guess is that management want a feature and they want it like yesterday
03:47	TEttinger	if you want to make sure your input validation actually only lets in valid input, that tester util may be especially handy
03:48	oddcully	Guest72028: test are written to make change/refactoring easy. if every little change breaks half your testbase, then there are problems with the architecture
03:48	magnars	Guest72028: the only way to go fast is to go clean
03:49	Guest72028	ah, ok, so you have to tell the requester be reasonable, you can't build a house in 1 day
03:49	oddcully	Guest72028: you can't have a kid in 1 month with 9 wifes
03:49	Guest72028	:-)
03:50	magnars	software development is a marathon, not a sprint
03:50	Guest72028	i guess im the only one here, using git with feature branches running unit tests after new feature then merging back to develop branch
03:51	Guest72028	i mean only one in my company
03:51	oddcully	also i bet the most expensive part of testing right now are the poor sods that test the app before release by hand (hopefully be protocol); start with automated tests there. granted it can be hard, but at least it tests the stuff that faces customer
03:51	slipset	oddcully: wouln't mind trying though
03:52	oddcully	haha
03:52	Guest72028	oddcully: they're the typical cowboy coder, the users tests the features ,hehe
03:52	magnars	slipset: sounds exhausting ;)
03:52	magnars	if your users are doing all the testing, and they are still your customers, that might be a deliberate strategy that is currently working ;-)
03:53	Guest72028	the users have no choice, the software is enforced by the company
03:53	oddcully	sure. and management likes it because the devs can bring up new stuff every minute
03:54	Guest72028	yip
03:54	Guest72028	oh, looky here i have this button when i press it features magically appear
03:55	oddcully	and unless they are eastwook-level cowboys this pile of next-age-spaghetti will explode in at least a year
03:55	oddcully	wook... s/wook/wood/
03:55	Guest72028	strange enough, i got a hold of the source code, they don't do oop in c++, theyre using c structs.. etc (atoi), software been running for 10 years
03:56	Guest72028	no inheretance, they use if statements for that
03:56	oddcully	inheritance is the root of all evil anyway
03:56	Guest72028	ok, so composition better?
03:57	magnars	separating data and functions is better :)
03:57	oddcully	(inc magnars)
03:57	lazybot	⇒ 4
03:58	Guest72028	magnars: do you have an example?
04:01	magnars	Guest72028: wish I could come up with a brilliant example, but it is one of the core tenets of Clojure, in that "It is better to have 100 functions operate on one data structure than to have 10 functions operate on 10 data structures." - Alan J. Perlis
04:02	Guest72028	maybe i'm strange but i prefer the functional style of code
04:03	TEttinger	clojure is rather strongly opposed to defining new data structures when existing ones can be used, it's interesting
04:04	TEttinger	one of the few data structures that you do sorta need an external lib for are insertion-ordered collections.
04:04	TEttinger	$google clojure ordered collections
04:04	lazybot	[Data Structures - Clojure] http://clojure.org/data_structures
04:04	TEttinger	hm
04:04	TEttinger	$google clojure ordered collections flatland
04:04	lazybot	[amalloy/ordered · GitHub] https://github.com/amalloy/ordered
04:04	TEttinger	there we go
04:04	Guest72028	recently got interested with data science, incanter seems to be great tool for this
04:05	Guest72028	sorry for the bad grammar, my english can improve (2nd language)
04:06	TEttinger	I wouldn't have guessed, your grammar seems fine
04:08	Guest72028	ok, cool. feel free to correct if needed , hehe
04:09	Guest72028	i think i'd enjoy doing clojure programming as a job, don't see a market for it in our country at the moment
04:22	Guest72028	magnars: why is inheritence evil, just curious
04:23	Guest72028	ok, i see overcomplication of things
04:24	Guest72028	http://learnpythonthehardway.org/book/ex44.html
04:25	Guest72028	if clojure uses the jvm, is the limits of concurrency due to underlaying jvm, could it be as highly concurrent as say erlang?
04:27	slipset	Guest72028:what
04:27	slipset	is your country?
04:27	Guest72028	south africa
04:28	slipset	so, no clojure in south africa either :(
04:31	Guest72028	i found 3 jobs on a job searching app
04:31	dstockton	slipset: there's clojure in ZA
04:32	dstockton	https://github.com/ptaoussanis wrote a load of popular libraries, is South African
04:34	oddcully	Guest72028: if inheritance is used to save you typing (as a cheap macro system) and not for reflecting the domain, this builds up bad code
04:34	oddcully	Guest72028: the quote was from http://channel9.msdn.com/Events/GoingNative/2013/Inheritance-Is-The-Base-Class-of-Evil
04:34	oddcully	(and i quoted wrong ;p)
04:35	jonathanj	oh hey, more saffers ;P
04:40	Guest72028	dstockton: interesting
04:40	dstockton	i don't know about jobs, but there's definitely a community/interest
05:17	acron^	in David's enlive tutorial, he uses the following syntax: (def ^:dynamic story-selector
05:17	acron^	a) what is the "dynamic" metadata?
05:18	acron^	b) idiomatically, what do the asterisk' indicate?
05:20	kungi	acron^: dynamic means you can change story-selector through with-redefs (for example) by using clojures dynamic scoping.
05:21	kungi	acron^: the asterisk' or earmuffs are a remnant of lisp inidcating a "special variable" that's not lexically scoped.
05:22	kungi	acron^: Have a look here: https://stackoverflow.com/questions/1986961/how-is-the-var-name-naming-convention-used-in-clojure
05:22	acron^	Thanks kungi
05:22	acron^	(inc kungi)
05:22	lazybot	⇒ 2
05:22	acron^	;)
05:22	magnars	kungi acron^ - the idiomatic way to bind dynamic vars is with `binding`
05:24	kungi	magnars: I dont't see the difference between `binding` and `with-redefs` (just had a look at the docs)
05:26	magnars	kungi: `with-redefs` is global, `binding` is thread-local
05:26	kungi	magnars: I see thank you
05:27	kungi	Why are new bindings made in parallel? This introduced a bug in my code lately and I had to nest binding forms.
05:33	dysfun	i can't say i'm terribly impressed with cljs compile times. it's just taken 18 seconds with no optimisations to build effectively "hello, world"
05:34	Empperi	you have an issue in your compilation settings
05:34	dysfun	howso?
05:34	Empperi	I have compilation times of 0.1-0.8 secs here and this is definetly way more complex than "hello world"
05:34	Empperi	a real reagent based webapp
05:35	dysfun	i'm generating a source-map. is that somehow mega slow?
05:35	Empperi	no if you have your settings correct
05:35	Empperi	you're using cljsbuild plugin for leiningen I assume?
05:35	dysfun	my settings are source map, no opts
05:35	dysfun	no, i'm using the boot cljs plugin
05:35	Empperi	ok
05:35	Empperi	do you get one big "main.js" or gazillion smaller files?
05:36	Empperi	if you get one main.js then you do have simple optimizations on which will slow down compilation considerably
05:36	Empperi	with no optimizations clojurescript compilation will generate one .js file per .cljs
05:36	dysfun	i have optimisations disabled
05:37	Empperi	so you get tons of js files as on output?
05:37	dysfun	and yes, i have one js file per cljs
05:37	Empperi	then you do not have optimizations disabled
05:37	Empperi	but you may think you have
05:37	Empperi	ah fuck
05:37	Empperi	sorry
05:37	Empperi	misread
05:37	Empperi	my bad
05:37	Empperi	shit, need more coffee
05:38	dysfun	i'm looking at the generated js file
05:38	Empperi	anyway, sourcemap generation will slow it down somewhat, but those numbers I'm seeing is with sourcemaps on
05:38	dysfun	there are 4 lines of code that actually do something
05:38	dysfun	it's all very human readable
05:38	Empperi	the main.js file should pretty much contain very little
05:38	Empperi	with no optimizations
05:39	Empperi	here we have 4 lines with fourth line being empty
05:39	Empperi	it just inserts some stuff to DOM to allow google closure dependency management to kick in
05:39	Empperi	and then it'll dynamically load all required js files one by one
05:40	dysfun	https://www.refheap.com/99457
05:40	Empperi	so that's your core.js I take it
05:40	dysfun	yes
05:40	Empperi	is that your output target file?
