2016-02-05
| 02:12 | macrolicious | how/can I access clojure language documentation inside emacs? ex. for 'recur' ? |
| 02:12 | justin_smith | ,(doc recur) |
| 02:13 | clojurebot | excusez-moi |
| 02:13 | justin_smith | macrolicious: anyway, there is a function clojure.repl/doc that is often helpful |
| 02:13 | justin_smith | there's some mode that makes it pop up automatically as part of cider |
| 02:14 | macrolicious | ok, thx |
| 02:43 | scottj | macrolicious: C-c C-d d when on a function name, if using cider. recur is a special form, so may not work like most functions/macros. |
| 02:43 | macrolicious | oh man, bam! |
| 02:43 | macrolicious | thanks, that's sweet. |
| 02:44 | macrolicious | every day, learning emacs is like finding something cool you didn't know you owned |
| 02:45 | ridcully_ | the same is true for vi, but there you discover it by typos |
| 02:48 | macrolicious | to be honest, I've found a few things by typos so far in emacs, too ;-) |
| 02:49 | macrolicious | well, key bindos |
| 04:11 | kungi | Anyone else experiencing problems when trying to connect to nrepl after updating to version 0.2.12? |
| 04:11 | kungi | I get the following error: SocketException The transport's socket appears to have lost its connection to the nREPL server |
| 04:13 | adam____ | I am using 0.2.12 and it seems to be working fine |
| 04:14 | adam____ | are you using cider? |
| 04:15 | kungi | adam____: I tried using cider and lein repl :connect |
| 04:15 | adam____ | yeah im not going to lie, the setup is a bit painful the first time |
| 04:15 | adam____ | do you have a .lein/profiles.clj |
| 04:15 | adam____ | {:repl {:plugins [[cider/cider-nrepl "0.11.0-SNAPSHOT"]] |
| 04:15 | adam____ | :dependencies [[org.clojure/tools.nrepl "0.2.12"]]}} |
| 04:16 | adam____ | those two version generally work for me |
| 04:16 | kungi | adam____: Ok wait a second. Let me clarify the problem a bit more. I have a working development setup. In my app I am using (nrepl/start-server) to start an nrepl server. |
| 04:17 | kungi | This one I cannot connect to any more |
| 04:17 | adam____ | hmm, sorry I don't know |
| 04:18 | adam____ | maybe check the github page and see if somebody had a similar issue? |
| 04:18 | kungi | adam____: tools.nrepl does not use github issues as far as I see |
| 04:19 | adam____ | http://stackoverflow.com/questions/20088908/problems-connecting-to-a-clojure-nrepl-with-ring-compojure |
| 05:01 | kungi | Downgrading the tools.nrepl version from 0.2.12 to 0.2.6 solved the problem. I will try to find the most recent working version. |
| 05:11 | CaptainLex | I am trying figwheel on a fresh lein install on the flappy bird demo and running "lein figwheel" hangs after dependencies are met |
| 05:11 | CaptainLex | Well, I don't know about hags, but it doesn't output anything |
| 05:13 | CaptainLex | Scratch that, lein is just misbehaving inside repos |
| 05:20 | CaptainLex | This is bonkers: |
| 05:20 | CaptainLex | '$ time lein version' |
| 05:21 | CaptainLex | 'Leiningen 2.6.0 on Java 1.8.0_72-internal OpenJDK Zero VM' |
| 05:21 | CaptainLex | 'real 0m23.441s' |
| 05:21 | CaptainLex | It takes 23 seconds just to output version |
| 05:21 | CaptainLex | What could possibly be causing lein to be so slow? |
| 05:21 | ridcully_ | -SNAPSHOT plugins? |
| 05:21 | MJB47_ | takes 4 seconds for me |
| 05:21 | MJB47_ | which while not as extreme, is still odd |
| 05:22 | jeaye | real0m2.064s |
| 05:22 | MJB47_ | is there a hidden progress bar we can disable? |
| 05:22 | CaptainLex | I can see why I would interpret this as hanging for anything more complex |
| 05:22 | CaptainLex | I'm going to try...restarting, I guess? |
| 05:22 | CaptainLex | Be right back |
| 05:27 | CaptainLex_ | Well, restarting definitely did not help |
| 05:27 | MJB47_ | out of interest, what hardware are you running? |
| 05:28 | MJB47_ | lets say cpu ram ssd/hdd to keep it short |
| 05:29 | CaptainLex_ | MJB47_: Nothing super respectable- it's a Chromebook. ARM processor, 2 gigs of RAM, 16 gigs of SSD |
| 05:31 | CaptainLex_ | MJB47_: Just running up Clojure's JAR on Java doesn't take hardly any time at all, for what that's worth |
| 05:32 | MJB47_ | thats to be expected |
| 05:33 | MJB47_ | i was just wondering why it took you 20 something and me 4 seconds |
| 05:33 | MJB47_ | i have a laptop i7, which might explain it |
| 05:33 | MJB47_ | but why it would take so long is rather odd |
| 05:34 | CaptainLex_ | MJB47_: Yeah, 36 seconds to run lein help is more than one can reasonably chalk up to hardware issues, at least to my mind |
| 05:34 | MJB47_ | indeed |
| 05:35 | CaptainLex_ | My guess is that something in the environment is out of date, but I grabbed lein from the website and it should be pulling in everything it needs except Java itself |
| 05:35 | CaptainLex_ | And I think Java's a good version: 'openjdk version "1.8.0_72-internal"' |
| 05:38 | MJB47_ | im using Java 1.8.0_25 Java HotSpot(TM) 64-Bit Server VM |
| 05:39 | CaptainLex_ | Wow, it just took it 10 seconds to tell me "fighwheel" was not a task. It is probably too much to hope that I can get anything done under these circumstances. What's the best way to go about reinstalling lein and its dependencies? |
| 05:45 | CaptainLex_ | Oh look, the figwheeling has picked up! I guess for the purposes of the project I need to do I can just be patient |
| 05:45 | CaptainLex_ | and just not worry too hard about doing any real dev until I get home to my real computer |
| 06:16 | gfdhjf | bad uncle is inserting his penis into 10 year old girl's vagina causing her to moan in agony because they both are niggers |
| 07:01 | hyPiRion | amalloy_: ping |
| 07:12 | bcoburn | I have a project with three namespaces, call them A, B, and C. A requires B and C. I have a single that I want to have available in all 3 namespaces. What's the easiest way to accomplish this? |
| 07:13 | hyPiRion | bcoburn: a single what? |
| 07:14 | bcoburn | ph |
| 07:14 | bcoburn | function |
| 07:15 | bcoburn | typing is hard |
| 07:16 | hyPiRion | The easiest way to solve that is to make a new namespace D (usually it's called utils) and require that in all the namespaces. |
| 07:16 | hyPiRion | But you can also pass the function in as a parameter if that's suitable for your use case. |
| 07:18 | bcoburn | I think I'll do the new namespace plan, just sad at having to make another file etc for ~4 lines of code |
| 07:20 | hyPiRion | yeah, it feels painful at first. It'll be better if/when you end up having more utility functions though |
| 08:49 | python476 | hi there |
| 08:52 | python476 | cider github page mention C-c C-i to inspect value, I believe they meant C-c M-i (using cider 0.10.2, maybe the main page document a previous version though) |
