Anyone have any experience using an ouath v1 client with clojure? I’m trying to translate an example from Ruby but running into some issues.

I find myself using something like (first (filter #(= (foo %) "some-val") xs)) a lot. Before I make it into a first-where/`first-with` function, I bet I'm overlooking a better way to do it. Is there one?

More general would be (first (filter pred xs)

(some pred? coll) is that what you are looking for? Edit: Uhh, I am not sure whether I am reading that correctly

yeah no ignore me

^^

@johb.maier There are several reasons why you probably shouldn't be doing that...

(this came up very recently in one of these channels)

In general filter is going to give you a chunked result so first doesn't mean that filtering stops on the first element -- up to 31 more elements might be realized.

Also, if you need to do a linear search, you might have the wrong data structure in the first place. Something to think about.

@seancorfield Thanks. It's in a testing context (I'm traversing a kind of "DOM") and a "linear" collection is what the library gives me to work with.

So it's not a data structure I control sadly.

Then, (first (filter ...)) is likely the best you're going to get.

I think it was @andy.fingerhut who said, when this came up elsewhere, if it isn't in core, that's not because the core team haven't thought about it 🙂

I guess the best I could do is write more utility functions to hide that functionality and ease traversing the "DOM".

Ha, that's a good quote 🙂

I usually ask because I'm often sure I just don't know some core functionality or interpret it wrong.

Are you on clojure or cljs?

This is cljs at the moment.

(.find #js[1 2 3 4] even?)

JavaScript arrays have a find method, that is really like find first

That's good to know, thanks! It's not a js array in my case though (pretty sure at least, I will check).

If it isn't, you're better off with (first (filter as converting to a js array using clj->js will probably be much slower

But if it is, it can be simpler

Oftentimes you can avoid clj->js and js->clj. For shallow conversion of an effectively immutable JavaScript array to ClojureScript, vec is much faster than js->clj. Going in the opposite direction, into-array is much faster than clj->js.

occurs java.lang.Exception: No namespace: reply.eval-modes.nrepl found

anyone knows the reason ?

It's caused by a version conflict -- a bug in Boot 2.8.2 https://github.com/boot-clj/boot/issues/720

If you downgrade to Boot 2.7.2 it will go away.

thx, i will try it

Is there any way to decorate protocol methods when defining a record? E.g.:

(defprotocol MyP
  (thing []))

(defrecord MyR
  MyP
  (thing (decorator (fn [_] 1)))

no, you can use a higher order function inside the implementation of course, but that syntax isn't valid

whats the best way to join collections and singles of things?

seems concat is bad ?

for example i want eveything to be in one vetor.. Out of these args: [1 23] 4 56 [5 8 9]

concat is fine in a lazy context, the problems happen when you mix eager evaluation with stacked concat invocations

for that specific example #(vec (flatten %&)) would work, I can't think of anything else simple that would handle it

but flatten has some gotchas

apply conj?

order dependent though

ah ok.

theres so many..

its very confusing

each with so many gotchas.

The best way of using socket in clojure is by java interop?

that’s what I usual do at least, I don’t know of any clojure socket libraries and socket interop is pretty painless

why does get only work on vector literal but not on a vector that is def?

(get [(t/now) (t/now)] 1) => #object[org.joda.time.DateTime 0x1dd01b46 "2018-11-06T22:11:10.695Z"

but (get mydates 1) => nil

What is mydates?

where (def mydates [(t/now) (t/now)])

That should work...

hmm its not working my repl

(! 509)-> clj -Sdeps '{:deps {clj-time {:mvn/version "RELEASE"}}}'
Clojure 1.9.0
user=> (require '[clj-time.core :as t])
nil
user=> (def mydates [(t/now) (t/now)])
#'user/mydates
user=> (get mydates 1)
#object[org.joda.time.DateTime 0x2cd2c8fe "2018-11-06T22:16:35.465Z"]
user=> 

Perhaps you have mydates shadowed by a local binding in the context where you call get on it?

Just used a different name and still get nil

Show code -- or your REPL session...

Does nth work?

get on anything that isn't associative simply returns nil

any chance you turned the vector into a lazy-seq with map, filter, etc??

nth returns UnsupportedOperationException nth not supported on this type: Unbound clojure.lang.RT.nthFrom (RT.java:888)

that means mydates is a var that hasn't received a value yet

is it somehow declared but not defined yet?

com.handler/start> (def testt [(t/now) (t/now)])
#'com.handler/start/testt

com.handler/start> testt
[#object[org.joda.time.DateTime 0x21712cdb "2018-11-06T22:23:00.389Z"] #object[org.joda.time.DateTime 0x6867ddbc "2018-11-06T22:23:00.389Z"]]
com.handler/start> (get testt 0)
nil
com.handler/start> 

your namespace is bogus

namespaces with "/" in the name are not going to work correctly

So the issue is happening because my namespace is busted?

that would be my guess, I mean, I am not sure what the mechanics going on behind the scenes that are causing that to happen are

but putting '/' in a namespace name will break a lot of things, so I am not surprised it breaks that

Just ran the app and it worked so you were right. The namespace does seem to be busted

Clojure 1.9.0
user=> (in-ns 'foo/bar)
#object[clojure.lang.Namespace 0x41709512 "foo/bar"]
foo/bar=> (clojure.core/refer-clojure)
nil
foo/bar=> (def a 1)
#'foo/bar/a
foo/bar=> a
1
foo/bar=> (identity a)
#object[clojure.lang.Var$Unbound 0x51972dc7 "Unbound: #'foo/bar/a"]
foo/bar=> 

I am kind of surprised you were able to create a badly named namespace like that, it takes a fair amount of work to do (the ns macro won't let you)

I didn't create it but I dont have a file with a namespace defined as such so not sure whats happening

I have a handler namespace with a function start in it though

you might be using an older clojure version without the checks for the ns macro

How did you start that REPL session? Clojure/Java or ClojureScript?

Clojure

I mean, no need to answer those questions if you can start a new REPL and the issue goes away, but if it is a repeating problem you are having, and you are interested in figuring out why, maybe we can figure it out.

In my project.clj I have a :clj { . . . :main com.handler/start . . . }

The value of the :main keyword in Leiningen's project.clj file should be namespace name, not a function

Right above that :main I have :ring { :handler com.handler/app, :init com.handler/init}

It is a little surprising to me that nothing blew up before the REPL started up and gave you a prompt, but there are enough functions in Clojure that don't throw exceptions on invalid input, that I am not completely surprised.

I don't know ring well enough to answer, but would guess that those values are expected to be function names, not namespace names.

what is :main used for? Is that what the REPL is going to fire up as the initial namespace?

yes

It is optional.

This heavily commented sample Leiningen project.clj file has lots of info on what various parts of a project.clj file are for, but I wouldn't recommend reading it from cover to cover unless you are bored or intensely curious. Use it when you see something unfamiliar, although since project.clj files can also contain library-specific add-ins like the one you showed for Ring, this file cannot document everything you might see: https://github.com/technomancy/leiningen/blob/master/sample.project.clj

com.handler> (require '[clj-time.core :as t])
nil
com.handler> (def mydates [(t/now) (t/now)])
#'com.handler/mydates
com.handler> mydates
[#object[org.joda.time.DateTime 0x6b9cf4ad "2018-11-06T22:41:15.525Z"] #object[org.joda.time.DateTime 0x7cce1e84 "2018-11-06T22:41:15.525Z"]]
com.handler> (get mydates 0)
#object[org.joda.time.DateTime 0x6b9cf4ad "2018-11-06T22:41:15.525Z"]
com.handler> 

Thanks @andy.fingerhut @hiredman