Presumably, it's to act as the recur target if the value has already been seen.

I mean, why not a loop?

Oh right

This is a neat hack. I’m going to steal it

I always wrote a loop in a condition instead of doing it this way. So elegant

To be honest, I'm not sure why it doesn't use a loop instead. A looping version might look like:

(let [step (fn step [xs seen]
             (lazy-seq
              (loop [xs (seq xs)
                     seen seen]
                (if xs
                  (let [x (first xs)]
                    (if (contains? seen x)
                      (recur (next xs) seen)
                      (cons x
                            (step (next xs) (conj seen x)))))))))]
  (step coll #{}))

Yes. I wrote it similarly too, which was why my question in the first place

I guess because the anonymous function has been named step and that name has been used in the call itself? Talking about the name given besides fn and not in the let binding.

Transducers, reductions, iteration.

thanks, iteration is interesting and I always forget about it

Ah, and, of course, anything that uses reduce itself and passes it a function that can return reduced. I see reduce-kv and run!, maybe there are others.

Loop/recur is also your good friend

*but unrelated to reduced, isn't affected by it.

Ended up just manually writing loop recursive

I don't know the actual reasoning, but doing it this way prevents fully realizing seqs.

i don’t think it returns a vector for take?

It does.

(defn split-at
  "Returns a vector of [(take n coll) (drop n coll)]"
  {:added "1.0"
   :static true}
  [n coll]
    [(take n coll) (drop n coll)])

this is the new 1.12 splitv-at

split*v*-at.

ah sorry

allows for infinite seqs i suppose

yeah that makes sense

(split-at 10 (range)) works, so should (splitv-at 10 (range))

Well, "X works so should Y" is by itself not a very good basis. :) map works on inf seqs, mapv doesn't. Well, it does work but it hangs, of course.

https://clojure.atlassian.net/browse/CLJ-2713

That doesn't really answer the question in the OP of "why not (into [] (drop n) coll)".

yeah it answers the "why does this exist" and i understand that the goal is efficiently splitting at a specific point, which the IDrop interface provides.

i guess it's just "implementation details" for why they're different types

vectors can efficiently drop, but cannot re-form a vector

it is designed for the case where you are splitting your way through an arbitrary collection (which may be lazy / infinite)

retaining the collection type for the remainder allows you to continue efficiently dropping from it without incurring the cost to copy the data into another collection

that makes sense, thank you for the explanation

partitionv[-all] is one way to do this, but splitv-at is better for the explicit loop/recur variant of that where you have more control over when to stop

both split-at and splitv-at get the drop advantages, it's just whether you want your partitions as vectors or seqs (vectors matches partitionv - that change is why we introduced a new function there, to avoid breaking existing assumptions)

that makes sense. the lack of a vector in the drop branch feels incongruent with the other -v functions but it's a small price to pay

it is a vector in the place where that's important

when-let doesn't examine the binding for the when- part of when-let

it only looks at the value on the rhs - and you gave a map

maps are always truthy

ah that makes sense, counter intuitive at first but more obvious if you deconstruct with more than one binding

clojure.core/when-let
([bindings & body])
Macro
  bindings => binding-form test

  When test is true, evaluates body with binding-form bound to the value of test

What's the stacktrace?