| 08:53 | shiranaihito | what's this about: "IllegalArgumentException: interface c2d.struct.IJavaize is not a protocol" -- "IJavaize" is very much a protocol.. i'm trying to extend it to a Record in another namespace |
| 08:54 | python476 | maybe I should "complain" on github ~_~ |
| 08:55 | hyPiRion | python476: yeah, I guess that's just a typo |
| 08:55 | python476 | aight |
| 08:55 | python476 | it's not an life critical feature, but as any documentation typo, it makes you question your sanity |
| 09:03 | hyPiRion | heh |
| 09:14 | dysfun | so has anyone tried using emscripten with clojurescript? |
| 09:14 | dysfun | (to use a c library from clojurescript) |
| 09:59 | dysfun | justin_smith: i've changed several projects to using group.package instead of package.core . are you happy now, free-speech-hater? ;) |
| 11:03 | kwladyka_ | justin_smith ha i understand loggers! thank you for you help and patient. Anyway what do you think about https://github.com/ptaoussanis/timbre to use for logs? |
| 11:50 | python476 | hi again |
| 11:51 | python476 | any idea why lein check would hangs at `Compiling namespace ...` ? |
| 11:51 | python476 | the codebase is two tiny files |
| 11:53 | python476 | almost no cpu usage, no io ~_~;; |
| 11:55 | dysfun | how long has it been hung? |
| 11:55 | python476 | forgot some effectul network code.. I noobed |
| 11:55 | python476 | dysfun: minutes |
| 11:55 | dysfun | well that's not supposed to happen, obviously |
| 11:55 | dysfun | but if you know your code did it... :) |
| 11:56 | python476 | the price of free lisp code, I tested a few stuff on the go, and forgot to comment / delete http requests... |
| 12:02 | blackbird_ | hi, maybe i'm screwing up my thinking here, but if i've def'd a map, and then call a macro with that def'd map, how do i get that map in the macro? is that possible? |
| 12:02 | python476 | no more idl-ing but still a good minute, odd |
| 12:04 | blackbird_ | here is an example: https://www.refheap.com/114516 |
| 12:17 | amalloy | hyPiRion: ? |
| 12:28 | rhg135 | blackbird_: macros run at compile-time and thus only gets a symbol instead of the map |
| 12:31 | blackbird_ | rhg135: so is it impossible to get the value of a symbol in a macro? shouldn't something that is def'd be there at compile time? |
| 12:34 | rhg135 | For eagerly using it with the unquote, no. But if you return a form that operates over the value at runtime, that would work |
| 12:36 | blackbird_ | rhg135: ok, thanks for the info! |
| 12:37 | rhg135 | ,(do (def x 1) (defmacro y [i] `(do ~(inc i) nil)) (y x)) |
| 12:37 | clojurebot | #error {\n :cause "clojure.lang.Symbol cannot be cast to java.lang.Number"\n :via\n [{:type java.lang.ClassCastException\n :message "clojure.lang.Symbol cannot be cast to java.lang.Number"\n :at [clojure.lang.Numbers inc "Numbers.java" 112]}]\n :trace\n [[clojure.lang.Numbers inc "Numbers.java" 112]\n [sandbox$y invokeStatic "NO_SOURCE_FILE" 0]\n [sandbox$y invoke "NO_SOURCE_FILE" 0]\n [clo... |
| 12:38 | rhg135 | ,(do (def x 1) (defmacro y [i] `(do (inc ~i) nil)) (y x)) |
| 12:38 | clojurebot | nil |
| 12:40 | blackbird_ | rhg135: ahh, i can see how that works, still unclear as to why x the value of x there isn't known at compile time, guess i need to read up on what happens when. |
| 12:44 | rhg135 | It might be known, but the macro gets a symbol. You can do resolve, or other such things, to look it up manually. But that probably is frowned upon. |
| 12:47 | blackbird_ | rhg135: ok thanks |
| 13:00 | WorldsEndless | Is my suspicion correct that {:keys} can't be used twice in the same let statement? |
| 13:00 | justin_smith | WorldsEndless: no |
| 13:00 | WorldsEndless | hmm... I guess I need to continue debugging |
| 13:01 | hiredman | http://clojure.org/reference/special_forms#binding-forms |
| 13:01 | justin_smith | ,(let [{:keys [a b c] {:keys [e f g]} :d} {:a 0 :b 1 :c 2 :d {:e 3 :f 4 :g 5}}] [a b c e f g]) |
| 13:02 | clojurebot | [0 1 2 3 4 ...] |
| 13:03 | WorldsEndless | Doh! yet another case of an upload coming through ring that appears to be a map, but is actually a string |
| 13:03 | justin_smith | WorldsEndless: sounds like you need a wrap-json middleware or something |
| 13:04 | WorldsEndless | Transit is what I'm going for... |
| 13:57 | KevinCorcoran | Is EuroClojure going to happen this year? |
| 14:11 | arcatan | I'd assume so |
| 14:11 | python476 | so, I learned about (shutdown-agents)... that's what was blocking lein check |
| 14:11 | python476 | also (prn ...) isn't threadsafe |
| 14:12 | justin_smith | python476: OK, nothing in your namespace should make shutdown-agents neccessary |
| 14:12 | LucidTortoise | Anyone here uses Light Table as their primary editor? |
| 14:12 | justin_smith | def shouldn't have side-effects |
| 14:12 | python476 | justin_smith: toying with (future ...) |
| 14:12 | hiredman | python476: that means you are launching new threads (agents or futures) as a side effect of code loading, that is generally a bad practice |
| 14:12 | justin_smith | python476: yeah, that should be inside your -main or an init of some sort |
| 14:13 | python476 | sure, but I just I put everything inside functions, and only -main is running them |
| 14:13 | justin_smith | python476: then why would you need shutdown-agents in order for lein check to return? |
| 14:13 | justin_smith | lein check doesn't run -main |
| 14:13 | hiredman | python476: either you didn't do that, or some library you are using didn't |
| 14:14 | python476 | right sorry, previously, these expressions weren't in (defn...), and blocking lein check |
| 14:14 | justin_smith | OK, that makes more sense |
| 14:14 | hiredman | ah |
| 14:14 | python476 | I thought it was a bad timer variable or something but now I see that was normal behavior for (future...) without (shutdown-agents). right ? |
| 14:14 | justin_smith | right |
| 14:15 | python476 | cool |
| 14:15 | hiredman | prn, println, print, etc don't do any synchronization on anything |
| 14:15 | python476 | I put side effects into my source code. I'm down to php3 level |
| 14:15 | python476 | hiredman: thanks, I assumed they were, scrambled output is a rare thing to witness :) |
| 14:16 | justin_smith | python476: (def outcome (launch-missiles)) |
| 14:18 | python476 | yum |
| 14:23 | python476 | ha, concurrency is so fun http://stackoverflow.com/questions/5181367/is-defn-thread-safe |
| 14:56 | binjured | is there a simple/idiomatic way to chain-filter a collection into different groups? for instance, i have a collection where certain predicates need to come before others and filter out values that would have matched later predicates but belong in the former group, if that makes any sense. it's like sifting, maybe? |
| 15:06 | irctc | Can I read in a file using clojurescript? |
