I'm confused about defmulti

#object[Error Error: No method in multimethod 'cljs.user/query-dispatcher' for dispatch value: ["Select" ["first_name" "like" "Sitges"]]]

I want to match on the "Select" in q but it does not work

I feel like I've followed the examples

If I have a list of things, and want to construct a map where the keys are each thing, and the values are some function applied to each thing, how do I do it? Like this in Python:

{k: f(k) for (k, _) in ['a', 'b', 'c']}

Why am I getting an arity exception when an empty string is passed in as a second argument to a function that takes two arguments? nucleotide-count/count \A, "" throws actual: clojure.lang.ArityException: Wrong number of args (1) passed to: nucleotide-count/count where it's defined as:

(defn count
  "counts instances of base in sq; throws if not a real nucleotide"
  [base sq]
  (count (filter #(
        (assert (contains? bases %))
        (= base %)) sq))
  )

@purohit the problem is that you're shaddowing the clojure.core/count function, but you're still trying to use that core function inside your function. The count which are you trying to invoke in the function body is your count function, not clojure.core/count

I'm doing all of the above in proto-repl, and thought I had copied code over directly from a colleague who had made it work in their environment (emacs I think)

@hotbelgo I tried your code and it works for me

oh, so I don't know how to make proto repl work then

Try it in plain clojure repl first. Are you sure you evaluated everything properly?

I've cut and paste the whole thing

I've ctrl+,+b each part

yikes, how do i do that

i'm actually trying to update someone elses code and I'm seeing the same error when I pass data to this multimethod

I don't use protorepl, so I don't know exactly how to eval things there. However, in most editors, you can easily eval the whole file/buffer. That way it should work. Maybe ask in #protorepl if you have specific problems with it. You can also try to select everything and eval it with ctrl-alt-, s (see https://atom.io/packages/proto-repl - Sending a selection).

Also notice, that once you define a multimethod, it's not that easy to redefine it (this is because defmulti uses defonce to define var that holds the multimethod). You can use ns-unmap to remove defmulti definition.

It's possile to reproduce your error with something like this:

;; Notice I incorrectly use sequence
(defmulti query-dispatcher sequence)
(defmethod query-dispatcher "Select" [[_select [lhs op rhs]]]
  (str lhs " " op " " rhs))

(def q  ["Select" [
                   "first_name" "like" "Sitges"]])
(query-dispatcher q)
;;=> 
IllegalArgumentException No method in multimethod 'query-dispatcher' for dispatch value: ("Select" ["first_name" "like" "Sitges"])  clojure.lang.MultiFn.getFn (MultiFn.java:156)

;; Now I want to fix it:
(defmulti query-dispatcher first)
;; But I have to unmap query-dispatcher first
(ns-unmap *ns* 'query-dispatcher)
;; and then re-eval everything.

ah, yes, what you say about renaming rings a bell

and i managed to make it work on a different computer

---- Could not Analyze <cljs form> line:1 column:1 ---- nth not supported on this type: Symbol 1 (ns-unmap ns 'query-dispatcher) ^--- ---- Analysis Error ----

when I was being taught we just renamed stuff

but I changed to query-dispatcher3

did ctrl+, l on each part

and end up with

----  Compiler Warning on   <cljs form>   line:1  column:2  ----

  Use of undeclared Var cljs.user/query-dispatcher3

  1  (query-dispatcher3 q)
      ^--- 

----  Compiler Warning  ----
#object[TypeError TypeError: Cannot read property 'call' of undefined]

ok, resarted everytihngf and it works

wow - this is hard

How do I map a function to a collection but also pass another argument? Like I want (map my_fn my_seq extra_arg) because my_fn takes two args

I thought about using partial functions but the argument I'm adding is in the second place instead of the first one

https://stackoverflow.com/questions/15140056/clojure-separating-comp-and-partial-arguments

Guess I have to use an anonymous fn

@purohit correct

How do I consume two sequences of equal length and run some function on each pair of objects? Like if I have

(def a '(1 2 3))
(def b '(4 5 6))

@purohit mapv should do it

Perfect, thanks!

i have these random brainfreezes regarding anything clojure. i want to evaluate/test/try out two sql libs, how should i do it? create a completely empty project for this? seems like a lot of work? using something that's already filled up seems weird

use repl

lein repl

I do all my brainstorm there

@ashnur I usually just have a project for screwing around in, made with lein new, It's the easiest way. There's also https://github.com/rkneufeld/lein-try and if you use boot it should be easy enought to add deps to the repl.

interesting that lein-try

thanks @madstap @rcustodio

@ashnur If you're on osx, I think the easiest thing now is to use the new official clj tool. brew install clojure

windows or linux

coming soon to a package manager near you

i am not a rich person 🙂

we're not there quite yet 😃

i just did lein new sandbox for now

{:deps {org.clojure/clojure {:type :mvn :version "1.9.0-beta1"}
        org.bouncycastle/bctsp-jdk14 {:type :mvn :version "1.38"}}}

that's the new deps.edn format for the clj tool

took me a moment to find, but the docs for that tool moved here: https://clojure.org/guides/deps_and_cli

fyi, a lot of stuff on that page is undergoing big changes atm

@ashnur Might be worth looking at boot since you can first up boot repl and then add new dependencies on the fly and test libraries without needing to restart your REPL (and without needing a project at all).

