2014-11-22
| 00:00 | justin_smith | marcuscreo: it's about midnight on a friday there, you may not find many people who are there on irc |
| 00:00 | cfleming | marcuscreo: I'm at the conj, but just about to go to bed. |
| 00:01 | ambrosebs | cfleming: great talk |
| 00:01 | cfleming | ambrosebs: Thanks! |
| 00:01 | cfleming | ambrosebs: You're not at the conj, right? |
| 00:01 | ambrosebs | cfleming: right |
| 00:02 | cfleming | ambrosebs: Actually, I wondered if I'd find you at the Scheme workshop these days :-) |
| 00:02 | ambrosebs | cfleming: I had too many conferences lately to even consider it |
| 00:03 | cfleming | ambrosebs: Bummer, would have been cool to meet up. |
| 00:03 | ambrosebs | yea definitely |
| 00:10 | ffwacom | people not smiling in photos is going to be a trend |
| 00:10 | ffwacom | fuck wrong channel |
| 00:22 | holo | hi |
| 00:23 | holo | jabba question: what's this {} in → 'new ConstraintMapping[] { cm }' and how to interop with it? |
| 00:25 | amalloy | jabbslad: that's a single-element array, whose only element is cm |
| 00:25 | amalloy | er |
| 00:25 | amalloy | sorry, holo |
| 00:25 | holo | amalloy, i see |
| 00:28 | holo | I just noticed state of clojure didn't ask if programmer has any java background |
| 00:30 | holo | (inc amalloy) |
| 00:30 | lazybot | ⇒ 198 |
| 00:30 | holo | inc could have print side effects saying "thanks" |
| 01:08 | teajoe | Hey all! When I println, I'm seeing this: #<core$rsc_files boot.core$rsc_files@5401e235> |
| 01:08 | teajoe | is this a java obj? |
| 01:09 | TEttinger | you are printing some object that doesn't have a toString implemented apparently |
| 01:09 | TEttinger | what are you passing to println? |
| 01:10 | teajoe | some object from boot-clj |
| 01:11 | teajoe | they call a juxt with (memfn getPath) on it, so its probably some file object |
| 01:11 | teajoe | but for future reference can i get type info on such a thing |
| 01:17 | marcuscreo_ | (+ 1 2 3) |
| 01:17 | clojurebot | *suffusion of yellow* |
| 01:17 | marcuscreo_ | (println ‘here") |
| 01:18 | marcuscreo | (+1 2 3) |
| 01:18 | marcuscreo | (+ 1 2 3) |
| 01:18 | clojurebot | *suffusion of yellow* |
| 01:18 | justin_smith | http://stackoverflow.com/questions/3987428/what-is-an-initialization-block |
| 01:19 | justin_smith | err |
| 01:19 | justin_smith | http://www.urbandictionary.com/define.php?term=A%20Suffusion%20of%20Yellow that is the one I was looking for |
| 02:05 | rritoch | I'm not sure if this is a bug or not, but clojure regular expresions aren't matching #"\s+" to non-breaking spaces... Does anyone know how I can change this to match any whitespace including nonbreaking spaces? |
| 02:09 | ivan | rritoch: stackoverflow says 'It's a non-breaking space. According to the Pattern Javadocs, \\s matches [ \t\n\x0B\f\r], so you'll have to explicitly add \xA0 to your regex if you want to match it.' |
| 02:10 | rritoch | ivan: Ok, thanks. |
| 02:11 | rritoch | ivan: Ok #"[\s\xA0]+" worked |
| 02:12 | rritoch | (inc ivan) |
| 02:12 | lazybot | ⇒ 1 |
| 02:17 | dysfun | wow. after fighting the swift build process for two days, i have a whole new appreciation for clojure |
| 02:17 | dysfun | also leiningen |
| 02:25 | TEttinger | if you want to be really in-depth, rritoch: |
| 02:25 | TEttinger | ,(re-seq #"\pZ|[\f\t\r\n\u000b]" "good\u00a0job !\f\t\r\n\u000b\u3000\u200a") |
| 02:25 | clojurebot | (" " " " "\f" "\t" "\r" ...) |
| 02:25 | TEttinger | \pZ is the category of all unicode separator characters that aren't also control chars |
| 02:26 | TEttinger | so space, nbsp, ideographic space in the CJK asian character block |
| 02:26 | TEttinger | several others |
| 02:27 | TEttinger | \f form feed, \r carriage return, \n line feed, \t horizontal tab, \u000b vertical tab don't count for some reason |
| 02:32 | justin_smith | wow - first time ever, I think I deadlocked my repl |
| 02:32 | justin_smith | word of advice - using (locking *out* ...) inside pmap is probably a bad idea |
| 02:32 | mj-0 | what is regex for separating words of any language? |
| 02:33 | rritoch | TEttinger: Thanks, that was helpful. I changed it to #"(\pZ|\s)+" which seems to be working |
| 02:33 | luxbock | I don't think there exists one for a lot of languages |
| 02:33 | TEttinger | ah, that works rritoch! |
| 02:33 | mj-0 | I tried "\W+" but for other languages its not working. |
| 02:33 | TEttinger | mj-0: in chinese writing, you often don't have spaces at all between characters, right? |
| 02:34 | TEttinger | also, \w is not so good, it only recognizes latin script |
| 02:34 | TEttinger | you can use \b to get a word boundary using \w rules, IIRC |
| 02:34 | mj-0 | TEttinger: yeah. it doesn't matter to me. I just want to split words. |
| 02:34 | justin_smith | some info here http://www.regular-expressions.info/unicode.html |
| 02:35 | luxbock | in Thai you separate sentences with a space, but there's nothing that tells you when a word ends other than recognizing what that word is and what its last character is |
| 02:35 | TEttinger | ah cool! Java supports Unicode for \b but not for \w. |
| 02:35 | TEttinger | http://www.regular-expressions.info/wordboundaries.html |
| 02:35 | justin_smith | nice, just a jump away from the one I found |
| 02:36 | TEttinger | I should bookmark regular-expressions.info's unicode page, it's so helpful |
| 02:36 | TEttinger | that and just DDG searching for stuff like "feff unicode" |
| 02:37 | TEttinger | https://duckduckgo.com/?q=feff+unicode |
| 02:40 | TEttinger | here's some sinhala that I can't read or print correctly |
| 02:40 | TEttinger | ,(re-seq #"\b" "බුදුරජාණන් වහන්සේගේ දන්තධාතූන්") |
| 02:40 | clojurebot | ("" "" "" "" "" ...) |
| 02:40 | TEttinger | err |
| 02:41 | TEttinger | &(re-seq #"\B+" "බුදුරජාණන් වහන්සේගේ දන්තධාතූන්") |
| 02:41 | lazybot | ⇒ ("" "" "" "" "" "" "" "" "" "" "" "" "" "" "" "" "" "" "") |
| 02:41 | TEttinger | errrrrr |
| 02:41 | justin_smith | TEttinger: yeah, I was trying to make re-seq work with \B / \b and couldn't |
| 02:41 | justin_smith | even with ascii |
| 02:41 | TEttinger | &(re-seq #"\b.+\b" "බුදුරජාණන් වහන්සේගේ දන්තධාතූන්") |
| 02:41 | lazybot | ⇒ ("බුදුරජාණන් වහන්සේගේ දන්තධාතූන්") |
| 02:41 | TEttinger | &(re-seq #"\b.+?\b" "බුදුරජාණන් වහන්සේගේ දන්තධාතූන්") |
| 02:41 | lazybot | ⇒ ("බුදුරජ" "ා" "ණන්" " " "වහන්ස" "ේ" "ග" "ේ " "දන්තධ" "ා" "තූන්") |
| 02:41 | TEttinger | there we go |
| 02:42 | justin_smith | woah, black magic |
| 02:42 | TEttinger | &(re-seq #"\b.+?\b" "The Quick brown fox jump'd ov'r the lazy dog.") |
| 02:42 | lazybot | ⇒ ("The" " " "Quick" " " "brown" " " "fox" " " "jump" "'" "d" " " "ov" "'" "r" " " "the" " " "lazy" " " "dog") |
| 02:42 | justin_smith | TEttinger: speaking of black magic, did you see when SagiCZ1 had a problem with a dataset that got a huge number of boms inserted into the middle of files |
| 02:43 | justin_smith | he was trying to parse it and getting maddening bugs |
| 02:43 | justin_smith | TEttinger: but thanks to your foolery, I was quickly able to identify his problem |
| 02:43 | TEttinger | lol |
| 02:43 | TEttinger | I did remember that |
| 02:43 | justin_smith | "oh, that's just that thing TEttinger does" |
| 02:43 | TEttinger | haha |
| 02:43 | mj-0 | (re-seq #"\W+\b" "බුදුරජාණන් වහන්සේගේ දන්තධාතූන්") |
| 02:44 | TEttinger | ,(let ["even number of forms, eh?" "let's make it even" "or not"]) |
| 02:44 | clojurebot | "or not" |
| 02:44 | TEttinger | mj-0: \W won't work |
| 02:44 | justin_smith | mj-0: use , or & if you want a bot to evaluate it |
| 02:44 | TEttinger | it finds non-word chars but only knows ascii |
| 02:44 | mj-0 | TEttinger: got it. |
| 02:44 | mj-0 | &(re-seq #"\W+\b" "බුදුරජාණන් වහන්සේගේ දන්තධාතූන්") |
| 02:44 | lazybot | ⇒ ("බුදුරජාණන් වහන්සේගේ දන්තධාතූන්") |
| 02:45 | mj-0 | ohh this regex is working for some of the languages |
| 02:45 | rritoch | Someone should petition the international criminal courts to make it a crime to use anything other than UTC |
| 02:45 | justin_smith | haha |
| 02:45 | TEttinger | &(re-seq #"\b(\S|\PZ)+?\b" "The Quick brown fox jump'd ov'r the lazy dog.") |
| 02:45 | lazybot | ⇒ (["The" "e"] ["Quick" "k"] ["brown" "n"] ["fox" "x"] ["jump" "p"] ["'" "'"] ["d" "d"] ["ov" "v"] ["'" "'"] ["r" "r"] ["the" "e"] ["lazy" "y"] ["dog" "g"]) |
| 02:46 | TEttinger | &(re-seq #"\b(?:\S|\PZ)+?\b" "The Quick brown fox jump'd ov'r the lazy dog.") |
| 02:46 | lazybot | ⇒ ("The" "Quick" "brown" "fox" "jump" "'" "d" "ov" "'" "r" "the" "lazy" "dog") |
| 02:46 | TEttinger | that looks good |
| 02:46 | justin_smith | rritoch: I was contracting for a company in England recently, and it was actually kind of cool to be keeping UTC hours. It felt authoritative :) |
| 02:46 | justin_smith | rritoch: I reset my computer clock to UTC and everything |
| 02:46 | TEttinger | &(re-seq #"\b(?:\S|\PZ)+?\b" "The Quick brown fox jump'd ov'r the lazy dog. බුදුරජාණන් වහන්සේගේ දන්තධාතූන්") |
| 02:46 | lazybot | ⇒ ("The" "Quick" "brown" "fox" "jump" "'" "d" "ov" "'" "r" "the" "lazy" "dog" "බුදුරජ" "ා" "ණන්" "වහන්ස" "ේ" "ග" "දන්තධ" "ා" "තූන්") |
| 02:47 | TEttinger | slick |
| 02:47 | justin_smith | ~unicode-wordsplit is #"\b(?:\S|\PZ)+?\b" |
| 02:47 | clojurebot | In Ordnung |
| 02:47 | justin_smith | ~unicode-wordsplit |
| 02:47 | clojurebot | unicode-wordsplit is #"\b( |
| 02:47 | TEttinger | awww |
| 02:47 | justin_smith | :P |
| 02:47 | TEttinger | it doesn't like facts with ? |
| 02:48 | TEttinger | ~unicode-wordsplit is just ask TEttinger |
| 02:48 | clojurebot | A nod, you know, is as good as a wink to a blind horse. |
| 02:49 | justin_smith | $learn unicode-wordsplit is #"\b(?:\S|\PZ)+?\b" |
| 02:49 | lazybot | My memory is more powerful than M-x butterfly. I won't forget it. |
| 02:49 | justin_smith | $whatis unicode-wordsplit |
| 02:49 | lazybot | unicode-wordsplit is is #"\b(?:\S|\PZ)+?\b" |
| 02:49 | justin_smith | (inc lazybot) |
| 02:49 | lazybot | ⇒ 36 |
| 02:49 | rritoch | justin_smith: Thats cool, I'm in SGT (GMT+8) So at least my development machine has a stable time to work with. I saw in the news that Russia just changed their time system again and I swear they're doing it just to cause trouble for programmers. |
| 02:50 | rritoch | Either way, any code dealing with dates properly is virtually incomprehensible now. |
| 02:51 | mj-0 | TEttinger: #"\b.+?\b" this also returns spaces, special characters. How can I get only words? |
| 02:51 | mj-0 | &(re-seq #"\b.+?\b" " Çarşi En büyük BEŞİKTAŞ gerisi yalan oyunda. super, otesi inanilmaz sariyor Yorumun Tamamı") |
| 02:51 | lazybot | ⇒ ("Çarşi" " " "En" " " "büyük" " " "BEŞİKTAŞ" " " "gerisi" " " "yalan" " " "oyunda" ". " "super" ", " "otesi" " " "inanilmaz" " " "sariyor" " " "Yorumun" " " "Tamamı") |
| 02:52 | TEttinger | &(re-seq #"\b(?:\S|\PZ)+?\b" " Çarşi En büyük BEŞİKTAŞ gerisi yalan oyunda. super, otesi inanilmaz sariyor Yorumun Tamamı") |
| 02:52 | lazybot | ⇒ ("Çarşi" "En" "büyük" "BEŞİKTAŞ" "gerisi" "yalan" "oyunda" "super" "otesi" "inanilmaz" "sariyor" "Yorumun" "Tamamı") |
