Hi, I am facing really weired issue. Maybe i am trying something very stupid

Here is piece of code

Works fine if i make a post json request

{ "text": { "text": "some text" }, "_id": "579ec8fd6c6f8134d72a14cd" }

but if I try to access the key, i.e. text , its null

I checked the type of bdy and its PersistentArrayMap

Any help is really appriciated

application uses liberator, compojure and monger

if I directly use bdy as in first snippet, it adds entry in mongodb correctly. however if i try to read property of the body and add (e.g. by replacing {:text bdy} by {:text text}, :text is added as null

@udhan: Is the key in your bdy map "text" or :text?

I am assuming json/read converts string keys to keywords, however i tried both

my request post body is simply

{ "text":"some text" }

can you print/log bdy to see what's actually in there?

anyways, was just a guess as I head to bed. good luck!

logger returns {text some text}

sry I had to set logger as app wasnt printing anything on console

@codonnell: no problem. Gn!

What's the best way to remove all keys but a specific few from a map?

Oh, select-keys

@pesterhazy: regarding coercion, let's take simple example of transforming {:a name :b age} to {:a age b: name}, but not manually, because I want all the good stuff that Schema gives me, like validation and handling more than this one example.

Perhaps this is a bug in clojure/data.json: When I convert / to json in Ruby I get this "/".to_json #=> "\"/\"" When I do the same in Clojure I get this (json/write-str "/") => "\"\\/\"" And when I apply Base64 encoding to the outputs I get different results as well. Am I missing something?

I'm trying uploading files to AmazonS3 from front-end. In order to do that I need to generate a base64 signature. And the signatures based on data which includes pathnames.

As I can check, cheshire does it in the same way as ruby:

(require '[cheshire.core :as json])
(json/encode "/")
;; => "\"/\""

seems like you’re trying to import a class that does not exist ?

where?

every freaking day is a battle with setup and tools

programming time 1%

how to fix this? why this appeared I didn’t change any dependency just reopened idea and now I have this trace with 0 info in it

it’s a matter of becoming more familiar with the ecosystem

@ilukinov: just by opening intellij? did you run anything inside it?

i think a good way to reduce to struggle with setup and tools is to limit the amount of automatic updates etc

this appeared when I trying to run repl

it throws this in my face and then still will start repl

then when I try to load any clj file from my project it with throw again Loading src/eem_rest_api/scenarios.clj... java.lang.NoClassDefFoundError: Could not initialize class clojure.core.async.impl.channels__init clojure.lang.Compiler$CompilerException: java.lang.NoClassDefFoundError: Could not initialize class clojure.core.async.impl.channels__init, compiling:(clojure/core/async.clj:1:1)

@ilukinov: Can we see your project.clj file and maybe some more of the project?

@ilukinov: maybe clean out your target/ directory.

@hans we have a winner! why the hell it uses something from target folder? Why this “something” is broken? How I could figure it out without slack?

how are you people staying sane being a programmers - I don’t know, it’s a mess, total mess.

if you don't like solving problems, then you should choose another profession.

1) everything is broken 2) it cannot be fixed 3) we can still get things done

i think it is fair to say that there are different types of problems, and dealing with package manager quirks is not high on my list of problems i like to solve either

i like solving problems, but I don’t like battling with tools and digging into error messages with 0 info, It’s like if I use hammer yesterday and it worked, and today it became screwdriver.

however, in defense of clojure, it’s a relatively new language and not as mature as, say, java. the toolset voor java (or even scala) is much more mature.

@ilukinov: it only feels like that because you're inexperienced. this is certainly a problem with clojure (or most lisps, for that matter)

@ilukinov: after only a few years, you can make sense of the error messages even though they're sometimes not very explicit (or, which is often worse, too explicit).

I think you just get tolerance tho pain thats it

so you saying that you figured out that I need to clean my target/ folder from that error message about core/async?

yes. if you get low-level class-not-found errors, they're often caused by outdated .class files.

this is pretty common stuff on jvm, not clojure strictly related

java will thow similar error if such thing happens

dependency detection is hard. all tools that try fail at some point.

