I can't remember the var name but try to toggle and look in the message area it'll tell you the name to set to t so it will automatically create one

thanks, don’t know how I could miss it, maybe I have to stop working in early morning (: the variable is projectile-create-missing-test-files

There is a clojure-lsp code action to create a teds for the current fn, creating the test ns and the function to be tested

yes, i saw such fn in clojure-lsp, but somehow it does nothing when i call it from emacs lsp-mode, either nothing appears both in lsp-log and messages buffers, even if the test ns already exists, maybe you know, where should I look also?

I believe clojure-lsp cannot recognise my tests because on lsp-clojure-show-test-tree it outputs No Clojure test tree data found. I’ve already added test to source paths (without alias) after i’ve tried to change classpath-cmd in lsp config.

This works well for me : https://github.com/rags/toggle-test

looks good, but is it better than projectile?

Less magic, easier to figure out if it doesn't work

@UJ1UZGJ75 if you have a simple repro, feel free to open a issue on clojure-lsp

Thanks, I should get rid of additional stuff in my config and try to reproduce it a in clean env.

I figured out the problem, lsp-mode incorrectly detected project root and started lsp in my projects directory, not in the directory of the project, where i do tests. It was fixed enabling lsp-auto-guess-option. thank you all for your help

More apt for #C0617A8PQ

the actual files are located in .emacs.d/local/.cache/workspace/.cache/file.clj

thanks, I probably have a better way of asking too, now!

Just so answer your question, the files are opened as read only by default, so toggle that mode

thanks @UK0810AQ2, I did that, made the change, but nothing seems to happen when I evaluate that file (the behavior doesn't change)

(probably because it's not considered part of the same CIDER session)

Do you use Leiningen? Did you access the library via a symlink in the "checkouts" directory?

No, I use deps.edn and reached the file via lookup-definition in doom emacs

I believe what makes vectors faster is that the values are colocated in memory (vectors store values in chunks, for each of which the underlying storage is an array). I think it's OK to start with an empty vector. In my experience, the vector is the go to data type for sequential data and I only use lists rarely

If you need more performance, you can look into transients and transducers

As an alternative to passing the empty vector as the initial value you could also (fnil conj [])

It would very much depend on your planned pattern of access. Vectors are optimized for adding & removing items at the end -- and have indexed access (as well as sequential). Lists are optimized for adding & removing items at the front -- and for sequential access.

(there's also a persistent queue if you need that kind of construction/access pattern)

Thx, all! Much more, compute intensive processing will be done on the results than in collecting them, so colocation souds good. I'll go with vectors for a start.

Buddy. It's even mentioned in the README.

Looks like Buddy is also "in maintenance mode" unless I'm looking in the wrong place.

> plase use the individual modules

https://github.com/funcool?q=buddy&type=all&language=&sort=

Checked there. The buddy-auth user guide also describes itself as "in a maintenance mode." https://funcool.github.io/buddy-auth/latest/user-guide.html

If people are out there using it I won't turn my nose up at it. Not everything has to have a commit in the last month. But gives one pause when starting a new project when a major dependency isn't being actively worked on.

People are using it. :) It's not described as deprecated. While the project is looking for a maintainer, the amount of open issues is small and "maintenance mode" means (at least to me) that its API is complete.

I'm down with that. I'll dig into the docs. Thanks!

There's also keycloak

Just looking now. Seems like Keycloak is more an SSO service. I'm looking for something reasonably light that can handle "is user X allowed to do Y."

How is user X abilities encoded/stored/retrieved? Is using JWTs and the https://www.rfc-editor.org/rfc/rfc7519#section-4.1.3 an option? Edit: Some people (including me) wrote https://github.com/sikt-no/clj-jwt, a small wrapper over buddy, for unsigning JWTs and fetching keys easily.

Ideally I'd like them to be stored in something similar to AWS IAM policies.

I’m biased, but you might want to consider decoupling authorization :) https://github.com/anderseknert/clj-opa

That's pretty interesting. I'll take a look.