| 02:52 | TEttinger | better? |
| 02:52 | mj-0 | TEttinger: Yep. Looks better now. Thanks! |
| 02:52 | TEttinger | it will still return apostrophes as a separate thing |
| 02:53 | TEttinger | the meaning of that particular regex: word boundary, followed by at least one non-space, non-separator character, then another word boundary |
| 02:54 | TEttinger | also, do not store the group of non-space chars, that's what (?:) does |
| 02:54 | mj-0 | TEttinger: Okay. I'll look into it. |
| 02:55 | TEttinger | hope it helps! |
| 03:05 | rritoch | ,(as-> (java.util.Calendar/getInstance) c (.setTime (int 0)) c) |
| 03:05 | clojurebot | #<IllegalArgumentException java.lang.IllegalArgumentException: No matching field found: setTime for class java.lang.Integer> |
| 03:05 | rritoch | ,(as-> (java.util.Calendar/getInstance) c (.setTime c (int 0)) c) |
| 03:05 | clojurebot | #<ClassCastException java.lang.ClassCastException: java.lang.Integer cannot be cast to java.util.Date> |
| 03:06 | rritoch | ,(as-> (java.util.Calendar/getInstance) c (.setTime (java.util.Date 0)) c) |
| 03:06 | clojurebot | #<ClassCastException java.lang.ClassCastException: java.lang.Class cannot be cast to clojure.lang.IFn> |
| 03:06 | rritoch | ,(as-> (java.util.Calendar/getInstance) c (.setTime (java.util.Date. 0)) c) |
| 03:06 | clojurebot | #<IllegalArgumentException java.lang.IllegalArgumentException: No matching field found: setTime for class java.util.Date> |
| 03:06 | rritoch | Ba, I just can't win this one |
| 03:08 | rritoch | ,(as-> (java.util.Calendar/getInstance) c (.setTime c (java.util.Date. 0)) c) |
| 03:08 | clojurebot | nil |
| 03:08 | TEttinger | ,(doto (java.util.Calendar/getInstance) (.setTime (java.util.Date 0))) |
| 03:08 | clojurebot | #<ClassCastException java.lang.ClassCastException: java.lang.Class cannot be cast to clojure.lang.IFn> |
| 03:09 | rritoch | Ok, thats better, how do I get this to return the modified calendar object? |
| 03:09 | TEttinger | ,(doto (java.util.Calendar/getInstance) (.setTime (java.util.Date. 0))) |
| 03:09 | clojurebot | #inst "1970-01-01T00:00:00.000+00:00" |
| 03:09 | TEttinger | there we go |
| 03:09 | clojurebot | Huh? |
| 03:09 | rritoch | aah, doto, cool |
| 03:13 | rritoch | ,(doto (java.util.Calendar/getInstance) (.setTime (.parse (java.text.SimpleDateFormat. "MM/dd/yyyy") "11/22/2014"))) |
| 03:13 | clojurebot | #inst "2014-11-22T00:00:00.000+00:00" |
| 03:13 | rritoch | Cool, exactly what I need |
| 03:19 | rritoch | Does clojure automatically compare calendar dates correctly? |
| 03:21 | rritoch | ,(> (doto (java.util.Calendar/getInstance) (.setTime (java.util.Date. 1))) (doto (java.util.Calendar/getInstance) (.setTime (java.util.Date 0)))) |
| 03:21 | clojurebot | #<ClassCastException java.lang.ClassCastException: java.lang.Class cannot be cast to clojure.lang.IFn> |
| 03:21 | rritoch | ,(> (doto (java.util.Calendar/getInstance) (.setTime (java.util.Date. 1))) (doto (java.util.Calendar/getInstance) (.setTime (java.util.Date. 0)))) |
| 03:21 | clojurebot | #<ClassCastException java.lang.ClassCastException: java.util.GregorianCalendar cannot be cast to java.lang.Number> |
| 03:21 | rritoch | Hmm, guess not |
| 03:23 | TEttinger | ,(compare (doto (java.util.Calendar/getInstance) (.setTime (java.util.Date. 1))) (doto (java.util.Calendar/getInstance) (.setTime (java.util.Date. 0)))) |
| 03:23 | clojurebot | 1 |
| 03:24 | TEttinger | ,(compare (doto (java.util.Calendar/getInstance) (.setTime (java.util.Date. 1))) (doto (java.util.Calendar/getInstance) (.setTime (java.util.Date. 2)))) |
| 03:24 | clojurebot | -1 |
| 03:24 | TEttinger | so compare works |
| 03:25 | rritoch | TEttinger: Cool, I've never run into compare before, but that's a better option than where I was heading :) |
| 03:25 | TEttinger | (doc compare) |
| 03:25 | clojurebot | "([x y]); Comparator. Returns a negative number, zero, or a positive number when x is logically 'less than', 'equal to', or 'greater than' y. Same as Java x.compareTo(y) except it also works for nil, and compares numbers and collections in a type-independent manner. x must implement Comparable" |
| 03:28 | rritoch | doto + compare almost makes this code readable :) |
| 03:32 | rritoch | ,(as-> 1 n (min (compare n 5) (compare 10 n)) |
| 03:32 | clojurebot | #<RuntimeException java.lang.RuntimeException: EOF while reading> |
| 03:33 | rritoch | ,(as-> 1 n (min (compare n 5) (compare 10 n))) |
| 03:33 | clojurebot | -1 |
| 03:33 | rritoch | ,(as-> 5 n (min (compare n 5) (compare 10 n))) |
| 03:33 | clojurebot | 0 |
| 03:33 | rritoch | ,(as-> 10 n (min (compare n 5) (compare 10 n))) |
| 03:33 | clojurebot | 0 |
| 03:33 | rritoch | ,(as-> 20 n (min (compare n 5) (compare 10 n))) |
| 03:33 | clojurebot | -1 |
| 03:33 | justin_smith | ,(compare [1 2] [1 3]) |
| 03:33 | clojurebot | -1 |
| 03:33 | justin_smith | :) |
| 03:33 | rritoch | I want to check if a date is in a range |
| 03:34 | justin_smith | yeah, compare should work directly on date |
| 03:34 | justin_smith | ,(compare (java.util.Date.) (java.util.Date.)) |
| 03:34 | clojurebot | 0 |
| 03:34 | rritoch | Yeah, I just used numbers to test with because it was simpler. |
| 03:34 | justin_smith | right |
| 03:34 | justin_smith | now I see the scrollback |
| 03:46 | amalloy | rritoch: i bet clj-time compares dates better than some hand-rolled code you or i would write |
| 03:48 | rritoch | amalloy: I ended up with (defn in-range [from to item] (< -1 (min (compare item from) (compare to item)))) which is pretty good and more general purpose than just dates. |
| 03:48 | amalloy | rritoch: that is like super-wrong, though |
| 03:49 | amalloy | (< -1 x) is a very bad way to write (neg? x), because it returns false for -1 |
| 03:49 | rritoch | 1 is the fail case so it should return false |
| 03:49 | rritoch | err, -1 is the fail case |
| 03:50 | rritoch | It's cleaner than (not (neg? x)) |
| 03:51 | amalloy | you're writing (< -1 x) instead of (>= x 0)? in that case it's not wrong, just confusingly hard to read |
| 03:52 | rritoch | Yes, basically |
| 03:54 | amalloy | so my point re clj-time is that, whether your code is right or not, there are a number of edge cases, and you've had to write code that is not obviously correct at first glance. i have no idea what (min (compare x y) (compare z x)) actually computes, whereas if you used some pre-existing library it's already written, tested, and credible |
| 03:56 | rritoch | amalloy: This works, its fast, and doesn't require importing code I don't need, so it's more efficient |
| 03:58 | rritoch | amalloy: I'm developing OSGi modules, so bringing in additional code is non-trivial and increases the load time of the OSGi module significantly. |
| 03:59 | TEttinger | I see your point, rritoch. I think this solution isn't bad and if documented well could be fine |
| 04:01 | rritoch | Well, I do see amalloy's point also, but while developing these OSGi modules I find myself re-inventing the wheel often to maintain efficiency. It's a nessisary evil. |
| 04:01 | rritoch | err, necessary |
| 04:03 | amalloy | i don't understand osgi. i get the promise behind it: basically it's about saving you from dependency hell, right? because you can have multiple versions of whatever modules on your classpath at once? but it seems to me like the people who use it are even sadder than the people who don't |
| 04:04 | rritoch | amalloy: It is about modularization, not only do you avoid conflicts between modules since they have separate classloaders, but they can also export functionality as services to provide shared resources. |
| 04:05 | rritoch | amalloy: With OSGi you can load and unload modules at runtime, you can't unload namespaces in a single classloader environment so it is useful. |
| 04:06 | rritoch | amalloy: But it does require a lot of additional framework code to keep things working properly and efficiently |
| 04:07 | rritoch | amalloy: Getting it to work with clojure also requires jumping through a lot of additional "hoops" but once it's functional it tends to be very stable. |
| 04:33 | daGrevis | hi! i have a vector and i want to make a hash-map from it that contains all possibilities how items can map to other items in it. like [:a :b :c] would be {:a :b, :a :c, :b :a, :b :c, :c :a, :c :b}. where do i look? |
| 04:35 | TEttinger | there may be a combinatorics lib somewhere |
| 04:35 | daGrevis | https://github.com/clojure/math.combinatorics/blob/67b2e2a38be69fa25e713594c6da74303d52042c/src/main/clojure/clojure/math/combinatorics.clj |
| 04:35 | daGrevis | sure |
| 04:36 | opqdonut | ,(let [v [:a :b :c]] (into {} (for [x v y v] [x y]))) |
| 04:36 | clojurebot | #<CompilerException clojure.lang.ArityException: Wrong number of args (3) passed to: core/for, compiling:(NO_SOURCE_PATH:0:0)> |
| 04:37 | opqdonut | ah, right, had a weird nbsp there |
| 04:37 | opqdonut | ,(let [v [:a :b :c]] (for [x v y v] [x y])) |
| 04:37 | clojurebot | ([:a :a] [:a :b] [:a :c] [:b :a] [:b :b] ...) |
| 04:37 | opqdonut | anyway, maps can't have repeated keys |
| 04:38 | opqdonut | so you can't have {:a :b, :a :c ... } |
| 04:38 | daGrevis | ahh right |
| 04:38 | daGrevis | i can go for tuples in vectors then |
| 04:38 | TEttinger | you are probably after a sequence of sequences yeah |
| 04:38 | daGrevis | *vector |
| 04:39 | TEttinger | &(let [v [:a :b :c]] (for [x v y v :when (not= x y)] [x y])) |
| 04:39 | lazybot | ⇒ ([:a :b] [:a :c] [:b :a] [:b :c] [:c :a] [:c :b]) |
| 04:39 | TEttinger | if you don't want :a :a |
| 04:39 | daGrevis | thanks! |
| 04:40 | daGrevis | any way that i can use function from math.combinatorics to look cooler? |
| 04:40 | daGrevis | :D |
| 04:40 | daGrevis | aren't these permutations? |
| 04:42 | opqdonut | nope, cartesian product |
| 04:43 | daGrevis | opqdonut, ty |
| 05:39 | hellofunk | why can you (conj {} [:k :v]) but not (conj {} '(:k :v)) ? a list satisfies seq? so wouldn't they both work? |
| 05:46 | donbonifacio | anyone sharing clj in java and js? cljx is the way to go? |
| 06:46 | hellofunk | donbonifacio a lot of people think quite highly of cljx and it seems quite stable |
| 06:54 | daGrevis | which style is more idiomatic? https://gist.github.com/30d4aba120a475ba7662 |
| 06:58 | hellofunk | daGrevis having closing parens on a line by themselves us not idiomatic |
| 06:58 | hellofunk | *is |
| 06:59 | daGrevis | so v2 is better, right, hellofunk? |
| 06:59 | hellofunk | yes, though whether the opening outside vector is on the top line or not is a stylistic choice, but the closing should not be on its own line, creates more whitespace than most prefer to look at |
| 07:01 | daGrevis | if I put opening paren of vector on the same level as def is, my Vim indents it like this https://gist.github.com/dc420498b93e38b0e720 |
| 07:01 | daGrevis | is it okay? |
