Re: [PATCH v2 03/10] reftable/basics: provide new `reftable_buf` interface

> > I don't feel very strongly about it, but I had suggested it because my
> > initial read of this patch confused me, and I had wondered if others may
> > be similarly confused.
> >
> > For what it's worth, I was thinking something on the order of the
> > following added to the patch message:
> >
> >     Note that the reftable_buf_add() function intentionally takes a "const
> >     void *" instead of a "const char *" (as does its strbuf counterpart,
> >     strbuf_add()) to emphasize that the buffer may contain NUL characters.
> >
> > But, as I said, I don't feel very strongly about it.
>
> You know: let me amend the function documentation itself. That feels way
> less out of place compared to putting this info into the commit message
> and has the benefit that a future reader of the code will know why we
> have types without digging into the commit history.