Hahaha! Eat that Dijkstra!

C translation: please test against bigger numbers for edge cases.

Awesome use of the REX prefix! Nice way to use NOP to form UTF8.

This was quite a challenge to pass all tests with distorted inputs. Took more time than expected. Really nice kata!

Wow! No need to get personal here about my translations. I did use pointers you say? No way!

I've always responded to issues raised about my translations, and I keep things clear for the end-user. That was what this comment is about.
If you can't handle criticism about how to improve upon such a simple thing, especially for more inexperienced users, you have bigger issues to deal with.

Relax man.

Before I erase the code, I read it, thank you.
It doesn't explain it as you claim:

#include <stddef.h>

//  assign values to pre-allocated *array
//  *array allocation fits up to length n
//  return your array length as an output

size_t off(unsigned n, unsigned array[n]) {
    size_t length = 0;
  
    //  <----  hajime!

    return length;
}

A quick read shows that array length is n. You mean something different, which took some time to discover. Please be a bit more clear. Also it's missing from the Description, which says:

Create the function off, that receives the nth switch as argument n. The function should return an ascending array

Clearly the function off doesn't return an array at all, so the description is inaccurate.
I don't care too much about "preferred way", as I've written plenty of C translations where I adapted the description for C specifics. There's not much that leads to "maintenance issues" with that.

Other than that, thanks for your efforts.

C translation has really strange API and expected return value is not mentioned in the description, and is the same as input parameter n??

delete...

delete...

delete...

in C: random tests don't excite all paths in solution, allowing incorrect code to pass all tests.
Some cases not covered by default:

  dotest (16,8448,8449);
  dotest (7,5935,4);

but others might also be missing. You can try to find them by running tests 10+ times.

Well done, thank you!

Adapting the description is also an option, if it was intended that non-ascending and non-descending are allowed.
However, I think that would make the kata even more difficult for arrays that would contain all equal values.

Otherwise I think random tests should generate arrays that match the description (which is not THAT hard...)

I didn't mention it's not sorted. I mentioned it's not descending. It's non-ascending. Which is something completely different.
You can check any mathematical reference on this.

(in C only?)
Description says that arrays are either ascending or descending, but random test has this array:

a        = { 7, 7, 2, 0, 0 }

which is NOT a descending or an ascending array.