CVS update: samba/source/libsmb

[Jean Francois Micouleau]

On Thu, 3 Jan 2002, Tim Potter wrote:

> OK - I'll fix it up when I get a chance.  So, as a matter of form do
> you prefer:
>
> 	a) doing a call with zero length buffer then making a second call
> 	   with the right buffer size
>
> or
>
> 	b) doing a call with a buffer size of 100 (the default at the
> 	   moment) and then re-issuing the call if that is too small
>
> The first option will always take two round trips but the first trip
> is always a short one.  The second option can complete in only one
> trip.

for the moment a). a buffer size of 100 is totally random, if you want to
optimize, it should be done on a info level by info level basis.
Optimizing that is way low on my todo list ;-) we're talking of rpcclient
or the future extension to the net command not something really important.

	J.F.