05:41	Empperi	for cljs compiler
05:41	Empperi	I'm guessing not
05:41	dysfun	i haven't explicitly specified one
05:42	dysfun	(cljs :source-map true :optimisations :none) ;; my boot config
05:42	Empperi	this is leiningen config but... https://www.refheap.com/4e6ee2a346dfbe075c35c7417
05:42	Empperi	you might want to enable :compiler-stats true and :verbose true to see what's really happening under the hood
05:43	Empperi	you'll most likely end up seeing some kind of optimization happening even if you say you shouldn't have any optimizations
05:43	Empperi	and oh
05:43	Empperi	dysfun: optimiZations, not optimiSations
05:43	dysfun	aha
05:43	Empperi	clojurescript compiler reads :optimizations keyword
05:44	Empperi	I'd guess boot plugin passes the parameters as-is to the cljs compiler like leiningen plugin does
05:44	dysfun	the one major benefit i've found of type systems is that it prevents locale-related spelling errors
05:44	Empperi	there are several others
05:44	dysfun	which is pretty niche i'll grant :)
05:44	Empperi	but then again dynamic typing gives you other perks
05:44	Empperi	there is no silver bullet :)
05:45	dysfun	of course
05:45	Empperi	but I'm guessing you'll see a dramatic speedup in compile times if you change that 's' to 'z'
05:45	dysfun	it's theoretically possible to get the benefits of both, and yet it's sufficiently difficult we haven't done it yet
05:45	Empperi	of course we have
05:45	dysfun	ah, my fans are going full pelt. i must have started a jvm...
05:45	Empperi	check Google's Dart for instance
05:45	Empperi	it has optional typing
05:46	Empperi	pretty nice implementation of static typing on top of dynamic typing
05:46	dysfun	core.typed is a bit like that i suppose
05:46	dysfun	Elapsed time: 24.063 sec
05:46	Empperi	dysfun: make a change to your file
05:47	Empperi	it had to compile everything from scratch since you changed optimization level
05:47	dysfun	ah, 3.187, that's better
05:47	Empperi	which means clojure.core etc
05:47	Empperi	still really slow
05:47	Empperi	but better
05:47	Empperi	"Successfully compiled "resources/private/generated/main.js" in 0.643 seconds."
05:47	dysfun	yeah, but it's fast enough to not make me want to give up clojurescript :)
05:47	Empperi	speed I'm seeing here
05:47	dysfun	this laptop is a bit slow
05:47	Empperi	we have something like 2k lines of clojurescript in this app
05:47	dysfun	and i have firefox running with $many tabs
05:48	dysfun	it takes on a good day about 15 seconds to get to a boot repl
05:49	Empperi	geesh, you have something seriously funky going on there :D
05:49	Empperi	but anyway, with this ~3 year old windows desktop I'm seeing compilation times like that
05:49	dysfun	this is a three year old macbook air
05:50	dysfun	it flies when i kill firefox
05:50	Empperi	my ~year old mbp gives me similar compilation times as this desktop
05:50	Empperi	kill the firefox and try to change a file then
05:50	Empperi	what's the compilation time then?
05:51	dysfun	yes, of course firefox has just hung when i tried to quit it from within itself
05:51	Empperi	just kill it :)
05:51	dysfun	i did
05:51	Empperi	and if it kills your performance that bad then I'd change to Chrome
05:51	dysfun	2.805
05:51	Empperi	not much of a difference there
05:51	Empperi	strangely slow
05:52	dysfun	yes, kernel_task has spun up now
05:52	Empperi	oh well, might be air's slow I/O
05:52	dysfun	i'll wait for it to die down and report back
05:52	Empperi	but anyway, the speed I'm getting is what you're supposed to see too
05:52	Empperi	and you definetly shouldn't think that kind of compilation times are normal
05:52	dysfun	and i've killed my VM too
05:52	Empperi	and dump cljs due to that
05:53	Empperi	you'd regret it, I postponed getting familiar with cljs for quite some time
05:53	dysfun	oh great, spotlight is running an index too. it's not my day to actually try and compile stuff fast
05:53	Empperi	and now I've been using it for over a month
05:53	Empperi	and shit I feel so stupid not going for it earlier
05:53	dysfun	well i've played with om a bit before
05:53	dysfun	it's just that clojure is really fun and compiling isn't
05:54	dysfun	0.992
05:54	dysfun	hooray
05:54	Empperi	well, these days if you do frontend stuff you'll have some kind of compilation phase anyway
05:54	Empperi	well that looks more like it :)
05:54	clojurebot	Excuse me?
05:54	dysfun	i try not to do frontend stuff for the most part
05:54	Empperi	clojurebot: you're excused
05:54	clojurebot	Cool story bro.
05:55	dysfun	and when i was doing it last, i didn't precompile css or js actually
05:55	Empperi	you can do that these days too but that's stupid way to do things :)
05:55	Empperi	if you're writing raw javascript then you'll most likely have something like browserify or similar
05:55	dysfun	it's less stupid than you might think
05:55	Empperi	and css is written via less or sass
05:56	Empperi	sure, if you're app is small then it's not a problem
05:56	dysfun	for example i checked out a competitor's app last year (it's open source!) and got fed up when the build process broke
05:56	Empperi	but with 10kloc and larger it gets really hairy
05:56	Empperi	I've written a javascript frontend which had nearly 40kloc of javascript code
05:56	dysfun	yes, i like sass. i just don't care enough about the reduced code size enough
05:57	Empperi	without any kind of compilation
05:57	Empperi	I tell you, that was pain :)
05:57	dysfun	however, now that boot seems to be doing a good enough job, it should be fairly easy for me to set it up and have it all managed by boot
05:57	dysfun	which means no build tool hell
05:57	Empperi	happy to help
05:57	Empperi	gotta do some work now :)
05:57	dysfun	frankly maintaining 40klo javascript is anybody's idea of hell
05:57	sobel	once i understood google's js closure compiler and why it's necessary, i realized the hand-written web is over
05:58	dunderproto	have any of you read Paul Graham's On Lisp?
05:58	dysfun	the hand written web has been dead for a while, but the corpse is still twitching
05:58	dunderproto	the book really piqued my interest but I wasn't sure how much of it would be applicable to clojure
05:58	sobel	sure, but let's not confuse the throes of death with viability ;)
05:59	dysfun	:)
05:59	Empperi	javascript is the new assembly
05:59	Empperi	you see ppl writing raw assembly code too
05:59	Empperi	:)
05:59	dysfun	dunderproto: it's been a long time since i read it. i'm happy to help you with specific questions
05:59	sobel	Empperi: yup, and there will always be people getting in deep to tweak things. i mean, someone implemented the logic gates that execute the machine code, right?
06:00	sobel	someone had to smelt the silicon and actually touch pre-hardware
06:00	dunderproto	dysfun: do you think it'd be worth reading? I was looking for a book that would teach me how to write macros for clojure. I saw some folks recommend On Lisp
06:00	sobel	(yuck!)
06:00	dysfun	dunderproto: frankly i'm a bit down on paul graham these days. and i really wouldn't recommend reading how to write macros in the wrong lisp in order to learn them in clojure
06:01	dysfun	reading about macros in lisps, sure, but if you want to learn how to write macros, no, i wouldn't say it's a good idea
06:02	sobel	is there a koans chapter just on macros? :)
06:02	dunderproto	dysfun: ah. I appreciate the advice. Maybe I will keep looking.
06:02	dysfun	i forget which lisp graham favours. clojure macros are a bit like hygienic macros in scheme
06:02	dysfun	we have different (imo better) syntax
06:03	sobel	i'm also useless at macros so far and figure i should eventually get introduced
06:03	dunderproto	Graham used common lisp I think
06:03	dysfun	yeah, common lisp macros are a bit special imo
06:04	dunderproto	dysfun: what book would you recommend for clojure?
06:04	dysfun	but hey, they look like function definitions. except your arguments are all the code between the parens instead of the values they represent
06:05	dysfun	and you return some code.
06:05	dysfun	book choice is highly personal. do you want a practical book or a theoretical one?
06:05	dunderproto	practical :)
06:05	dysfun	how quickly do you need to feel you've done something to keep going?
06:05	kris_	In Clojure(Script) how do I dereference an atom which is a different namespace?
06:06	dysfun	kris_: the same way? @other-ns/atom ?
06:06	dunderproto	I have a lot of patience
06:07	slipset	dunderproto: why do you want to learn how to write macros?