| 15:11 | dnolen | irctc: ClojureScript has not I/O facilities you have to rely on whatever the JS environment you're targeting provides |
| 15:11 | dnolen | side note there is a #clojurescript specific channel |
| 15:11 | dnolen | s/not/no |
| 15:16 | ridcully_ | there is noe there ?! |
| 15:20 | rhg135 | any reason why my `lein repl` would time out on a slow fs if target is on a fast one (it doesn't write elsewhere afaik). |
| 15:21 | rhg135 | as an aside launching a clojure.main repl with that classpath loads really fast. so I assume lein is loading something. |
| 15:21 | ridcully_ | network? |
| 15:21 | rhg135 | yes, sshfs |
| 15:21 | rhg135 | target is symlinked to my home dir |
| 15:26 | rhg135 | hmm, apparently if I require seesaw.cor., a dep, java segfaults |
| 15:29 | python476 | wasn't there a lein plugin to avoid clojure bootstrapping time ? |
| 15:30 | rhg135 | no, but there is LEIN_FAST_TRAMPOLINE |
| 15:30 | python476 | I don't wanna be a circus employee |
| 15:31 | justin_smith | python476: you can not use lein, then you'll notice that most of that "clojure bootstrapping time" was actually lein startup |
| 15:31 | ridcully_ | python476: are you anti circi? |
| 15:32 | rhg135 | and ragequit life when trying to gater dependencies by hand |
| 15:32 | python476 | ridcully_: I'm somehow noobish on clojure (especially the ecosystem) |
| 15:33 | python476 | justin_smith: that's also possible indeed, let's go standalone |
| 15:33 | rhg135 | hey, trampoline repl starts. justin_smith, I'm fairly certain the splash is a lie |
| 15:33 | justin_smith | python476: there's a middle ground, where you let lein calculate deps once, then you reuse those deps without running the whole lein thing again |
| 15:33 | rhg135 | nrepl timesout normally |
| 15:34 | justin_smith | python476: that is what that FAST_TRAMPOLINE option mentioned above does |
| 15:34 | rhg135 | mhm |
| 15:34 | justin_smith | ~faster |
| 15:34 | clojurebot | faster is https://github.com/technomancy/leiningen/wiki/Faster |
| 15:34 | justin_smith | see above for elaborate documentation of these issues ^ |
| 15:35 | rhg135 | idk why it's not the default |
| 15:35 | python476 | justin_smith: thanks, also, any reference doc about clojure dev without lein (should I dig clojure's website ?) |
| 15:35 | justin_smith | rhg135: there's subtle issues, for example I've seen "lein trampoline repl" eat stack traces that are shown with "lein repl" on startup |
| 15:35 | rhg135 | if you do want to excise lein for production, you'd make a uberjar |
| 15:36 | justin_smith | python476: "java -jar clojure.jar" will give you a repl, "rlwrap java -jar clojure.jar" gives you a nicer repl |
| 15:36 | justin_smith | clojure.jar can be found via maven, or the clojure.org site |
| 15:37 | python476 | justin_smith: this part I knew, I meant running namespaced -main(s), tests |
| 15:37 | python476 | (sorry, I sounded impatient) |
| 15:37 | rhg135 | or lein trampoline repl lying about nrepl |
| 15:37 | python476 | I'm actually reusing lein .m2 cache |
| 15:38 | justin_smith | python476: java -cp clojure.jar:src clojure.main -m my.ns; java -cp clojure.jar:src:test clojure.main -m my.ns.test |
| 15:38 | python476 | ha, -m |
| 15:38 | python476 | of course u_u |
| 15:38 | justin_smith | yeah, the output of lein cp can be cached and provided to the -cp arg |
| 15:39 | python476 | it's indeed a tad more reactive |
| 15:40 | python476 | I can reuse emacs+make to wrap things up and run with a shortcut |
| 15:40 | python476 | rhg135: I'll take a loop at that a bit later |
| 15:41 | justin_smith | python476: this is really what trampoline with lein does, when you have the LEIN_FAST_TRAMPOLINE env var |
| 15:41 | rhg135 | ^ |
| 15:42 | rhg135 | I just set it in my .zshenv |
| 15:43 | rhg135 | I can just unset it if I ever need the default |
| 15:43 | python476 | no value attached to LEIN_FAST_TRAMPOLINE, just its existence is enough ? |
| 15:43 | justin_smith | python476: the value doesn't matter, but I think it needs to be non-empty |
| 15:44 | python476 | LEIN_FAST_TRAMPOLINE='#golang' |
| 15:44 | justin_smith | lol |
| 15:44 | rhg135 | heh, that works |
| 15:46 | python476 | well, it's not uber faster |
| 15:46 | python476 | but I've learned a few things |
| 15:46 | phorse | is compojure-api meant to be a standalone server sort of deal, or is it typically grafted onto a regular application server? |
| 15:46 | python476 | gonna replenish electrolytes, my scales are all cracky |
| 15:47 | rhg135 | lein trampoline run -m clojure.main |
| 15:48 | justin_smith | python476: trampoline is faster after the second time |
| 15:48 | justin_smith | python476: the first time it's still caching |
| 15:48 | python476 | justin_smith: I'm not that noobish |
| 15:48 | justin_smith | phorse: the great thing is it doesn't care |
| 15:48 | python476 | I did run it a few time, it's a cache thing |
| 15:48 | python476 | I'll forgive you, here https://pbs.twimg.com/media/CZLC9yPW0AArfnl.png |
| 15:48 | justin_smith | it can be standalone (a few different embedded server options), or it can run in a container |
| 15:48 | clojurebot | No entiendo |
| 15:49 | python476 | (Super No entiendo) |
| 15:53 | koshmar | is the clojure a good lang for scientific computing - solving PDE in parallel. are numerical libraries mature? if so, what is the best tutorial to start picking numerical oriented clojure? (I have no experience with java or lisp, however I do know haskell if it helps) |
| 15:55 | justin_smith | koshmar: core.matrix has a protocol for numeric stuff, with a few implementations you could check out |
| 15:55 | justin_smith | koshmar: honestly clojure can be tricky because of auto-boxing issues, but core.matrix can help with that |
| 15:55 | justin_smith | *good numeric performance in clojure |
| 15:57 | koshmar | what about maturity of core.matrix? |
| 15:58 | koshmar | could I start learn core.matrix skipping clojure tutorials, or at least learning very basik clojure |
| 16:01 | rhg135 | why is '(try (clojure.core/require (quote nil)) (catch java.lang.Exception t__11836__auto__))' in the init script lein makes?? |
| 16:06 | justin_smith | koshmar: I think so, seeing as you have haskell experience |
| 16:07 | justin_smith | koshmar: some of the exercises on 4clojure.com might help you figure out what's what |
| 16:07 | justin_smith | koshmar: core.matrix has been around a while, at the least it's the most mature clojure option for numerics |
| 16:08 | justin_smith | the implementations are not all the same - there's tradeoffs for performance vs. safety, but it should be easy enough to swap them out |
| 16:34 | algal | Hello, chaps. |
| 16:34 | algal | I've got a possibly dumb question. |
