🚀 go-pugleaf

In article <wuth.75se@castrov.cuc.ab.ca> wuth@castrov.cuc.ab.ca (Brett Wuth) writes:
> : ODU, in and of itself, is not slow unless the lock is contended. What
> : version of ODU are you running?
>
> It says it's 2.0.  3384 bytes.  Should I be looking out for a newer
> version?

-----rw-d   20-May-93 06:28:33	 73    37230 Archives:OwnDevUnit33.lha

available for BMS from adastra. For a couple of years. :)

> : >	* Have a separate log file for each concurrently running
> : >	  program.  If it could be done, there would be no need
> : >	  for lock/unlock or open/close.
>
> : This would disallow you from having multiple sendmails', rnews',
> : etc. running. Not a good idea, IMO.
>
> Yeah, unless we could somehow have a different log file for each
> separately running program (i.e. multiple invocations get multiple
> log files).

Sorry, but _yuck_.

> : My testing doesn't agree with yours completely. I see the speed loss
> : in the seek() to EOF of the logfile, not in ODU.
>
> Here's the test code I used.  It was tacked onto the end of log.c.
> I then in various tests ifdef'd out the the lock/unlock and forced
> the code to open only once.  Can you try it with this test code?

I will try it. I won't have time for a week or so (I'm on the road
traveling this week).
--
  //   Michael B. Smith
\X/    mbs@adastra.cvl.va.us