On Thursday 11 October 2001 02:36 pm, Aaron Bannert wrote: > On Thu, Oct 04, 2001 at 03:23:00PM -0700, Ryan Bloom wrote: > > On Thursday 04 October 2001 02:37 pm, Aaron Bannert wrote: > > > On Thu, Oct 04, 2001 at 03:34:23PM -0600, Brad Nicholes wrote: > > > > As I was implementing the apr_thread_cond_*() functions it appears > > > > that we are missing one function. I propose that we add the > > > > following function: > > > > > > > > APR_DECLARE(apr_status_t) apr_thread_cond_timedwait(apr_thread_cond_t > > > > *cond, apr_thread_mutex_t *mutex, unsigned long interval); > > > > > > > > > > > > What do you think?? > > > > > > If it can be implemented reliably on all of the platforms, I'm all for > > > it. I'll even write the patch for the stubs. Can we get some > > > reassurance that it'll work out on: beos, os2/win32? > > > > I can make it work on win32. > > cool. > > Two questions: > > - What about OS/2 and Beos? (I know David is nursing his RSI, but what > about OS/2?)
I would bet that it is portable, go ahead and create the API. > - How should we handle the timeout, as an offset or absolute, and in > what units? POSIX uses "const struct timespec *abstime" but that will > work with Brad's proposal above too. The timeout should be handled just like all other timeouts in APR, namely the number of microseconds until it should pop. Ryan ______________________________________________________________ Ryan Bloom [EMAIL PROTECTED] Covalent Technologies [EMAIL PROTECTED] --------------------------------------------------------------