| 07:02 | hellofunk | daGrevis: "okay" is really subjective, here. i prefer your secret santa over the others, personally. |
| 07:07 | daGrevis | hellofunk, thanks and thanks :D |
| 07:07 | hellofunk | daGrevis if your pairs-banned is really so small then i'd probably just put the whole thing on one line anyway |
| 07:08 | daGrevis | that's an possibility too |
| 07:08 | puppeh | in a bit string mutation mutation (genetic algorithm), the "probability" the chance that *one* of the digits will be flipped? or is it applied to every digit? |
| 07:08 | daGrevis | but it can change and get it bigger |
| 07:09 | hellofunk | daGrevis yes but will those changes happen due to program logic, or are they bigger at the time you define the structure initially? |
| 07:09 | daGrevis | hellofunk, people may change it by hand before script is run |
| 07:10 | daGrevis | is it 'by hand' or 'with hand', by the way? |
| 07:11 | TEttinger | puppeh, I could see it either way, but the generations would make the most sense if it is one bit |
| 07:11 | hellofunk | why can you (conj {} [:k :v]) but not (conj {} '(:k :v)) ? a list satisfies seq? so wouldn't they both work? |
| 07:12 | rurumate | can I give an alias to current namespace using require? |
| 07:12 | rurumate | so I can refer to vars within it the same way I would from other namespaces |
| 07:12 | puppeh | TEttinger: so when Implementing such a function, which accepts a bitstring and a probability, what would that probability be about? kinda confused :S |
| 07:13 | puppeh | TEttinger: it would be for ex. that at a chance of 20%, at most 1 of the bits will be flipped? |
| 07:14 | TEttinger | do you have a link to a question that you're trying to solve? |
| 07:14 | rurumate | say I have a namespace com.foo and I use it in my other namespaces like this (require '[com.foo :as foo]) (foo/bar) |
| 07:15 | rurumate | and bar is also used in other functions in com.foo, is it possible to also refer to com.foo/bar as foo/bar within com.foo |
| 07:15 | puppeh | it's a code kata actually, that doesn't give much information, and I'm reading about mutation in order to understand how it works |
| 07:15 | puppeh | I only have the function signature and input values |
| 07:16 | puppeh | I think I've found the answer though: http://stackoverflow.com/a/5444676/1242778 |
| 07:20 | TEttinger | ok, this is a bit confusing |
| 07:21 | TEttinger | ,(reduce #(if (= %2 42) (reduced %1) (inc %1)) (iterate #(bit-xor % (bit-shift-left 1 (rand-int 8))) 12)) |
| 07:21 | clojurebot | 670 |
| 07:21 | TEttinger | that works fine, but when I change to reductions to see how it did it... |
| 07:21 | TEttinger | ,(reductions #(if (= %2 42) (reduced %1) (inc %1)) (iterate #(bit-xor % (bit-shift-left 1 (rand-int 8))) 12)) |
| 07:21 | clojurebot | (12 13 14 15 16 ...) |
| 07:21 | TEttinger | &(reductions #(if (= %2 42) (reduced %1) (inc %1)) (iterate #(bit-xor % (bit-shift-left 1 (rand-int 8))) 12)) |
| 07:21 | lazybot | java.lang.ClassCastException: clojure.lang.Reduced cannot be cast to java.lang.Number |
| 07:22 | Bronsa | TEttinger reductions still doens't handle reduced |
| 07:22 | daGrevis | why isn't this working? :( https://gist.github.com/0050af6586e7f2d05e70 |
| 07:23 | TEttinger | daGrevis, your #() fn accepts two args |
| 07:23 | Bronsa | daGrevis: replace [[.. ] [..]] with [..] [..] |
| 07:23 | TEttinger | but it's getting one, a collection |
| 07:24 | daGrevis | ahh, right :@ |
| 07:25 | TEttinger | ,(map inc nil) |
| 07:25 | clojurebot | () |
| 07:25 | TEttinger | <3 clojure |
| 07:26 | daGrevis | TEttinger, is that a good thing? |
| 07:26 | TEttinger | clojure treats nil as an empty seq if it ends up where it shouldn't |
| 07:26 | TEttinger | it counts as falsy for truthiness tests |
| 07:28 | daGrevis | TEttinger, but I mean... is that a good thing? you know, javascript treats number as string in some cases and vice-versa and it is terrible :) |
| 07:28 | TEttinger | it seems to be, nullpointerexceptions are probably more common in java than clojure |
| 07:29 | daGrevis | ahh, yes. |
| 07:29 | daGrevis | ,(> 42 nil) |
| 07:29 | clojurebot | #<NullPointerException java.lang.NullPointerException> |
| 07:29 | Bronsa | &*clojure-version* |
| 07:29 | lazybot | ⇒ {:major 1, :minor 7, :incremental 0, :qualifier "alpha1"} |
| 07:29 | Bronsa | TEttinger reductions supports reduced since alpha2/3 |
| 07:29 | TEttinger | lol |
| 07:30 | TEttinger | ,*clojure-version* |
| 07:30 | clojurebot | {:interim true, :major 1, :minor 7, :incremental 0, :qualifier "master"} |
| 07:30 | arav93 | Hi, does anybody know what miniKanren is? |
| 07:30 | TEttinger | &(clojure.string/join " " (take 40 (reductions #(if (= %2 42) (reduced %1) (inc %1)) (iterate #(bit-xor % (bit-shift-left 1 (rand-int 8))) 12)))) |
| 07:30 | lazybot | ⇒ "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" |
| 07:31 | TEttinger | hm |
| 07:31 | broquaint | arav93: http://minikanren.org/ ? |
| 07:32 | arav93 | I saw that link, but could you please tell me what it is? broquaint |
| 07:32 | TEttinger | &(clojure.string/join " " (take 40 (reductions (fn [[ctr oldval] newval] (if (= %2 42) (reduced [ctr oldval]) [(inc ctr) newval])) (iterate #(bit-xor % (bit-shift-left 1 (rand-int 8))) 12)))) |
| 07:32 | lazybot | java.lang.RuntimeException: Unable to resolve symbol: %2 in this context |
| 07:33 | TEttinger | &(clojure.string/join " " (take 40 (reductions (fn [[ctr oldval] newval] (if (= newval 42) (reduced [ctr oldval]) [(inc ctr) oldval])) (iterate #(bit-xor % (bit-shift-left 1 (rand-int 8))) 12)))) |
| 07:33 | lazybot | java.lang.UnsupportedOperationException: nth not supported on this type: Long |
| 07:34 | broquaint | arav93: It provides logic programming facilities, see also core.logic for clojure. |
| 07:35 | arav93 | broquaint: Thanks :-) |
| 07:35 | broquaint | np |
| 07:36 | hellofunk | why does (conj {} [:k :v]) work but not (conj {} '(:k :v)) ? a list satisfies seq? so wouldn't they both work? |
| 07:37 | Bronsa | hellofunk: vectors are mapentry, seqs aren't |
| 07:37 | TEttinger | ah cool, but puppeh quit |
| 07:37 | TEttinger | [0 12] [1 12] [2 13] [3 141] [4 140] [5 132] [6 164] [7 180] [8 182] [9 150] [10 182] [11 166] [12 38] [13 34] [14 162] [15 170] [16 174] [17 172] [18 140] [19 136] [20 200] [21 202] [22 200] [23 202] [24 206] [25 238] [26 254] [27 222] [28 94] [29 222] [30 94] [31 86] [32 214] [33 246] [34 214] [35 246] [36 182] [37 166] [38 164] [39 36] [40 44] [41 40] [41 42] |
| 07:37 | hellofunk | Bronsa, oh I thought all maps were seqs over their keys and values Bronsa, thus any seq could also be used to go the other direction. i guess not, only vectors can become kv pairs? |
| 07:38 | Bronsa | hellofunk: correct |
| 07:38 | TEttinger | that is a sample bit mutation thing that finds 42 |
| 07:38 | hellofunk | Bronsa I wonder what the reason for restricting map entry into vectors is, when the seq abstraction would seem to make more sense in general |
| 07:39 | Bronsa | actually nevermind about vectors being mapentries, the code for conj explicitely tests for IPersistentVector |
| 07:40 | hellofunk | Bronsa and since into uses conj, then making a map out of a seq, whether using conj or into, is only possible if that seq is a vector? |
| 07:41 | hellofunk | i guess this is possible (apply hash-map (list 1 2)) |
| 07:41 | hellofunk | i would think conj would test not for a vector but for a seq in general |
| 07:52 | rurumate | ,*ns* |
| 07:52 | clojurebot | #<Namespace sandbox> |
| 07:53 | rurumate | ,(alias 'socks 'sandbox) |
| 07:53 | clojurebot | nil |
| 07:53 | rritoch | One step forward... two steps back... I finally have a servlet that serves clj files selected by the URL, but for some reason nothing short of restarting tomcat will update my changes, even with my src folder on the classpath |
| 07:54 | rurumate | ,d an alias in the current namespace to another |
| 07:54 | rurumate | namespace. Arguments are two symbols: the alias to be used, and |
| 07:54 | rurumate | the symbolic name of the target namespace. Use :as in the ns macro in preference |
| 07:54 | clojurebot | #<CompilerException java.lang.RuntimeException: Unable to resolve symbol: d in this context, compiling:(NO_SOURCE_PATH:0:0)> |
| 07:54 | rurumate | whooops |
| 07:54 | rurumate | ,(count (ns-publics *ns*)) |
| 07:54 | clojurebot | 0 |
| 07:54 | rurumate | ,(def a 1) |
| 07:54 | clojurebot | #'sandbox/a |
| 07:54 | rurumate | ,(count (ns-publics *ns*)) |
| 07:54 | clojurebot | 1 |
| 07:55 | rurumate | will the vars in public be cleared periodically? |
| 07:55 | rurumate | *in sandbox |
| 07:55 | TEttinger | yes |
| 07:55 | TEttinger | every 10 min or something |
| 07:56 | rurumate | so it can happen that I define something and it gets cleared a second later, because there are fixed intervals? |
| 07:57 | rurumate | or are the 10 minutes a real TTL of each individual var |
| 07:58 | rurumate | nvm I found the source on github |
| 08:01 | rurumate | ,(:status (clj-http.client/get "http://www.google.com")) |
| 08:01 | clojurebot | #<CompilerException java.lang.ClassNotFoundException: clj-http.client, compiling:(NO_SOURCE_PATH:0:0)> |
| 08:01 | rurumate | ,(require '[clj-http.client :as client]) |
| 08:01 | clojurebot | #<FileNotFoundException java.io.FileNotFoundException: Could not locate clj_http/client__init.class or clj_http/client.clj on classpath. Please check that namespaces with dashes use underscores in the Clojure file name.> |
| 08:41 | rurumate | I just watched Tim Baldrige's "Data all the ASTs" video. Apparently ioc/state-machine now uses tools.analyzer for initial AST parsing. But then in ioc/parse-to-state-machine transforms the map-AST to a mix of maps and records. Would it be wrong to say that only some ASTs in core.async were "data'ed"? |
| 08:43 | hiredman | depends who you ask, I suspect tim would be in the camp that considers records to be data as well |
| 08:45 | rurumate | hiredman: so there are camps on this? can you elaborate what the camps disagree on? |
| 08:46 | rurumate | or point to some resource |
| 09:05 | SagiCZ1 | in regex, how do i match "everything until ," ? |
| 09:06 | SagiCZ1 | (re-find #"match" "some-string45821:.,hello") --> "some-string45821:.," |
| 09:07 | TEttinger | ,(re-find #".+," "some-string45821:.,hello") |
| 09:07 | clojurebot | "some-string45821:.," |
| 09:07 | TEttinger | ,(re-find #".+," "some-string45821:.,hello,") |
| 09:07 | clojurebot | "some-string45821:.,hello," |
| 09:07 | TEttinger | ,(re-find #".+?," "some-string45821:.,hello,") |
| 09:07 | clojurebot | "some-string45821:.," |
| 09:07 | schrotti | greetings |
| 09:07 | TEttinger | so you want #".+?," to stop just after reaching a comma |
| 09:07 | TEttinger | hey schrotti |
| 09:07 | SagiCZ1 | TEttinger yeah |
| 09:07 | SagiCZ1 | +? |
| 09:08 | schrotti | i'm trying to create a simple tcp socket and a writer: (def socket (Socket. "hostname" 1234)) and for the writer (def out (io/writer (.getOutputStream socket))) |
