On Fri, Jun 20, 2025 at 10:32:27AM +0200, Marc Chantreux wrote:
On Thu, Jun 19, 2025 at 06:09:49PM +0100, Pádraig Brady wrote:
> > we should aim for ACID file replacement functionality.
> it seems to me that's the whole point of sponge
> so out of curiosity: is there another possible usage ?
You're right. The atomicity issue in method 6
was due to the use of truncate and write (cp and rm in my example).
I have to admit I didn't look at your examples carrefully but
what I was saying is
* if it is not for acidity, what's the point of sponge?
* if the answer is "none" (as I believe) so is there a reason
to be explicit about it?
I agree with this, and question why the austin group is considering
it at all. The atomicity guarantees in sponge are too contingent and
not discoverable (at least without as much work as it would take to
reimplement sponge).
