dumblob

> So I feel a modified quicksort could be a reasonable choice ... Even then, I'm pretty sure that stable quicksort is slower than Timsort :( I'm somehow against the...

Timsort seems to be slightly slower (say 85% of the speed) than stable quicksort for fully random data with small variance. For all other data Timsort is faster and also...

I found out that python has quite reasonable test suite for sort algorithms. It's mentioned in an [article describing Timsort](http://svn.python.org/projects/python/trunk/Objects/listsort.txt) which written by the author itself.

Just for curiosity, Julia has also a Timsort implementation (but without some useful optimizations used in the genuine Python implementation) written in a very well readable way https://github.com/JuliaLang/SortingAlgorithms.jl/blob/master/src/SortingAlgorithms.jl and it's...

It seems research in the field of unstable sorting algorithms is still ongoing. The state of the art **unstable** sort for parallel systems (which Dao with kind of is :wink:)...

@MusicTheorist thanks! That sounds like a valuable initiative. I hope something measurable will come out of it. 2020 seems to have brought some new wind into the topic of sorting....

> EDIT: is it normal that 'load os' in script crashes? No, it's not (at least it works for me). Could you please put here a valgrind or gdb stack...

Should we also mark the `nk_color_picker/pick` functions as deprecated and add to comment a very approximate estimation when they'll be removed (e.g. `will be removed any time after 2023-12-14` i.e....

> Instead of a timeframe, maybe remove it at next major version bump? So the comment would be `will be removed in v6.0.0` or something like that. Yep, that's easier....

> Hmm good question. I'm not sure how `-Werror` effects deprecation warnings. We don't wonna break people who use `-Werror` but it would be nice if people who don't got...