@juanjo.lenero What, specifically, are you trying to achieve? There's probably an easier, more functional way to do it (without a macro).

I’m going through the clojure for the brave and true book and there is an example with a vector of maps that contain body parts.

(def asym-hobbit-body-parts [{:name "head" :size 3}
                             {:name "left-eye" :size 1}
                             {:name "left-ear" :size 1}
                             {:name "mouth" :size 1}

You have to then define a function that constructs a symmetric hobby.

So I wanted to do something like:

(reduce (fn [full-hobbit part] (into full-hobbit [part (only-if (left-part? part) (right-part part)])) [] half-hobbit)

But I extracted that whole [part (only-if …] expr into a function that returns either 1 or two parts depending on if the part has a counterpart.

@seancorfield

So I guess you were right, I didn’t need it.

Macros are very rarely needed. Always try to write a function first. Sometimes you need a macro, but it's not often -- and they don't compose like functions. Sometimes a macro can provide useful syntactic sugar -- but those are usually just wrappers around functions that do the real work @juanjo.lenero

BTW @juanjo.lenero You might want to look at cond->: (cond-> [part] (left-part? part) (conj (right-part part)))

@seancorfield Thank you, I did run into it, I’ll keep it in the back of my mind, but I’ll avoid it in the meantime.

I am referring to an example from http-kit. I am new to clojure

(defn -main [& args]
  ;; The #' is useful when you want to hot-reload code
  ;; You may want to take a look: 
  ;; and 
  (reset! server (run-server #'app {:port 8080})))

#'app

@gdrte What is your question?

(and, btw, if you're new to Clojure, #beginners might be a useful channel for you)

#'app is a Var reference -- so it provides a level of indirection over just app -- does that answer your question?

I'm a bit confused about how to exactly set up honeysql and postgres depedencies in a project

I added at the moment

[org.clojure/java.jdbc "0.7.0"]
                 [org.postgresql/postgresql "9.4-1201-jdbc41"]
                 [honeysql "0.9.0"]
                 [migratus "0.9.8"]
                 [nilenso/honeysql-postgres "0.2.3"]
                 [clj-postgresql "0.7.0"]

but following the docs from honeysql and this code

(def db
  {:connection-uri (get env :database-url LOCAL-DB-CONN)})

(def all-people
  {:select [:first-name] :from [:people]})

(j/execute! LOCAL-DB-CONN (honey/format all-people))

PSQLException A result was returned when none was expected.  org.postgresql.jdbc2.AbstractJdbc2Statement.executeUpdate (AbstractJdbc2Statement.java:371)

If I instead do (j/execute! db (honey/format all-people)) I get No suitable driver found for postgresql://....

and the last thing I don't get is that looking up org.postgresql in Clojars returns nothing, and what actually instead comes back is

which I'm not sure if it's the same thing or not, but surely it's a different version

@andrea.crotti have you tried using query instead of execute!?

IIRC execute! is used for things that doesn’t return results

@schmee query seems to work, and I understand the error

but that's in the honeysql docs so I thought it was correct

hmm, according to jdbc docs execute! “…performs a general (non-select) SQL operation”, so I guess the example is just wrong

and I don't get the driver error though

ah ok I can open an issue about that then

the driver error though I think it's just misleading then

these are the dependencies I have:

[org.clojure/java.jdbc "0.6.2-alpha2"]
                 [org.postgresql/postgresql "9.4.1212.jre6"]

and those work for me at least

uhm jdbc there is 0.7.0 stable, and you have "jre6" instead of "jdbc41"

yeah I haven’t updated those in a while

but those two are enough to make everything work

anyway if with query it's fine then I don't really mind, I thought it was related with my dependencies

then I think you’re good 🙂

it also looks like I never really need to use {:connection-uri (get env :database-url LOCAL-DB-CONN)} but I'm passing LOCAL-DB-CONN directly

which is a DATABASE_URL, and also seems to differ from what the docs say

but well as long as it works

thanks

hello all. I'm new to clojurians and wondering, do I need to pay money to see the full slack history or is that not even an option?

@gcast have you looked at https://clojurians-log.clojureverse.org/clojure/index.html ? I don’t know if it’s complete though

Slack’s search feature should work too?

@hmaurer no I have not. thank you.

hey folks, how do you use clojure.spec.test.alpha/check with clojure.test ?

