:clojure-lsp: Release https://github.com/clojure-lsp/clojure-lsp/releases/tag/2021.02.07-03.04.31 add support for a new code action, alias suggestion require 🎉 It'll require a full qualified ns if present on classpath suggesting an alias
I added clojure-lsp
to scoop-clojure
here: https://github.com/littleli/scoop-clojure/pull/151
I would do it earlier, but I didn't know it exist. Great project ❤️
Oh, that's great news @ales.najmann, Thank you! We can add a entry for Windows here: https://clojure-lsp.github.io/clojure-lsp/#installation
Nice 🙂
btw I have a question what kind of Visual C++ Redistributable library it depends on? Usually there is one if you use native-image. And there is so many of them 😞
Yeah, I don't know exactly which, but I needed to install this one on CI to build successfully for windows: https://github.com/clojure-lsp/clojure-lsp/blob/master/.github/workflows/release.yml#L213-L214
@ales.najmann probably we should close this now? https://github.com/clojure-lsp/clojure-lsp/issues/28 I think scoop replaces chocolatey right?
If scoop is a good package manager to install things on WIndows, we could close that issue after adding docs 🙂 Willing to open a PR?
Chocolatey is a different project and different approach. I can only speak about scoop installer which I use and I create packages for, and not just for Clojure.
Let me check how to add docs entry. I'm willing to open PR of course 🙂
Yeah I know, I was just saying that having a good way to install clojure-lsp on Windows, for example with scoop, would be enough to close that issue 🙂
great!
I see 🙂 Yeah, I personally believe scoop way is superior on (native) Windows.
Perfect!
it seems that this line answers my question: VC++ 2015 Compiler Toolset
so it's going to be vcredist2015
Yeah, probably 🙂
I have to add it otherwise on machine where no Visual Studio is installed there is a good chance that binary would not run.
oh, makes sense, I tested on my windows but probably already had that installed
Also, I'd love to hear some windows user if clojure-lsp graalvm binary is working as expected 😉
I didn't test it for windows that much
I would like to actually start development on Windows in Clojure. But I'm not there yet...
I believe Windows users (also devs) deserve Clojure 🙂
Sure! You are helping they with your contribution 😄
PR merged 🙂 my quest is done for now. With another release Scoop automation should take care of the new version and update automatically. I hope. Your versioning scheme is quite different, but should be good I tuned the manifest a bit.
Nice! LMK if need any help, we have a https://github.com/clojure-lsp/clojure-lsp/pull/286 for homebrew too for MacOS users. @andre.peric is hellping with it
We (as a community) we should start conversation about having own homebrew tap with all the tools so there is no friction in terms of having dedicated tap for every tool. Also it helps people to find tools if they want to know what's new in the ecosystem. But that's a bit offtopic here.
yeah, I agree. But I'not even a Mac or Windows user :man-shrugging:
Autocompletion seems to be a little inconsistent for me. If I type quickly, it fails to find any suggestions at all. Is that expected? Any suggestions for how I can debug further or adjust any configuration? In the below gif, I type slowly and the suggestions come up in a reasonable timeframe. If I type quickly, I don't get any suggestions
Hey @rahx1t, this is a known issue that happens when your code is "unparseable" (not a valid clojrure code) We have this isssue to track that: https://github.com/clojure-lsp/clojure-lsp/issues/270
Basically, if you try to complete something like set/
it'll not work, but set/u
it will
Not sure @borkdude has any suggestion here
Thanks Eric for the quick response! I'm not sure the linked issue fully explains the behavior I'm seeing -- sometimes set/
will return matches and sometimes it will not. In both cases wouldn't it be treated as unparseable code? What would explain the different behavior?
Are you using Emacs with lsp-mode?
• lsp-mode caches the results if on same line/ same code to complete • you are using CIDER repl with override lsp completion I think
Yep! Using lsp-mode on Emacs. I've been trying to get the LSP experience right so I have been intentional about not connecting to CIDER.
Yeah, so that should be probably the lsp completion cache, it finds the completion while you are typing, that s why some times works before you type the slash
Oh interesting. So the LSP completion cache also explores the set namespace before I've typed the slash?
Yeah, if what you typed has more characters than company-min-length then lsp-mode will request completion items
You can disable this behavior also
Oh cool, thanks. I think this all explains the behavior I'm seeing. When I type too quickly and get to the unparseable code state, LSP-mode did not yet have a chance to request any completion items and once the code is unparseable, it cannot request any completions; when I type slowly, LSP-mode is able to request some completions before it gets to the unparseable state and once the code is unparseable, lsp-mode can still use results from the LSP completion cache. Is my understanding correct?
That's exactly how it works 🙂
:clojure-lsp: We improved documentation about all available features 😄 Check it out: https://clojure-lsp.github.io/clojure-lsp/features/
clojure-lsp
reached 500 ⭐ on GitHub 🎉 :clojure-lsp:
Also, FYI, clojure-lsp
issues were reduced by half after clj-kondo integration @borkdude 😄