Hi,

Yukihiro Matsumoto wrote:
> In message "Re: [Patch] String and (repost) MemLeak"
>     on 03/03/04, Michal Rokos <michal / rokos.homeip.net> writes:
> 
> |> I consider (or designed) str_new(0, len) as allocation of string
> |> buffer, so that I think garbage is OK, just like malloc().
> |
> |You're probably right - if programmer wants garbage he gets it... But I
> |still think that setting proper len cost nothing and could make some
> |tired programmers happier.
> 
> Cost is nothing, I agree.  But if a programmer gets a zero length
> string even when I specified len to str_new(), he/she may confuse
> sometimes.

	I think it's correct - lenght is zero, but capa is correct. The
reality is that string, that you created, is 0 bytes long - not X. So
when you do rb_str_cat (or any other) you'll get the right* result.

		Michal

* If you change rb_str_cat and some others as well :)

PS: You may also remove MEMZERO from rb_str_cat and maybe we should test
ptr in rb_str_buf_cat for NULL.