@Matt – haven’t done anything with constraints as yet because it didn’t seem necessary for what we were doing so far. I guess if it was a completely generic function that was used all over the code then constraints would make sense.

@Michael – in fact in this case the two services don’t implement the same interface otherwise your suggestion would certainly be the way to go.

nice to see other poeple doing this as well.

I find myself doing this stuff all the time now though. It’s nice.

Would you ever add any type constraints to the generic function, i.e. “where T : ISomeInterface” or where T : class, new()” or do you not go that far because you’re passing in the Func(..) that gets the values, so it can handle that?

IEnumerable values = new List();

C# has a perfectly good coalescing operator (??) that you can use:

return values ?? new List();

This way the allocation only happens when you haven’t been able to get the list (or if your getValues() function returns null)

I more and more end up with similiar code.
Some times my code look like

…. => {
….
….
…. => {
etc.
}
}

BTW: why use the IEnumerable instead of plain T?
In some cases the C# intference is not able to get it right but in most cases it works for me just fine.

Heres one of my favs from my common DAL:

public virtual IEnumerable<LazyReadonly> GetAll()
{ return GetEnumerable(
env =>
UseDataAdapter(
env,
adp =>
GetAllRows(adp)
.Select(r =>
LazyReadonly.Create(
GetKeyFromRow(r),
() => RowToItem(r))
)));
}

(ok it messed up the spacing … thats something I didn’t figure out for my code yet – how to space the lines)