| 16:34 | algal | If I want to trigger some side-effect only when a swap! succeeds in modifying the value of an atom, what's the recommended way to do that? |
| 16:34 | algal | Should I put the side-effecting code inside the update function I am passing to swap? Or should I add a watch to the atom? |
| 16:35 | algal | Actually, now that I think on it, a watch would not work, because the parameter for the side-effecting function also depends on the call to swap. Hmm. |
| 16:36 | hiredman | swap! always succeeds |
| 16:36 | algal | So this suggests I should put the side-effecting fn call inside the update fn which I pass to swap!, but that seems odd, because I think of that function as something that should be pure as the driven snow. |
| 16:36 | hiredman | in that swap! retries the function you give it until the compare and swap succeeds |
| 16:36 | algal | Right, sorry. I meant that I am passing an update function to swap that sometimes returns a modified value and sometimes returns an unmodified value |
| 16:37 | algal | And I only want to call the side-effect when that update function is returning a modified value. |
| 16:43 | justin_smith | (let [a @v b (swap! v f)] (if (= a b) ... ...)) ? |
| 16:43 | python476 | is 6 seconds normal for LEIN_FAST_TRAMPOLINE ? |
| 16:43 | justin_smith | python476: what task |
| 16:44 | python476 | lein run, still that same small clj script |
| 16:44 | justin_smith | repl booting to user? way too long, repl booting to your ns? depends on what your deps are |
| 16:44 | justin_smith | lein run, depends what you code is doing of course |
| 16:44 | python476 | a tiny 10 iteration loop + prn |
| 16:45 | python476 | and no, i'm not running on a 386 :) |
| 16:45 | justin_smith | yeah, 6 seconds is way too long there |
| 16:45 | justin_smith | python476: how are you running it? |
| 16:45 | python476 | sad trampoline |
| 16:45 | justin_smith | "lein trampoline run" ? |
| 16:45 | python476 | ha ffsssss |
| 16:45 | justin_smith | ? |
| 16:46 | python476 | down to 1s |
| 16:46 | python476 | I'm facepalming hard |
| 16:46 | justin_smith | oh did you forget the part where you actually trampoline? |
| 16:46 | python476 | I thought the env var was enough for lein to trigger the rest |
| 16:46 | python476 | I hope you appreciate fixing noobs typing before they understand things |
| 16:47 | python476 | while we're at that, what's your favorite(s) book / tutorial (on any part of clojure) |
| 16:47 | justin_smith | clojure applied is great |
| 16:48 | python476 | real problems no 'yet another language reference' ? |
| 16:49 | justin_smith | it gets a little deeper into how to do clojure right than most books I've seen |
| 16:49 | python476 | very nice |
| 16:49 | python476 | 25$ e-book |
| 16:50 | justin_smith | python476: and while still learning, http://conj.io is a great thing to leave open at all times while developing |
| 16:50 | justin_smith | and to skim occasionally so you don't forget the awesome stuff in core |
| 16:51 | python476 | yeah, even if 'tiny' it's still large in some way |
| 16:52 | phorse | I'm trying to use compojure.api.sweet, but none of the symbols defined in that library are being recognized by the compiler. I've got the correct dependency in my project.clj and in my namespace declaration, is there anywhere else I should check? |
| 16:52 | algal | justin_smith: So the thing about that idea << (let [a @v b (swap! v f)] (if (= a b) ... ...)) ? >>, is that there could be changes in v in between the binding (a @v) and the binding (b (swap! v f)] (if (= a b) ... ...)) . Right? |
| 16:52 | justin_smith | python476: even now I sometimes accidentally re-implement some function from core just because I forget it's there |
| 16:52 | justin_smith | algal: yes, this is unavoidable |
| 16:53 | justin_smith | perhaps you want a ref |
| 16:53 | justin_smith | refs can do coordinated changes, and this sounds like something you would need coordination to be sure of |
| 16:53 | algal | But I am not trying to produce coordinate updates. I'm trying to produce a side-effect if there is a certain kind of udpate. I'm only updating the one atom. |
| 16:54 | python476 | justin_smith: that's how fun clojure is |
| 16:54 | algal | hmm.. |
| 16:54 | justin_smith | algal: but you need coordination in order to know whether your swap! actually changed the object or not |
| 16:55 | algal | Well, the update function which I pass to the swap knows if it changes its arguments, so it knows. So that was why I was wondering about sticking the side-effecting call in there. But that seemed odd.. |
| 16:55 | justin_smith | algal: (dosync (let [a @v b (alter v f)] (= a b)) |
| 16:56 | justin_smith | that actually allows you to know if the altering function changed the value |
| 16:56 | justin_smith | there's no equivalent with an atom (or at least nothing simple) |
| 17:03 | amalloy | watchers? |
| 17:04 | aztak | g'd evening all. |
| 17:04 | amalloy | seems like exactly what algal wants, not refs |
| 17:04 | justin_smith | amalloy: oh, yeah, I guess a watcher could do that, since it gets the before and after |
| 17:04 | amalloy | you can also do it by hand on an atom, by changing what the atom stores: instead of just storing the current value, you can make it store both the previous and current value in a tuple |
| 17:05 | amalloy | that means you have to change all readers and writers, but it can be done without the coordination of a ref |
| 17:05 | justin_smith | amalloy: "I hate the other people who use this value anyway" |
| 17:05 | aztak | anyone "going to" clojure-remote next week? |
| 17:05 | justin_smith | oh, right |
| 17:05 | amalloy | (honestly you can do a surprising amount with atoms if you make them into tuples) |
| 17:06 | justin_smith | simplifying update logic, at the cost of an extra lookup with every usage |
| 17:06 | justin_smith | (every non-update usage) |
| 17:07 | algal | amalloy: ah, interesting. |
| 17:08 | amalloy | but if your goal is to just sometimes do a side effect based on what happened in the swap, that's exactly the case watchers were designed for |
| 17:10 | algal | amalloy: So I need the parameters I pass to my side-effecting function to be different for different calls to swap! So that makes it harder to use a watcher. |
| 17:10 | justin_smith | algal: why would that matter? |
| 17:10 | algal | But I guess I can (again) pass more information into the atom, like a tuple, just to expose that extra piece of info to the watcher. |
| 17:10 | sdegutis | justin_smith: What's your favorite Clojure HTTP library? |
| 17:10 | justin_smith | algal: the extra parameters change how the side effect should happen? |
| 17:11 | phorse | When I use (dir compojure.api.sweet) I get a list of 18 functions. When I check the api docs for compojure.api.sweet, they list 21 functions with only some overlap between the two. Is the documentation out of date? |
