@trailcapital Thanks for pitching in. The result still is empty. I am trying to loop over a collection and accumulate the result in a vector. Your code snippet also prints an empty result.

my code snippet definitely returns a value (`[2]`). Can you post more context so I can see why it might be returning empty? or are you just running the snippet and it is empty? is it nil or []?

(let [result []] result (conj result 2) (if true result)) is giving my a []. In your snippet I think there is a misplaced ]

ah true, my bad!

(let [result []
       result (conj result 2)]
  (if true result))

@trailcapital that works. Thanks

Happy Thanksgiving

my thanksgiving was back in October (Canada!), but thanks! you too.

hey guys, i am newbie, getting myself to learn clojure. I need to write test and want to know the difference between clojure.spec and clojure.test?

Generally speaking, a spec says what code should do, a test checks that it does

So I could say that a square function always returns natural numbers, thats a spec

And then check that, given 3, it will return 9, thats a test

Is there anybody use Carmine for operating Redis? How to iterate over all items in Redis list?

@lennart.buit thank you very much, things make sense now. so that mean i can use both 🙂

Yes, if you write your specs correctly you can generate your tests ^^!

@stardiviner you can use (car/llen) to find the length and (car/lrange start end) to get the values as a clojure list. They you can map over it.

i need to represent two different entities and considering using map or defrecord. Which one should i pick? Also both are immutable. So how can i am record changes on my entities?

@rahul080327 Good idea.

using deps.edn i'm trying to include a package from maven, and also a classifier from the same package. it looks like including the :classifier key works but doesn't load the main package, and i can't add the dependency twice because deps.edn is a map:

{:deps {
        {remote.package/parser {:mvn/version "1.0.0"}
         remote.package/parser {:mvn/version "1.0.0"
                                :classifier  "additional-package"}}}}

Anyone in here ever had any issues with shadow-cljs? When running (shadow.cljs.devtools.api/nrepl-select :app) I get the following error:

ClassNotFoundException shadow.cljs.devtools.api  java.net.URLClassLoader.findClass (URLClassLoader.java:382)

@vitorstipanich looks like shadow-cljs is not on your classpath. what did you use to start the JVM?

you mean the repl? I'm using the proto-repl on atom

I don't know how proto-repl works but here are some instructions to use it with shadow-cljs https://shadow-cljs.github.io/docs/UsersGuide.html#_proto_repl_atom

I'm definitely gonna check this out (thanks!!). I just choose proto-REPL cause it's my first time on clJS and earlier I was checking a tutorial. What is the most recommended one?

this might also help https://medium.com/@jacekschae/slick-clojure-editor-setup-with-atom-a3c1b528b722

atom is totally fine to use. I personally use Cursive and many other people use emacs

vscode is also ok

Good morning

Can someone help me with this schedule wrapper, it keeps giving me errors

(defmacro schedule [pool time func & body]
  (let [pool (vary-meta pool assoc :tag `ScheduledExecutorService)
        unit TimeUnit/MILLISECONDS
        unit (vary-meta unit assoc :tag `TimeUnit)]
    `(.schedule ~pool
                (fn [] (~func ~@body))
                ~time
                ~unit)))

CompilerException java.lang.ClassCastException: java.util.concurrent.TimeUnit$3 cannot be cast to clojure.lang.IObj

@thheller thats the tutorial I was following earlier. the problem is that he goes with keybinds and is using a mac. Im using a windows lol

I'm on windows too but no idea about atom so no idea what would be different 😛

Yeah I'm already stuck on the step5 of the tutorial lol. It says that I should run an command but it doesnt do anything

I mean the shadowcljs with protorepl one

I think it should show a dialog with host/port inputs?

@rcustodio just emit TimeUnit/MILLISECONDS directly instead of ~unit. can't set metadata on the TimeUnit instance itself. the above vary-meta sets the :tag on the pool symbol only. not the instance.

@thheller thanks, and without the ~???

yes without

so it becomes

`(.schedule ~pool
                (fn [] (~func ~@body))
                ~time
                TimeUnit/MILLISECONDS)

Thanks @thheller

@thheller yeah it should... but it doesnt do anything xD I think I think Im doing something wrong. Im running it with ctrl + r

ctrl+r doesn't do anything for me? ctrl+shift+p works for me

LOL yeah you're right

sorry xD

And how do I change this (fn [] (~func ~@body)) to runnable only, error More than one matching method found

^Runnable (fn ...) should do it

Thank you again

Now.. I have 2 files, client.clj and client/ping.clj The client needs to call ping/handle and the ping needs to call client/write, BUT it gives me Cyclic load dependency, what is recommended in this cenario?

I'm trying to come up with a rule to memorize about the use of functions with the word "some" in them, since there are a bunch of them and there is an inconsistency. - some and some-fn define "some" as "at least one result is truthy" (not false or nil, as usual) and you pass a predicate fn to return the result. - all the others (`some?`, if-some, when-some, some->, some->>) define "some" as non-nil (false is considered "something") and you don't pass a predicate fn. Sound right?

Which is more performatic, case or cond?

case is constant-time, cond is linear-time

@rcustodio Honestly, I wouldn't even hope to find a general rule for that - there's just too many unpredictable ways both the Clojure Compiler and the JVM could optimize (or not) this stuff. Although cond is supposedly linear time and case supposedly constant-time, you usually won't use them with enough args that you can rely on an asymptotic argument about performance.

But good news: benchmarking is super easy in Clojure 🙂 https://github.com/hugoduncan/criterium

YEah… I dunno what you mean by constant-time / linear-time, but I will try the bench, thanks

@rcustodio do note that they are not interchangeable. Case can only deal with constants, whereas cond can deal with pretty much anything.

I have (:x val), in this case the case would be better

For what I’m reading

Bozhidar Batsov's guide mentions it: https://github.com/bbatsov/clojure-style-guide#case

check out the docstring for it: >>> Unlike cond and condp, case does a constant-time dispatch, the clauses are not considered sequentially. All manner of constant expressions are acceptable in case, including numbers, strings, symbols, keywords, and (Clojure) composites thereof.

If you have a compile-time constant, you can use case. It wold not only be faster but IMHO you gives an information to the reader that you have a compile time constant.

Hmmm… if its not compile-time, the variable comes from a databse, then case would not be that performatic

Or the compile-time is the value to test

(case (:prefix message)
  nil
  "jtv"

when test expressions are compile-time constants

Not better, but more recommended

Thanks for all the help

Halp! I'm suffering from turkey left-over brain freeze, but isn't there a built-in function for this somewhere:

(if (pred value)
    value
    else-value
    )

or ?

there is no core function for what you're doing

that I know of, at least

There might be a bigger problem if you're having to do this very often.

Excessive branching is a smell IMHO.

or would work if I'm only interested in the truthiness of value

in this case I'm manipulating hiccup structures, which have an optional second attr-map