Subject: Re: [boost] [Boost.Lockfree] Interest in a ringbuffer class
From: Gavin Lambert (gavinl_at_[hidden])
Date: 2017-01-26 18:05:30
On 26/01/2017 23:56, Raphaël Londeix wrote:
> That's right, but that's not the main selling point. The idea is to be able
> to store raw buffers of different sizes, without any external storage or
> indirection. When you do
>
> buffer* buf = rb.start_write(10);
>
> The variable `buf' is stored inside the ring buffer, so that the commit
> operation only change the position of the write cursor. One possible
> way to do that is to define buffer as
>
> struct buffer {
> size_type size;
> byte_type data[];
> };
That sounds a bit different from what I was imagining. I have a
somewhat similar ringbuffer interface (in terms of separated reserve and
commit) but it assumes single-copy inside the methods; it never exposes
internal buffer memory (because zero-copy gets tricky if you want to
span the ringbuffer's wrap point).
I've only ever used it as a block-of-bytes buffer, typically in cases
where at least one end doesn't care about message boundaries and can
just process bytes in arbitrary chunks.