| 17:11 | justin_smith | sdegutis: I kind of hate all of them right now sorry, I'll let you know if I start liking one |
| 17:11 | algal | yes. But the extra parameters are not strictly needed to compute the change in the atom's state. |
| 17:11 | sdegutis | :D |
| 17:11 | sdegutis | justin_smith: Any reason? |
| 17:12 | justin_smith | sdegutis: my app needs to suck up the internet and the internet is terrible |
| 17:13 | sdegutis | :D |
| 17:13 | sdegutis | I just need to interface with MailChimp so I'm gonna stick with clj-http. |
| 17:17 | amalloy | justin_smith: re the internet being terrible: today at lunch my company had joel spolsky come give us a talk entitled "how to have nice things" (in contrast to the usual "this is why we can't have nice things") |
| 17:17 | justin_smith | amalloy: interesting |
| 17:18 | justin_smith | I've learned that scraping data for urls and trying to check those urls en mass can lead to lots of stupid stuff (who would have guessed, right?) |
| 17:19 | amalloy | no kidding. even if they are all valid urls to sites with actual data |
| 17:19 | amalloy | why do you want to slurp up the whole web, again? |
| 17:19 | justin_smith | amalloy: we do social analytics, our client wants to know what's happening regarding a certain topic, finding and following rss and blogs from social media is part of this |
| 17:20 | justin_smith | eg. figuring out the hot content, and part of that is figuring out two people were talkign about the same content even if via two different clickbait sources... |
| 17:21 | justin_smith | amalloy: and then some guys' gonna be lick "this is my rss feed" and it's actually like an xml bomb or something |
| 17:22 | amalloy | yep |
| 17:22 | lockdown | amalloy: material things? |
| 17:22 | lockdown | you guys are not giving in to slack? ;) |
| 17:23 | justin_smith | ~slack |
| 17:23 | amalloy | the talk was about why creating sustainable internet communities is hard, and what kind of stuff you should expect to have to do if you want to try it |
| 17:23 | clojurebot | It's greek to me. |
| 17:23 | amalloy | i think slack is great. it's a way to trick people without neckbeards into using an irc-like |
| 17:24 | aztak | This is good: https://drive.google.com/file/d/0B8ZX1RoWHuiJSnZHZFJqOERqaDg/view |
| 17:26 | rhg135 | I didn't know I had a neckbeard |
| 17:27 | aztak | where's the largest Clojure community then, on IRC or on Slack? (They are obviously not exclusive) |
| 17:28 | rhg135 | I've never been on slack, but it seems to be quite active |
| 17:29 | justin_smith | aztak: 4859 signed into clj slack, usually about 800 here |
| 17:29 | aztak | justin_smith: so metrics says "Slack" then? :) |
| 17:29 | aztak | too bad, I've just learnt so configure ERC in Emacs :( |
| 17:30 | justin_smith | also too bad because I don't trust slack, and prefer irc |
| 17:30 | lockdown | justin_smith: of those 4859 normally less than 500 are online |
| 17:31 | lockdown | so this channel is bigger! |
| 17:31 | justin_smith | lockdown: oh, where's that number? |
| 17:31 | rhg135 | and 'hipsters' |
| 17:31 | aztak | hehe |
| 17:32 | lockdown | justin_smith: click in the number |
| 17:32 | lockdown | slack is so laggy, don't know if it is because I use firefox instead of chrome |
| 17:32 | rhg135 | probably |
| 17:33 | Bronsa | can we just stop talking about slack? :) |
| 17:33 | rhg135 | it's horribad on some js |
| 17:34 | lockdown | Bronsa: be we feel slighted |
| 17:34 | lockdown | :P |
| 17:35 | aztak | either way - it's great that there is a big and diverse Clojure community. |
| 17:35 | aztak | For neckbeards *and* for hipsters! |
| 17:35 | aztak | :) |
| 17:36 | rhg135 | yay neckbeards :) |
| 17:36 | rhg135 | s/neckbeards/clojure/ |
| 17:37 | lockdown | when you do in-ns in a repl to switch to an existing ns, the ns macro of that ns is not parsed right? you have to use ns functions in the repl to replicate the ns macro? |
| 17:37 | lockdown | justin_smith: so yeah, at any given time there are more users here |
| 17:37 | rhg135 | nope, the repl just switches to it |
| 17:37 | rhg135 | doesn't even have to exist |
| 17:37 | justin_smith | lockdown: the ns macro is a wrapper for in-ns |
| 17:38 | justin_smith | lockdown: you probably want to make sure you have required your ns before doing the in-ns |
| 17:38 | justin_smith | or use ns, yes |
| 17:38 | amalloy | (ns foo) is no better than (in-ns 'foo) |
| 17:38 | justin_smith | lockdown: also, solution to a common problem (clojure.core/refer-clojure) |
| 17:39 | justin_smith | saves you from the situation of having switched to an empty ns that doesn't even have access to clojure.core |
| 17:39 | rhg135 | I just stay in user and alias it |
| 17:39 | lockdown | ok thanks |
| 17:39 | rhg135 | keeps dev utils handy that way |
| 17:40 | lockdown | yeah, looks staying in user and using require seems the way to go |
| 17:41 | lockdown | so if you require a ns from user, all the ns macro of the namespace is loaded? |
| 17:41 | rhg135 | I wish CIDER had a better way to do that than typing it manually in the repl |
| 17:41 | rhg135 | yes |
| 17:41 | lockdown | and attached to the repl? |
| 17:41 | lockdown | ok |
| 17:42 | justin_smith | rhg135: doesn't cider have a "load this file" keystroke? |
| 17:42 | rhg135 | I may as well use inf-clj if not for the debugger |
| 17:43 | amalloy | rhg135: of doing what? |
| 17:43 | rhg135 | yeah, justin_smith, I meant to send '(alias <something> <ns name>) |
| 17:43 | justin_smith | ahh |
| 17:43 | lockdown | justin_smith: yeah, I hit that problem already (the refer-clojure) ;), the repl only loads clojure.core for user |
| 17:44 | rhg135 | when the ns name is 50 chars long it gets tedious |
| 17:44 | amalloy | rhg135: i don't understand what tedious task you're doing that you can't fix with ten seconds of elisp |
| 17:44 | amalloy | or by using require :as instead of require + alias |
| 17:44 | lockdown | justin_smith: makes sense since user is only for trying things out |
| 17:45 | amalloy | ,(macroexpand-1 '(ns foo)) |
| 17:45 | rhg135 | I'm new to emacs though, amalloy, and elisp is greek to me |
| 17:45 | clojurebot | (do (clojure.core/in-ns (quote foo)) (clojure.core/with-loading-context (clojure.core/refer (quote clojure.core))) (if (.equals (quote foo) (quote clojure.core)) nil (do (clojure.core/dosync (clojure.core/commute (clojure.core/deref (var clojure.core/*loaded-libs*)) clojure.core/conj (quote foo))) nil))) |
| 17:45 | amalloy | it's nothing to do with the user ns |
| 17:45 | amalloy | (ns) expands to code that calls (refer-clojure) |
| 17:45 | lockdown | ~emacs |
| 17:45 | clojurebot | emacs is an out-moded belief system |