| 09:08 | TEttinger | was that a question, or the actual term? |
| 09:09 | schrotti | but i get an exception: CompilerException java.lang.IllegalArgumentException: No matching method found: getOutputStream for class java.net.Socket |
| 09:09 | schrotti | ok my bad :/ |
| 09:09 | TEttinger | hm? |
| 09:09 | schrotti | actually had (def out (io/writer (.getOutputStream socket true))) |
| 09:09 | TEttinger | oh ok |
| 09:09 | schrotti | -true works |
| 09:31 | _segfault | Hello ! Is anybody there ? |
| 09:32 | SagiCZ1 | _segfault: yup |
| 09:32 | TimMc | _segfault: There are like 500 people in the channel. I bet some of them are at their keyboards. :-) |
| 09:33 | SagiCZ1 | we are not sleeping, just merely dozing |
| 09:33 | _segfault | SagiCZ1: Do use emacs ? I needed some help with respect to emacs ! :) |
| 09:33 | SagiCZ1 | SagiCZ1: it's a coincidence that i don't but it seems that everyone else here does |
| 09:33 | SagiCZ1 | why am i talking to myself |
| 09:34 | _segfault | TimMc: Do you use emacs ? :) |
| 09:34 | boyscared | i do. ask your question |
| 09:34 | TimMc | Actually, 750. |
| 09:35 | _segfault | boyscared: Oh thanks ! I'm on windows & and on emacs 24.3 |
| 09:35 | _segfault | I wanted to use monore repl |
| 09:35 | TimMc | _segfault: Just ask your question, you don't need to find someone to ask it of. |
| 09:35 | _segfault | instead of Cider |
| 09:35 | _segfault | But Monroe doesnt seem to be able to connect to the repl |
| 09:36 | boyscared | i use cider, so can't help with that one |
| 09:37 | _segfault | Here's the steps : I manually start the repl (lein repl in the root of the project), switch to emacs and M-x monroe, I put in host:port and it says "Unable to Connect" |
| 09:39 | supersymmetry | I don't use emacs either.. havent had the time yet to learn :P |
| 09:40 | supersymmetry | did you provide a host port? else the nrepl is bound to local sockets only |
| 09:41 | yeahnoob | Tools is the tool ,not the Target. |
| 09:41 | _segfault | supersymmetry: yeah ! |
| 09:41 | supersymmetry | not sure if this was a remote connection or on the same machine, but that had me puzzled for a bit trying to hook up Light Table remote nrepl |
| 09:41 | supersymmetry | kk, just checking ;) |
| 09:49 | rurumate | watching rich hickey's "Transducers" video, at 13:10 there's this example: (transduce (comp process-bag (mapping weigh-bag)) + 0 pallets) |
| 09:49 | afhammad | how do you escape characters in a string? |
| 09:50 | rurumate | shouldn't it rather be (transduce (comp (mapping weigh-bag) process-bag) + 0 pallets) ? |
| 09:50 | rurumate | afhammad: same as in java? |
| 09:51 | afhammad | rurumate: i probably should be asking this on the appropriate channel but in clojurescript (str "\"") returns "\"" |
| 09:51 | afhammad | rurumate: also not everyone comes from java |
| 09:52 | rurumate | sorry. have you tried writing like you would in javascript? |
| 09:52 | afhammad | rurumate: in js "\"" returns """ |
| 09:53 | rurumate | looks like a bug to me, but such a severe bug should have been found earlier |
| 09:54 | afhammad | rurumate: actually it looks like its a eval bug in lighttable, when i print it to the browser console its fine. |
| 09:54 | rurumate | oh, that's a different story then |
| 09:55 | afhammad | yes, i put too much trust in it |
| 09:57 | rurumate | back to transducers question |
| 10:00 | rurumate | my intuition about comp would say that (comp process-bag (mapping weigh-bag)) applies process-bag to the result of (mapping weigh-bag), which would not work because (mapping weigh-bag) produces numbers, not bags |
| 10:00 | rurumate | how can I test the transducers stuff? |
| 10:01 | rurumate | it's not in clojure 1.6 afaik |
| 10:01 | luxbock | rurumate: is this the Strange Loop or the conj talk? |
| 10:02 | rurumate | strange loop, here's the link: https://www.youtube.com/watch?v=6mTbuzafcII&t=13m10s |
| 10:02 | luxbock | I think he talks about the ordering of transducers in comp a bit later |
| 10:03 | luxbock | yep, at 25:31 |
| 10:03 | luxbock | hold onto your seat |
| 10:03 | rurumate | so when will scala get transducers |
| 10:04 | luxbock | there's https://github.com/knutwalker/transducers-scala |
| 10:04 | rurumate | this is really funny: https://www.youtube.com/watch?v=6mTbuzafcII&t=9m11s |
| 10:05 | rurumate | amazing. where is this transducers idea coming from? did rich just think this up? |
| 10:06 | luxbock | I think so, in his hammock I'm sure |
| 10:06 | n0n3such | its an obvious abstraction |
| 10:07 | mavbozo | evolution of reducers |
| 10:10 | mavbozo | n0n3such: yeah, It takes a very unusual mind to undertake analysis of the obvious -Whitehead |
| 10:10 | clojer | Using Yesql and Cheshire to generated some JSON from a Postgres query, ip addresses are returned as :ip #<PGobject 86.148.167.17> which Cheshire doesn't like. What's the best way of dealing with these objects? |
| 10:12 | legittalon | My main editor is VIM, are there any convincing arguments for switching to Nightcode to do clojure work? I'm not entirely opposed to it. (I'm also new to clojure) |
| 10:12 | mavbozo | clojure: create custom encoder |
| 10:12 | mavbozo | clojer: create custom encoder for cheshire |
| 10:12 | n0n3such | “In all affairs it's a healthy thing now and then to hang a question mark on the things you have long taken for granted.” |
| 10:12 | n0n3such | ― Bertrand Russell |
| 10:13 | clojer | mavbozo: Ah, just noticed this at the bottom of the docs. Missed it first time round. |
| 10:17 | mavbozo | legittalon: a coworker of mine who's new to clojure and using Windows decided to use Nightcode and still going good |
| 10:18 | clojure_beginner | hello I have some trouble understandign this function http://pastebin.com/R5N2u8Ne |
| 10:18 | clojure_beginner | it's recursive. But i want to start at the let binding. It looks unusuall to me |
| 10:19 | legittalon1 | I have qualms about switching to an IDE, that I'm willing to question. That and I'm stuck in a Stockholm situation with vim commands I guess I'll just try it out and see what happens. |
| 10:20 | clojure_beginner | or is the order in let binding not important is this the same (let [a [1 2 3]...) (let [ [1 2 3] a..) |
| 10:22 | clojure_beginner | i know that [x & xs] n should say like x is the head and xs is the tail of sequence n |
| 10:22 | mavbozo | legittalon1: why don't you use vim? do you find clojure integration very poor there? |
| 10:22 | clojure_beginner | but it looks weird to me. |
| 10:22 | clojure_beginner | legittalon, use fireplace plugin for vim |
| 10:23 | mavbozo | clojure_beginner: it's clojure's destructuring argument |
| 10:24 | legittalon1 | Oh right, plugins! |
| 10:24 | clojure_beginner | mavbozo..Ahh, I totaly forgot about destructuring. |
| 10:24 | clojure_beginner | I've beend doing exercises on 4clojure.com and didn't have to use destructuring yet, so i forgot all about it |
| 10:25 | clojure_beginner | I didn't know zou can use destructuring in let expressions I thought only in fn arguments |
| 10:27 | mavbozo | clojure_beginner: I learn clojure through Clojure Programming book and the authors introduce destructuring through let. |
| 10:28 | rurumate | will transducers be compatible to reducers framework also? |
| 10:29 | clojure_beginner | mavbozo: I am also using this book. :) Now I need to wrap my head around recursion in this function |
| 10:30 | clojure_beginner | I think i got it now. The base clause returns empty list |
| 10:31 | clojure_beginner | and to that list the function compositions "roll out" :) |
| 10:32 | clojure_beginner | so all the work is done in the :else part |
| 10:32 | clojure_beginner | god, this thing is beutiful |
| 10:32 | clojure_beginner | But I am wondering, is it a clojure way to write recursive functions, or is this not the best practice ? |
| 10:33 | clojure_beginner | As I can see so far, you can live without recursion just fine in clojure. |
| 10:34 | rurumate | clojure_beginner: the way it's done in the example you gave is probably not idea |
| 10:34 | rurumate | have a look at clojure.core/flatten |
| 10:35 | clojure_beginner | rurumate, I am aware of the flatten function, this was just an exercise from 4clojure :) |
| 10:36 | rurumate | clojure_beginner: writing a recursion can often be avoided by using map or reduce, but these use recursion internally |
| 10:36 | clojure_beginner | rurumate, could I create a flatten alternative only with for list comprehension ? |
| 10:36 | rurumate | clojure_beginner: when you find you have to recurse, consider using (recur) instead of having the function call itself by name. that way you get tail recursiveness |
| 10:37 | clojer | mavbozo: I tried: (add-encoder org.postgresql.util.PGobject encode-str) |
| 10:37 | clojure_beginner | rurumate, tnx, will do that. I heared about tail recursion not being supported on jvm |
| 10:37 | rurumate | clojure_beginner: no I don't think it can be done with for alone |
| 10:37 | clojer | mavbozo: ... but this failed: (encode org.postgresql.util.PGobject. (:ip (first (recent-logs conn "2014-09-16 00:00:00")))) |
| 10:37 | clojure_beginner | I see. |
| 10:38 | rurumate | clojure_beginner: tail recursion is just a fancy way to write a loop, which sure is supported |
| 10:38 | clojure_beginner | rurumate: Sorry I meant tail call optimisations. |
| 10:39 | rurumate | isn't that the same thing? |
| 10:39 | clojer | rurumate: I think tail call elimination is different |
| 10:39 | jaaqo | There's no tail call optimisation. Recursive functions consume stack space like mad. Provided loop/recur forms are exception, which consume static space. |
| 10:40 | rurumate | anyway just remember to try use a (recur) to avoid stack overflow. the other way to avoid stack overflow in a recursion is to use lazy-seq buiit that's kind of advanced |
| 10:41 | clojure_beginner | rurumate: Yes it's the optimisation of tail recursion, I am not a compiler guy, so I don't know much about this. It's more low level stuff, what jvm enables you and what not. |
| 10:41 | rurumate | also you don't need a loop to go with the recur. it will recurse to the (defn [...] ) above it by default |
| 10:42 | clojure_beginner | yes, so just (recur with_args_of_defn) |
| 10:42 | rurumate | yes. a good way to think about (recur) is: the compiler will make a while loop out of this. it's really not more than that. nothhing scary :D |
| 10:43 | rurumate | and the compiler will tell you if it can't compile the thing into a while loop |
| 10:44 | clojure_beginner | rurumate: I know lazy seqs, that is in my opinion one of the best things in clojure i encountered so far. |
| 10:44 | rurumate | because you didn't use the (recur) in "tail position" |
| 10:44 | rurumate | yes lazy seqs can be really cool |
| 10:56 | mavbozo | clojer: here's my encoder |
| 10:56 | mavbozo | (add-encoder org.postgresql.util.PGobject |
| 10:56 | mavbozo | (fn [c jsonGenerator] |
