What are you thinking?

sorted-maps are not entirely drop in replacements for maps, some things will throw exceptions on sorted-maps that work on maps

so my guess is things aren't frozen, but a thread has died and stopped doing work because of exception

Usually exceptions show up on the terminal, but nothing shows up after I change to the sorted map. I'm partitioning the map and the summing the partitions. I'll play with that, though.

When in doubt, search through the transcripts of Rich Hickey's talks. :) The word "complect" is mentioned here, with a definition and a discussion: https://github.com/matthiasn/talk-transcripts/blob/9f33e07ac392106bccc6206d5d69efe3380c306a/Hickey_Rich/SimpleMadeEasy-mostly-text.md

Definition of complect
obsolete. : intertwine, embrace especially : to plait together : interweave.

to separate things that were woven

Yes. complect is tangle.

more like "weave" than tangle, but yes - combining things is the core concept

I find that distinction useful because a lot of "complecting" in software looks relatively simple - eg. combing the concept of a value with the concrete detail of a storage location. it is not a tangle in the code to do that. the aspects can be skillfully handled together. but they are woven together and this has larger scale consequences.

I'm 80% certain it's supposed to be read as "f should be a (reducing step) function [...]". So, just a reducing function - something that you'd pass as the first argument to reduce, nothing fancy (except for the fact that it must also have a 1-arity).

Nevermind, I'm wrong, it's "(step function)": https://clojure.org/reference/transducers#_early_termination

Although this section makes me think that it doesn't matter much how you group the words together: https://clojure.org/reference/transducers#_creating_transducers

Yep, although I don't use the word "step" myself, reading other discussion here after searching for "step function" tells me that that's a synonym for "reducing function".

there is a hack I did awhile ago https://github.com/vlaaad/tweet-defhttp://github.comhttps://github.com/vlaaad/tweet-def

not sure if it helps 🤷

It's complicated to say the least. If you're trying to make a business out of it I say run away

I co founded a company that was relying quite a bit on their API. Every now and then they change the rules and make it difficult to adapt unless you get exceptions, most of the time through partnerships or just knowing people inside.

If it's for non profit I have no clue. In our cases we also leveraged the fact some of our clients were fortune 500, so they would get a "call" from them for particular use cases. The only times it was easy is when Twitter itself was the client

I left that world 3 years ago, so not sure how it is now.

Another trick was to re-use existing apps

Since we were working with their api for a long time we had this option for exploratory work

And lastly you can pay to get access, via gnip, at least back then

Seems now it's part of Twitter biz

I do recall ppl having to make videos of the use case for approval towards the end, and even then it was hit/miss

@U050SC7SV @U47G49KHQ thanks a lot! Will take a look!

hi, yes, this is an issue with older versions of graal vm.

we should document this, please try upgrading to a later version of graal vm or use any regular (non-graal) JDK

Upgraded to Java 17 (Azul Zulu version) and the same error shows up

is this still graal vm?

what version of graal vm are you using? Can you paste the output of java -version ?

we use the graal js scriptengine as a library, see https://github.com/nextjournal/viewers/blob/a20c8b54d1f0b15008d101d7a1ed86dea296d451/modules/markdown/deps.edn#L5-L6

unfortunately the graal JDK ignores the library and uses it’s version of the scriptengine that’s built-in, even when it’s an older one and doesn’t understand all options the newer lib understands

openjdk version "17.0.2" 2022-01-18 LTS OpenJDK Runtime Environment Zulu17.32+13-CA (build 17.0.2+8-LTS) OpenJDK 64-Bit Server VM Zulu17.32+13-CA (build 17.0.2+8-LTS, mixed mode, sharing)

huh, that’s very strange, never seen the error with a non-graal JDK

how do you start your clojure process?

I don't even know what graal is doing at all, this is the first time I've heard about it

these are the libraries on the classpath that contain the word graal:

~/.m2/repository/org/graalvm/js/js-scriptengine/20.0.0/js-scriptengine-20.0.0.jar: ~/.m2/repository/org/graalvm/js/js/20.0.0/js-20.0.0.jar: ~/.m2/repository/org/graalvm/regex/regex/20.0.0/regex-20.0.0.jar: ~/.m2/repository/org/graalvm/sdk/graal-sdk/20.0.0/graal-sdk-20.0.0.jar: ~/.m2/repository/org/graalvm/truffle/truffle-api/20.0.0/truffle-api-20.0.0.jar

ok, something is bringing in an older version

I'm using intellij/cursive

is this project using deps.edn?

no, leiningen

can you paste your project.clj?

ahhh got it!

there was this as dependency: [metasoarous/oz "1.6.0-alpha34"]

this is what brought in that 20.0.0 version of graalvm

thank you for your help!

pleasure, glad you figured it out!

I’m still a bit surprised that the newer version didn’t take precedence in that case

Yes, me too. What is graalvm about anyway? How does clojure use it?

our markdown parsing library is written in javascript, and we call it from Clojure via the graal vm polyglot capabilities

I see, so this is your way of reusing some JS code in clojure

never heard of anything/anyone using JS code from Java or Clojure, but times change 🙂

our goal is having one markdown parser that works in the browser and in Clojure. That’s how we ended up going with https://markdown-it.github.io

aha I see

are you the author of clerk btw?

ah ok sure 🙂 I just see on the website

Is #nextjournal the official channel for it?

I've worked my way through the demos and have some questions, I'll ask there

yes

yes, it's called Cursive 😄

oh, I'm not sure how it handles namespacedness when renaming, but I saw it changing local bindings in :keys [...] block