Re: lynx-dev connect_time patch

[Leonid Pauzner]

12-May-2003 06:46 Thomas E. Dickey wrote:
> On Mon, 12 May 2003, Leonid Pauzner wrote:

>> Subj, see below.
>> This doesn't help (15d snapshot).
>>
>> We are still busy waiting for select() in HTCheckForInterrupt() for 0.1 sec.
>> Do we need this? Perhaps this is an atrefact before DontCheck() invention.
>> We may test stdin once nicely, man select():

> perhaps - I seem to recall that there was some problem with the program
> using too much CPU time.  (We may have fixed that in a different part
> of the code, but it's the first thing I would check).


 -    socket_timeout.tv_usec = 100;
 +    socket_timeout.tv_usec = 0;  /* test once */

Huh!!!  That was 0.0001 sec (not 0.1 sec as I thought previously).

Reading docs more closely, DJGPP implementation of gettimeofday()
gets time with 1/20 sec granularity (55msec timer tics),
so we effectively get 1/3 time spent in a waiting loop,
exactly as I stated in my original post ("50% performance lost").

Anyway, we could test stdin once.




; To UNSUBSCRIBE: Send "unsubscribe lynx-dev" to address@hidden