can you disassemble an inner class with no.disassemble?

yeah

usually I just do (disassemble (Foo.)), but (disassemble (Foo$Bar.)) doesn’t seem to work

inner classes don't exist at the jvm level

@schmee define doesn't work

at the bytecode/jvm level a java inner class is nothing different than a regular class, with an $ in the name

well, I get an empty string

¯\(ツ)/¯

I don't know how to help with that but I wouldn't focus on it being an inner class

aight, I’ll dig around some more

it might be that there's some bug in no.disassemble

So I’m quite sure this is not in the standard library but I’m trying to write a function that takes the output of partition and combines the overlapping parts with a function

so for example I have (partition 1024 768 (repeat 0) coll) and I want to combine the overlaping partitions with *

so you want to reduce each partitioned list to a numeric value?

or do you mean something else by "overlapping"

wait let me get my ipad, I’ll draw out what I mean

Is the image visible or do I have to host it on some other image service

that's visible

seems like you have to take the partitioned input, then make your "un-partition" by counting the items in the first element

no, wait, you need to know the second arg to partition I think

You would have to know the “overlap rate” do to speak

oh, right, the first and second arg are needed for that

My current setup for providing https:// and wss:// is: 1. provide http:// and ws:// via clojure http-kit/ring 2. use some node.js wrapper to do magic with *.pem, open up https:// and wss:// and redirect to clojure's http:// and ws:// question: is there a nice way to do https/wss directly in clojure? I have all the certificates/private keys, I just need the webserver to do the encryption stuff

I'm running some Clojure code that counts the number of words in a wikipedia dump. When I run it with lots of pages (500+) get a NullPointerException [trace missing]. How can I debug this further?

Never mind; manually dug into the code and found that one of the seqs was nil

@qqq there is https://github.com/nginx-clojure/nginx-clojure which should allow to use the nginx way of doing, and should also give you http2 easily. Still have to try for myself, no priority so far.

12001 user here 🎉

there was a question in #cider about how to turn off the prepended namespace from maps with namespaced keys like

clj -e "{:a/a 1 :a/b 2}"
#:a{:a 1, :b 2}

user=> (clojure.repl/doc *print-namespace-maps*)
-------------------------
clojure.core/*print-namespace-maps*
  *print-namespace-maps* controls whether the printer will print
  namespace map literal syntax. It defaults to false, but the REPL binds
  to true.
nil

@zalky ^

thanks @noisesmith

can I def a primitive byte?

(def ^byte ^:const ubyte (byte 0x01)) doesn’t seem to work when I look at the bytecode

yeah no

a def can't hold a primitive

clojure doesn't really support primitives other than long and double that well

even a primitive long can't go in a def though, right?

not in a def

but if the def is const

then a primitive long will be inlined at the usage sites

oh, right

so technically the def would hold a Long, but usages of that var compatible with primitive values would only see an inlined primitive long

I find it quite annoying how clojure ignores the existence of primitives other than long and double

for no good reason IMO

combinatorial explosion of IFn method signitures

is it a some sort of fundamental limitation or just priorities?

no that's not what I'm talking about

@noisesmith I understand the rationale for supporting only long/double as primitive function arguments

but clojure doesn't really deal well with other primitive in contexts where there's no technical limitation for it

@schmee no fundamenal limitation

just a lack of interest in supporting it from what I can tell

I've made a few patches enhancing support but they've been rotting in jira for a while so it's clear that there's not a great deal of interest about it

link them and I’ll give ’em a vote 🙂

so, if I have a byte[] and want to do a bunch of stuff with it, I guess I’m better off using long[] in Clojure and shed a tear for all my wasted bits?

there's e.g. https://dev.clojure.org/jira/browse/CLJ-1905

jira's search sucks, there's others but I don't want to spend my time looking tbh

@schmee well, depends

if you stay in a local scope, primitives other than longs and doubles are somewhat supported

as long as you don't escape a fn context or loop or...

:)

pick your poison

@schmee an array of byte still works - it's just standalone unboxed bytes that have this issue

it's the "doing a bunch of stuff with it" part that might be a problem

oh, right, of course

but you can stil use an array of byte - worst case converting to long, but in the local scope it might even do the right thing

yeah, just to give some context, I wrote a board game thing in Clojure, and just to see how fast it could go I wrote another implementation in Java that uses bitboards

now I want to see how closely I can emulate the Java bitboard version in pure Clojure

@noisesmith eh, clojure will sometimes do weird stuff

like byte -> Byte -> long

yeah, if you’re using native bytes, you’re doing it because you want something to be predictably fast

if something gets auto-boxed in any place everything just goes down the toilet

but by all means, you should try and see for yourself

it's possible that for your usage you'll get the right unboxed behaviour

just beware

yeah, thanks for the advice

it would be cool if the compiler had some additional support for primitives, that would mean that I could use Clojure to replace Java even for the low-level stuff 🙂

@schmee I forget - did you mention using no.disassemble?

nope, but I do and it’s indispensable 🙂

I'm still an old dinousaur and compile my stuff and decompile it using java c p

it’s a fools errand to do primitives in Clojure without it 🙂

well, I have an emacs mode that does it JIT when I open aclassfile

or javap, of course

but I shoud start using no.disassemble

@schmee sometimes I fear that that sentence could end at "Clojure"

I don't remember 100% anymore but I believe I made t.e.jvm work with all primitive types as you'd expect

@noisesmith speaking of combinatorial explosions, I just looked at the IFn interface… that’s not something you see in Java everyday 😄

is there really no builtin for: (def atom? (fn [obj] (instance? #?(:clj clojure.lang.Atom :cljs cljs.core.Atom) obj))) ?