| 10:56 | mavbozo | (.writeString jsonGenerator (.toString c)))) |
| 10:57 | mavbozo | (encode #<PGobject 86.148.167.17/32>) returns "\"86.148.167.17/32\"" |
| 10:58 | clojer | mavbozo: thanks. Will try it. |
| 11:02 | clojer | mavbozo: Still no luck: (encode (:ip (first (recent-logs conn "2014-09-16 00:00:00")))) |
| 11:02 | clojurebot | Excuse me? |
| 11:02 | clojer | 1. Unhandled com.fasterxml.jackson.core.JsonGenerationException Cannot JSON encode object of class: class org.postgresql.util.PGobject: 92.0.164.246 |
| 11:03 | mavbozo | clojer: i have a table web_address which has column named ip |
| 11:04 | mavbozo | clojer: i query it (jdbc/query pgdb ["SELECT ip FROM web_address"]) |
| 11:04 | mavbozo | clojer: returns ({:ip #<PGobject 86.148.167.17/32>}) |
| 11:05 | mavbozo | clojer: suppose i def the result to a var named res (def res *1) in my repl |
| 11:05 | clojer | mavbozo: Sorry, real problem is: Unable to resolve symbol: add-encoder in this context |
| 11:06 | mavbozo | clojer: (encode (:ip (first res))) returns "\"86.148.167.17/32\"" |
| 11:06 | clojer | But I have this added to :require [cheshire.generate :refer [add-encoder encode-str remove-encoder]] |
| 11:08 | mavbozo | clojer: different cheshire version? |
| 11:10 | clojer | mavbozo: Required repl reboot. Working now. |
| 11:10 | mavbozo | clojer: great |
| 11:11 | sdegutis | I'm using this function in production: https://gist.github.com/sdegutis/b376cf100e337941d826 |
| 11:11 | clojer | mavbozo: this simpler version now works: (add-encoder org.postgresql.util.PGobject encode-str) |
| 11:12 | mavbozo | clojer: that's greater! |
| 11:14 | sdegutis | It's very useful, it should be part of clojure.core |
| 11:28 | justin_smith | sdegutis: you should really be doing arglist matching by count. I mean that's what clojure does, matches the arglist by count. |
| 11:28 | sdegutis | justin_smith: You're right. This is a rough draft of the function, that's why it doesn't have that. |
| 11:28 | sdegutis | justin_smith: Once I add that in, it should be added to clojure.core for its usefulness. |
| 11:32 | clojer | Why does the Used Heap size of a dormant app gradually increase? In jvisualvm I hit "Perform GC" and Used Heap drops to next to nothing but not it's over 200MB RAM while the single page app is doing nothing, ie. dev mode on my laptop. |
| 11:33 | sdegutis | Any guesses on what https://github.com/weavejester/clofor/blob/master/src/clofor/core.clj is for yet? |
| 11:33 | clojer | Now at almost 250Mb RAM |
| 11:33 | clojer | Heap size is 1,750Mb |
| 11:51 | justin_smith | clojer: I know the jvm will ask for RAM preemptively (within its limit) so that when it needs it it can utilize it quickly |
| 11:51 | justin_smith | clojer: are you seeing any unexpected growth in used memory? |
| 11:52 | justin_smith | or only in the amount allocated/ |
| 11:53 | clojer | justin_smith: I'm aware that the JVM allocates total heap size but curious as to why used heap increases when nothing is happening to the app |
| 11:53 | emacs | how do you change nick |
| 11:54 | clojer | justin_smith: I thought total heap size (orange in jvisualvm) was the pre-allocated ram? |
| 11:54 | justin_smith | clojer: look at the threads that are running. There are maintinance loops, that allocate small amounts of memory that is quickly freed. That freed memory is being put back in the pool, not freed from the app. |
| 11:54 | clojer | justin_smith: It's the increasing (blue) used heap that perplexes me |
| 11:54 | Guest34275 | hey, why don't records follow the lispy syntax in examples? |
| 11:55 | justin_smith | clojer: OK. Your answer is in one of those threads, I think I misenterpreted you above. |
| 11:55 | justin_smith | Guest34275: what lispy syntax? |
| 11:55 | justin_smith | and which examples? |
| 11:56 | Guest34275 | or naming convetions rather |
| 11:56 | clojer | I notice now, the blue approached the heap size limit then dropped by about 150Mb |
| 11:56 | justin_smith | Guest34275: they create objects that are usable from java |
| 11:56 | Guest34275 | http://clojuredocs.org/clojure.core/defrecord |
| 11:56 | clojer | I set a 500Mb limit in this example |
| 11:56 | justin_smith | Guest34275: so they should hve names that make sense from java |
| 11:56 | Guest34275 | oh, so thats why they use camelcase |
| 11:56 | justin_smith | ,(munge "my-record") |
| 11:56 | clojurebot | "my_record" |
| 11:57 | justin_smith | that's what the java class name is for my-record |
| 11:57 | Guest34275 | oh yeah, java doesn't support dashes and stuff |
| 11:57 | justin_smith | right |
| 11:57 | justin_smith | ,(munge "my-record'") |
| 11:57 | clojurebot | "my_record_SINGLEQUOTE_" |
| 11:57 | clojer | justin_smith: 69 threads started, live: 53, live peak: 64, daemon: 8 |
| 11:57 | justin_smith | I'd love to see java code accessing that :P |
| 11:59 | Guest34275 | :justin_smith |
| 11:59 | sharms | /allchan chanopt confmode on |
| 11:59 | clojer | justin_smith: I was running: wrk -c 100 -d 60s -t 2 http://localhost:3000 |
| 11:59 | Guest34275 | justin_smith: records should be used over structs, right? |
| 12:00 | justin_smith | yeah, structs are depricated I think |
| 12:01 | justin_smith | clojer: oh, so wrk is what was doing it? |
| 12:02 | justin_smith | (well, the request handling that was spawned by wrk) |
| 12:03 | d0ky | hello how can i read file by 2 lines ? i need read file and if i find proper line i need read values from next line but im using line-seq and dont know how to get on next line in cycle ... can someone help ? |
| 12:04 | justin_smith | d0ky: maybe (partition 2 1 (line-seq src)) |
| 12:04 | justin_smith | that gives you every pair of lines |
| 12:04 | justin_smith | ,(partition 2 1 (range 10)) |
| 12:04 | clojurebot | ((0 1) (1 2) (2 3) (3 4) (4 5) ...) |
| 12:04 | luxbock | there's something demoralizing about solving a tough 4clojure puzzle and then seeing that all the other solutions are 1/3 as short |
| 12:04 | d0ky | (inc justin_smith) |
| 12:04 | lazybot | ⇒ 144 |
| 12:05 | d0ky | thanks it works :) |
| 12:05 | justin_smith | luxbock: that's where half the learning happens :) many of my 4clojure answers are much too long |
| 12:05 | justin_smith | luxbock: though personally, I think the more intelligible answer is better |
| 12:06 | justin_smith | if only "intelligibility" had as clear a metric as "byte count" |
| 12:06 | luxbock | I have yet to come to a solution where my answer would be unique and elegant at the same time |
| 12:07 | justin_smith | luxbock: there are some extremely smart people on 4clojure, treating it as a competition may not be as productive as treating it as a chance to learn... |
| 12:07 | luxbock | yeah, I was just joking around :) |
| 12:07 | justin_smith | luxbock: I will have a sense of humor after I get my coffee, sorry :) |
| 12:08 | luxbock | it's just funny how in that moment I click to see the link for other solutions my feelings go from feeling proud to being humbled |
| 12:08 | justin_smith | it's funny, but also humility can be an important part of learning |
| 12:10 | luxbock | https://gist.github.com/luxbock/deebf89b83feaa3fa9e3 |
| 12:10 | luxbock | this is the one I just did |
| 12:13 | justin_smith | eww, regex |
| 12:14 | justin_smith | luxbock: though clearly your loop could be a reduce, and that would be much clearer |
| 12:15 | luxbock | true, at first I was toying around with having one more loop-local variable |
| 12:16 | justin_smith | luxbock: you can have a map or vec as an accumulator, in order to effectively have multiple accumulators in a reduce |
| 12:17 | luxbock | is that cleaner? I often wonder about that and end up using a loop in those cases |
| 12:17 | justin_smith | ,(reduce (fn [[o e] v] (if (even? v) [o (conj e v)] [(conj o v) e])) [[] []] (range 10)) |
| 12:17 | clojurebot | [[1 3 5 7 9] [0 2 4 6 8]] |
| 12:17 | justin_smith | two accumulators |
| 12:17 | justin_smith | destructuring makes it work nicely |
| 12:18 | luxbock | I actually only recently learned about `reduced` from the most recent Hickey transducers talk |
| 12:18 | justin_smith | luxbock: reduced is what makes reduce a general replacement for loops that iterate across a collection |
| 12:18 | luxbock | which would make writing most of my loops as reduce-operations more appealing |
| 12:19 | justin_smith | right |
| 12:24 | luxbock | justin_smith: tried to translate this to using reduce, but got stuck with trying to get the first letter of the next grouping so use with `cmp` |
| 12:25 | justin_smith | luxbock: see my above suggestion of (partion 2 1 col) to d0ky :) |
| 12:25 | luxbock | ah hmm |
| 12:26 | justin_smith | (reduce f (partition 2 1 ...)) |
| 12:32 | justin_smith | or, more generally (reduce f (take-while not-empty (iterate rest coll))) |
| 12:32 | justin_smith | ,(take-while not-empty (iterate rest (range 4))) |
| 12:32 | clojurebot | ((0 1 2 3) (1 2 3) (2 3) (3)) |
| 12:33 | justin_smith | in fact, that is useful enough to give a name like rests |
| 12:33 | justin_smith | tails? |
| 12:45 | luxbock | pfft, finally: https://gist.github.com/luxbock/deebf89b83feaa3fa9e3 |
| 12:46 | luxbock | justin_smith: yeah that looks cleaner |
| 12:47 | sdegutis | $ lein repl |
| 12:47 | sdegutis | Retrieving org/clojure/clojure/1.3.0/clojure-1.3.0.jar from central |
| 12:47 | sdegutis | UHHH..... |
| 12:48 | sdegutis | wat |
| 12:49 | luxbock | sdegutis: you have a dependency somewhere that was written using 1.3.0, and you don't have it on your machine yet |
| 12:49 | luxbock | it's not a big deal |
| 12:49 | sdegutis | Oh. |
| 12:49 | sdegutis | Dang. |
| 12:49 | surrealanalysis | Does anyone know if there is an easy way to pass a variable number of arguments to another function? |
| 12:49 | surrealanalysis | For example (defn function1 [a & bs] ...) |
| 12:49 | surrealanalysis | (defn function2 [a & bs] |
| 12:49 | surrealanalysis | (function1 a bs) |
| 12:49 | surrealanalysis | ) |
| 12:50 | luxbock | surrealanalysis: apply works |
| 12:51 | surrealanalysis | Ah, thanks. I misread it last night, but reading it now makes more sense. Thanks |
| 12:51 | luxbock | ,(letfn [(f1 [a & bs] [a bs]) (f2 [a & bs] (apply f1 a bs))] (f2 1 2 3)) |
| 12:51 | clojurebot | [1 (2 3)] |
| 12:52 | surrealanalysis | Yeah, that worked. Don't know why I missed that. Thanks again luxbock |
| 12:52 | luxbock | np :) |
| 13:04 | justin_smith | luxbock: just got back from grabbing coffee - did you forget to update the gist? |
| 13:04 | luxbock | it's there but I added it as a new file to the same gist |
| 13:05 | justin_smith | luxbock: as a minor style thing, if you know order is a hash-map by context, (order a) can be better than (get order a) |
| 13:06 | justin_smith | but that's something not everyone agrees on |
| 13:06 | justin_smith | ,(map {:a 0 :b 1 :c 2} (range)) |
| 13:06 | clojurebot | (nil nil nil nil nil ...) |
| 13:06 | justin_smith | err |
