Yeeees? Inconsistency was not enough reason? (It is to me!)

I mean, I get that there might be more important things to worry about than abstractions over the host language leaking here and there when language users do weird things… But, like you said, why should anyone expect having to deal with an undocumented exception in one case, and not the other?

Please share

The suspense is killing us!!

it's interesting. sorted-set behaves like a homogenous set, whereas set behaves heterogenous... but sorted-set implies homogeneity (or at least the ability to compare all pairs of elements) it seems more reasonable that (conj (sorted-set 0) "") should fail i have to admit that i'm very used to the idea that get is safe and should never fail harder than returning nil

So, I think I consider myself a Clojure programmer (that’s stupid, you shouldn’t be married to tech, they say), and also a Clojure advocate if not evangelist. I present at (non-clojure) conferences saying that Clojure is the greatest thing since sliced bread, and that dynamic typing actually works in real life, in production. In doing so, I don’t really want to be met with/having to defend stuff like:

user> (get {9 0} :wat)
;; => nil
user> (get "wat" :wat)
;; => nil
user> (get nil :wat)
;; => nil
user> (get 9 :wat)
;; => nil
user> (get (hash-map 0 9) :wat)
;; => nil
user> (get (sorted-map 0 9) :wat)
Execution error (ClassCastException) at user/eval573567 (form-init8054768168167100120.clj:27117).
class java.lang.Long cannot be cast to class clojure.lang.Keyword (java.lang.Long is in module java.base of loader 'bootstrap'; clojure.lang.Keyword is in unnamed module of loader 'app')
user> 

And, I’d be very happy to settle for yes, it’s a bug, but we ain’t fixing it anytime soon.

There was something unspoken about this that was bothering me, too, and I think I’m with you. The language implementor attitude towards things like this can make all the difference.

Or, at least, a lot of difference 🙂

I think I’ve seen other, similar examples of unexpected exceptions due to abstraction leakage, with similar responses.

I think it's a little bit of hubris mixed, a little bit of "meh, no big deal", mixed with a slightly obsessive attitude towards maintaining backwards compatibility in every respect.

But lately they do seem to fix more of these papercuts

dynamic != weak typing 👀 also, some ex-colleague would humorously use the expression "bag of shit" to describe anything-goes-in hashmaps. I ended up feeling the same - hashmaps are cool but bags o' shit aren't

are you sure you’re not a professional photographer impersonating ray?

😂

Wowsers

obviously you should flip it so the sun is in the right position though. 😉

I’m an idiot :man-shrugging::skin-tone-2:

I like how you often get some interesting path through your photos, @U04V5V0V4. I've gotten inspired and think of it when I take a picture of something. It's fun to just experiment a bit with the perspectives, and also I notice that I start to see my surroundings more in this light. ”That's a nice line towards the horizon” and similar.

Is this something you think about when you take pictures, or does it just happen that way?

Yes, I do think about that. I'm a big fan of Nigel Danson who is an amazing and generous landscape photographer (that was in software before he turned pro) https://www.youtube.com/watch?v=XTuw_KvMY7U

👋

I just got my plane tickets 😁 I'm in Amsterdam Friday to Monday. Staying in a hotel near the event. No plans other than Saturday.

c'est moi, and @ULE3UT8Q5 and @U0HJ7CX6H

👍:skin-tone-3:

My family schedule doesn't allow for much travelling at the moment but I'm looking forward meeting you all again eventually!

+1 🙂

I’m on #C047C0BE3HC

I think it is this: https://www.techradar.com/news/got-an-intel-laptop-be-warned-new-linux-bug-can-literally-break-your-screen

:( https://twitter.com/Iinux/status/1583065507057618948

(ubuntu stable 20 + liquorix kernel, still in the setting up phase and was stuck in setting up nvidia gpu driver last month, reverting to backup kernel/changing to llvmpipe doesn’t seem to help either)

reading how that bug can cause physical damage to the hardware, I dare not even boot up my laptop now…

…maybe I should just declare today to be “Black Tuesday” and call it a day

Bummer!

I like stack traces too, but I did java for ages.

https://twitter.com/borkdude/status/1404016344245415939

I was definitely confused in the beginning, but now it’s blondes, brunettes, redheads…

I remember first doing some Haskell: some of the type warnings weren't particularly beginner friendly either :P

"Clojure is optimized for correct programs" - Stu Halloway

I think the presentation of stackframes is the problem. the fact that you can deduce exactly the "path" the execution of your code took is actually beginner friendly.

My microcontroller code written in C will typically just fail with random unexpected behaviour.

I also liked that you could inspect the source of core functions, especially in the beginning

oh, yes!

Agree, "here's exactly how you got to something exploding in your face" is actually great The only conclusion is people panic and don't just stop and read the damn traces. Slowly

close and slow reading is one of the more frustrating things to learn in clojure coming from more verbose languages

I love the information density of Clojure

I think the stack traces are fine as long as you know Java or Javascript or whatever host platform you're on, but there's no doubt in my mind that it's a significant hindrance to anyone who hasn't already done enough to Java to understand what's going on. It's basically a path dependency of being a hosted language, so the issue can never truly be resolved.

I actually (ab)used Clojure to learn more about the JVM (as Java was a too boring language to me)

if you compare them to rust, elm & co, sure. Otherwise they are just on par with most mainstream languages

personally I don't mind having too much info, you can always filter out what you don't need at another level

The lowest hanging fruit must surely be some more descriptive error messages for those "x can't be cast to IFn" and the like that are actual problems for beginners. There is some obvious tension created by Clojure's dynamism inside the Java host, but forcing people to find what the hell an IFn is, is really jusy forcing people to both learn Java and examine how Clojure is implemented before being able to use Clojure. Maybe it's a feature, not a bug? I dunno.

What is meant by https://clojurians.slack.com/archives/CBJ5CGE0G/p1666703246217989?thread_ts=1666702845.935019&cid=CBJ5CGE0G is that there aren't any beginner friendly checks in the cast-paths because of performance. People who see this error usually know what it means, unless you're a true beginner

This is one problem that a statically typed programming language doesn't have because these checks happen at compile time rather than at run time

Unless you build a dynamic programming language in them 😈

well luckily we have some level of static analysis available through external tools

I personally think the error message is a thing you get over quickly. Clojure might not be optimized to be a first language to teach, but given the bandwidth clojure core has it's more than ok

I'd rather have stuff like creepingly slow satisfies? & co fixed than more work on error messages

this is OT but why would you ever use satisfies? and not just extend?

type vs value

they act on different things

With forms being read by the REPL, it's more difficult to give a nice printout of the offending code with "here's where bad things happened"

We need something file-independent which keeps history. Basically copy what Unison did?

We're winning

Slowly... migrating... old iMac... from... ... to... 🐧

Loading your computer with penguins, please stand by

🐧

I use MacOS and Linux on the daily using Synergy

It’s total crap on MacOS though

My issue was that my iMac is getting old. Old enough to stop running certain programs or give dire warnings when doing so. So I'm switching my iMac to run Linux.

Ah

Nice