qsort implementation to minimal stdlib

[JonBruchim]

added musl implementation of qsort
http://git.musl-libc.org/cgit/musl/tree/src/stdlib/qsort.c

related to this issue i've opened
#28896

Signed-off-by: Jonathan Bruchim [email protected]

[andrewboie]

The implementation from musl seems reasonable at first glance.

The license may be a problem.

It would be good to have a test, although I am not completely sure what the policy is for minimal libc as I know we don't have full test coverage, and since it implements a standard I think the direction was to find third-party test suite. But we have at least some tests in tests/lib/c_lib.

Probably need some input from the TSC on this.

[github-actions]

This pull request has been marked as stale because it has been open (more than) 60 days with no activity. Remove the stale label or add a comment saying that you would like to have the label removed otherwise this pull request will automatically be closed in 14 days. Note, that you can always re-open a closed pull request at any time.

[cfriedt]

I have some reservations about just copying from another project, although I haven't looked at the implementation.

1. qsort does not actually specify which sorting algorithm should be used under the hood, which leaves a lot of possibilities open for the implementation. I've read that musl has implemented Smoothsort
2. if we add a sorting algorithm to the minimal libc, let's add one that has decent worst-case performance.
3. if we add a sorting algorithm to the minimal libc, let's add one that has all of the other random features that we could want, such as

• in place sorting (no external memory required)
• stability (i.e. the input order of elements is preserved in the output)

A decent reference for some of these features are here:
https://en.wikipedia.org/wiki/Sorting_algorithm#Comparison_of_algorithms
https://en.wikipedia.org/wiki/Qsort

I mistakenly opened #31107 for this same topic since I've been carrying qsort.c in my greybus module for a while (from the BSD C library). However, due to the aforementioned concerns, I never bothered to make a PR.

@JonBruchim - would you be able to provide some analysis of this vs possibly other algorithms?

Just at a quick glance, Block Sort looks promising too.

https://en.wikipedia.org/wiki/Block_sort

• O(N) best case
• O(NlogN) worst / average case
• O(1) space
• Stable
• In-place

I haven't looked to see if it can be implemented easily though, or what kind of stack space it might require. That's also obviously a concern.

I guess the only slight advantage that Block Sort might have is that it's stable. Smoothsort looks to have a reasonably simple implementation though.

[github-actions]

This pull request has been marked as stale because it has been open (more than) 60 days with no activity. Remove the stale label or add a comment saying that you would like to have the label removed otherwise this pull request will automatically be closed in 14 days. Note, that you can always re-open a closed pull request at any time.