[java] Exception in thread "main" java.lang.ExceptionInInitializerError [java] at dynapath.defaults__init.load(Unknown Source) [java] at dynapath.defaults__init.<clinit>(Unknown Source) [java] at java.base/java.lang.Class.forName0(Native Method) [java] at java.base/java.lang.Class.forName(Class.java:467) [java] at clojure.lang.RT.classForName(RT.java:2207) [java] at clojure.lang.RT.classForName(RT.java:2216) [java] at clojure.lang.RT.loadClassForName(RT.java:2235) [java] at clojure.lang.RT.load(RT.java:453) [java] at clojure.lang.RT.load(RT.java:428) [java] at clojure.core$load$fn__6824.invoke(core.clj:6126) [java] at clojure.core$load.invokeStatic(core.clj:6125) [java] at clojure.core$load.doInvoke(core.clj:6109) [java] at clojure.lang.RestFn.invoke(RestFn.java:408) [java] at clojure.core$load_one.invokeStatic(core.clj:5908) [java] at clojure.core$load_one.invoke(core.clj:5903) [java] at clojure.core$load_lib$fn__6765.invoke(core.clj:5948) [java] at clojure.core$load_lib.invokeStatic(core.clj:5947) [java] at clojure.core$load_lib.doInvoke(core.clj:5928) [java] at clojure.lang.RestFn.applyTo(RestFn.java:142) [java] at clojure.core$apply.invokeStatic(core.clj:667) [java] at clojure.core$load_libs.invokeStatic(core.clj:5985) [java] at clojure.core$load_libs.doInvoke(core.clj:5969) [java] at clojure.lang.RestFn.applyTo(RestFn.java:137) [java] at clojure.core$apply.invokeStatic(core.clj:667) [java] at clojure.core$require.invokeStatic(core.clj:6007) [java] at clojure.core$require.doInvoke(core.clj:6007) [java] at clojure.lang.RestFn.invoke(RestFn.java:421) [java] at dynapath.util$loading__4920__auto__.invoke(util.clj:1) [java] at dynapath.util__init.load(Unknown Source) [java] at dynapath.util__init.<clinit>(Unknown Source) [java] at java.base/java.lang.Class.forName0(Native Method) [java] at java.base/java.lang.Class.forName(Class.java:467) [java] at clojure.lang.RT.classForName(RT.java:2207) [java] at clojure.lang.RT.classForName(RT.java:2216) [java] at clojure.lang.RT.loadClassForName(RT.java:2235) [java] at clojure.lang.RT.load(RT.java:453) [java] at clojure.lang.RT.load(RT.java:428) [java] at clojure.core$load$fn__6824.invoke(core.clj:6126) [java] at clojure.core$load.invokeStatic(core.clj:6125) [java] at clojure.core$load.doInvoke(core.clj:6109) [java] at clojure.lang.RestFn.invoke(RestFn.java:408) [java] at clojure.core$load_one.invokeStatic(core.clj:5908) [java] at clojure.core$load_one.invoke(core.clj:5903) [java] at clojure.core$load_lib$fn__6765.invoke(core.clj:5948) [java] at clojure.core$load_lib.invokeStatic(core.clj:5947) [java] at clojure.core$load_lib.doInvoke(core.clj:5928) [java] at clojure.lang.RestFn.applyTo(RestFn.java:142) [java] at clojure.core$apply.invokeStatic(core.clj:667) [java] at clojure.core$load_libs.invokeStatic(core.clj:5985) [java] at clojure.core$load_libs.doInvoke(core.clj:5969) [java] at clojure.lang.RestFn.applyTo(RestFn.java:137) [java] at clojure.core$apply.invokeStatic(core.clj:667) [java] at clojure.core$require.invokeStatic(core.clj:6007) [java] at clojure.core$require.doInvoke(core.clj:6007) [java] at clojure.lang.RestFn.invoke(RestFn.java:436) [java] at bultitude.core$loading__5569__auto____3.invoke(core.clj:1) [java] at bultitude.core__init.load(Unknown Source) [java] at bultitude.core__init.<clinit>(Unknown Source) [java] at java.base/java.lang.Class.forName0(Native Method) [java] at java.base/java.lang.Class.forName(Class.java:467) [java] at clojure.lang.RT.classForName(RT.java:2207) [java] at clojure.lang.RT.classForName(RT.java:2216) [java] at clojure.lang.RT.loadClassForName(RT.java:2235) [java] at clojure.lang.RT.load(RT.java:453) [java] at clojure.lang.RT.load(RT.java:428) [java] at clojure.core$load$fn__6824.invoke(core.clj:6126) [java] at clojure.core$load.invokeStatic(core.clj:6125) [java] at clojure.core$load.doInvoke(core.clj:6109) [java] at clojure.lang.RestFn.invoke(RestFn.java:408) [java] at clojure.core$load_one.invokeStatic(core.clj:5908) [java] at clojure.core$load_one.invoke(core.clj:5903) [java] at clojure.core$load_lib$fn__6765.invoke(core.clj:5948) [java] at clojure.core$load_lib.invokeStatic(core.clj:5947) [java] at clojure.core$load_lib.doInvoke(core.clj:5928) [java] at clojure.lang.RestFn.applyTo(RestFn.java:142) [java] at clojure.core$apply.invokeStatic(core.clj:667) [java] at clojure.core$load_libs.invokeStatic(core.clj:5985) [java] at clojure.core$load_libs.doInvoke(core.clj:5969) [java] at clojure.lang.RestFn.applyTo(RestFn.java:137) [java] at clojure.core$apply.invokeStatic(core.clj:667) [java] at clojure.core$require.invokeStatic(core.clj:6007) [java] at clojure.core$require.doInvoke(core.clj:6007) [java] at clojure.lang.RestFn.invoke(RestFn.java:3659) [java] at cloverage.coverage$loading__5569__auto____1.invoke(coverage.clj:1) [java] at cloverage.coverage__init.load(Unknown Source) [java] at cloverage.coverage__init.<clinit>(Unknown Source) [java] at java.base/java.lang.Class.forName0(Native Method) [java] at java.base/java.lang.Class.forName(Class.java:467) [java] at clojure.lang.RT.classForName(RT.java:2207) [java] at clojure.lang.RT.classForName(RT.java:2216) [java] at clojure.lang.RT.loadClassForName(RT.java:2235) [java] at clojure.lang.RT.load(RT.java:453) [java] at clojure.lang.RT.load(RT.java:428) [java] at clojure.core$load$fn__6824.invoke(core.clj:6126) [java] at clojure.core$load.invokeStatic(core.clj:6125) [java] at clojure.core$load.doInvoke(core.clj:6109) [java] at clojure.lang.RestFn.invoke(RestFn.java:408) [java] at clojure.lang.Var.invoke(Var.java:384) [java] at clojure.lang.Util.loadWithClass(Util.java:250) [java] at cloverage.coverage.<clinit>(Unknown Source) [java] Caused by: java.lang.ClassNotFoundException: sun.misc.Launcher$ExtClassLoader [java] at java.base/jdk.internal.loader.BuiltinClassLoader.loadClass(BuiltinClassLoader.java:641) [java] at java.base/jdk.internal.loader.ClassLoaders$AppClassLoader.loadClass(ClassLoaders.java:188) [java] at java.base/java.lang.ClassLoader.loadClass(ClassLoader.java:525) [java] at java.base/java.lang.Class.forName0(Native Method) [java] at java.base/java.lang.Class.forName(Class.java:375) [java] at dynapath.defaults$fn__68.<clinit>(defaults.clj:12) [java] ... 100 more