06:07	dysfun	in that case, have a look at one of the o'reilly's. i think there's a roughly 50/50 split on preference. read up about them both
06:07	slipset	It seems to me that the clojure community favors data over functions over macros
06:07	dunderproto	dysfun: thanks, will check it out
06:08	dunderproto	I was mostly drawn to lisps because of PG's essay about "beating the averages"
06:08	dysfun	slipset: well it's more about not using a hammer to crack a nut
06:08	dysfun	slipset: the further down the chain you go, the more you lose in terms of composability
06:08	dysfun	so "if you can do without a macro, do" is a good guideline
06:08	dunderproto	but yeah, I do understand that some of the techniques may not be as ideal or idiomatic for clojure
06:08	slipset	dunderproto: same thing for me, but I've yet to figure out the macros are great thingy
06:09	dysfun	macros are pretty great
06:09	oddcully	dysfun: that's how i usually do it. well last time i used a big skrewdriver for the chestnuts
06:09	Empperi	"the first rule of macro-club is that you don't write macros"
06:09	slipset	But I can see that in 1995, when there was little support for web programming, you could probably do a lot with macros to help that situation
06:10	dysfun	for example, how many times have you been tidying up code and thought "i wish i could generate the try/catch" ?
06:10	Empperi	"the second rule of macro-club is you'll still avoid writing that macro"
06:10	ddellacosta	dunderproto, if you're not familiar with macros and want to learn how to write them in Clojure, take a look at the chapters in Clojure Programming and Joy of Clojure
06:10	Empperi	"the third rule of macro-club is you should write that macro since you got this far in the rules"
06:10	kris_	@dysfun thanks, I was doing other-ns/atom (no @)
06:10	slipset	Empperi: sounds like Knuths optimization rules.
06:10	dysfun	i see macros largely as "the last stage of boilerplate removal"
06:10	Empperi	pretty much
06:10	Empperi	macros give you insane power but they come with a cost
06:10	dunderproto	thanks dysfun and everyone else
06:11	dysfun	but you know what the real thing is?
06:11	slipset	dunderproto also remember, you cannot pass macros around
06:11	dysfun	it's that you're writing programs that read and write programs. and it's not that hard
06:11	slipset	(map some-macro myvec) will not work
06:11	dysfun	i can write a macro without a second thought to do what i need because it's easy when you've done it for a while
06:11	Empperi	there are some clojure.core macros which are macros for a good reason and still they bite me in my arse time to time
06:12	Empperi	my favourite two are 'and' and 'or' macros
06:12	dysfun	anyone who's ever written macros for a non-lisp will know what i mean here
06:12	Empperi	wish I could sometimes do: (apply and list-of-predicates)
06:13	Empperi	but you can't
06:13	Empperi	not with 'and'
06:13	dysfun	however, every-pred does that
06:13	Empperi	yes
06:13	Empperi	there are other cases too though
06:14	Empperi	but anyway, a good example of one of the prices to be paid when you go the macro-way
06:14	dysfun	you should go read up about f-expressions. slightly different than macros, but pretty equivalent
06:14	dysfun	and yes, not being able to use macros in higher order stuff is a bit of a pain
06:15	slipset	But back to Graham, I still fail to see how my writing macros make my life so much easier.
06:16	slipset	I can see the value of eg the go-macro, but it would not be for me to write.
06:17	slipset	And so, to me the go-macro could just as well be a feature of the compiler
06:17	slipset	I can also see that as a designer of libraries, macros can be nice, but I'm just a lowly consumer of such.
06:19	dysfun	well for example, i've got a simple macro 'safely' that wraps something in a try/catch
06:19	dysfun	it's just a nice little touch to tidy it up
06:20	Empperi	I use macros especially with tests
06:20	Empperi	(with-clear-database ....) etc
06:23	dysfun	right, now to see how well boot-reload works with a custom ring handler
06:23	sobel	so, how long until i'm one learning all the basic functions like every-pred?
06:24	dysfun	never?
06:24	sobel	one=done
06:24	dysfun	keep https://clojure.org/cheatsheet/ open in a tab
06:24	sobel	fair
06:24	sobel	it's already up
06:24	sobel	hadda ask
06:24	dysfun	:)
06:25	dysfun	it's quite pleasing when you realise you have a useful working set of them in memory
06:36	sobel	i'm close but i still write more code than i have to, just because i don't always have the right ones yet
06:37	stian	it's a lisp, there will always exist a shorter way of expressing your forms - no matter your level ;)
06:40	TEttinger	yep
06:41	TEttinger	I can still probably code-golf some of my one-liners further
06:41	TEttinger	but clojure is an amazingly good language for code golf
06:47	dysfun	yes, except that unlike golfing e.g. perl, your code tends to become cleaner because golfing is more about figuring out how to do the same thing with better use of builtins
06:52	TEttinger	dysfun, excuse me, but my golf game must be on another level, since my clojure golf code starts out getting more readable and then gets much, much less readable once I start treating strings as sequences of chars and other useful but confusing tricks
06:52	TEttinger	:)
06:56	TEttinger	,(map partition [1 2 3] (repeat "The Quick Brown Fox"))
06:56	clojurebot	(((\T) (\h) (\e) (\space) (\Q) ...) ((\T \h) (\e \space) (\Q \u) (\i \c) (\k \space) ...) ((\T \h \e) (\space \Q \u) (\i \c \k) (\space \B \r) (\o \w \n) ...))
06:56	dysfun	your output is unreadable, but that's a different thing :)
06:56	TEttinger	heehee
06:57	TEttinger	still, clojure is loaded with really handy features for golf that turn out to be useful in many more cases
06:57	dysfun	heh
06:57	Foxboron	o20aføsldkfjsadx20
06:57	Foxboron	wops
06:57	TEttinger	though I doubt myself that golf is an intended goal, some other lisp users have suggested that
06:58	dysfun	Foxboron: you might want to change that ;)
06:58	TEttinger	Foxboron, especially since this channel is publically logged
06:58	Foxboron	Not a password actually
06:58	Foxboron	me typing randomly as irssi hung up a few sec
06:59	TEttinger	haha
06:59	stian	or just golfing :)
07:00	dysfun	i think he missed if that was golf :)
07:00	oddcully	i find codegolfing a good way to actually learn about all that fancy functions
07:15	borkdude	I have a problem that I would like to discuss. I have a form with a rich text editing field in which users can upload pictures. I want to use a uuid to keep track of which form post the pictures belong to. The pictures are posted before the form. Is this a good approach?
07:15	dysfun	it's not an unreasonable approach. why uuids though?
07:16	borkdude	dysfun because I need some identifier
07:16	dysfun	you can create one of those by obtaining the next sequence id for the post
07:16	dysfun	or by inserting a partial record with some flag
07:18	borkdude	dysfun the first option: this can pose a race condition. the second option seems ok to me, but more complex than using a uuid, because I won't have to clean up records if users decide not to finish their submission
07:18	dysfun	actually, you're going to have to clean out the images table
07:19	borkdude	dysfun true
07:20	dysfun	there are merits to both approaches
07:20	dysfun	i'm not going to pretend to know more about your code than you :)
07:21	borkdude	dysfun well, I bet I'm not the first dealing with this specific problem :) Maybe other just upload the pictures and leave them lingering without even touching upon this problem?
07:21	dysfun	i expect in most cases they insert a partial record and then never clean up
07:22	borkdude	ok
07:25	mavbozo	borkdude, that's a good idea. btw, what db do you use?
07:25	borkdude	mavbozo mysql (legacy choice)
07:26	mavbozo	borkdude, and the uuid is just a standard uuid without indexing on the table?
07:26	borkdude	mavbozo yeah
07:27	mavbozo	borkdude, don't you a little a bit worried about the query performance?
07:27	zot	i'm trying to use a cljx-based project and run lein install to test the jar locally with another package, but the actual source files never make it into the jar. project file is here: https://github.com/benfleis/datascript/blob/feature/clojure-port-v2/project.clj . anybody see where i'm going wrong?
07:28	borkdude	mavbozo well, the uuid is just to prevent garbage. For example I could check if there are pictures with an unknown uuid on S3 and clean then. in the database I use an integer id also.
07:28	borkdude	mavbozo in other words, there won't be a query using the uuid
07:29	zot	and at the moment, i don't even care whether the cljs makes it in, just the clj would be enough
07:30	acron^	I have an atom set, which holds hashmaps. Is there any way for me to alter a single hashmap in the set without swap! disj, swap! conj combo ?
07:32	zot	acron^: you mean you have sth like (atom #{ { … } { … } })? if so, how do you match elements of the set? (or is it a sorted set and you do some range matching or sth?)
07:32	acron^	zot: each hm has an :id
07:33	mavbozo	borkdude, well, for once then, i thought you use uuid as primary key, or index on the table.
