I've been using linked for a few years. Mostly for an ordered set for CLJS - ordered has it only for CLJ.

My use case only requires ordered maps. The support for transients looks promising (but I don't need that either, just the ability to assoc-in and keep insertion order). Given that you've used the library for a few years, I assume you haven't ran into any bugs or serious issues with linked. I think I'll try using it, thanks!

I've been using tiara in clj. It has cljs support, but I haven't tried. The comparison section might be helpful, https://github.com/quoll/tiara?tab=readme-ov-file#comparison

Heh, I wish this library were easier to find. I didnt see any CLJS benchmarks, but the JVM ones look promising. My use case does not involve any dissoc or disj operations, just assoc-in. So anything optimized for insertion and lookup is ideal for me Overall, I think I will switch to tiara now.

@U7RJTCH6J Maybe Dewey could also have a toggle that makes it filter based not only on name/description but also on the text contents of READMEs? But that'd probably require a completely different approach to avoid downloading whole READMEs content at page load.

@U2FRKM4TW That’s a good idea. I hadn’t considered indexing readmes, but that seems pretty useful. My intention for the next upgrade to dewey is to do some basic aggregations as well as build some databases like sqlite or datalevin as part of the weekly release. I’ll probably also combine the normal search stuff with cloogle which has an actual server and can use a database.