Windows Updates Process
Fri, 04/24/2009 - 13:55
So I was reading a rant about the design choices in Microsoft Windows, ranging from 95 to vista (and a bit of talk about 7). One of the things that was focused on, was that:
a) windows had no way of differentiating criticalsystem.dll that I wrote, and criticalsystem.dll that microsoft wrote.
b) patches simply write within the c:\windows folders overwriting files, that have to be reloaded into memory, which happens during boot.
Can anyone confirm if this is actually what happens? and Is this the reason why so many patches have to restart the computer?
and for all the mac geeks, is this the same reason why all of apples software updates result in you having to restart your mac as well? or is there another reason?
I thought it was more to do with the way windows locks files in use, where as *Nix loads them into memory.
That way the file can be upgraded in *Nix and then a restart of the process is all that is really needed. Of course you can't restart the kernel so you need to reboot then.
In Windows it basically sets it up so that the file(s) updated are actually updated at next boot. They are getting better, iirc Volume Shadow copy is used to allow updates with less reboots, probably by updating the shadow copy, restarting the service and overwriting the original with the shadow?
There are bad things to the way *Nix works, if you do an rm -rf / it will work most of the way...
--
"Put me here and I am all yours, not for the money and not for the appluase", Melanie Safka