try this once

[sainath thota]

public sealed class singelton
{
private static readonly singelton instance;
private static Object syncroot=new Object();
private singelton();
public static singelton instance
{
get
{

if(instance==nu ll)
{
Lock(syncroot)
{
if(instance==nu ll)
instance=new singleton();
}
}return instance;

}
}
}

[Marc Gravell]

Re: try this once

There is a very good discussion of Singleton patterns on Jon's site:

http://www.pobox.com/~skeet/csharp/singleton.html

Note that this double-checked/locked approach is exactly the 3rd
version, which Jon then justifies as "// Bad code! Do not use!"

Marc

[Mythran]

Re: try this once



"Marc Gravell" <marc.gravell@g mail.comwrote in message
news:uTyJt1Z5IH A.3480@TK2MSFTN GP03.phx.gbl...

There is a very good discussion of Singleton patterns on Jon's site:

http://www.pobox.com/~skeet/csharp/singleton.html

>
Note that this double-checked/locked approach is exactly the 3rd version,
which Jon then justifies as "// Bad code! Do not use!"
>
Marc

One of the comments Jon writes in the 2nd bullet of the 3rd version:

"I tend to try to avoid situations where experts don't agree what's right
and what's wrong!"

Hmm, Jon = expert....does he not agree either/or? Or is he not an expert in
the 'explicit memory barrier' field? :P

Mythran

[Marc Gravell]

Re: try this once

Or is he not an expert in

the 'explicit memory barrier' field?  :P

I think the key point here is that there are people who truly are
experts in this area (far deeper than language level) who can't agree;
it doesn't help that part of this is theory (the CLR/CLI spec) rather
than practice (the MS .NET CLR implementation) , which means you can't
strictly prove it just by example, because that could be an
implementation detail. So since it is so easy to work around the
ambiguity...?

Marc

[Jon Skeet [C# MVP]]

Re: try this once

Mythran <Mythran@commun ity.nospamwrote :

One of the comments Jon writes in the 2nd bullet of the 3rd version:
>
"I tend to try to avoid situations where experts don't agree what's right
and what's wrong!"
>
Hmm, Jon = expert....does he not agree either/or? Or is he not an expert in
the 'explicit memory barrier' field? :P

I believe that the version with a volatile field is probably okay.
However, I don't regard myself as a real expert in this field (or
virtually any field, to be honest). When people who definitely know
more than I do disagree, it makes life tricky.

I *think* that the volatile version is largely regarded as safe, but
there are lock-free ways of doing it which are more debatable.

--
Jon Skeet - <skeet@pobox.co m>
Web site: http://www.pobox.com/~skeet
Blog: http://www.msmvps.com/jon_skeet
C# in Depth: http://csharpindepth.com