| 13:06 | justin_smith | ,(map {0 :a 1 :b 2 :c} (range)) |
| 13:06 | clojurebot | (:a :b :c nil nil ...) |
| 13:06 | luxbock | I like the explicit get |
| 13:06 | Bronsa | it's also safer. |
| 13:07 | Bronsa | ,((fn [a] (a :foo)) nil) |
| 13:07 | clojurebot | #<NullPointerException java.lang.NullPointerException> |
| 13:07 | justin_smith | Bronsa: the hash map is defined on the preceding line, I did mention context |
| 13:07 | Bronsa | right, I missed that |
| 13:07 | Bronsa | then +1 for not using get |
| 13:08 | luxbock | hmm true, here it's not so needed, but I wrote it there out of habit |
| 13:08 | justin_smith | right, it's not a correctness thing at all |
| 13:08 | luxbock | I just like to be able to tell at a glance when I'm using a function and when I'm retrieving something from a map |
| 13:10 | justin_smith | luxbock: something that changed my mind on this front was reading some higher math, where functions were defined in terms of maps |
| 13:10 | justin_smith | in that sense, a clojure map is more like a real function than most of our functions are |
| 13:11 | justin_smith | "real function" meaning "function as talked about in mathematics" here of course |
| 13:11 | luxbock | right |
| 13:11 | justin_smith | and when I say "functions were defined in terms of maps" I don't mean particular function instances, I mean the very concept of a function started with the concept of a map, and evolved from that. |
| 13:13 | luxbock | I'd definitely like to learn some more high level math, but I need a problem where I can apply it to grok it |
| 13:13 | luxbock | I quit school when I was 19 so there's definitely a lot for me to learn |
| 13:13 | justin_smith | luxbock: I found my clojure experience definitely helped me with category theory |
| 13:13 | luxbock | doing the Cryptography and Machien Learning courses on Coursera helped quite a bit on that front |
| 13:13 | justin_smith | luxbock: I quit school at 18, and my last real math class was algebra II :) |
| 13:14 | luxbock | nice |
| 13:14 | justin_smith | but I was a stubborn kid who refused to learn from others, and spent a lot of time teaching myself |
| 13:15 | justin_smith | but recent stuff I have read about eg. cramming and memory retention makes me think I didn't have it so bad |
| 13:16 | justin_smith | skimmed a book three times over a course of years ~= crammed really hard, took a test, and never thought about it again |
| 13:16 | luxbock | yeah, that's why I think it's important to have a practical problem to work out with the theory for it to stick |
| 13:17 | luxbock | give those ideas some neural pathways to hold on to |
| 13:17 | justin_smith | luxbock: I found the mental models for category theory very applicable to program design in general. |
| 13:17 | justin_smith | the general domain / mapping concepts (this is low hanging fruit in that domain, it goes so much further of course) |
| 13:18 | luxbock | I think learning some Haskell might be a good step towards that direction |
| 13:18 | justin_smith | also, it was nice to know how far haskell is from normal category theory |
| 13:18 | justin_smith | hehe |
| 13:18 | justin_smith | though maybe if I learned more of both I would see more direct connections |
| 13:19 | justin_smith | monoids / endofunctors were these quick asides in the ct books I found (though to be fair, functors come up a lot and are used in haskell a lot too) |
| 13:20 | luxbock | there's a functional programming course that uses Haskell that's running now on edX I believe |
| 13:20 | luxbock | I'd like to take that but I'm a bit too busy at the moment, so will have to leave that for later |
| 13:23 | engblom | Haskell is one of the most nice languages you can learn. Of all languages I have coded in, I have least problem reading other peoples Haskell code. |
| 13:24 | engblom | However, there is a big problem: Cabal and hackage. The libraries are changing way to often and whatever you want to take into use by cabal and hackage might not compile because something changed. |
| 13:24 | engblom | Lein is so much more robust. |
| 13:25 | justin_smith | engblom: have you tried cabal sandboxes? I have only briefly played with them, but they seem to eliminate most of these issues (maybe all?) |
| 13:25 | justin_smith | engblom: agreed global installs are stupid (hell, at this point I think even shared libraries are a bad idea - RAM is cheaper to come by than my sanity) |
| 13:25 | engblom | justin_smith: No, I have not been hearing about the sandboxes. I should look it up |
| 13:26 | justin_smith | engblom: my only problem with cabal sandboxes is that they are not the only option available :) |
| 13:26 | justin_smith | (so far...) |
| 13:26 | justin_smith | engblom: it is similar to the lein paradigm, where deps are per-project, as they should be |
| 13:32 | justin_smith | So I am considering using a template to build the static part of a query (before getting any user input) and then using that as a prepared statement with the user input provided args following (clojure.java.jdbc/query [(build-format) user-arg-a user-b user-c]) |
| 13:32 | justin_smith | is this reasonable? |
| 13:33 | justin_smith | build-format has embedded loops based on data about the table structure and the fields defined, queried from the db |
| 13:33 | justin_smith | (realistically it would not be called for every query) |
| 13:34 | justin_smith | if only sql were a query api rather than a textual language... |
| 13:36 | engblom | I have not done anything with SQL together with clojure so far... But in all other languages I have used template system. I do not know if it is good or not though. But I would probably do as you are thinking to do, justin_smith |
| 13:39 | justin_smith | engblom: thanks. I know there are a lot of pitfalls with this stuff. |
| 13:39 | justin_smith | (in terms of code correctness and maintainability too, not to mention security) |
| 13:43 | kenrestivo | for implementing a state machine for a network protocol, what's the current best practice? just do it all in a core.async go-loop? |
| 13:44 | justin_smith | kenrestivo: I think ztellman has a nice lib for this |
| 13:45 | engblom | He seem to have very much of useful libraries. I often stumble upon recommendations to his repositories. |
| 13:45 | kenrestivo | i fear zach's libs. every time i dive in i disappear for days trying to understand them. |
| 13:45 | engblom | Yes, documentation seem to not be the strong side for his libraries. |
| 13:45 | kenrestivo | and i eventually discover that what i was trying to do was so straightforward i could have found a much simpler alternative |
| 13:46 | justin_smith | kenrestivo: maybe one of these? https://github.com/ztellman/manifold https://github.com/ztellman/manifold |
| 13:46 | justin_smith | err |
| 13:46 | kenrestivo | guy's brilliant, just, i'm wary. |
| 13:46 | justin_smith | https://github.com/ztellman/lamina |
| 13:46 | kenrestivo | yeah those in particular :-P |
| 13:46 | justin_smith | hehe |
| 13:47 | kenrestivo | core.async makes sense to me. manifold makes my brain melt |
| 13:49 | engblom | A couple of week ago, as an introduction project to Clojure I made it possible to fly Flightgear with two mice. It was a simple program parsing the mouse devices and then controlling flightgear by telnet. Gloss was a good library by ztellman, but his network library (I do not remember the name) was too difficult to get working as the documentation lagged. It was easier to do in other ways. |
| 13:52 | justin_smith | engblom: thanks for the feedback on that - did you end up successfully parsing the event structures with gloss? |
| 13:52 | justin_smith | engblom: iirc I may have even suggested that |
| 13:52 | engblom | justin_smith: Yes, I got that fully to work. And thanks for recommending gloss! |
| 13:53 | engblom | Now I can use one mouse for ailerons and elevators and the other for rudder, if I want so. |
| 13:53 | justin_smith | sweet! |
| 13:53 | justin_smith | years ago I wrote a small c program so I could attach N mice, and use one mouse to control each pair of parameters in a puredata audio patch |
| 14:11 | gfredericks | lein plugin idea |
| 14:11 | gfredericks | `lein repl-uberjar` creates an uberjar with all your user/repl deps, and that does something similar to `lein repl` when you run it. |
| 14:12 | arrdem | hum... |
| 14:12 | arrdem | not convinced that'd ever be useful |
| 14:12 | gfredericks | for running repls in contexts where leiningen is inconvenient |
| 14:13 | justin_smith | gfredericks: nice idea - doesn't the LEIN_FAST_TRAMPONOLINE option do something effectively similar? |
| 14:13 | gfredericks | I don't know what it does |
| 14:13 | justin_smith | ahh, but your idea doesn't use lein at all when run |
| 14:13 | gfredericks | exactly |
| 14:13 | justin_smith | gfredericks: https://github.com/technomancy/leiningen/wiki/Faster |
| 14:14 | justin_smith | "Setting the LEIN_FAST_TRAMPOLINE environment variable causes the bin/lein script to memoize all trampoline calls by saving off the java process invocation to disk upon the first run." |
| 14:14 | justin_smith | it means 'lein now runs really fast (until you change your project.clj) |
| 14:14 | justin_smith | ' |
| 14:14 | luxbock | damnit, 4Clojure seems to be runnning Clojure version < 1.5 |
| 14:14 | luxbock | can't use reduced |
| 14:14 | justin_smith | luxbock: how about take-while and reductions? |
| 14:15 | justin_smith | (not as nice, admittedly) |
| 14:15 | luxbock | yeah that works, I was just excited to use reduced for the first time |
| 14:15 | justin_smith | :) |
| 14:16 | gfredericks | ,(nth (iterate reduced 42) 10) |
| 14:16 | clojurebot | #<Reduced@429664: #<Reduced@19fa96d: #<Reduced@12c3c25: #<Reduced@120a5d7: #<Reduced@67e287: #<Reduced@492e19: #<Reduced@719ebf: #<Reduced@16d0dd3: #<Reduced@a6a9d0: #<Reduced@1d6be00: 42>>>>>>>>>> |
| 14:16 | gfredericks | ^ protip |
| 14:16 | justin_smith | rofl |
| 14:17 | gfredericks | ~protipping is a marginally weirder variant of cowtipping |
| 14:17 | clojurebot | Ok. |
| 14:17 | justin_smith | &@@@@@@@@@@(nth (iterate reduced 42) 10) |
| 14:17 | lazybot | ⇒ 42 |
| 14:18 | justin_smith | that @ chain tho |
| 14:18 | gfredericks | lightning talk idea: implementing church numerals with reduced |
| 14:19 | gfredericks | probably titled "You won't believe what this man did with clojure.core/reduced" |
| 14:20 | justin_smith | & (first (drop-while reduced? (iterate deref (nth (iterate reduced 42) 10)))) |
| 14:20 | lazybot | ⇒ 42 |
| 14:20 | morfeen | (+ 1 1) |
| 14:20 | clojurebot | 2 |
| 14:20 | morfeen | cool |
| 14:21 | gfredericks | was that a clojurebot magic trick? |
| 14:21 | justin_smith | gfredericks: repl is magic |
| 14:21 | gfredericks | (+ 1 1) |
| 14:21 | clojurebot | 2 |
| 14:21 | gfredericks | (+ 1 1 1) |
| 14:21 | clojurebot | 3 |
| 14:21 | gfredericks | (* 2 3 7) |
| 14:21 | clojurebot | *suffusion of yellow* |
| 14:21 | Bronsa | lol |