07:33	acron^	(atom #{ { :id 4 … } { :id 6 … } etc })
07:33	acron^	it's not sorted, the ids are uids
07:33	zot	acron^: in your shoes i would mod it to be hm -> hm, with the id being the key, and do swap! update-in … (or assoc-in …)
07:34	borkdude	mavbozo well no. I would generate a uuid client side. Upload pictures in the rich text editor with the uuid as an extra parameter to keep track. The pictures would end up on S3. Then the user submits the form. The uuid is just an extra field, so I know which picture belongs to which form post.
07:34	zot	unreferrable data inside the atom just sounds needlessly tricky to me
07:34	borkdude	Another option is just to analyse the html and generate a list of used and unused pictures.
07:34	zot	err, un-path-map-able
07:34	acron^	so you mean i shouldn't use a set?
07:35	zot	meaning that there's no obvious sequence of keys that tells you what's what. even a vector would be an improvement, where you could do update-in with the vector index.
07:35	zot	acron^: imho
07:35	zot	if you're updating those things, are you expecting it to keep uniqueness for you as you make changes? that would be the only reason i can imagine, but that sounds unlikely to me
07:37	acron^	ok, i am a novice clojure users so my motive for using a set is flaky
07:37	acron^	lets say I change it to vec
07:38	dysfun	borkdude: there's another option. you could embed the image directly in the html with src="data://"
07:38	acron^	is there still no better way than removing and re-adding an element to the atom? (swap! disj, swap! conj)
07:38	dysfun	then when the user posts it, strip them out and save them properly
07:39	oddcully	swap! #(conj (disj %)) ish
07:40	acron^	:o
07:41	zot	acron^: if your structure looks like [ {:k1 1, :k2 2}, {:k3 3, :k4 4} ], (swap! <thing> assoc-in [0 k3] 1000) will work
07:41	zot	oops, s/0/1/
07:41	zot	for the array index
07:41	zot	although in my example k3 would be added to the 0th element with a new val of 1000
07:42	zot	using 1 would update k3 in the second
07:42	acron^	zot
07:42	acron^	that's perfect, thank you
07:43	acron^	(inc zot)
07:43	lazybot	⇒ 1
07:43	zot	yw
07:43	oddcully	,(doc replace) maybe then
07:43	clojurebot	"([smap] [smap coll]); Given a map of replacement pairs and a vector/collection, returns a vector/seq with any elements = a key in smap replaced with the corresponding val in smap. Returns a transducer when no collection is provided."
07:46	oddcully	,(into #{} (replace {{:id 1} {:id 42}} #{{:id 1}}))
07:46	clojurebot	#{{:id 42}}
08:06	hellofunk	anyone have first-hand experience with the first and second editions of Joy of Clojure and can comment on the significance of their differences?
08:33	nicferrier	if I use a gensym in a macro it's compiled to the evaled name right?
08:33	nicferrier	it won't gensym in the byte code... it'll be the name.
08:33	nicferrier	I've written a macro to wrap a function in a child process
08:34	nicferrier	so it starts the process and then calls the function that you specify in the child process.
08:34	justin_smith	nicferrier: like future but in a new vm?
08:34	nicferrier	it does that by calling the same clojure program you are running ... but with an argument.
08:34	nicferrier	yeah. similar.
08:34	nicferrier	it's just to protect code that crashes.
08:35	nicferrier	you could use a bash while loop... but trust me... there are reasons why not.
08:35	nicferrier	it's just the basic supervisor pattern
08:35	justin_smith	I'm all too familiar with craziness from bash while loops
08:35	nicferrier	anyway, I'm using a gensym for this arg
08:35	nicferrier	justin_smith: well, it's more the organization I'm in. but anyway.
08:36	nicferrier	and it appears the gensym is different on either side of the child process.
08:36	justin_smith	gensym is not deterministic
08:36	blashyrk	hi! is there a way to do something like a "case-let" ? in terms of regular java it would be some thing like int x; if(something) x = someFunction() else x = someOtherFunction();
08:36	justin_smith	,(gensym "hello")
08:36	clojurebot	hello27
08:36	justin_smith	,(gensym "hello")
08:36	clojurebot	hello52
08:36	justin_smith	,(gensym "hello")
08:36	clojurebot	hello77
08:37	nicferrier	justin_smith: yeah. but surely if I make a macro with the gensym it'll be gen'd at compile time and the gen'd value will be used in the bytecode?
08:37	justin_smith	blashyrk: ##(let [a (if (> (rand) 0.49) 42 :else)] a)
08:37	lazybot	⇒ :else
08:39	justin_smith	nicferrier: I'm less sure of that
08:39	blashyrk	@justin_smith I know I could do it that way I was just checking if there was a shorter way, macro or something. thanks :)
08:39	nicferrier	justin_smith: yy. it's tricky enough at the best of times.
08:39	justin_smith	blashyrk: I can't think of anything that would be more concise than if
08:40	blashyrk	something like
08:40	blashyrk	(case-let [foo condition set-if-true set-if-false])
08:41	blashyrk	it's not that big of a deal really, I am just constantly trying to learn the tricks and was wondering if something like this was implemented
08:41	justin_smith	(let [foo (if condition set-if-true set-if-false)]) has exactly the same character count
08:41	justin_smith	it's literally the same number of symbols
08:42	justin_smith	blashyrk: but if you really want that syntax, it would be a relatively easy macro to write
09:44	hellofunk	anyone have first-hand experience with the first and second editions of Joy of Clojure and can comment on the significance of their differences?
09:49	not-much-io	Has anyone had indentation problems with Clojure in the graphical variant of emacs? (24.4.1) It displays that the indentation is off although opening in emacs in the terminal or another text editors shows the indentation is correct.
09:51	f3ew	not-much-io: tabs vs spaces?
09:53	borkdude	nice that this works: (assoc nil :foo "bar") ;;=> {:foo "bar"}
09:53	borkdude	clojure is so forgiving
09:54	mavbozo	,(get get get get)
09:54	clojurebot	#object[clojure.core$get "clojure.core$get@bb699c8"]
09:54	not-much-io	f3ew: I've never understood the tabs vs. spaces debate, I've used tabs. In the recent Stackoverflow survey there was a almost 50/50 split... What's the deal? Some press tab others click space multiple times?? :D
09:55	nicferrier	not-much-io: no. their editors press space multiple times.
09:55	justin_smith	not-much-io: sounds like you have a bad default font (maybe not a fixed width)
09:55	nicferrier	not-much-io: tabs are "bad" because you can vary what the tab is and that might change how the file looks.
09:56	not-much-io	justin_smith: the font is the same in both the terminal and GUI
09:56	mavbozo	,(get get nil get)
09:56	clojurebot	#object[clojure.core$get "clojure.core$get@bb699c8"]
09:56	justin_smith	,(get (gensym) (gensym) 42)
09:56	clojurebot	42
09:56	not-much-io	nicferrier: If their editors press space for them, then how many times it presses can also vary right?
09:57	justin_smith	not-much-io: the difference is varying on creation vs. varying on display
09:57	justin_smith	when it varies on creation, the author can fix any errors
09:57	justin_smith	when it varies on display, the author just looks like they are incompetent because the indentation is broken
09:58	mavbozo	,(get nil nil 42)
09:58	clojurebot	42
09:58	nicferrier	but some people say "but then I can adjust the indentation"
09:58	justin_smith	clearly we need code editors with css support
09:58	not-much-io	justin_smith: Ahso, okei, well that makes sense. I have to re-evaluate my life now. Spaces it should be then?
09:58	mavbozo	wait til you see clojure's java source code
09:58	nicferrier	not-much-io it really should, yes.
09:59	not-much-io	Is there some easy step to switch from tab to space in emacs or should I just start googling?
09:59	tomjack	https://books.google.com/books?id=n8AOBwAAQBAJ
09:59	tomjack	kind of frightening
09:59	justin_smith	not-much-io: untabify-region
10:00	justin_smith	tomjack: yuck
10:00	justin_smith	tomjack: that look-inside is so terrible
10:01	oddcully	is it me not allowing google executing javascript or is this just MD-ish code printed as book?
10:01	justin_smith	tomjack: and the authors who make crap books like that do review spam too
10:01	justin_smith	to make the books look popular
10:02	tomjack	:(
10:02	justin_smith	oddcully: MD-ish?
