LSP client in Clojure in 200 lines of code

https://vlaaad.github.io/lsp-client-in-200-lines-of-code

by vlaaad • 1 year ago

164 10 1 year ago

vlaaad.github.io

This is the most Java-y Clojure I’ve probably ever read. Just use Java? It’s so verbose and complex for what it is doing. Breaking this down into smaller functions and using core.async would make it even more succinct.

Just want to emphasize this because clojure is indeed a small, lesser known language that has a hard enough time attracting users. This is not what anyone would consider an idiomatic example of using clojure.

roenxi• 1 year ago

Would it be 200 lines of Java? It'd be 200 lines of just for the boilerplate. It isn't really a selling point of Clojure because it is subjective, but low-syntax high-terseness look of the code is in itself a reward for using the language.

And there isn't anything especially wrong with sticking to Java primitives if someone is comfortable with them. They work fine for Java programmers. The dude doesn't need to learn a new async library to write an LSP client if he doesn't feel like it. Code works, its easy to read, easy to understand and modify.

koito17• 1 year ago

Line count is not very useful to compare without the context of standard library size, third-party dependencies, etc. The code in TFA depends[1] on a JSON library[2] that is about a thousand lines of code (excluding tests) wrapping a Java library for JSON decoding.

Then there's other things to consider, like the fact that this LSP client, while succinct, pays not only the cost of loading Jackson, but also the cost of loading clojure.core, which is quite non-trivial[3]. Startup time for LSP servers and clients definitely matters to some, considering that e.g. even clojure-lsp recommends running native executables over JAR files[4]. Can't find documentation proving it's for quick startup, but it's a plausible rationale for their recommendation of a binary over a JAR.

Note: I have used Clojure professionally and in hobby projects. I think it's nice that one can interactively develop a minimal LSP client and the resulting amount of work is roughly 200 lines of code. I say "minimal" because it's unclear how this client deals with offsets reported by LSP servers, which are all given as offsets in a UTF-16 encoded string. In any case, I still think advertising "LSP client in 200 lines of code" hides valuable information regarding functionality, implementation, "actual" code size, and trade-offs made in the choice of technology stack.

[1] https://github.com/vlaaad/lsp-clj-client/blob/a567e66/deps.e...

[2] https://github.com/metosin/jsonista/blob/c8f2b62/project.clj...

[3] https://clojure-goes-fast.com/blog/clojures-slow-start/#cloj...

[4] https://clojure-lsp.io/installation/#embedded-jar-legacy-exe...

pron• 1 year ago

Between records and compact classes [1] Java's boilerplate isn't what it once was.

[1] https://openjdk.org/jeps/512

newlisp• 1 year ago

It's idiomatic "low-level" Clojure, though. Not everything is a happy place where you're just manipulating maps and vectors like in most examples.

0x1ceb00da• 1 year ago

lesser known language that has a hard enough time attracting users

For very good reasons.

dig1• 1 year ago

I don't see why this wouldn't be considered idiomatic clojure code; it makes proper use of all the facilities provided by the language and the main intention of this code is to follow the article. Additionally, the clojure core team often encourages not to shy away from using java code directly, as this approach strikes a good balance between performance and language expressivity.

It’s so verbose and complex for what it is doing. ... and using core.async

I think this code is actually quite straightforward and easy for a clojure developer to understand. In fact, using core.async in this case would be overkill and could complicate things further.

daveliepmann• 1 year ago

This looks like the other completely normal, idiomatic Clojure programs I've seen which manipulate a StringBuilder. And as Clojurians go I'm far to the succinctness/concision-preferring end of the spectrum.

I'm curious to see your core.async-based version :)

askonomm• 1 year ago

Holy crap is this unreadable or what (notably the lsp-base fn). There's a reason why in most Clojure companies I've worked at we try to make as small functions as possible, because otherwise it very very quickly becomes an unreadable mess, and you write code after all for humans to read, because if you didn't, you might as well just write binary. But, I'm not surprised many people don't want to get into Clojure or Lisps in general, because it takes a boatload of conventions and active discipline to make it a good experience.

slifin• 1 year ago

To me something unreadable is code that I cannot statically make any assertions about the runtime behaviour of the code

This function you're complaining about looks like 2 virtual threads doing program input reading and output writing for the LSP client given some ArrayBlockingQueues in about 25-30 lines

If I wanted the complete story I could use Clojure's inbuilt test runner to slip some ArrayBlockingQueues in there and run it under record with Flowstorm

Then leisurely seek through the entire state of the program, to get the play-by-play of how this works

There are so many good design choices in this language and a good 30% of colleagues I run into are not even doing the basics of like running a REPL, I think some people just need to clock in with a decade of C# or PHP or TS or JS or Python or whatever to get a taste of a language with next to no inbuilt immutability, statements instead of expressions, no reload-ability in the language semantics and just crapshot debuggers that run in lockstep with the program execution

LSP client in Clojure in 200 lines of code

Related Stories