Re: [Moq] Is Expression<Func<T, bool> supported?

[Daniel Cazzulino]

The IsAny<T> is the only way around it now, yes.

APIs that receive expressions are not supported unless you do that. It looked quite complicated to support those, as I didn't have a good way of knowing when to stop "resolving" the expression (as moq itself also receives an Expression!). I think it would be non-trivial.

On Sun, Oct 5, 2008 at 5:45 PM, gd34 <gareth...@googlemail.com> wrote:

I have this method on my repository interface:

interface IRepository
{
       IQueryable<TEntity> GetAll<TEntity>(Expression<Func<TEntity,
bool>> predicate) where TEntity : class;
}

and I want to set up an expectation that when I pass in a particular
predicate it returns X.

So I tried:

repositoryMock.Expect(r => r.GetAll<Absence>(a => a.UserId ==
1)).Returns(absences.AsQueryable());

When I do that I get a NotSupportedException. So I tried this to get
around it:

Expression<Func<Absence, bool>> arg = a => a.UserId == 1;
repositoryMock.Expect(r =>
r.GetAll<Absence>(arg)).Returns(absences.AsQueryable());

This works but has no effect on my mock repository. I did a little
test to see why this was and found that this test fails, which
explains why my expectation doesn't work.

Expression<Func<Absence, bool>> arg = a => a.UserId == 1;
Expression<Func<Absence, bool>> arg2 = a => a.UserId == 1;

Assert.AreEqual(arg, arg2); // Fails

I guess the compiler is generating different types so when Moq calls
Equals() it doesn't see the delegates as the same??

I was just wondering if there was:

a) any way around this
b) if anyone knows whether lambda expressions as arguments will be
supported in a future release
c) whether anyone has any idea how hard this might be to implement? I
might try to give implementing it a go myself and submit a patch.

In the end I got around it like this, but this isn't really ideal:

repositoryMock.Expect(r =>
r.GetAll<Absence>(It.IsAny<Expression<Func<Absence,
bool>>>())).Returns(absences.AsQueryable());

I'm using Moq on a fairly simple project at the moment, but if I was
doing something more complicated this work around probably wouldn't be
sufficient!

Any help/suggestions would be appreciated.

Thanks

Gareth

[Daniel Cazzulino]

btw, this is further aggravated by the fact that two *expressions* (which are NOT delegates until you compile them) will not be Equal even if they have the same lambda. Also, if you compile both, they're gonna be diffrerent delegates to.

So matching the arg would be quite complicated I think (would require full tree traversal for comparison)