10:02	tomjack	maybe before long the books will actually be valuable
10:02	justin_smith	oddcully: it's a terribe book, probably written by a crappy mail-merge macro in microsoft word
10:02	oddcully	section 3 -> `* Groovy (programming language)\|Groovy`
10:02	justin_smith	yeah
10:03	justin_smith	macro-generated
10:03	justin_smith	put on the store in the hopes they can trick someone into buying it
10:03	justin_smith	I bet every single book they publish uses the same stupid template
10:03	oddcully	should have used TeX ;P
10:03	borkdude	is this an autogenerated bok?
10:03	borkdude	*book
10:03	justin_smith	yeah - it's a whole genre of crap
10:03	not-much-io	justin_smith: I think it might have fixed it but, configuring the space to indent like tab did?
10:03	oddcully	section 4 -> headline on page end
10:03	borkdude	it indeed looks like crap
10:04	oddcully	that looks like what my firefox generates when i generate a pdf
10:04	justin_smith	they generate books, make it look like something that might be valuable, and hope someone buys it. They use sockpuppet accounts to make fake reviews to make the books look popular.
10:04	justin_smith	huge scam
10:04	justin_smith	I mean, if it was just a formatting issue, you would find some content in the book
10:05	justin_smith	but no, it's not just badly formatted, it has no useful information
10:05	justin_smith	there are literally people who generate books with markov chains, then combine randomized keywords to generate a title
10:06	justin_smith	there's no cost to publishing the stuff, and if anyone is stupid enough to buy, jackpot
10:06	tomjack	sorry, hope I didn't ruin your morning :)
10:06	justin_smith	haha no, I'm sorry for the rant
10:06	justin_smith	I just really hate people that do that
10:06	oddcully	rant away
10:07	oddcully	clears the system
10:10	tomjack	I found the book in my search for a solution to this hilarious problem: https://www.refheap.com/2ab934158d7df2615f7bdf6fb
10:10	oddcully	did it help?
10:11	tomjack	(org.eclipse.jdt.core)
10:11	justin_smith	tomjack: wait, is that a thing that actually works?
10:11	tomjack	why on earth would someone make a .java -> AST library where the AST nodes print _as .java source_??
10:11	lazybot	tomjack: Uh, no. Why would you even ask?
10:11	justin_smith	oh yeah, eclipse basically has its own java parser, huh
10:11	tomjack	java brain poisoning I guess
10:11	justin_smith	haha
10:12	justin_smith	but you can still access the ast data, right?
10:12	tomjack	yeah, but I have to apparently write a library which converts their AST into data, which just makes me "ugh"
10:14	justin_smith	tomjack: just because the printed form is java source doesn't neccessarily make the data harder to access though does it?
10:14	tomjack	AFAICT it means you have to go to the javadoc or inspector to see what's available, instead of just printing
10:15	justin_smith	tomjack: but most printed forms for java objects are totally useless
10:15	tomjack	yes, I guess I should not be surprised :(
10:29	not-much-io	I am confused with how to switch from using tabs to using spaces in Emacs. From here -> http://emacsblog.org/2007/09/30/quick-tip-spaces-instead-of-tabs/ It says just use '(setq-default indent-tabs-mode nil) but for me tabs keeps working and space is still a single space. Is there no easy way to pretty much keep using space just let it use spaces?
10:29	not-much-io	keep using the tab button *
10:30	dysfun	is there a clojurescript equivalent of http://www.clojure-toolbox.com/ yet?
10:35	justin_smith	not-much-io: after you do that, the tab key should insert multiple spaces
10:35	justin_smith	as apropriate for your mode's indentation
10:36	justin_smith	not-much-io: something very helpful is M-x whitespace-mode
10:37	not-much-io	justin_smith: That is very helpful, I now see that thespaces in the GUI emacs are shorter than the spacen in the CLI
10:38	justin_smith	yeah, like I was saying, proportional font
10:38	justin_smith	even in a terminal, a proportional font will display as fixed-width, it's a limitation of the CLI
10:40	not-much-io	justin_smith: If I look really close, they might really be different, but really similar. Alright thanks, I'll experiment with the font.
10:40	not-much-io	Have I understood correctly that if someone helps you out you '(inc that-user) for clojurebot? :)
10:41	justin_smith	without the quote, and it's lazybot that does it
10:41	justin_smith	but yeah
10:41	not-much-io	I quoted to not trigger the inc :D
10:41	not-much-io	(inc justin_smith)
10:41	lazybot	⇒ 237
11:34	tomjack	,(:foo #{:foo}) ; whoa
11:34	clojurebot	:foo
11:38	calvinfroedge	Can anyone tell me if something like this is possible? (defn foo [a & [b c :as args & d]](prn d))
11:39	calvinfroedge	I want to bind b and c to :args, but also have d
11:45	mavbozo	,(defn foo [a & [b c :as args & d]](prn d))
11:45	clojurebot	#error{:cause "Unable to resolve symbol: d in this context", :via [{:type clojure.lang.Compiler$CompilerException, :message "java.lang.RuntimeException: Unable to resolve symbol: d in this context, compiling:(NO_SOURCE_PATH:0:0)", :at [clojure.lang.Compiler analyze "Compiler.java" 6535]} {:type java.lang.RuntimeException, :message "Unable to resolve symbol: d in this context", :at [clojure.lang.Ut...
11:47	calvinfroedge	mavbozo: See = ) Great mystery.
11:49	justin_smith	,(defn foo [a & rst] (let [[a b & d] rst args [a b]] d)
11:49	clojurebot	#<RuntimeException java.lang.RuntimeException: EOF while reading>
11:50	justin_smith	,(defn foo [a & rst] (let [[b c & d] rst args [b c]] {:a a :b b :c c :args args :rst rst}))
11:50	clojurebot	#'sandbox/foo
11:50	justin_smith	,(foo (range 10))
11:50	clojurebot	{:a (0 1 2 3 4 ...), :b nil, :c nil, :args [nil nil], :rst nil}
11:50	justin_smith	,(apply foo (range 10))
11:50	clojurebot	{:a 0, :b 1, :c 2, :args [1 2], :rst (1 2 3 4 5 ...)}
11:52	calvinfroedge	justin_smith: Nice work! Thanks a lot.
12:17	borkdude	ordnungswidrig I'm using :handle-request-entity-too-large (fn [ctx] {:error "My error"}) and I get this error: No method in multimethod 'render-map-generic' for dispatch value: null. What could it be?
12:19	borkdude	ordnungswidrig I'm reading this https://github.com/clojure-liberator/liberator/issues/94
12:40	favetelinguis	what am i not understanding with features: http://lpaste.net/130495 why is this code not printing anything 4sec?
12:40	favetelinguis	futures that is :)
12:48	justin_smith	favetelinguis: "I will print after 4 sec" is not a function
12:48	justin_smith	favetelinguis: if you deref the future, you'll see the exception thrown
12:48	favetelinguis	hehe i saw it now, thanks
13:27	favetelinguis	does core.reducers support trancducers? if not will they in the future?
13:48	zanes	Is there any reason why `(go (<! (async/timeout 1000)))` should behave differently than merely `(async/timeout 1000)`?
13:53	whodidthis	not sure if im using splice correctly but if i run #?@(:clj [(def a 1) (def b 2)]) how come only a gets defined
13:54	Bruce_Wayne	does anyone know how to render a link in reagent?
13:56	Bronsa	whodidthis: wrap it in a (do)
14:00	dnolen	zanes: those those don't mean the same things
14:01	zanes	dnolen: Sorry, (<!! (async/timeout 1000)) and (<!! (go (<! (async/timeout 1000))).
14:02	dnolen	zanes: what do you mean by "behave differently" then?
14:03	zanes	Yeah, let me get you a more complete example.
14:07	zanes	dnolen: False alarm.
14:08	zanes	I’m writing a test and one of the go blocks was using Thread/sleep.
14:16	zanes	I thought the test was deadlocking with `(<!! (go (timeout…` and not with `(<!! (timeout`, but I was wrong.
14:19	ncthom91	hi all. When I use a `go` block in the middle of my function body, does the go block execute in a separate thread while the execution of the remainder of my function continues?
14:19	ncthom91	or does the remainder of my function wait for that go-block to finish
14:19	ncthom91	I think it's probably the former, yea? Cause the go block immediately returns a channel
14:22	zanes	ncthom91: The former.
14:22	ncthom91	zanes cool, thanks
14:22	zanes	If you want to block until it’s finished you can call `<!!` on the channel returned by `go`.