| 14:22 | Bronsa | (inc clojurebot) |
| 14:22 | lazybot | ⇒ 47 |
| 14:22 | gfredericks | (+ 1 1 1 1 1 2 1 1 1 1 1 1 1) |
| 14:22 | clojurebot | *suffusion of yellow* |
| 14:22 | gfredericks | oooookay. |
| 14:22 | Bronsa | that made me laugh way too much |
| 14:22 | justin_smith | gfredericks: the urban dictionary for suffusion of yellow is actually useful and not at all obscene |
| 14:22 | justin_smith | imagine that |
| 14:23 | gfredericks | I was just injesting it |
| 14:23 | gfredericks | (- 2 50) |
| 14:23 | clojurebot | -48 |
| 14:23 | gfredericks | ((fn [x] x) -3) |
| 14:23 | gfredericks | (- (+ 3 5) 20) |
| 14:23 | gfredericks | (- 2 8888888888888888888888888888888) |
| 14:23 | gfredericks | (- 2 8888888) |
| 14:23 | clojurebot | -8888886 |
| 14:24 | gfredericks | (* 2 -3 7) |
| 14:24 | gfredericks | well this has been an educational experience yet again in #clojure |
| 14:25 | mavbozo | (reduced? []) |
| 14:25 | justin_smith | gfredericks: I think I showed you that super low level clojure code in the pink audio engine - the Steven Yi talk was good, and helped me make a bit more sense of all of it |
| 14:26 | justin_smith | mavbozo: try & or , or ## |
| 14:26 | mavbozo | ##(reduced? []) |
| 14:26 | lazybot | ⇒ false |
| 14:26 | mavbozo | &(reduced? (seq [])) |
| 14:26 | lazybot | ⇒ false |
| 14:27 | justin_smith | ,(reduced? (reduced [])) |
| 14:27 | clojurebot | true |
| 14:27 | justin_smith | ,(reduced? @(reduced [])) |
| 14:27 | clojurebot | false |
| 14:27 | justin_smith | the deref gets the value from the reduced |
| 14:28 | justin_smith | @(reduced []) |
| 14:28 | justin_smith | ,@(reduced []) |
| 14:28 | clojurebot | [] |
| 14:28 | justin_smith | ,(reduced []) |
| 14:28 | clojurebot | #<Reduced@8d9915: []> |
| 14:28 | gfredericks | justin_smith: pink? what? |
| 14:29 | mavbozo | &(reduced? 1) |
| 14:29 | lazybot | ⇒ false |
| 14:29 | kenrestivo | (tan 74) |
| 14:29 | justin_smith | gfredericks: oh, I thought I had shared it with you. pink is Steven Yi's synthesis engine, made in pure clojure. Like overtone without the external supercollider program. |
| 14:29 | justin_smith | kenrestivo: ##(Math/tan 74) |
| 14:29 | lazybot | ⇒ -5.737022539278999 |
| 14:31 | gfredericks | justin_smith: interesting |
| 14:31 | justin_smith | gfredericks: Yi is one of the core csound devs, and pink is highly informed by that expertise |
| 14:32 | justin_smith | it's interesting to see super low level / real time code in clojure |
| 14:32 | justin_smith | (and also weird in many places) |
| 14:37 | rurumate | ,(reduce ((map inc) conj) () [1 2 3]) |
| 14:37 | clojurebot | (4 3 2) |
| 14:38 | justin_smith | (inc transducers) |
| 14:38 | lazybot | ⇒ 1 |
| 15:18 | sveri | Hi, how can I generate sequential uuids in clojure? |
| 15:23 | justin_smith | sveri: isn't sequential contradictory to the point of UUID? |
| 15:23 | jeremyheiler | sveri: you can use java.util.UUID and generate the bits yourself |
| 15:24 | noncom | hi! what is the current recommended clojure web stack? i want to deploy a little free webserver on heroku... |
| 15:24 | justin_smith | "Since the identifiers have a finite size, it is possible for two differing items to share the same identifier. The identifier size and generation process need to be selected so as to make this sufficiently improbable in practice. Anyone can create a UUID and use it to identify something with reasonable confidence that the same identifier will never be unintentionally created by anyone to identify something else." |
| 15:24 | justin_smith | sequential generation violates that |
| 15:25 | jeremyheiler | justin_smith: i think sequential for uuids means there's an order, not that the bits are incremented by one each time |
| 15:25 | jeremyheiler | for example, time based uuids give you an order so you can sort |
| 15:25 | justin_smith | jeremyheiler: ahh - new concept for me, thanks |
| 15:28 | jeremyheiler | sveri: you can also look here: http://commons.apache.org/sandbox/commons-id/uuid.html |
| 15:28 | jeremyheiler | justin_smith: np |
| 15:33 | justin_smith | I guess I was underinformed, partly because java doesn't support creating the time-sequence based uuids directly |
| 15:33 | justin_smith | (you can construct one from the right bits, but nothing in the java standard libs specifically generates those bits) |
| 15:34 | SagiCZ1 | how do i do t his? i want it to return [true false true true] |
| 15:34 | SagiCZ1 | ,(drop-while #(%) [false false false true false true true]) |
| 15:35 | clojurebot | #<ClassCastException java.lang.ClassCastException: java.lang.Boolean cannot be cast to clojure.lang.IFn> |
| 15:35 | ambrosebs | SagiCZ1: #(%) is (fn [%] (%)) |
| 15:35 | justin_smith | SagiCZ1: idenitty |
| 15:35 | SagiCZ1 | ambrosebs: i know, i just didnt know how to do it |
| 15:35 | SagiCZ1 | justin_smith: thanks |
| 15:36 | justin_smith | *identity, of course |
| 15:36 | SagiCZ1 | ,(drop-while identity [false false false true false true true]) |
| 15:36 | clojurebot | (false false false true false ...) |
| 15:36 | hyPiRion | justin_smith: isn't that the opposite? |
| 15:36 | gfredericks | ,(drop-while #(do %) [false false false true false true true]) |
| 15:36 | clojurebot | (false false false true false ...) |
| 15:36 | gfredericks | ^ or for golfing |
| 15:36 | SagiCZ1 | identity seems to do the opposite |
| 15:36 | hyPiRion | ,(drop-while false? [false false false true false true true]) ; is completely fine |
| 15:36 | clojurebot | (true false true true) |
| 15:36 | SagiCZ1 | hyPiRion: thanks |
| 15:37 | justin_smith | (inc hyPiRion) |
| 15:37 | lazybot | ⇒ 55 |
| 15:41 | gfredericks | drop-while-not-non-positive-or-twelve |
| 16:04 | andyf | Bronsa: Looking through tools.reader some more, and had some questions about possible small simplifications of code. I can file tickets if you like. These lines look like they could be removed, leaving only the ?else? part of the if: https://github.com/clojure/tools.reader/blob/master/src/main/clojure/clojure/tools/reader.clj#L159-L162 |
| 16:05 | andyf | And this line looks like it is already covered by the comment reader in the ?;? key of the map ?macros?: https://github.com/clojure/tools.reader/blob/master/src/main/clojure/clojure/tools/reader.clj#L747 |
| 16:05 | andyf | dang smart quotes |
| 16:05 | sveri | justin_smith: I think its nice to have semi sequential uuids like this: http://docs.datomic.com/javadoc/datomic/Peer.html#squuid%28%29 |
| 16:07 | arrdem | &(java.net.uuid/Uuid. 1 1) |
| 16:07 | lazybot | java.lang.ClassNotFoundException: java.net.uuid |
| 16:07 | justin_smith | sveri: yeah, that concept was new to me |
| 16:07 | justin_smith | arrdem: UUID |
| 16:07 | justin_smith | and it's util |
| 16:08 | arrdem | &(java.util.uuid/UUID. 1 1) |
| 16:08 | lazybot | java.lang.ClassNotFoundException: java.util.uuid |
| 16:08 | arrdem | third times' a charm... |
| 16:08 | arrdem | &(java.util.UUID/UUID. 1 1) |
| 16:08 | lazybot | java.lang.IllegalArgumentException: No matching method: UUID. |
| 16:08 | arrdem | fuckit |
| 16:09 | hyPiRion | ,(java.util.UUID. 1 1) |
| 16:10 | clojurebot | #uuid "00000000-0000-0001-0000-000000000001" |
| 16:10 | justin_smith | (inc hyPiRion) |
| 16:10 | lazybot | ⇒ 56 |
| 16:11 | hyPiRion | Seems like I'm harvesting karma today |
| 16:12 | justin_smith | hyPiRion: you keep having the clear answer to things I try and fail at :) |
| 16:12 | Bronsa | andyf: it's possible that those are fast-paths there for performance reasons, it's been a while since I've worked on t.reader and I honestly don't remember what that does, I'll take a look though |
| 16:13 | andyf | ok. There is definitely no bug here that I've found. I can do some performance testing to see if it makes any difference. |
| 16:14 | Bronsa | andyf: looks like the comment-prefix? line can go, let's see the other one |
| 16:15 | dcunit3d | hey whats up guys |
| 16:17 | dcunit3d | finally got irssi configured again, using the OSX Terminal app for my IRC app and iTerm for my terminal |
| 16:18 | Bronsa | andyf: yeah the read-delimited line looks like it's a microoptimization to avoid an useless unread+read, not sure if that has an actual performance impact though |
| 16:19 | andyf | The other change might be bad, according to a test case I just found, e.g. "(defn [x ;comment\n ])" |
| 16:22 | Bronsa | andyf: ah right. we can't use read there because `read` needs to read a value and read-comment doesn't return any |
| 16:24 | andyf | good to know. I won't be filing any tickets :) |
| 16:24 | andyf | well, unless it is to add test cases for things like that. |
| 16:25 | justin_smith | this Goetz talk is awesome, and gets me excited for value types |
| 16:35 | pdurbin | justin_smith: sorry, which talk? |
| 16:36 | justin_smith | brian goetz talk from the conj |
| 16:37 | justin_smith | https://www.youtube.com/watch?v=ADUrGfINXdI from yesterday I think |
| 16:38 | justin_smith | he ended with "allowing tail recursion in the vm is on the radar, is not a priority but it will happen" |
| 16:45 | andyf | arrdem: It might be nice to get another dump from ClojureDocs.org every few months or so for the parts of Grimoire that originally came from there. I don't think they have the web API set up to pull the new examples from there in an automated fashion yet. I can let you know if/when that happens. |
| 16:46 | andyf | Not sure if that was part of your plans. |
| 16:49 | arrdem | andyf: hum... that's not a bad idea but I should take a look at the legal on ClojureDocs before I continue straight up stealing their data |
| 16:49 | andyf | sure |
| 16:49 | andyf | explicit permission and/or attribution is good |
| 16:50 | arrdem | yeah. I'll throw that on the todo list for post-0.4.0, but right now getting the API and the new datastore online is the top priority |
| 16:50 | arrdem | a friend of mine has a cleaned UI in the works as well |
| 16:50 | sveri | justin_smith: I like it, however, in a "normal" database you ahve id's and timestamps and maybe a UUID column, which together are some kind of a semi sequential UUID pattern |
| 16:51 | pdurbin | justin_smith: thanks! |
| 16:53 | justin_smith | sveri: we added UUID columns to caribou so that migrations that have to happen in both directions between production / staging could stay sane (we still have timestamps and incremental ids, but content going from one db to another drops the incremental id, and keeps the UUID) |
| 16:54 | justin_smith | sveri: if development stays active, I may even experiment with doing the joins based on UUID rather than numeric incremental id |
| 16:58 | arrdem | andyf: if you've got a minute, I realized earlier today that someday god willing Grimoire will go multi-lingual and that the lib-grimoire API should have lang in there somewhere. Commentary on where would be appreciated. |
| 16:59 | andyf | Just because thalia has a placeholder for lang in there does not mean I have thought deeply about the problem, nor have any experience with multi-lingual apps :) |
