I'm looking for an Edn library for .net.

We evaluated https://github.com/edn-format/edn-dot-net but found it unusable.

@olivergeorge almost all non Clojure(Script) edn readers are unusable. That's one of the key reasons Transit was written: https://github.com/cognitect/transit-format

Nevermind that Transit is a lot faster than EDN

I had asked here about the fine print at https://clojure.org/reference/transducers yesterday, and got a link to a youtube video featuring very heavy accent and in general hard to follow. Can anyone recommend a good resource to learn about these caveats of transducers? > If you have a new context for applying transducers, there are a few general rules to be aware of: > • If a step function returns a reduced value, the transducible process must not supply any more inputs to the step function. The reduced value must be unwrapped with deref before completion. > • A completing process must call the completion operation on the final accumulated value exactly once. > • A transducing process must encapsulate references to the function returned by invoking a transducer - these may be stateful and unsafe for use across threads.

I am concerned at going into transducers without apt understanding of these issues, as the guy with the heavy accent said in the beginning of his video, not understanding these cause him great grief and frustration, something I cannot afford as they translate to zero productivity.

Perhaps you can recommend a book or resource specifically addressing these or more broadly the safety rules or must knows for transducers?

Can I serve files with compojure, which are not in the resources folder?

@matan I'd start with "Transducers by Rich Hickey": https://www.youtube.com/watch?v=6mTbuzafcII (If you haven't seen it yet) and then continue with "Inside transducers": https://www.youtube.com/watch?v=4KqUvG8HPYo One good example that I've recently seen is also https://tech.grammarly.com/blog/building-etl-pipelines-with-clojure

needed to strip some non-printable things from an ex-data thing, ended up with this but listing things like that feels like a certain way to miss things:

(defn strip-non-serializable [m]
  (walk/postwalk (fn [x]
                   (if-not (or (string? x)
                               (keyword? x)
                               (number? x)
                               (symbol? x)
                               (map? x)
                               (vector? x)
                               (list? x))
                     :ommitted-not-serializable
                     x))
                 m))

@martinklepsch I suspect that's the best you can do

(eventually it goes into cheshire but cheshire doesn’t seem to have a jsonable? type thing

java.util.Date is also printable

I've needed a function like this many times by the way

You can have it 😛

to "sanitize" a data structure for serialization/priting

thanks!!!! 🎁

would be nice if serialization libs (I’ll also need this for nippy) expose a “serializable?” predicate to let you do this kind of thing

(maybe nippy has this, haven’t checked)

alternatively a flag to just skip over/replace all non-serializable values

that’d be “nicer” but a predicate would just get the job done w/o much consideration of impact on public APIs etc https://github.com/dakrone/cheshire/issues/118

it would also be faster..

true

@danielgrosse IIRC, yup!

@martinklepsch I think nippy is protocol based, so in theory you can use satisifes?

(didn't check)

@mpenet yeah could be. Didn’t get to nippy yet anyways

@martinklepsch Use seq? instead of list? since lots of data structures end up as seqs that aren't actually lists

@stuartsierra good point

Is there a way to get a .cljc file to hot reload while developing? prior to adding it, we had our .clj code hot reloading anytime we made a change. We're using lein and immutant, among lots of other libraries. Using emacs/cider to develop, and even C-c C-k doesn't recompile the cljc code.

Can anyone recommend a good Cursive cheatsheet. I’m helping someone who uses Intellij and a quick google didn’t produce anything I liked. Intellij is not my goto ide…

They’d prefer to use the default bindings

Anyone here have a favourite way of working with elasticsearch through Clojure? https://github.com/clojurewerkz/elastisch seems a bit outdated. https://github.com/mpenet/spandex ? Raw http requests? Something else?

https://crossclj.info/ns/cc.qbits/spandex/0.3.4/project.clj.html

https://crossclj.info/ns/timgluz/elastisch/3.0.0-snapshot/project.clj.html

based on crossclj I’d say elastisch is more mature and better supported / more widely used

Wrong assumption :) but I am biaised

elastisch seems to lag behind in elasticsearch versions though

I am not sure elastish will work with 2.4+ too

Spandex is quite new, so crossclj doesnt paint a clear pic. It s been running millions of queries per day in prod without hiccups

@mpenet I hope it didn’t look like I was comparing their merits. Of course “better supported” is speculation but clojurewerkz has been around for a while and that counts for something. No disparagement meant toward your project of course.

what differences does the transport protocol vs the HTTP protocol offer?

i've never actually used transport

Persistent cnx and binary proto plus few other things. But it's going the way of the dodo for clients

And perf is nearly identical between the new http client and legacy native now

The readme of spandex links to a blog post from elastic talking about that

(On my phone sorry)

@mpenet I'll take a look at spandex then! 🙂

Been using spanex, works great! I'd choose it over elastisch again (Which I also used).

have only used elastisch, but it was quite nice

@curlyfry I have also only used elastich and found it worked great, but it’s been about a year since I was doing that work.

@noisesmith @mpenet @rauh @joshjones @bballantine Thanks for the input!

@ztellman Do you ever sleep? I thank you very much for the effort you put into your software. I really do enjoy #aleph and #manifold. Thank you very much!

@ddeaguiar There are a couple here: https://www.cheatography.com/tag/cursive/

Note that I don’t think his windows bindings are the defaults, although I may make them the defaults due to https://github.com/cursive-ide/cursive/issues/1139

has anyone considered a development mode for clojure where when you evaluated

(def some-user {:first-name "Adam"})

...

(:firrst-name some-user) ;; => Exception

Obviously you could do something similar with spec and making sure that maps conform to having those keys, but I would like to try something with minimal effort on the programmer's part.

the IFn interface for Keywords is in the clojure source as Java.

because things read, then evaluated

so the data returned by the data reader there is passed through the compiler after the read is done

x/y1 reads 3 and evals it to 3 x/y2 reads (+ 1 2) and evals it to 3 x/y3 reads (quote (+ 1 2)) and evals it to (+ 1 2)

@hiredman @bronsa I see. I was looking at the built in stuff (uuid and inst) and none of those seems to be quoted anywhere

but, based on the above, they could be quoted once and still be correct, right?

because they return data that evaluates to itself

no, if they returned quoted data, it could change what macros see

when foo3 is invoke, it does return a quoted thing, it returns a list

err. foo2

the quotation, when foo2 is compiled, tells the compiler to compile generate code that creates a list object at runtime, not compile the list as a function call

so what foo2 is returns is a list

and data returned by a data reader is as if the reader returned that data

so if you evaluate a list returned by the reader, well we do that all the time and we know how lists evaluate

@hiredman I think I understand

what seems strange to me is that a macro then, if passed #inst ..., sees not some literal data, but an object

reader-macros are expanded before regular macros are

we need to go deeper. pre-read time 😄

pre-read, what you have is an inputstream