| 17:45 | lockdown | ~cursive |
| 17:45 | clojurebot | cursive is that thing what colin wrote |
| 17:46 | justin_smith | ~ed |
| 17:46 | clojurebot | ed is the standard editor! |
| 17:46 | rhg135 | when cursive doesn't require X call me |
| 17:46 | lockdown | amalloy: yeah, ns does the refer-clojure implicitly |
| 17:46 | justin_smith | rhg135: it'll run without X on osx or windows no problem :P |
| 17:47 | rhg135 | a real masochist uses a steady hand and a needle, justin_smith |
| 17:47 | rhg135 | not ed ;) |
| 17:48 | rhg135 | justin_smith: true heh |
| 17:48 | lockdown | ~vim |
| 17:48 | clojurebot | Gesundheit! |
| 17:48 | lockdown | ~atom |
| 17:48 | clojurebot | I don't understand. |
| 17:48 | rhg135 | ~editor |
| 17:48 | clojurebot | Huh? |
| 17:48 | tolstoy | Ugh. Having the worst lock with eastwood. |
| 17:49 | amalloy | i just realized: do they have something like clojurebot in the slack channel? |
| 17:49 | lockdown | tolstoy: clint is a tough guy |
| 17:49 | justin_smith | amalloy: yup |
| 17:49 | justin_smith | and on slack you don't even need one liners |
| 17:49 | tolstoy | lein eastwood: Caused by: java.lang.ClassNotFoundException: leiningen.core.project$reduce_dep_step |
| 17:49 | amalloy | requiring one-liners is a feature, to stop people pasting giant essays |
| 17:49 | justin_smith | tolstoy: oh man, clearly reduce-dubstep was a terrible idea |
| 17:50 | lockdown | amalloy: yeah, you normally see walls of code there |
| 17:50 | rhg135 | I have a char limit on mine |
| 17:51 | rhg135 | and if you trip the alarm you get kicked/banned |
| 17:51 | tolstoy | justin_smith: Yeah. What's even calling it? Oy. ;) |
| 17:55 | tolstoy | Eastwood is broken using Lein 2.6.0? Oy. I don't even care anymore. If my code doesn't bomb, it's perfect! |
| 17:55 | rhg135 | 2.6.0 is out?? |
| 17:56 | tolstoy | Yep. |
| 17:57 | tolstoy | https://groups.google.com/forum/#!topic/clojure/XlROLH76yNg |
| 17:57 | justin_smith | tolstoy: "lein upgrade 2.5.3" is also an option |
| 17:57 | justin_smith | 2.6.0 also broke lein-ring |
| 17:57 | tolstoy | Or downgrade? |
| 17:58 | tolstoy | Yeah, but I was having issues with eastwood anyway (a protocol with two many methods?). And there are 62 issues. I'll come back next year. |
| 17:58 | justin_smith | tolstoy: the command is "upgrade", ¯\_(ツ)_/¯. |
| 17:59 | tolstoy | Even when the announcement says, "As usual, for those who manually installed, `lein upgrade` will pull in |
| 17:59 | tolstoy | the latest. `lein downgrade 2.5.3` will back it down to the previous |
| 17:59 | tolstoy | version if you run into any issues." |
| 17:59 | justin_smith | oh, well why remember two commands when one command works |
| 17:59 | sdegutis | Hi. |
| 18:04 | tolstoy | Indeed! |
| 18:06 | Somelauw | I'm a bit of a beginner. When using Emacs, I get the following warnings on start-up https://pastee.org/xn98p telling me nREPL is outdated. How do I get this up to date? |
| 18:06 | justin_smith | Somelauw: https://github.com/clojure-emacs/cider-nrepl |
| 18:07 | justin_smith | that has comprehensive instructions for making sure you have the right version |
| 18:07 | lockdown | Somelauw: make sure cider and nrepl-cider versions are in sync |
| 18:08 | lockdown | Somelauw: also, use the latest version of leiningen, which pulls nrepl 0.2.12 |
| 18:08 | lockdown | so cider doesn't complain |
| 18:08 | justin_smith | lockdown: as a plugin, cider overrides nrepl version anyway |
| 18:09 | Somelauw | so might just updating lein be enough? |
| 18:09 | justin_smith | that won't fix your cider-nrepl version, no |
| 18:09 | rhg135 | can someone explain the advantage of nrepl middleware over just sending code to eval? |
| 18:09 | lockdown | justin_smith: yeah, that's another way to do it |
| 18:09 | Somelauw | rhg135: probably autocompletion |
| 18:10 | rhg135 | you can def. autocomplete without it. |
| 18:10 | lockdown | Somelauw: if you installed cider from melpa you also need nrepl-repl master branch |
| 18:10 | rhg135 | afaik, neither cursive nor fireplace need middleware |
| 18:10 | justin_smith | rhg135: how would you get the auto-complete results without nrepl? |
| 18:11 | lockdown | errr nrepl-cider I mean |
| 18:11 | justin_smith | the idea is to get the data from clojure itself, and nrepl lets you separate that request from the data going to your repl window |
| 18:11 | rhg135 | I don't mean no nrepl, although you can, I mean without custom middleware |
| 18:11 | justin_smith | OK |
| 18:13 | Somelauw | Instead of hard-coding the version number as "0.10.2" would it be wise to use something like :latest? |
| 18:14 | justin_smith | Somelauw: then it will break when it's newer than your cider version |
| 18:14 | justin_smith | (or might) |
| 18:14 | rhg135 | justin_smith: https://github.com/clojure-emacs/inf-clojure/blob/master/inf-clojure.el#L634 |
| 18:14 | justin_smith | better to leave it explicit so you can upgrade the two halves together |
| 18:15 | justin_smith | rhg135: yeah, I had misunderstood what you were saying, I just meant nrepl was useful for sending completions to elisp while having a separate repl not full of those things |
| 18:15 | rhg135 | ah, indeed. |
| 18:16 | rhg135 | maybe the tools.nrepl readme would help |
| 18:17 | Somelauw | My cider version is even 0.11.0-snapshot |
| 18:18 | seangrove | Hey all, what's the clojure equivalent of `jse = (JavascriptExecutor)wd;` ? |
| 18:18 | seangrove | Not sure about the (...) casting part |
| 18:19 | rhg135 | the only thing I can think of is bandwith |
| 18:19 | justin_smith | seangrove: since our bindings are not typed, we don't need casts |
| 18:20 | seangrove | justin_smith: That's what I thought, but this object is missing a method, thought that might be related |
| 18:20 | seangrove | Probably just an outdated dep though - thanks |
| 18:20 | justin_smith | seangrove: how would telling the compilar "treat x as a y" make a method show up on x? |
| 18:21 | amalloy | justin_smith: well, that *is* a thing you sometimes need to do, and there is a way to do it |
| 18:21 | justin_smith | amalloy: oh? |
| 18:21 | amalloy | the usual place this comes up is when trying to submit a function to an executorservice |
| 18:21 | amalloy | which is overloaded to take either a Callable or a Runnable, and functions implement both |
| 18:21 | justin_smith | ahh, and you would handle that with a hint, right? |
| 18:21 | amalloy | right |
| 18:23 | rhg135 | what is clojure.core/cast for then |
| 18:23 | justin_smith | rhg135: read the doc |
| 18:23 | justin_smith | it's a silly function |
| 18:23 | justin_smith | ,(doc clojure.core/cast) |
| 18:23 | clojurebot | "([c x]); Throws a ClassCastException if x is not a c, else returns x." |