14:23	ncthom91	zanes i took a slightly similar, but essentially equivalent (i think?), approach by removing the (go) block and changing my >! call inside the go block to a >!!
14:23	zanes	Yeah, that works!
14:24	saik0	having some trouble improrting clojure.data.priority-map https://gist.github.com/anonymous/cc6ffaa0c58442890f95
14:24	ncthom91	sweet. zanes my other question is, this function itself actually runs inside a go block from the calling function, so all of this happens in a separate thread right?
14:25	zanes	ncthom91: So, like, `(f (go ((fn [] (>!! c :thing)))))`?
14:25	zanes	Something like that?
14:25	ncthom91	Well I have a function `update` and inside update I make my >!! call
14:26	ncthom91	But `update` gets called from a different function with (go (update x y))
14:26	zanes	But regardless: Yes. If a function calls another function inside a go block the called function is executing in one of the go threads.
14:26	zanes	Yes, totally. That’s happening in one of the go threads.
14:26	ncthom91	sweet.
14:26	ncthom91	What happens if my `update` function does IO?
14:27	ncthom91	in terms of what code runs in what threads?
14:35	mavbozo	saik0, priority-map function call should've been (clojure.data.priority-map/priority-map :a 1 :b 2)
14:36	saik0	mavbozo ah, got it. thanks
14:38	mavbozo	saik0, because the namespace is too long, you might want to alias it (:require [clojure.data.priority-map :as pm])
14:38	mavbozo	saik0, might help the eye spotting what's wrong
15:09	lemonodor	if you’re into gorilla (or planespotting), this has been a fun worksheet I’ve been working on: http://viewer.gorilla-repl.org/view.html?source=github&user=wiseman&repo=orbital-detector&path=stats.cljw
15:17	kwladyka	what is the right architecture in functional language like clojure? i am talking about things like MVC, Enitity, Boundary, Interactor etc. What is the right way to separate UI, use cases and DB? How to do that right in Clojure?
15:19	kwladyka	It will be greate if somebody can give some real example of project like web app where the UI is undepended code, and the core is the core, and DB is independed, easy to change.
15:19	pbx	kwladyka, what do you mean by "undepended" and "independed"?
15:20	pbx	(not that i have an example handy)
15:20	ecelis	kwladyka: check luminus
15:21	ecelis	http://www.luminusweb.net/
15:21	ecelis	there is no such thing as "right way"
15:22	aaelony	how does luminus facilitate transfer of dat structures from clojure to clojurescript?
15:22	kwladyka	pbx, sorry, it should be "independent"
15:24	mavbozo	kwladyka, see http://modularity.org/index.html. malcolm sparks is in the process of writing a book which i think would be a good answer to your question. It's still a draft for now but there are some ready-made leiningen templates.
15:24	kwladyka	ecelis hmm so another quesion is: should use web framework in Clojure or just ring+compojure? What more gives thing like luminus? I am sorry about that question but i am learning Clojure and it is my first functional language, so i am confuse sometimes.
15:26	kwladyka	mavbozo thx
15:27	mavbozo	kwladyka, you wrote Interactor, that's interesting choice of term. Where did you know about that term?
15:27	mavbozo	kwladyka, i only heard it from uncle bob videos
15:27	kwladyka	mavbozo clean code, episode 07 about architecture
15:27	kwladyka	:)
15:29	kwladyka	i feel how i am doing this now it is bad... and i am looking solution everywhere... and the worst thing i don't understand how to write in functional language yet. but i am pretty sure how i did this in OOP it was bad and i never liked that :)
15:29	kwladyka	i feel myself all should be divded as much as possible
15:29	mavbozo	kwladyka, ah, so you have previous experience building webapps. If that's the case, then I recommend you to dive straight in making website in clojure.
15:30	kwladyka	the on thing is use for example AngularJS, Clojure to write backend, and DB. It is good step. But... how to do that without AngularJS? I am not sure how to do right architecture then.
15:30	mavbozo	kwladyka, yogthos's book Web Development in clojure will guide you https://pragprog.com/book/dswdcloj/web-development-with-clojure
15:33	kwladyka	Adding things liek AngularJS makes this easier, because i have separated UI in natural way and i can communicate by JSON. But how to do that if i will use static pages?
15:33	kwladyka	this book will answer for that question?
15:34	kwladyka	and even if i use AngularJS or static pages... details about core are important too.... is there any conception like MVC, but for functional language?
15:35	mavbozo	kwladyka, the book will help you get real experience creating websites in clojure. besides, the book starts with a familiar MVC pattern.
15:36	kwladyka	mmm but i feel MVC is really bad way to do projects. I need solution for architecture now, not how to use tools in Clojure.
15:36	mavbozo	kwladyka, i think it helps having some practical coding with clojure first before continuing on with answering your question
15:37	kwladyka	i will start do some web app anyway to learn, this process of learning is obviously. But i am not sure how to learn creating of good architecture.
15:38	kwladyka	my main think is: architecture should be true image of business model, easy to read, easy to write, extermly elastic - it has to be done in some way.
15:38	kwladyka	especially with Clojure :)
15:41	mavbozo	kwladyka, there are good books-- https://leanpub.com/building-a-system-in-clojure and this https://pragprog.com/book/vmclojeco/clojure-applied, and video presentations to help you answer your question about architecture in clojure application such as
15:41	mavbozo	web application
15:41	bacon1989	I've been meaning to grab that book
15:41	kwladyka	thx, added to read :)
15:41	mavbozo	or business model
15:41	bacon1989	but I like physical books :(
15:41	mavbozo	but i don't think it's not for clojure beginners
15:42	bacon1989	do you know if it's coming out to paperback anytime soon?
15:42	bacon1989	mavbozo double negative
15:42	mavbozo	correction, i think it's not for clojure beginners
15:42	bacon1989	;)
15:44	kwladyka	mavbozo there is only one way to not be beginner, i has to read things like that and try to understand :)
15:44	mavbozo	kwladyka, i still think you learn creating good architecture by studying -> doing -> review -> discuss
15:44	bacon1989	ah, the paperback release is at the end of the year
15:44	mavbozo	kwladyka, you discuss it here
15:44	bacon1989	I might get it
15:45	kwladyka	mavbozo yes but i don't want research whole wheel again.
15:47	ecelis	kwladyka: luminus includes bootstrap for the front end, but its my guess it can be easily changed to use angular
15:47	ecelis	or any other front end framework for that matter
15:48	ecelis	but you can also choose to use bare compojure or ring, it depends on your needs really
15:48	kwladyka	ok so another question is: what should use for web apps? In the Internet it is a lof of informations but out of date. Use rings+compojure? Or lumius or something else? Mayby right question should looks: when use which one? Example if i want write e-store app as independent core and connect with that web intefrace or mayby later even mobile phone interface. Which tools should i use?
15:48	ecelis	kwladyka: try yourself, there is no absolute answer of what tools to use
15:49	bacon1989	kwladyka: most people use compojure+ring
15:49	ecelis	it depends on programmers needs, taste, background, etc
15:49	weavejes_	kwladyka: Luminus is a project template, rather than a framework in the same sense of Rails or Django.
15:50	saik0	i have a collection of infite seqs, I'd like to "make" str print something like '((first seq) "LazySeq@xxxxxx"), it would be handy for experimenting.
15:50	kwladyka	weavejes_ what do you mean "project template"? I should understand that like get other libraries as dependency but zero lumius code?
15:51	mavbozo	kwladyka, starts with instructions from yogthos's book web development with clojure. It makes you focus building your web app quickly without overwhelmed by various ways and toolings in clojure.
15:51	mavbozo	kwladyka, and then you can review it based on your previous knowledge about building good architecture.
15:51	kwladyka	mavbozo but tools are very important, i don't really want miss that
15:51	weavejes_	kwladyka: If you run "lein new luminus foo", then Luminus will generate a new project for you called "foo". It'll include a number of dependencies, such as Compojure and Ring.
15:52	kwladyka	i want do this in 80% or something like that and later finish that research :)
15:52	mavbozo	kwladyka, we can not hope to understand everything in this world.
15:52	mavbozo	kwladyka, if you want to learn good architecture, then focus on that.
15:54	kwladyka	just i have to start with something, so i am doing quick research about tools and architecture to decide which tools i will use first and in what way :)
15:54	weavejes_	kwladyka: Luminus is essentially a way of getting started. It's not really a framework, since it doesn't include its own code in the template, as far as I'm aware.