Is there an official way of testing expression equality in clojure? https://stackoverflow.com/questions/45663629/testing-equality-of-clojure-expressions

@beoliver no, there's nothing that does this in clojure itself, and I'm not aware of libraries that have the feature either.

@noisesmith ok cool... Im writing a function to do it now. Just wanted to check that I wan't reinventing the wheel 🙂

interesting - I guess that would involve a deterministic renaming of symbols for all locals?

so that (let [a 0] a) and (let [b 0] b) would be equal

Yeah, just implemented it. Seems to work... It also means that (-> foo :blah) = (:blah foo)

@noisesmith if you are interested a minimal "working" example is

(defn normalize [expr]
  (let [hygenic-form (-> expr jvm/analyze emit-form/emit-hygienic-form)]
    hygenic-form))

(defn unify
  "throws an assetion error if impossible to unify"
  [mappings expr1 expr2]
  (if-let [bound (get @mappings expr1)]
    (assert (= bound expr2))
    (swap! mappings assoc expr1 expr2)))

(defn zipCompare [mappings expr1 expr2]
  (assert (= (type expr1) (type expr2)))
  (cond
    (symbol? expr1) (unify mappings expr1 expr2)
    (coll? expr1)
    (do (assert (= (count expr1) (count expr2)))
        (map (partial zipCompare mappings) expr1 expr2))
    :else (assert (= expr1 expr2))))

(defn unifiable
  [expr1 expr2]
  (let [mappings (atom {})]
    (if (zipCompare mappings (normalize expr1) (normalize expr2))
      true)))

and

(:require [clojure.tools.analyzer.passes.jvm.emit-form :as emit-form]
            [clojure.tools.analyzer.jvm :as jvm]))

not sure that would work, tools.analyzer uses the base name to generate hygienic let symbols

e.g a becomes a__0# and b becomes b__0#

it can be tweaked to do what you want, but that's how it behaves by default

@bronsa, yeah, so I just create a series of mappings

i pass an atom that keeps track so if a__0# ~> b__0# then it must always hold

(unifiable '(let [a 10] (-> {:a a :b 20} :a)) '(let [b 10] (:a {:a b :b 20})))
true

ah, gotcha

although something is going wrong... as the following should not be equal

(unifiable '(let [a 20] (-> {:a a :b 20} :a)) '(let [b 10] (:a {:a b :b 20})))
true

but when I step debug after instrumenting zipCompare, it throws the correct assertion error on the branch :else (assert (= 20 10))

just removed the assert and using bools has fixed it.

Hi! Quick question: what would be the idiomatic way in Clojure to pipe data through a number of function, each of which may or may not be applied based on a condition (the condition does not depend on the data). Something like this:

(-> data
    (#(if condition1 (f %) %))
    (#(if condition2 (g %) %))
    ...)

juxt might be what you’re looking for. It takes a number of functions and returns a new function that applies the arguments to all, returning a vector of the results.

user=> ((juxt inc dec #(str "hello, " %)) 123)
[124 122 "hello, 123"]

@chrisjd thanks but not quite! I want the data to be piped through: the result of applying the first function (if the condition passes) should be used when applying the second function, etc

or cond->

Ah, sorry.

The example I added to my message works, but it’s ugly

@gcast cond-> seems to be exactly what I was looking for; thank you!

@hmaurer no problem! never actually used it but seen it enough on

Hi guys. I have faced such problem. I write my personal ai assistant(chatbot), using http://api.ai, clojure as backend and deploy it to aws lambda. My problem is, that on cold requests, it takes up to 5 seconds to respond. I understand that it happens because of 2 reasons. First is aws internal container/server start up time. And second, jvm start up time. So i want to optimize my side - jvm. I don’t want to write backend in ClojureScript and then use Node.js as host platform. Do someone know, some methods of optimisation of jvm start up time. GC params or something. I appreciate any answers.

Another question: is there an “if-let” construct in Clojure that short-circuits? if-let only allows one binding. I would like a version which allows multiple bindings but short-circuits if any of those bindings are nil

Nevermind, found answers on stackoverflow…

do i need to manage serialization/deserialization of requests on server/client manually? (from transit/json to clj/cljs data structures) i am using luminus template for web apps meaning do i use libs like transit.clj/cljs, chashire.. it's not magically done for me via some built in thing?