If you'd actually like to be removed I can just change it to "anonomous contributer" and remove the context.
It's your source code so you can mention whomever you want. I simply meant that it's not necessary to give credit for a few observations produced in my leisure time.
The name spiral-anon reminds me to ask if you've had the time to work on anything for your self lately?
I've been trying to make a flower crown out of composing sine waves in cairo. It's not going well.
- I see you changed your mind about the 'proc' of naive-append-vector!. As one of the benefits, the copy-append-recur! helper is now iterative, as is vector->radix-recur.
- The final-vector-remainder of radix-append is 32 for a 32-aligned length. While the lambda no longer attempts to overflow its vector, the make-radix call receives a length that is incremented regardless of the lambda's decision, so the length does not match the no-copy case. The copy-append! also receives a read-start that does not match the no-copy case, so the first 32 radix-2 elements are skipped when (radix-length radix-1) is 32-aligned.
- In radix-append, the second argument of the make-radix call contains an inlined version of the fifth argument of the vector-copy! call, which should thus be computed once and reused.
- The second branch of radix-slice retains "or equal to", and the third branch test has an extra 0. The full-slice case should also be handled faster than via find-head.
- With the proliferation of (make-radix #() 0 0) calls, they should be factored out.
- The second branch of radix-delete retains "or equal to". A find-head call can still be triggered with start=0. The full-delete case should also be handled.
- With the advent of find-head it seems that a global decision should be made on whether an empty radix is a first-class radix or just an internal convenience targeting the first branch of naive-append-vector!. If it is decided that it is first-class, then all radix consumers should be prepared to receive it, among them radix-append.
- A radix-insert call with position=0 will cause problems for the current radix-append.