| 18:24 | justin_smith | rhg135: I mean maybe that's what you want, but most times when people ask how to cast something in clojure that is not the functionality they are asking for |
| 18:26 | rhg135 | so, it's more of a specialized assertion (not actually an assertion) |
| 18:27 | rhg135 | basically says x is a c or die |
| 18:27 | justin_smith | right |
| 18:28 | rhg135 | cool, that's good for defensive code |
| 18:32 | justin_smith | yeah, an assertion would throw an error (no way to stop things, app dies now), but it throws an exception (something might be catching class casts and dealing with them I guess) |
| 18:32 | lockdown | Somelauw: then you need 0.11.0-snapshot for nrepl-cider |
| 18:32 | Somelauw | lockdown: I put a new version of cider-nrepl in profiles.clj. Now it complains that my refactor-nrepl is still out of date. |
| 18:33 | amalloy | justin_smith: assertions are perfectly catchable too, if you want |
| 18:33 | amalloy | just catch AssertionError |
| 18:33 | amalloy | it's just...a pretty bad idea |
| 18:33 | rhg135 | I just throw WorldIsDoomedBecauseValueIsNotRightError |
| 18:33 | Somelauw | lockdown: https://pastee.org/bgsy2 |
| 18:33 | justin_smith | fair enough, but error family vs. exception family carry different implications |
| 18:34 | rhg135 | then I watch it all burn to the ground, metaphorically of course |
| 18:35 | amalloy | i think if you catch Error you deserve to have it non-metaphorical |
| 18:35 | lockdown | Somelauw: I haven't use refactor-nrepl but looks like you have sync isssue again, between refactor-nrepl and clj-refactor |
| 18:36 | rhg135 | it doesn't even subclass Error, it's Throwable |
| 18:37 | Somelauw | I not sure if I need refactorings, but autocomplete doesn't work either or not as I expect (which is just pressing tab) |
| 18:37 | lockdown | Somelauw: if you are just learning can you do without the refactor package? |
| 18:37 | lockdown | remove it so you don't get the warning |
| 18:38 | rhg135 | but for real, code with catch Throwable is terrifying |
| 18:38 | Somelauw | I can ignore the warning. My priority is figuring out how autocompletion works. |
| 18:38 | lockdown | Somelauw: did you read the auto-completion docs for cider? |
| 18:39 | justin_smith | rhg135: the classy thing would be (throw (Throwable. "Updog")) |
| 18:41 | lockdown | clojure doesn't need autocomplete, you should keep your modules and in your head ;P |
| 18:41 | rhg135 | justin_smith: hopefully that's not from your code base |
| 18:41 | justin_smith | rhg135: heh, I just made it up |
| 18:41 | rhg135 | good, good |
| 18:42 | Somelauw | a bit hacky, but I manually ran company-mode, and it worked after the second launch |
| 18:42 | lockdown | err your modules small I mean |
| 18:42 | justin_smith | lockdown: auto-complete can save you manually typing (dir some-stupid-ns) or going and opening the source file |
| 18:43 | rhg135 | ,(doc dir) |
| 18:43 | clojurebot | "([nsname]); Prints a sorted directory of public vars in a namespace" |
| 18:43 | rhg135 | I didn't know that one |
| 18:44 | justin_smith | dir+doc+apropos are a nice combo |
| 18:45 | justin_smith | ,(apropos #"^ns") |
| 18:45 | clojurebot | (clojure.core/ns clojure.core/ns-aliases clojure.core/ns-imports clojure.core/ns-interns clojure.core/ns-map ...) |
| 18:45 | Somelauw | the docs say (global-set-key (kbd "TAB") #'company-indent-or-complete-common), but I think I would only want this in clojure-mode or something, so I can bind tab differently in other filetypes. |
| 18:47 | amalloy | Somelauw: you can "globally" add it to the keymap for a particular mode. or you can, in the hook for a particular mode, use local-set-key instead |
| 18:48 | amalloy | see https://github.com/amalloy/dotfiles/blob/master/.emacs.d/init.el#L40-L55 for an example of the latter |
| 18:50 | amalloy | (define-key slime-repl-mode-map (kbd "<C-return>") nil) ;; is an example of the former |
| 18:50 | justin_smith | haha, amalloy's a slimer |
| 18:50 | amalloy | i unbound C-RET for some reason or other |
| 18:50 | justin_smith | which is a term I just made up for people who still use slime, I hope it sticks |
| 18:51 | amalloy | never upgrade. never surrender |
| 18:52 | Somelauw | amalloy: thanks |
| 18:52 | lockdown | justin_smith: sometimes I think we just obsess to much with tooling |
| 18:53 | justin_smith | lockdown: it's an evolutionary thing, our ancestors who did not obsess about tooling never compiled and shipped |
| 18:54 | amalloy | that's what a public dotfiles repo is for. although i'm surprised my default-frame-alist is the thing you'd want to steal |
| 18:54 | amalloy | like pls, i wish more people would steal my rebinding of C-x C-c |
| 18:54 | lockdown | justin_smith: I bet amalloy would be just as productive with slime as with cider |
| 18:55 | lockdown | amalloy: was using spectacle (an osx window resize thing) for setting the size of emacs |
| 18:55 | lockdown | amalloy: do you code on a laptop? 42 seems low |
| 18:55 | justin_smith | lockdown: I was making a bad evo-psych joke |
| 18:56 | amalloy | i have vision problems |
| 18:56 | amalloy | so the font size is jacked way up, and the resolution way down |
| 18:56 | lockdown | justin_smith: oh ;) |
| 18:56 | amalloy | i'm going to an Evo Psych concert this weekend |
| 18:56 | lockdown | amalloy: I see |
| 18:57 | justin_smith | lockdown: I gave up on cider myself, and just use clojure in a repl, sometimes lein repl and sometimes a jar launched with java |
| 18:57 | amalloy | you and hiredman should start a club |
| 18:57 | justin_smith | so beleive me I have plenty of skepticism about tooling |
| 18:57 | hiredman | hmmm? |
| 18:57 | justin_smith | amalloy: I totally would but he has had me on ignore for years |
| 18:58 | amalloy | people who advocate use of clojure from a repl rather than editor integration |
| 18:58 | lockdown | justin_smith: yeah, if the buil-in repl had readline support I wouldn't even use lein sometimes |
| 18:58 | amalloy | lockdown: rlwrap |
| 18:59 | justin_smith | ~rlwrap |
| 18:59 | clojurebot | excusez-moi |
| 18:59 | justin_smith | ~rlwrap is good enough sometimes. |
| 18:59 | clojurebot | Ok. |
| 18:59 | lockdown | interesting |
| 19:00 | justin_smith | rlwrap is great for making any unixy thing that expects lines of text more usable |
| 19:00 | justin_smith | rlwrap ed is like 1% of the way to being vim |
| 19:01 | lockdown | heh |
| 19:02 | lockdown | justin_smith: inf-clojure hits a sweet spot I guess |
| 19:03 | justin_smith | I might try it one of these days |
| 19:12 | arrdem | I should get a CIDER tatoo |
| 19:12 | justin_smith | arrdem: tattoo upgrades suck, wait until it goes stable |
| 19:14 | arrdem | nah man -SNAPSHOT for lyfe |
| 19:14 | arrdem | what's the point of living if you aren't on the bleeding edge |
