On Tue, 15 Jul 1997, Jukka Marin wrote:

> On Mon, Jul 14, 1997 at 07:39:15PM -0400, Bill Sommerfeld wrote:
> > It's worth noting that Jim Gettys and Bob Scheiffler sweated bullets
> > in order to minimise the number of round-trips in the X protocol.  in
> > particular, the protocol is asynchronous; graphics requests are
> > buffered and sent over to the server in decent-sized batches.  You
> > don't do one IPC per line or text string...
> 
> This brings my Applixware problem into my mind again.  Using ktrace, I
> found out that Applixware (the linux version on i386) is doing lots of
> small reads and writes to the X socket.  After a write, it calls
> oldselect() which takes 50 ms to return.  A word processor screen update
> consists of a thousand (well, it sure looks like that) write/oldselect/read
> chunks of code, most of which take 50 ms.. so the screen update takes 15
> seconds or more.  This only happens after Applixware has been running for
> a while.  Any new ideas, anyone? :)
> 
> I'd like to add some debug printf()s in the oldselect() code, but the
> amount of data would flood the logs immediately :-/
> 
	Maybe log every 1000th call (use a static int), paying particular
	interest to the struct timeval * value?

    "Its a bright sunny california day outside, and you're inside."
    "And your point would be...?"