A tentative guess - an old version of CIDER nREPL?

Hum, I don't think it's in the dependency chain. I am runing cloverage though, and error is coming from that

The root cause is at [java] at dynapath.defaults$fn__68.<clinit>(defaults.clj:12) and before, which is not shown. dynapath.defaults used to be a vendored namespace in CIDER nREPL in 0.18.

(Maybe still is, dunno).

Hum...

Nah, probably not that. That line seems to only extend classloader types to some protocols.

Oh, no - that is that. I was looking at a newer version. Here it is: https://github.com/tobias/dynapath/blob/0.2.5/src/dynapath/defaults.clj#L40

There was a check for a wrong class in 0.2.4. How ancient is your tooling? :)

Ancient lol

It's using 0.2.4 of dynapath

Ya, so I guess I need to upgrade

Wait, how do you trace back it was checking the wrong thing in 0.2.4?

Oh, nevermind, I see it

I could literally just shell out to diff if I really have to, but...

Yeah, when I needed to diff two JSON files printed out for humans, I ended up with shelling out to diff. :) Couldn't find a more reasonable solution.

Hmm if that's a gap in the ecosystem... adds to list of yaks to be shaven later

It might've easily be a gap in my ability to search. Or a gap in the ecosystem w.r.t. my own very particular needs at the time. It could easily be that your needs are different enough, or that enough time has passed since then, that now there's a thing you can use instead of bin/diff.

I should add that my needs included the best performance available since it was about relatively large JSON objects. For smaller data, maybe even diff2 by LambdaIsland can be enough, despite it being a deep diff.

I could make a deep diff work by doing something like stringifying the data, as long as it can detect insertions. Really I just need the right hand side of the diff. I am looking for items on the right side which are unaccounted for on the left side, where the items are not necessarily unique but are sorted.

And I'm looking for them by position in the original right hand side list, because I'll be stripping some data away for the comparison which I must then reconstitute. Configurable equality semantics could achieve that too I suppose

There's also this, maybe helpful, dunno: https://github.com/juji-io/editscript

Oh hmmm. I'd passed over that, but yeah that might do it. I'll do some tinkering. Thanks!