even though it wasn't a patch. ooops

I'm curious if you have any thoughts on how this compares to https://github.com/davidmoten/rtree2. I've wrapped it for previous projects, but having a cljc lib seems great.

@U04V70XH6 thanks !

@U7RJTCH6J I've haven't used r-trees as my spatial partition. I'm mostly pretty new to this kind of stuff. I really wanted quadtree to be cljc because I found most of the existing options weren't. I'm current building a game in Clojurescript and i'm pretty sure the only reason it works is because of quadtree-cljc, though it's quite a bit slower in the browser.

Neat!

Love the name and version number, but what does it do? 🙂

Sorry 😬 wasn't finished. Hit enter instead of shift-enter. Slack on personal laptop hasn't been configured.

Is it not open-source?

It is (MIT): https://github.com/raystubbs/zero

Ah, didn't realize the first link is to GitHub. You should add SCM info to both Clojars and cljdoc. :) Right now they're both saying that there's no SCM info.

Just noticed that 😬 gotta figure out how that stuff works. But it'll be there soon.

Also, not noticing the link wasn't your bad; I put it in after realizing I'd linked Clojars but not the repo.

scm should be added automatically if you use leiningen. otherwise you can set up a template pom.xml file (first generate it with clojure -Spom) then add an tag yourself and provide a :src-pom when invoking b/write-pom to keep your deps.edn and pom.xml template in sync i recommend making a script that generates the pom template then run it in your CI builds (assuming you use CI to deploy)

Thanks @U0479UCF48H 🙏 Got it working with :pom-data , so don't need to manage a separate pom file. Cljdocs API docs generation is broken though, seems like it doesn't like ClojureScript's catch :default, so will see if I can do something about that soon.

Ah yeah, :pom-data is a recent addition and works well, too. Probably more reliable than using template POMs

:z/on should this be :zero.core/on instead? Then you can use an alias to make it ::z/on if you wish to use such a short alias. Projects that adopt this may have an established pattern of aliasing something else as z, and some people may have a preference for shorter or longer aliases. With :z/on you're not giving people an option to choose their own alias, with :zero.core/on they can alias it however they want.

Hmm... yes you may be right. I don't so much like tying the keyword to the module's ns/location... but after reading your tip above, I think the advantages win here. Will makehttps://github.com/raystubbs/zero/issues/8 to switch these out. Already have quite a bit of code at work depending on :z/... keywords, so will probably phase them out a bit slowly. Thanks for the tip @UCQGNA673.

> I don't so much like tying the keyword to the module's ns/location That's why aliases exist. :) And note that you can also use :as-alias instead of :as in your :require vectors if you don't want to load that namespace (e.g. for avoiding circular references).

Fair enough. Perhaps just a natural aversion that I should get over. I guess my thought is that if :zero.core is the namespace exposed/recognized by an API, and the namespace is tied to file location... then you're kind of locking in your file structure (tying it to the API). But I guess in practice it probably isn't that big of a deal. And everything else (functions, etc) are tethered up the same way so :man-shrugging: I'm probably wrong on this.

You're not locking anything. :as-alias allows users to refer to namespaces that don't exist:

user=> (require '[does.not.exist :as-alias dne])
nil
user=> ::dne/hello
:does.not.exist/hello

Touché. Glad to be corrected.