| 16:59 | arrdem | hehe |
| 17:00 | andyf | Most of my job coding work never even touches Unicode, believe it or not. |
| 17:00 | arrdem | I mean.. neither does mine. I just thought it'd be a good idea to have it in there somewhere :P |
| 17:03 | justin_smith | arrdem: easter egg idea: provide a hidden Zalgo localization option |
| 17:04 | arrdem | justin_smith: depending on when I have time to drink and work on Grimoire, text/zalgo may or may not become a supported content type |
| 20:05 | justin_smith | $ping |
| 20:05 | lazybot | justin_smith: Ping completed in 0 seconds. |
| 20:07 | kenrestivo | finally found what i was looking for http://crossclj.info/ns/com.gearswithingears/async-sockets/0.1.0/project.clj.html tho i might mod it to remove dependency on tools.logging and use timbre instead. |
| 20:08 | justin_smith | kenrestivo: timbre uses tools.logging |
| 20:08 | justin_smith | https://github.com/ptaoussanis/timbre/blob/master/project.clj#L28 |
| 20:09 | justin_smith | err wait that's a test time config... |
| 20:09 | kenrestivo | yeah, i was gonna say |
| 20:10 | justin_smith | clearly that indicates they can at least coexist if they have to |
| 20:10 | kenrestivo | the disadvantage to catching exceptions in a library is then logging becomes your problem. |
| 20:10 | kenrestivo | instead of letting the user choose which logging substrate they want to use |
| 20:10 | justin_smith | kenrestivo: the README for timbre mentions tools.logging support |
| 20:11 | kenrestivo | i vaguely remember it having the ability to use different back-ends |
| 20:11 | justin_smith | kenrestivo: that, and users deal with all sorts of insanity because they don't find out about their basic errors except through odd symptoms |
| 20:12 | kenrestivo | not sure if timbre intercepts tools.logging calls tho. will look into it. if i can avoid hacking/forking someone else's library i'll do that. |
| 20:12 | justin_smith | kenrestivo: someone was having an issue with an environment wrapper lib, and nothing was making sense until we saw a try/catch/nil hiding the fact that their config file was unparsible (and simply returning empty config, with no message) |
| 20:12 | justin_smith | kenrestivo: well, since timbre advertises tools.logging compatibility, I would assume you can start using timbre without removing tools.logging |
| 20:13 | kenrestivo | this crossclj is amazing though. i think i've already said that, but i'm still amazed at how useful it is. |
| 20:13 | rkneufeld_ | Feel like I'm going crazy, clj-http is raising a "java.lang.IllegalStateException: Attempting to call unbound fn: #'clj-http.client/update" exception for me when I make a get. Anyone ever run into this? |
| 20:13 | justin_smith | kenrestivo: yeah, it is great |
| 20:14 | justin_smith | rkneufeld_: did you change your project.clj recently? that sounds like a weird interaction between compiled bytecode and a new version of a lib |
| 20:14 | justin_smith | rkneufeld_: can't hurt to exit, run lein clean, and run things again |
| 20:14 | rkneufeld_ | justin_smith: it's certainly possible its changed since the last time I made a get. I'll clean (thanks for reminding me) |
| 20:14 | kenrestivo | +1 for lein clean. also lein deps :tree |
| 20:15 | justin_smith | yeah, that is the kind of bizarro symptom old compiled files have (calling code in it's own package that doesn't exist) |
| 20:16 | rkneufeld_ | Fixed. Thanks for rubber-ducky-ing with me ;) |
| 20:17 | justin_smith | rkneufeld_: the first 10 times I dealt with that I was totally stumped |
| 20:17 | justin_smith | but the symptoms are very clear once you know them :) |
| 21:42 | rritoch | Is potemkin a viable alternative to gen-class to provide inheritance in clojure?c |
| 22:01 | kenrestivo | why, other than showing off use of squiggly macro-y things, would one do `(~@path :foo) instead of (-> path vec (conj :foo)) |
| 22:02 | justin_smith | or (concat path [foo]) |
| 22:02 | kenrestivo | even better, yeah |
| 22:03 | amalloy | kenrestivo: i don't find (-> path vec (conj :foo)) terribly easy to read |
| 22:03 | kenrestivo | i agree. (concat path [foo]) is much better |
| 22:03 | amalloy | `(~@path :foo) is pretty clear about what it's building, because it looks like what it is |
| 22:04 | amalloy | (concat path [:foo]) is fine too |
| 22:04 | kenrestivo | my point is, why would one use swear characters in something like that instead of easy-to-read functions like concat? |
| 22:05 | kenrestivo | is there some other good reason i'm missing? |
| 22:05 | amalloy | because why not? you're starting from this weird assumption that using ~@ is hard to read; from that perspective, of course using it looks dumb |
| 22:07 | kenrestivo | i just find it jarring to see splicing and unquoting occurring in places other than macros. yes it's a personal preference. |
| 22:08 | amalloy | kenrestivo: that's a fine personal preference to have, although my advice is that when there's a style that you find hard to read, the answer is to work on that, rather than to request that nobody use it |
| 22:09 | amalloy | ` is a great tool for building lists, and it can sometimes yield code that's easier to read than masses of concat and vector literals |
| 22:09 | kenrestivo | wasn't requesting that. was asking if my preference is going to get me into trouble, i.e. if using the macro characters is better |
| 22:09 | amalloy | kenrestivo: not at all, except that sometimes ~@ is more readable |
| 22:09 | kenrestivo | gotcha, thanks. |
| 22:10 | amalloy | kenrestivo: after all, the reader converts `(~@x :foo) to (seq (concat x (list ':foo))) or something similar to that |
| 22:10 | amalloy | ,'`(~@x :foo) |
| 22:10 | clojurebot | (clojure.core/seq (clojure.core/concat x (clojure.core/list :foo))) |
| 22:11 | yguan | Hi, there, get a question about defprotocol. How can I use deftype in another namespace? (ns this.ns (:require [nnss :refer :all])) seems not working |
| 22:11 | kenrestivo | hah, cool. it's doing the concat thing under the hood anyway. |
| 22:11 | amalloy | so you can't really be making a technical error by writing it out longhand |
| 22:12 | amalloy | yguan: deftype creates a java class, which you have to import like any other java class (ie, with import) |
| 22:12 | yguan | I see, thanks, amalloy |
| 22:16 | amalloy | i wonder if it would be better for `[~x ~@y ~z] to produce something like (conj (into [x] y) z), instead of (vec (concat [x] y [z])). is there much expense in building those intermediate sequences? |
| 22:17 | amalloy | presumably it doesn't matter because ` is mostly used for macros, but it looks like a free improvement for any runtime uses |
| 22:19 | amalloy | i suppose vec and into aren't available early enough in the bootstrapping phase |
| 22:21 | amalloy | kenrestivo: on the topic of "get better at code you find hard to read": years ago, i had trouble thinking of 'if as a statement rather than an expression: writing (f x (if foo y z)) confused me. so, i decided to make an effort to do the confusing thing myself until i had it handled. i think i actually did the same thing with using ` ~ ~@ in non-macro contexts, but i don't recall for sure |
| 22:22 | kenrestivo | thanks. macroexpand-1 is helpful, as is '. also i find it helpful when dealing with swear characters to translate them, i.e. (quote (unquote (splice x))) |
| 22:28 | kenrestivo | still takes me extra effort to puzzle the stuff out sometimes though. |
| 23:11 | kenrestivo | any advice for mnemonics to remember which direction <! is versus >! ? keep getting 'em confused, kind of wish they'd been called "put" and "take" |
| 23:11 | kenrestivo | (on the topic of my mental block on special chars) |
| 23:12 | justin_smith | kenrestivo: one comes from the thing on the left, the other goes to it |
| 23:12 | justin_smith | think about how it looks in a let block |
| 23:13 | justin_smith | (let [received (<! c)] (>! recieved outc)) |
| 23:13 | justin_smith | err I messed up the arg orders on that last one |
| 23:13 | kenrestivo | see what i mean |
| 23:13 | justin_smith | heh |
| 23:13 | justin_smith | I got the syms right, just not the arg order |
| 23:14 | justin_smith | anyway - with <! expect a binding to the left (that's where the data goes) |
| 23:14 | justin_smith | with >! expect a channel to the right (that's where the data goes)( |
| 23:15 | kenrestivo | ok, arrow is pointing towards or away from the channel (1st arg), that works. |
| 23:15 | justin_smith | right |
| 23:15 | justin_smith | it's pointing in direction of data's flow |
| 23:15 | kenrestivo | i think i need to watch the talk on swearjure in a loop until i just don't care anymore :-) |
| 23:16 | justin_smith | haha |
| 23:23 | amalloy | kenrestivo: i think of the ! as the channel, and > as drawing a picture: !> looks like taking something out of the channel, and !< like putting something into it |
| 23:24 | kenrestivo | that works too, thanks. |
| 23:24 | amalloy | although maybe it's actually >! and <!? i don't actually use core.async; my mnemonic helps me read it, not write it :P |
| 23:25 | justin_smith | amalloy: you mean >! and <! right? |
| 23:25 | justin_smith | heh |
| 23:27 | kenrestivo | ,(+ (*) (*)) |
| 23:27 | clojurebot | 2 |
| 23:27 | kenrestivo | who needs numbers. swearjure is amusing. |
| 23:28 | kenrestivo | at some point i will seriously walk through swearjure until my aversion to weird symbols is gone |
| 23:32 | justin_smith | &@@@@@@@@@@(nth (iterate reduced 42) 10) |
| 23:32 | lazybot | ⇒ 42 |
| 23:34 | kenrestivo | i saw that the other day and lol'ed |
| 23:35 | kenrestivo | back to core.async. is there some way to check if there's anything available, without blocking? |
| 23:35 | justin_smith | speaking of usage of symbols that approach meaninglessness :) |
| 23:35 | kenrestivo | i'm thinking about a kill-chan, at the top of a go-loop, before doing things, and before recurring, that'll shut down the loop if anything is available in the kill-chan |
| 23:35 | kenrestivo | but obviously i don't want to block. |
| 23:36 | justin_smith | kenrestivo: why not alts on a kill chan and your input at the top of the loop? |
| 23:36 | justin_smith | I think that's the typical way to do it |
| 23:36 | kenrestivo | i guess an alt and a channel that returns immediately would do the trick |
| 23:40 | justin_smith | a typical go loop will have input from a channel at the top anyway, so you just alt that with a poison channel... no need to make a special immediately returning channel for that to work |
| 23:41 | kenrestivo | sadly, crossclj has failed me for the first time: "No usages found for clojure.core.async/alts!" |
| 23:49 | justin_smith | yeah, we need to start submitting everything that gets used to crossclj.info |
| 23:55 | kenrestivo | oddly, i think it does spider, since my utility lib got added somehow. |