apr-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Aaron Bannert <aa...@clove.org>
Subject Re: [VOTE] proposal for names of new binary usec impl(s)
Date Fri, 12 Jul 2002 18:57:25 GMT
On Fri, Jul 12, 2002 at 07:47:36PM +0100, David Reid wrote:
> Can someone simply restate what issue needs fixing. No more hand waving or
> IRC chats, a simple email explaining the issue and what needs fixed.

As I see it, these are the issues:

1) apr_time_t is inefficient, especially for systems that only require
   per-second granularity. binary useconds have been proposed to solve
   the problem.

2) apr_time_t is similiarly named to time_t, but is not so similiarly
   implemented. Reducing or removing this confusion has been proposed.

3) What do we name the new time system? [I have vetoed to use of
   apr_time_t for this purpose. See previously discussed technical
   justification.]

4) How do we introduce this new time representation to APR and the
   numerous places where it is:
   - Operated on (with simply scalar operations -- +,-,*,/,%,etc..)
   - Passed with an APR function (select(), timeouts, etc...)
   - Printed


What did I miss?

-aaron

Mime
View raw message