@seancorfield how do you add dependencies on the fly? set-env! with the new vector of dependencies?

i was afraid of boot because most of the tools i need are explained in terms of lein

What is the meaning of () in the expression (){:x 42} ?

(this was generated by honeysql)

Oh wait, nevermind. Just some repl anomaly

@hmaurer I use merge-env! but, yes, you add new :dependencies to the environment and Boot loads them, then and there.

@ashnur I find boot much simpler to understand than lein, fwiw

@ashnur At work we switched from lein to boot about a year ago. No regrets 🙂

it's less magical

☝️:skin-tone-2: Yes, this.

@seancorfield neat. By the way, I’ve a bit of an unrelated question. I have been told that Spring Boot uses a different technique than standard “uberjars”. Instead of merging all class files into a single jar, they generate a jar containing the librairies jars in a folder, and load them dynamically

is that something doable with boot?

(and yes, I am a beginner and also found boot much simpler to understand after switching to it)

@hmaurer I don't know enough about Spring Boot to understand what you're suggesting.

@seancorfield https://docs.spring.io/spring-boot/docs/current/reference/html/executable-jar.html

> Java does not provide any standard way to load nested jar files (i.e. jar files that are themselves contained within a jar). This can be problematic if you are looking to distribute a self-contained application that you can just run from the command line without unpacking. > To solve this problem, many developers use “shaded” jars. A shaded jar simply packages all classes, from all jars, into a single ‘uber jar’. The problem with shaded jars is that it becomes hard to see which libraries you are actually using in your application. It can also be problematic if the same filename is used (but with different content) in multiple jars. Spring Boot takes a different approach and allows you to actually nest jars directly.

shrug No idea... nor do I have any idea why that would matter...

both approaches have downsides. really, creating uberjars at all is just a bad thing. if only it weren’t so damn convenient.

the shaded uberjar runs the risk of clobbering, particularly with manifests and other similarly tricky parts of jar files

using a jar-in-a-jar requires special classloader and url support for the jvm to understand what you’re doing

I’m not aware whether the latter works in a clean way with Clojure (you’d need something more than just Clojure to make that work)

@alexmiller I see; thanks for the details!

A simple pool, who uses lein and who uses boot?

I started to read about boot now, but for libraries I think lein is better

@rcustodio lots of data at http://blog.cognitect.com/blog/2017/1/31/state-of-clojure-2016-results

we do this poll every year and it’s a good sample size

Thanks

It'll be interesting to see how much ground Boot has made in the next survey. Leiningen will continue to be the default choice for a lot of people coming into Clojure because it's what (nearly?) all the books recommend right now.

All of our corporate tooling assumes lein, so boot would require a lot of support work for us, and so we’ll probably stay with lein unless there’s a huge advantage. I’m sure a lot of other places are the same.

I am getting mugged by startup issues. I have a project.clj:

(defproject voting "0.1.0-SNAPSHOT" :description "Cabal voting back end" :url "http://example.com/FIXME" :license {:name "Eclipse Public License" :url "http://www.eclipse.org/legal/epl-v10.html"} :dependencies [[org.clojure/clojure "1.8.0"][http-kit "2.2.0"]] :main ^:skip-aot voting.core :aot [voting.core] :target-path "target/%s" :profiles {:uberjar {:aot :all}})

(you can use triple-backtick to format code when pasting small fragments here @jrootham)

I ran lein deps after putting in the [http-kit "2.2.0"] and it said it downloaded them

Although the time it took was a clue that it didn't

I cannot find http-kit in my project (created with lein app voting)

JAR files are downloaded to your local Maven cache in .m2/repository in your home directory. They are not added to the project itself.

`(defproject voting "0.1.0-SNAPSHOT" :description "Cabal voting back end" :url "http://example.com/FIXME" :license {:name "Eclipse Public License" :url "http://www.eclipse.org/legal/epl-v10.html"} :dependencies [[org.clojure/clojure "1.8.0"][http-kit "2.2.0"]] :main ^:skip-aot voting.core :aot [voting.core] :target-path "target/%s" :profiles {:uberjar {:aot :all}})

@jrootham You'll need to adjust a Slack preference so pressing enter inside backticks does not send the message: https://www.dropbox.com/s/z9hsl1vtxn509uk/Screenshot%202017-09-19%2012.27.59.png?dl=0 (I don't know why this isn't Slack's default -- it's very counter-intuitive)

I think he just missed the closing triple back tick

Sorry, not familiar with that idiom

Back to your problem... The JAR files don't get added inside your project tree.

(and the time taken is because a lot of JAR files will be downloaded the first time you reference some new libraries)

OK, I have found the library, so it did download, I now have the newbie issue of how to use them in my code, in the lein repl should (:require 'http-kit) work, and how do I tell?

When you start a REPL with lein repl inside the voting project directory, it will have the dependencies declared in project.clj and you can (require 'http-kit.whatever) namespace.

(note: no : for the require function -- it has a : inside the ns form in your source code tho')

(require 'http-kit) gets me a file not found message, your last answer suggests I should be looking for a namespace inside http-kit

confusingly, the name of the artifact and the names of the namespaces inside an artifact, do not necessarily have any correlation (although usually the artifact name is part of the namespace name)

in this case, the namespaces of interest are either org.httpkit.server or org.httpkit.client

Docs have examples: http://www.http-kit.org/server.html http://www.http-kit.org/client.html

Ah, it looks closer, use instead of require. The demo works in the repl. Now for actual code.

Thank all