15:54	mavbozo	kwladyka, agree with weavejes_
15:55	mavbozo	kwladyka, luminus is a good way to start
15:56	kwladyka	i did lein new luminus foo and looking on code now :)
15:56	mavbozo	kwladyka, great!
15:56	kwladyka	is ther more tool wich i should be interested in?
15:57	ecelis	besides leiningen and your text editor, I don't think you need any other tool :)
15:57	mavbozo	kwladyka, git as scm and heroku to deploy your webapp
15:59	kwladyka	Do you think Light Table is good? Or is it not ready yet?
15:59	mavbozo	kwladyka, Light Table is good for clojure beginners
16:00	kwladyka	mavbozo but not for clojure pro?
16:00	justin_smith	kwladyka: it has some cool features that can definitely help you figure out clojure, but eventually most people move to a more normal editor or ide
16:00	justin_smith	kwladyka: no reason not to try it
16:01	mavbozo	kwladyka, Light Table is used as editor in yogthos's instructions in web development with clojure
16:02	mavbozo	kwladyka, i dont' want you get distracted from your architecture focus
16:03	kwladyka	mavbozo yes, anyway i think i will use light table for a while, becuase too many things too learn
16:04	bacon1989	kwladyka: no no no, you should jump right into emacs
16:04	mavbozo	and ends up studying emacs architecture instead
16:04	arohner	bwhaha! leinception! https://github.com/arohner/lein-docker#lein
16:05	kwladyka	ok, i checked how luminus build start architecture but still i don't have answer about how to build architecture to be image of business model, not technical separation. I think it is really hard issue :)
16:05	mavbozo	kwladyka, try put all your business model code in separate namespaces first.
16:06	kwladyka	mavbozo i will try but... it will be glad to see how sombody did this
16:06	kwladyka	from my live experience it is always better to first look how other did things :)
16:15	kwladyka	ok, i am tired today, its time to a little rest and play bloodborne game :) thank you for your time and see you later! :)
16:17	dfinninger	Gah, 2 hours left until I can get back to Bloodborne!
16:21	bacon1989	is bloodborne like dark souls?
16:21	bacon1989	I don't know anything about that franchise
16:29	bacon1989	already i'm enjoying this applied clojure book
16:29	bacon1989	I didn't know about the map->record function
16:30	bacon1989	it was kind of why I shyed away from using records that were more than 3-4 parameters
16:57	brainproxy	any examples of a (relatively) simple transducer stack that terminates early via (reduced ...)
17:00	dfinninger	bacon1989: yup, basically a more action-heavy dark souls
17:12	sidhartha	anyone have a nice clojure based web user registration and cookie system they'd like to share knowledge about ?
17:17	sidhartha	ya know.. i'm looking at mozillas' persona ...
18:11	danielglauser	In Emacs I have a handy piece that highlights an expr when I'm on the open or close of it. Works for (, [, {, etc. Is there anything similar in Intellij/Cursive land?
18:13	cfleming	danielglauser: You mean it highlights the whole expr?
18:13	danielglauser	Yes
18:14	cfleming	danielglauser: Cursive will highlight the matching paren, and show the extent in the left gutter, but I don't think there's anything that will highlight the expr. Could you file an issue for that?
18:15	cfleming	danielglauser: I've actually been considering something similar for if forms, that would highlight the then and else clauses slightly differently, perhaps in subtle green/red
18:16	danielglauser	cfleming: Funny you mention that, this question was born out of that complaint, folks would like an easier way to see the else part of an if
18:16	cfleming	danielglauser: Which I'm hoping will catch some of my stupid "used when instead of if" bugs
18:16	cfleming	danielglauser: Oh nice
18:16	danielglauser	cfleming: Link to place to file an issue? Google is not my friend.
18:17	cfleming	danielglauser: https://github.com/cursiveclojure/cursive/issues
18:17	danielglauser	Oh cool, it's on Github. Sweet.
18:17	cfleming	danielglauser: Just the issues, not the source, sorry
18:17	danielglauser	Got it
18:19	cfleming	Some of the source will be up there sometime soon when I get the extension API out there
18:28	cfleming	danielglauser: Thanks
18:29	danielglauser	cfleming: np
18:29	cfleming	danielglauser: Nice github profile pic too :)
18:30	danielglauser	Thanks! Back when California still had water in it.
18:30	cfleming	Hehe
18:31	cfleming	I'm in New Zealand, kayaking is pretty huge here - I did some white water stuff a long time ago, but nothing like that
18:33	danielglauser	cfleming: Well, it has been a while since I've done drops like that. New Zealand is awesome for paddling, I've seem some great footage from there.
18:38	cfleming	danielglauser: Yeah, lots of nice spots. Lots of sea kayaking around where I am.
18:39	danielglauser	cfleming: Nice! I'm sure hiredman would enjoy that too.
18:55	wei	what’s an idiomatic way to update one element in a vector into two elements? e.g. [1 2 3 4 5] => [1 2 :a :b 4 5]
18:55	wei	(flatten (assoc-in [1 2 3 4 5] [2] [:a :b])) is one solution, but I’d rather not use flatten
18:58	TimMc	I think a general-purpose splice fn (as in JS) is called-for.
19:07	rukor	hi, pls how can one disable source maps for advanced compilation
19:14	dnolen	rukor: source maps are not enabled by default for advance compilation, also there is a :clojurescript channel
19:14	dnolen	er #clojurescript channel
19:16	justin_smith	wei: ##(mapcat (fn [[i x]] (if (= i 2) [:a :b] [x])) (map-indexed list [1 2 3 4 5]))
19:16	lazybot	⇒ (1 2 :a :b 4 5)
19:16	justin_smith	wei: with a helper function or maybe a macro that can be made a bit more comprehensible, of course
19:17	justin_smith	maybe this is better ## (mapcat (fn [i x] (if (= i 2) [:a :b] [x])) (range) [1 2 3 4 5])
19:17	lazybot	⇒ (1 2 :a :b 4 5)
19:19	rukor	dnolen: thanks, i actually thought i was typing in #clojurescript
19:20	wei	##((fn [i x coll] (concat (take i coll) x (drop (inc i) coll))) 2 [:a :b] [1 2 3 4 5])
19:20	lazybot	⇒ (1 2 :a :b 4 5)
21:30	Lewix	whats the meaning of seq can be infinite?
21:32	justin_smith	Lewix: a lazy-seq is represented so that you don't realize all the items, you just have a function that you know will get the next item
21:32	justin_smith	so for example iterate returns a seq that doesn't end
21:32	justin_smith	or cycle
21:33	justin_smith	,(nth (cycle [1 2 3]) 1000000)
21:33	clojurebot	2
21:33	Lewix	justin_smith: If I understand correctly seq that are operations such as map, filter, reduce, some , replace are lazy and infinite
21:34	justin_smith	Lewix: iterate is always infinite, cycle is always infinite, map filter and replace are all as long as their input, or shorter
21:34	justin_smith	reduce is not infinite except in very special circumstances
21:35	justin_smith	I've yet to use an indefinite return value from a reduce
21:36	Lewix	justin_smith: Are sequence backed by functions infinite?
21:36	justin_smith	they can be
21:36	justin_smith	lazy-seqs are backed by functions, they are not always infinite
21:36	justin_smith	reduce is not lazy, the others you mentioned are
21:36	Lewix	why is (def a-range (range 1 4)) considered infinite?
21:36	justin_smith	it isn't
21:36	justin_smith	but (range) is
21:37	Lewix	why is reduce not lazy while the others are?
21:37	justin_smith	because the point of reduce is to get a single value from a collection
21:38	justin_smith	if you put your mind to it, you could ensure that single value is actually an infinite sequence, but I haven't seen that in practice, and it wouldn't be derived 1->1 from the inputs like the other higher level transforming functions are
21:39	Lewix	(def a-range (range 1 4)) as opposed to (range 1 4) doesn't get created until it's called
21:39	justin_smith	that's false
21:39	justin_smith	both are equally lazy
21:39	justin_smith	neither are realized until the values are accessed
21:39	justin_smith	in the repl, def returns nil so you don't access the value at first
21:40	Lewix	,(range 1 4)
21:40	clojurebot	(1 2 3)
21:40	justin_smith	the side effect of printing is why that gets realized
21:40	justin_smith	,(do (range 1 4) nil)
21:40	clojurebot	nil
21:40	Lewix	,(def a-range (range 1 4))
21:40	clojurebot	#'sandbox/a-range
21:40	justin_smith	that sequence cannot ever get realized ^
21:40	justin_smith	the one in the do that is
21:41	Lewix	ok
21:41	justin_smith	better example: ##(do (map println (range)) nil)
21:41	lazybot	⇒ nil
21:41	Lewix	(second (range 1 4)) is different from (second a-range) - the former is invoked once while the other twice
21:41	justin_smith	no
21:41	justin_smith	that's wrong
21:42	Lewix	How?
