flush_all caveat

[Brad Fitzpatrick]

On Tue, 17 Jan 2006, Peter van Dijk wrote:

> A real fix would be to use some incremental counter instead of
> a timestamp, or other implementations I can't think of right now -
> but this is not worth the trouble I think.

Or, somebody else reported using a higher-granularity timestamp.  An
atomic counter would work, except there are already paths that depend on
"keep this deleted for 4 seconds" (an option on the delete command) and we
were reusing that field for flush_all.

It's worth fixing because enough people get confused by it.

I just haven't had time lately.

- Brad