LINQtoSQL , repository pattern and lazy load

Question

How do you use LINQtoSQL with the repository pattern?

I’m new to L2S and find its lazy loading to be a real impediment to using the repo pattern.

Usually, I think of the repository pattern like this:

var myCustomer = null;

using (var myRepo = new Repo()){
    myCustomer = myRepo.GetCustomerForCustomerId(123);
}

if(myCustomer.Orders.Any()){
 //do something
}

Trouble is, won’t L2S attempt to make a data connection when myCustomer.Orders is interrogated? Doesn’t this lead to unpredicatable database access issues?

I mean, yes, I could tell my repo to verify orders inside of the repo confident that our complete test coverage verifies that developers never call an entity we didn't explicitly load, but I would rather just get rid of the lazy loading/object-datacontext persistence.

So I have 4 options

1. Create domain objects that get created from the L2S objects – lots of work and maintainance
2. Create derived verisons of my L2S object that break he linkage (http://www.codeproject.com/KB/linq/linq-to-sql-detach.aspx)
3. Use LLBLGenPro instead.
4. Appeal to the wisdom of stack overflow readers

I’m going with 4 for now.

How do I ensure that my objects wont call the db after my repo is closed?

And yes, I did read every stack question that talks about L2S and Repos and none of them answer this question.