C#

<br />
function Populate()<br />
{<br />
// Load table<br />
// populate all parent rows<br />
PopulateChildren(Table, ID)<br />
}<br />
<br />
function PopulateChildren(Table, ParentID)<br />
{<br />
// Filter for parent id's (which gives you the children)<br />
// Populate child rows<br />
PopulateChildren(Table, ID)<br />
}<br />

Luc Pattyn wrote:

So typical code sequence is: test, lock, test again, do, unlock

private static object locker = new object();
private static Configuration _configuration;
public static Configuration Current
{
   get
   {
       if(_configuration == null)
       {
            lock(locker){
               if(_configuration == null){
                    _configuration = LoadConfiguration();
               }
            }
       }
       return _configuration;
   }
}

private static readonly Configuration _configuration = LoadConfiguration();
public static Configuration Current
{
   get
   {
       return _configuration;
   }
}

Luc Pattyn wrote:

it always does the load

Luc Pattyn wrote:

it may execute sooner than required (e.g. by accessing something else in the
same class, hence class load and execution of static code)

Luc Pattyn wrote:

so overall my preference is
doing this kind of initialization in a background thread,

N a v a n e e t h wrote:

Means it calls the LoadConfiguration method each time when I access the Current property ?

N a v a n e e t h wrote:

Do you mean the readonly kind or the first one ?