docs(ipfs.add): flag behavior around cidv1

[lidel]

Closes #8396

@Jorropo would this make things clean enough?

[Jorropo]

@lidel you got the things the wrong way around.

My issue is that CIDv1 implies raw leaf unless the user specify it otherwise. (CIDv1 -> Rawleaves)
Your PR explain that raw leaf implies CIDv1 which is false, I mean just test it raw leaves on CIDv0 just works. (Rawleaves -> CIDv1 which is false)

I hope this clears it out :)

[Jorropo]

See the comment

[Jorropo]

By the way, when I've said : "Your PR explain that raw leaf implies CIDv1 which is false" that is partialy wrong.
If the target file is smaller than the blocksize CIDv1 is implied, else (if the file gets chunked) CIDv0 is implied. Which makes no sense to me, just pick one and stick with it.

[lidel]

Applied suggestions from review and dropped (experimental) from cid-version and raw-leaves. CIDv1 has been used in production for more than a year now, no need to keep this until it becomes the default.

I sympathize with @Jorropo that behavior is confusing, but realistically the best we can do to solve it is to continue work towards CIDv1 as default (#4143, we have ipfs/fs-repo-migrations#95 scheduled to land by EOY).

Let's merge this as-is and invest time towards CIDv1.