This page is not created by, affiliated with, or supported by Slack Technologies, Inc.
- # adventofcode (91)
- # announcements (7)
- # aws (3)
- # babashka (69)
- # beginners (46)
- # calva (30)
- # cider (12)
- # clj-kondo (88)
- # cljs-dev (11)
- # cljsrn (1)
- # clojure (195)
- # clojure-dev (21)
- # clojure-europe (2)
- # clojure-italy (13)
- # clojure-nl (56)
- # clojure-spec (4)
- # clojure-sweden (6)
- # clojure-uk (27)
- # clojurescript (179)
- # core-async (2)
- # cryogen (1)
- # cursive (2)
- # data-science (1)
- # datomic (57)
- # fulcro (15)
- # graalvm (9)
- # instaparse (6)
- # joker (18)
- # juxt (9)
- # leiningen (6)
- # off-topic (20)
- # other-languages (10)
- # pathom (5)
- # re-frame (20)
- # reitit (2)
- # rewrite-clj (5)
- # shadow-cljs (78)
- # sql (34)
- # tools-deps (128)
- # uncomplicate (16)
- # vim (6)
Next week I’m giving a talk at Dutch Clojure Meetup but I only have a bunch of unorganized notes so far. Today I’m gonna structure them, come up with a fun name, write an abstract and @skuro also asked for a short bio
I also want to make my slides in clojure to prove some points, so I might experiment with this a bit
if you have nice diagramming tips from clojure, I'd be interested in that. I want to make some diagrams for clj-kondo and babashka to help people contribute
preferably I would write data and the tool would generate a diagram without me messing with guis
In https://yannvanhalewyn.github.io/boyscout/, if you click on “Change” next to the alg name, there are little graph diagrams for every option. Those are actually accurate for each algorithm, I used clojure to generate them
there was a nice presentation about diagramming on the last clojure conj where he compared graphviz, plantuml and c4
http://draw.io is a nice open source tool, but you still have to do the drawing yourself
Probably not what you need, but for others: This is the raw code I wrote to generate those graphs in boyscout. Not pretty code, but it did the trick for me. I feel like you could easily make a little functional wrapper around it to fit your needs.
I've also used graphviz with clj-kondo: https://github.com/borkdude/clj-kondo/tree/master/analysis#namespace-graph
What am I exactly looking for: probably something in the middle of plantuml and fc4, using EDN
this could also be nice: https://danlebrero.com/2017/04/06/documenting-your-architecture-wireshark-plantuml-and-a-repl/
I've been considering just doing contributions to clojure projects recently, because that way I get feedback about the merge/pull request and it should help me get better without having to start an entire project that I'll probably abandon. You know kinda like short in-and-out adventures.
@hobosarefriends clj-kondo has a lot of open issues, if you're looking for something
Spring boot is a bit memory hungry 😛 https://graphql.gklijs.tech/results/graphql-servers/average-ge-mem
It now limited, might try again with more memory.. rust is much easier that respect, just takes what it really needs. Java takes much if it can..
I think Rust might be better now that it was since I last tried it, with the question mark thing instead of the weird error handling boilerplate it was then
There is still some boiler plate to transform errors into you 'own' errors. And not having a classpath to dynamically create instances makes it hard to get it on par with the Java version. Basically you need a function taking the data, and creating the correct instance. Which might be done with a macro, but didn't touch macro's yet.
repository link, https://github.com/gklijs/schema_registry_converter
I can't as a library. It's kind of neat with java. You just put your generated data structures on the classpath. And with the binary information from kafka, with the schema from the registry, it can create an instance with the values encoded in the binary.
maybe that's one of the things where Java is pretty dynamic which is what Rich Hickey likes about it
I'm more interested in a comparison on throughput. Should give it an other try some say. Before I tried Kafka things and there was a known bug making it impossible. At least with micronaut that's no issue anymore.
throughput is possibly worse with graalvm, but startup and memory usage are probably more friendly
Yes, it will be worse, they (Oracle) are clear on that. But I would like to know just by how much.
Hmm, actually seems the other way around, not sure when doing io through, https://medium.com/graalvm/stream-api-performance-with-graalvm-be6cfe7fbb52