runiq VS orderless

How does it compare with huniq and runiq?

A-ha! I knew I had one more.

How many times have you wanted to dedup a (text) file, but definitely didn't have enough memory to perform the task? I found this one day when I had to dedup a set of .ndjson.gz files which totaled a cumulative 312 GBs. Utilizing the bloomfilter option, I was able to dedup the records without any large investment on my part.

Anyways, runiq[1], "[an] efficient way to filter duplicate lines from input, à la uniq".

It provides several ways to filter of which I almost always default to utilizing the bloomfilter implementation (`-f bloom`).

---

[1] https://github.com/whitfin/runiq

[2] https://whitfin.io/filtering-unique-logs-using-rust/

orderless. A highly configurable completion style that matches multiple patterns in any order against minibuffer completion candidates.

An example relevant to your list would be some changes many people are taking with their completion framework - using package that leverage core emacs functionality rather than replacing it with a complete package that 'overrides' it. Consult, vertico, orderless and associate packages come to mind here. If you do a bit of a search you'll find plenty of info. Here is a video from Prot on the subject, but there are many others as well. I think Prot actually went on to write his own completion system to overlay native emacs functionality as well.

You didn't mention, so I'll ask, are you using the orderless (https://github.com/oantolin/orderless) completion style?

Fuzzy completion ala Orderless would be awesome: hitting space during completion acts as a pattern separator.

I dunno—I like how Vertico+Counsel feel. I'm not sure how good the support for Orderless and Embark are in Ivy, but I really like how those packages compose so nicely with the Vertico+Consult ecosystem.

In general, you want either orderless or prescient, with my personal preference being the former.

You probably want to investigate completion styles. There are many builtin styles, from basic, which just does prefix completion, on up. But there are also 3rd party styles. One of the most powerful is called orderless. Considering all these styles, there really is a ton of flexibility in how you can get to a completion candidate like some-named-something (some, s-n-s, sns, soso, [a-z]{4}-na, e\b \bs, ...). You can even configure more than one style at a time (and usually do).

Well, I've written some general purpose Emacs packages (orderless and embark) that I use a lot, but I also write Emacs Lisp for one-off tasks.

I like it. Would be nice to see orderless-style (https://github.com/oantolin/orderless) completion, and a config not to enter the directory by narrowing the completion to one, requiring enter to be pressed.