pretty sure httpkit can do it

user=> (.(Random.)nextInt(100)) ClassCastException java.lang.Long cannot be cast to clojure.lang.IFn user/eval150 (NO_SOURCE_FILE:4)

I'm trying to use the Random class from Java

But I'm getting the above error

I have done (import java.util.Random)

Guys.. Got the answer

@suryapjr There is also a beginners channel.

Yea.. Sorry

Was in a hurry

:)

Guys..clojure is interpreted?

no

it’s JIT compiled

just in time compiled ?

yes

so its fast enough for GUI's ,rught ?

sure

Great !!

Hope this is a legit question on this channel, I wonder what would be the state-of-the-art web server library to use in clojure, for a not-very-complicated small web application? in simple, I mean not too many pages or views belonging in the web application, and no huge traffic. Quite a devops oriented web application...

@matan A lot of people start with luminus — it’s a quick way to pull in a lot of standard libraries and get a server up and running with minimal fuss. http://www.luminusweb.net/

If it's fairly basic, I might skip Luminus and do more of a bare-bones ring handler function with some ring middleware, with hiccup to create html from data.

I’m doing something similar, but more of an API server. Looking at yada, compojure-api, and others. Thoughts?

I’m interpreting that as you @alexmiller didn’t like it! 🙂

I think it’s highly unlikely that Rich would like it :)

esp because :or in e.g. destructuring and or have different semantics

this would introduce a third

@teng - Do you know about fnil?

on the simple server topic, did any of the revolutionary ones (that like should have been what datomic is to data) end up very usable and stable? I think there was a project om.next going, which was somewhere between odd and experimental at the time

@jeff.terrell Thanks for the tip. The alternative was to introduce a let-statement, because the incoming parameter was used in several places in the code and I didn’t want to introduce code duplication by using the ‘or’ or ‘fnil’ function. Clojure supports destructuring in the parameter list for maps that also supports :or, my idea was to support the :or part even for other types than maps.

Is there some better (as in more idiomatic and/or faster) way in clojure to get the file line/col of the calling function (when in a function) than traversing (.getStackTrace (Thread/currentThread)) or (.getStackTrace (Throwable. ))?

Is there something, which can be dereffed to get the current value. Where the current value is updated by a periodic call. I kind of want a (sliding-buffer 1) that's dereffable.

an atom with a background thread that swaps it?

@ghadi For my case, the contention would potentially be prohibitive. There could be a lot of updates in bursts.

My initial thought was an agent, but the lack of backpressure is meh.

agents handle contention by queueing instead of retrying, which might be worth it if your retries are expensive

ahh

how could a single writer produce contention ?

@leonoel that's a good point. I was thinking about that wrong. swap! is sync isn't it?

yes, but in your case you likely want a dumb reset!

or a volatile!

Well, I'm doing step calculations, so it's a good fit for swap!

sliding-buffer takes values, whereas what you are describing is an update function

if you can ensure a single writer, you don't need more than volatile

how would atom / swap! apply any backpressure btw?

if backpressure is a problem you probably want a proper queue or channel setup

@noisesmith it doesn't, other than that the thing consuming the queue to perform the swap! has backpressure.

I have a channel setup, don't worry.

OK but I don't see how any of this leads to an agent being a problem

TDEPS: Is there an easy way to add a dir to the classpath? Without deps.edn but on the command line.

@rauh not in tdeps, you could port the add-classpath! function from pomegranate, or use dynapath

@noisesmith Because if I am consuming from a queue, and immediately sending it to an agent faster than the agent can consume that queue, that will fill memory.

and atom/swap! doesn't have this problem? I'm not picturing this

@noisesmith swap! is sync, so the thing consuming the queue in the first place is blocked until the swap! is complete. send-off is async.

oh, right, so it's about keeping the backpressure on the queue you consume from, I get it

what kind of queue?

@ghadi I'm using manifold. The queue is handed to me. (But in practice I'm only using manifold.stream/stream for now).

@dominicm Turns out {:paths ["foo/bar"]} works fine for adding to the CP on the command line

@rauh I totally misread, and thought you were trying to do it dynamically.

Hey Clojurians! I've posted this on the #clojure channel on Freenode, but posting here as well in case someone has some ideas about my question. I've started playing out with Compojure and built a simple API, and was wondering how would I go if I wanted to create new clojure.spec's dynamically by simply POST'ing for example some payload to my API endpoint? How would I serialize the spec, so that the API knows how to build a spec out of it? Is this something feasible to do? The idea is to allow users to create new clojure.spec's on demand.

@dnaeon specs are code, so technically load-file is the best you can do.

Or load-string even.

@dominicm that would be ideal I guess, but the end goal would be to allow non-dev users create specs. Having an API that can create specs and a nice frontend is what I'm thinking about. The users should not really know and care that behind all this is a clojure.spec.

Does this sound even reasonable to try implement? 🙂

@dnaeon so you're looking for a front-end which makes it easy for non-developers to build specs?

@dominicm yep.

It would be great if the users were able to write the specs, but they are not really developers, thus the need for a frontend which makes this simple for them.

The hardest part I guess would be to provide a frontend which covers all that clojure.spec provides in a nice and intuitive way to use.

@dnaeon It may be easier to build something that's more constrained to your domain.

Then have a general-purpose predicate which takes 2 inputs, the DSL & the input to validate against.

Has someone worked with AWS Lambda? I'm trying to read a file from the resources folder and it's not finding, I'm using serverless for packaging

@pfeodrippe the most common error here is that you use io/file or java.util.File to open it - if it's in your resources, it will be inside your jar at runtime, which means you should use http://clojure.java.io/resource to access the contents

or you can extract the jar to disk, but you probably don't want to do that in lambda, even if you could

@noisesmith Thank you, noisesmith, gonna try it, and yes, I'm using clj-pdf and it must use io/file, gonna try your way o/

It's working, thank you 😃

What's a good shorthand for (when (pred? (expensive-call)) (expensive-call))

basically a single element filter

but I don't want to calculate things twice

https://clojuredocs.org/clojure.core/when-let

when-let and if-let don't cut it because they bind the result of predicate

which is true but not the value being tested

oh shoot ya

a couple of these functions don't play nice with predicates that return booleans

like some

i guess just wrap with a let to store the expensive call

isn't that useful for "find an element that satisfies predicate x" when predicate returns a boolean

(let [x (expensive-call)] (when (pred? x) x))

edit: oops, focus of the cursor. At least it wasn't a password 😛

yeah, well I was looking for something shorter

Hello guys, any idea how to setup working Sentry appender for Timbre logging (from 3rd party appenders)? I have a configuration like

{:level :info
 :ns-blacklist ["com.mchange.*"] ;disable Korma DB pool logging
 :appenders {:spit (rotor-appender {:path "mylog.log"
                                    :backlog 10})
             :sentry (sentry-appender {:dns "url"})}}

Ohhh, is this nifty metadata handled by Emacs?

When I get an error like this at the repl, is there any way to find out what line number in the source file caused it? ArithmeticException Divide by zero clojure.lang.Numbers.divide (Numbers.java:158)

@ajs clojure.repl/pst will show the last stack trace nicely if you call it

there's also the value *e which always holds the most recent uncaught exception, including its stack

perfect, thanks