I found similar stuff in dependency detection in almost all languages that I have used. Although, clojure error are not the best on the world, I found similar stuff and cryptic errors in a bunch of other languages (in majority more mature that clojure...)

here is another one WARNING: boolean? already refers to: #'clojure.core/boolean? in namespace: clojure.tools.analyzer.utils, being replaced by: #'clojure.tools.analyzer.utils/boolean? WARNING: boolean? already refers to: #'clojure.core/boolean? in namespace: clojure.tools.analyzer, being replaced by: #'clojure.tools.analyzer.utils/boolean? WARNING: bounded-count already refers to: #'clojure.core/bounded-count in namespace: clojure.core.async, being replaced by: #'clojure.core.async/bounded-count WARNING: Inst already refers to: #'clojure.core/Inst in namespace: schema.core, being replaced by: #'schema.core/Inst WARNING: boolean? already refers to: #'clojure.core/boolean? in namespace: fipp.visit, being replaced by: #'fipp.visit/boolean?

what does this mean? how the hell definition was replaced? what about name spaces? Does this mean that boolean? is no longer boolean? clojure’s native boolean? was replaced? why?

Yeah, that’s because they just added boolean? to clojure.core — it wasn’t there before so a bunch of libs all implemented it, and now it’s generating warnings

I’m pretty sure it’s safe to ignore for now, I haven’t heard of any problems with it

to much to ignore for me

If you revert back to clojure version 1.8.0, you can avoid that, but you won’t get any of the clojure.spec stuff

so I guess the more you ignore the less pane you have when you programming 😀

lol

@ilukinov: tbh, running a beta version of clojure is not a good idea if you don’t like these kind of problems

heck, alpha version even

@lmergen: but clojure.specs is so tasty, that I will ignore those errors

My REPL keeps getting into a bad state where things like this happen:

(def first-of-month "2015-01-01")
=> #'user/first-of-month
(str first-of-month)
=> "Unbound: #'user/first-of-month”

huh, weird

is the namespace changing?

How did you get that repl?

It’s a Cursive REPL and I’m attempting to use the ‘Reloaded’ pattern

mostly successfully

but I have to restart the REPL a lot more often than I want to, because it gets into this state

it’s starting to get annoying, so I thought I’d try and solve it

I found a reproduction case: calling (reset) causes it, but only when (set-refresh-dirs “src” “dev”) includes user.clj. I’ve clearly not got to ‘the bottom’ of it. But I’m not updating user.clj very often so I can lose that for now.

oh. you've got a user.clj file

and user is the default "blank" namespace loaded

there are two things called user its sounds like

yep - its a convenient way of loading a bunch of (development / non-prod) code into the REPL

well, there’s the namespace, and there’s the code that defines it - in that sense there are 2 things

i think the repl's define a default namespace called user, and you have also defined a namespace called user

those are the two things

so its possible that it's getting interned in the other user when you are in a different user

I would be surprised if the REPL were maintaining two separate namespaces with the same name.

that (IMHO) would violate the principle of least astonishment

can you post the namespace declaration from user.clj?

not easily at the moment - I can try to produce a minimal repro case later

well, is it (ns projectname.user) or just (ns user)?

sorry:

(ns user
  (:require
    [clojure.repl :refer :all]
    [clojure.tools.namespace.repl :refer [refresh refresh-all set-refresh-dirs]]
    [com.stuartsierra.component :as c]
    [clojure.string :as string])
  (:import (java.sql BatchUpdateException)))

ah ok

you are definitely running two namespaces

https://github.com/technomancy/leiningen/blob/master/leiningen-core/src/leiningen/core/user.clj

and you said you can reliably reproduce the problem when you have a user.clj file?

anyone used https://github.com/Factual/s3-journal before? have couple of questions

(inst/parse-timestamp (java.util.Date ) "2016-07-29T13:43:00") is throwing an error saying that it can't java.lang.Class cannot be cast to lang.IFn, which seems valid, but what "constructor" or date creation function can i use here

(java.util.Date.) <— note the . at the end

hmm. same issue

you're supposed to pass it a function

(fn [years months days ...] ...)

right. so which function can i pass it.

you'd write your own

Ah, wasn’t sure what inst was...

what are you actually trying to do?

i didn't think that i needed a full fledged library, but i looked out and about and I see seancorfield with a library 🙂

inst is the clojure.instant namespace in this instance

i am trying to get a datetime object from a string

try inst/read-instant-date

well i'll be damned

i got confused as i don't understand why these are def's instead of defn's

but i appreciate it

hackier thing that also would have worked: (read-string "#inst \"" time-string "\"")

or (read-string (str "#inst " (pr-str time-string)))

haha that is some gross looking stuff there 🙂

@seancorfield: , appreciate you chiming in so quickly. do you also maintain clj-time?

and much appreciated gfredericks. that code got super dense inside there pretty quickly

np

@dpsutton: Well, "maintain" is a broad term… I’m one of the two official maintainers (with Michael Klishin) but mostly we just review and merge other people’s PRs these days.