mem.sliceTo on C pointers should return an optional slice

[daurnimator]

https://github.com/ziglang/zig/pull/8698/files#r632922210

Also makes the compile errors from SliceTo more consistent

[ifreund]

I'm not 100% sure this makes sense with the current semantics of C pointers, if #6597 were accepted I would think otherwise.

If this is the way to go, I think we need to update std.mem.span() to match this behavior, and you need to correct the doc comment for std.mem.sliceTo().

[daurnimator]

If this is the way to go, I think we need to update std.mem.span() to match this behavior,

hmm. I think I'd only like to do that if #6597 is accepted.

you need to correct the doc comment for std.mem.sliceTo().

done.

[ifreund]

hmm. I think I'd only like to do that if #6597 is accepted.

I think having consistent behavior between the two is pretty important, what's your reasoning here?

[daurnimator]

I think having consistent behavior between the two is pretty important, what's your reasoning here?

span looks like it shouldn't accept C pointers at all; and users should instead be using sliceTo.
Rather than change it, I'd rather remove the .C branch entirely.

[andrewrk]

Please re-open against master branch with passing tests