Reorganizing internal_getlogin() patch is in

[Pierre A. Humblet]

At 01:31 PM 6/17/2002 +0200, Corinna Vinschen wrote:
>The original reason was to speed up things in domain environments.
>The local machine has buffered the user information so it's called
>first.  Only if that fails we fallback to calling the logon server
>(a PDC probably).  This should avoid unnecessary net access.
>
>I'm curious, too, what you mean by "name aliasing".  Are you talking
>about having a local and a domain user of the same name?

Yes, precisely.

About caching, I did some experiments on NT.
The SID doesn't seem to be cached, in the sense that calling 
LookupAccountSid() twice in a row, with the Ethernet unplugged the 
second time, returns a failure after a very long delay.
I also tried calling NetUserGetInfo() with domain\user argument,
but that doesn't work.
I don't really see how Windows could cache the username (without 
knowing the domain), due to the aliasing problem. However
I haven't tried it.

Pierre