| 19:15 | amalloy | haha, i was jsut playing around with gist.el, and M-x gist-list breaks because i apparently have too many gists |
| 19:15 | amalloy | if: Variable binding depth exceeds max-specpdl-size |
| 19:16 | arrdem | hah |
| 19:19 | amalloy | i tried tracking it down, but the implementation is apparently a maze of generic functions or something and i'm no good at elisp |
| 19:21 | justin_smith | and elisp is no good at generic functions |
| 19:21 | justin_smith | so there we go |
| 19:27 | tolstoy | If a component only appears once in a system map, is there some reason Component would call start twice? |
| 19:28 | justin_smith | tolstoy: if you had your system in an atom and used swap! to start it and the calls overlapped |
| 19:29 | tolstoy | No, not doing that. Very weird. Starts two instances. |
| 19:30 | justin_smith | tolstoy: I don't see how it would do that |
| 19:32 | tolstoy | Weird. The component recvs messages on a topic, and, sure enough, it's doing it twice. So it's not a spurious log thing. |
| 19:32 | justin_smith | tolstoy: what if your whole system got started up twice? |
| 19:33 | tolstoy | Why wouldn't the other recv comps duplicate? Hm. |
| 19:33 | tolstoy | Heh. Just checked to see if I didn't name two records the same thing.... nope. |
| 19:34 | justin_smith | what about copy/paste error giving two names to instances of the same record? |
| 19:34 | Somelauw | i got emacs to freeze |
| 19:35 | tolstoy | That's what I just checked for. If so, can't find it. |
| 19:35 | justin_smith | there's a few ways to do that, a bunch of them involve very long lines |
| 19:35 | tolstoy | lein clean, etc. |
| 19:40 | justin_smith | tolstoy: can you share a paste of the system map? |
| 19:40 | justin_smith | or, rather the system/using call etc. |
| 19:41 | tolstoy | https://gist.github.com/zentrope/72820037e77062607f44 |
| 19:45 | justin_smith | I'm more familiar with system/using rather than inline component/using calls, but if component/using works like I expect that looks fine |
| 19:46 | tolstoy | Seems to have worked in lots of place. |
| 19:47 | justin_smith | tolstoy: what about writing your component such that starting the same one twice won't do anything? you could put a key in the system map that marks whether it's started |
| 19:47 | tolstoy | I'll give that a shot. I've just removed a bunch of deps to see if somehow that affects it, but, no.... |
| 19:47 | justin_smith | (start [component] (if (:this-started component) component (do .... (assoc component :this-started)))) |
| 19:48 | justin_smith | something like that |
| 19:49 | justin_smith | next step if that doesn't help is a delay or promise outside start that ensures the component is a singleton I guess |
| 19:49 | tolstoy | Yeah. I wonder _why_ the problem is there. I guess I'll have to try to repro with a minimum case, if I'm up for some madness. |
| 19:49 | justin_smith | maybe even throw an exception if starting after already started, and then see from the stack trace who is fucking up |
| 19:50 | justin_smith | (def started (promise)) (start [component] (assert (not (realized? started))) ... (deliver started true) (assoc component ...))) |
| 19:51 | tolstoy | Interesting. Setting a :starting? prop, it actually gets set twice. |
| 19:59 | tolstoy | Heh. Renamed the file "foo.clj": solved. |
| 19:59 | justin_smith | wat |
| 20:00 | Somelauw | I'm getting nuts if some "Unmatched bracket or quote" error |
| 20:01 | tolstoy | Slowly renaming things back to see what's what. |
| 20:02 | Somelauw | found it, I had an unmatched ] instead |
| 20:03 | tolstoy | Interesting. If my component is in health.clj, or the ns ends with .health (my-app.component.health) I see the prob. |
| 20:03 | justin_smith | Somelauw: many editors have tools that make this much easier to find and fix |
| 20:04 | justin_smith | tolstoy: that's weird and bad |
| 20:04 | tolstoy | Yep. |
| 20:04 | justin_smith | something clashing on the classpath? |
| 20:05 | tolstoy | I don't see how. Moving the health.clj file to another part of the tree. |
| 20:06 | Somelauw | justin_smith: I was told emacs is the most potent editor for clojure. I do have rainbow-coloring and such. I was looking for a round bracket and all my round brackets matched, it was a square bracket missing. |
| 20:06 | Somelauw | Sorry for using bracket and parenthesis interchangeably. |
| 20:08 | justin_smith | Somelauw: if you use highlight parens your mismatching parens would be a distracting / obvious color. Even with rainbow parens they should have been. |
| 20:09 | justin_smith | delimiters, I like the word delimiters (which I forgot to use just now) |
| 20:11 | tolstoy | justin_smith: Okay, we can restore our faith. It's because I inadvertantly had two loggers (myapp.component and myapp.component.health) defined in logback. |
| 20:11 | justin_smith | hahaha |
| 20:11 | justin_smith | so it was only starting once, but logging too much |
| 20:11 | tolstoy | justin_smith: Given that, it "seemed" as if I was processing dual copies of the message. |
| 20:11 | tolstoy | Right. |
| 20:11 | justin_smith | (dec loggers) |
| 20:11 | tolstoy | That's why the :started? scheme didn't work. |
| 20:12 | Somelauw | Ok, red isn't yet distracting for me |
| 20:12 | justin_smith | Somelauw: yeah, maybe you can change the color |
| 20:12 | justin_smith | maybe set a fast blink on the face :P |
| 20:40 | Somelauw | emacs + cider freeze yet again :-( |
| 20:41 | justin_smith | Somelauw: what makes it happen? |
| 20:42 | Somelauw | justin_smith: I'm not sure. I think "Mark set" is the last thing I read in the minibuffer. |
| 20:42 | justin_smith | Somelauw: what happens if you hit Control-g a bunch of times? |
| 20:43 | justin_smith | (while it's locked that is) |
| 20:43 | Somelauw | I've already killed it. I'll try next time. |
| 20:44 | justin_smith | Somelauw: anyway, in the future you could try turning on debug-on-error and then using Control-g when it locks up and the debugger should give an idea of what emacs was trying to do, and at least the cider folks could do something with that info maybe |
| 20:53 | scottj | justin_smith: Somelauw: I think you want M-x toggle-debug-on-quit (not error) for C-g to bring up debugger. |
| 20:55 | justin_smith | scottj: I assumed that after the quit an error would be exposed |
| 20:55 | justin_smith | scottj: but yes, that is more direct I guess |
| 20:56 | justin_smith | scottj: I find debug-on-quit annoying because I often quit from things... |
| 21:52 | Kamuela | Clojure becomes so easy on the eyes after a while |
| 22:06 | rhg135 | Most lisps do |
| 22:10 | stardiviner | I want to update CIDER-nrepl dependency version. (Try new feature cider-enlighten-mode). How to update it? |