21:42	justin_smith	lazy-seqs store their realized values
21:42	justin_smith	and def does not invoke any of the values
21:42	justin_smith	nor does using seq directly
21:43	justin_smith	it's only consuming values (eg with second, nth, print, count) that anything is realized
21:43	justin_smith	and the values are guaranteed to be realized exactly once
21:43	justin_smith	,(def foo (map println (range)))
21:43	clojurebot	#'sandbox/foo
21:43	justin_smith	,foo
21:43	clojurebot	(0\n1\n2\n3\n4\n5\n6\n7\n8\n9\n10\n11\n12\n13\n14\n15\n16\n17\n18\n19\n20\n21\n22\n23\n24\n25\n26\n27\n28\n29\n30\n31\nnil nil nil nil nil ...)
21:43	justin_smith	,foo
21:43	clojurebot	(nil nil nil nil nil ...)
21:44	justin_smith	see, the println only happens once
21:44	justin_smith	it's only invoked once
21:44	Lewix	justin_smith: so Alan Dipert got it wrong
21:44	justin_smith	I'd more likely believe you are misreading
21:44	justin_smith	but possibly
21:45	Lewix	justin_smith: thanks I'll go over it again
21:47	Lewix	justin_smith: so you're arguing that (second a-range) is invoked once
21:48	Lewix	just like (range 1 4)
21:48	Lewix	just like (second (range 1 4) )
21:49	Lewix	*
21:49	justin_smith	wait, what exactly do you mean by "invoked" here?
21:49	justin_smith	as in the thunks being realized? yes, in both cases invoked exactly once
21:50	Lewix	ok thanks
22:09	Lewix	justin_smith:
22:10	Lewix	justin_smith: "instead of being backed by data like linked lists are clojure sequences are backed by functions so they can be lazy and infinite" - this is a verbatim copy of what I heard
22:11	justin_smith	right
22:11	justin_smith	but the results of those functions are cached
22:11	justin_smith	so it's true, but perhaps incomplete
22:13	justin_smith	and they can be lazy and they can be infinite, but some seqs are neither
22:23	Lewix	justin_smith: that means that sequences are invoked twice
22:23	Lewix	justin_smith: (second (range 1 4))
22:23	justin_smith	I literally have no idea what you are talking about
22:23	justin_smith	where is the second invocation you are talking about here?
22:24	Lewix	in linked list you have to pass through the first item to get to the second
22:24	justin_smith	yes
22:24	justin_smith	so calling second realizes two elements
22:24	Lewix	invoked once for first the a second time for the second item
22:25	justin_smith	range is chunked, you would need to consume at least 33 items iirc for two distinc invocations to happen
22:25	Lewix	I meant that it's invoked twice in that specific case
22:25	justin_smith	no
22:25	justin_smith	it isn't
22:25	justin_smith	because range realizes 32 items at a time
22:25	Lewix	it what the same guy is saying ^^
22:25	Lewix	it got me thinking
22:25	justin_smith	it's an implementation detail
22:25	justin_smith	OK
22:25	justin_smith	,(first (map print (range)))
22:25	clojurebot	012345678910111213141516171819202122232425262728293031
22:26	justin_smith	note that it gave us the first item, but it realized 32 items
22:26	tomjack	whew. looks like I was a bit unfair to eclipse jdt: https://www.refheap.com/29f0551b9dab5c11bd8299f2f
22:26	Lewix	,(set! print-length 1000000000000)
22:26	clojurebot	1000000000000
22:26	Lewix	,(first (map print (range)))
22:26	clojurebot	012345678910111213141516171819202122232425262728293031
22:26	justin_smith	it's going to be 32 every time
22:27	justin_smith	like I said, it's an implementation detail, but unless you use anything beyond the 32nd item, there is exactly one invocation that happens
22:27	justin_smith	but if you rely on stuff like that you're doing it wrong anyway
22:27	Lewix	right justin_smith - i care about details though , so you're sayin gthat the 32 items are cached
22:28	justin_smith	right, there's exactly 1 invocation that realizes 32 items, in that case
22:28	Lewix	buok thanks
22:28	Lewix	for clarifying
22:28	justin_smith	which is one reason why laziness and side effects don't mix, and if your code relies on a specific number of items from some lazy-seq being realized, it's going to be buggy code
22:29	tomjack	user> (first (map print (range)))
22:29	tomjack	0
22:30	tomjack	user> clojure-version
22:30	tomjack	{:major 1, :minor 7, :incremental 0, :qualifier "beta1"}
22:30	tomjack
22:30	justin_smith	tomjack: yeah, I was careful to say "implementation detail" for a reason :)
22:30	justin_smith	tomjack: this was merely a counterexample to the assertion that you invoked a lazy-seq twice if you get the second item
22:31	justin_smith	if clojurebot was using beta1, I would pick a different example :)
22:31	tomjack	my point was, 'if you rely on it, your code may break' is not just a warning now
22:31	tomjack	it just broke! :)
22:31	justin_smith	oh, yeah
22:31	justin_smith	haha
22:31	justin_smith	true!~
22:31	tomjack	(also, I didn't know what the result was going to be..)
22:31	justin_smith	heh
22:32	justin_smith	tomjack: now try ##(first (map print [:a :b :c :d :e :f :g]))
22:32	lazybot	⇒ :a:b:c:d:e:f:gnil
22:33	justin_smith	gnil
22:33	justin_smith	gnil is a new nil that takes extra arguments including the standard --help command line flag
22:35	tomjack	same result
22:35	justin_smith	interesting
22:35	justin_smith	the purge of chunks
22:35	tomjack	no, I mean, same result as before 1.7
22:35	justin_smith	ahh
22:40	fitz	Normal Java interop can't see private fields, can it?
22:40	justin_smith	fitz: you can see them via reflection
22:42	fitz	justin_smith: that's different than just Classname/staticField, right?
22:42	justin_smith	right
22:42	justin_smith	it's trickier (and you can do the same thing in java too)
22:43	Lewix	juancate: tomjack ;i guess previous versions didn't have that constrain if I may call it so
22:45	Lewix	,clojure-version
22:45	clojurebot	{:major 1, :minor 7, :incremental 0, :qualifier "master", :interim true}
22:45	tomjack	ah, interesting
22:46	tomjack	user> (first (map print (range 100)))
22:46	tomjack	012345678910111213141516171819202122232425262728293031
22:46	tomjack
22:46	tomjack	(still 1.7.0-beta1)
22:46	justin_smith	oh, so (range) is special cased, but (range 100) is not
22:46	tomjack	(range) is now (iterate inc' 0)
22:47	justin_smith	aha
22:47	justin_smith	that explains it
22:49	Lewix	user=> (first (map print (range)))
22:49	Lewix	012345678910111213141516171819202122232425262728293031nil
22:49	Lewix	{:major 1, :minor 6, :incremental 0, :qualifier nil}
22:57	Lewix	justin_smith: what if i want to print 44 items with range
22:58	justin_smith	then it would realize 64 of them
22:58	justin_smith	well, depending on clojure version etc. of course
22:59	Lewix	justin_smith: I mean how do i do that since it's set to 32 and (set! print-length 1000000000000) wouldnt work
22:59	justin_smith	Lewix: clojurebot has weird settings
22:59	Lewix	I see
22:59	TEttinger	&(doseq [n (range 44)] (print n " "))
22:59	lazybot	⇒ 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 nil
23:00	TEttinger	lazybot has no print length
23:00	TEttinger	&(doseq [n (range 444)] (print n " "))
23:00	lazybot	⇒ 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 6... https://www.refheap.com/99487
23:00	TEttinger	lazybot also has no def, a tradeoff
23:12	tomjack	(inc cgrand)
23:12	lazybot	⇒ 2
23:13	tomjack	http://stackoverflow.com/questions/15013458/clojure-zipper-of-nested-maps-repressing-a-trie/15020649#15020649
23:53	Lewix	I'm looking for a benchmark of programming languages - slowest to fastest
23:54	Lewix	and any advantages/disadvantage of each programming languages. i would have thought that in 2015 we could easily find such a study ^^
23:55	justin_smith	there's the alioth.debian.org benchmarks game