Behaviour of ExecuteAndCapture with PolicyWrap and with Fallback

I’m wrapping a retry policy in timeout policy. The action returns null so I want to retry.

With retry policy alone, I get result.Outcome as Failure as expected. But with timeout policy, it is Successful with result.Result as null. Shouldn’t it also be a Failure?

If the action throws, then I see Failure for both retry and timeout policies.

class Holder
{
	public bool done;
	public Holder holder;
}

[Test]
public async Task Polly_Result_Outcome_Should_Be_Failure()
{
	Policy<Holder> retry = Policy
				   .Handle<Exception>()
				   .OrResult<Holder>(h =>
				   {
					   var flag = !(h?.holder?.done) ?? true;
					   return flag;
				   })
				   .WaitAndRetryAsync(
					   2,
					   retryAttempt => TimeSpan.FromMilliseconds(1000));
	var timeout = Policy
				  .TimeoutAsync(10)
				  .WrapAsync(retry);
	var result = await timeout.ExecuteAndCaptureAsync(async (ct) =>
	{
		ct.ThrowIfCancellationRequested();
		//throw new Exception();
		return await Task.FromResult((Holder)null);
	}, CancellationToken.None)
		.ConfigureAwait(false);
	if (result.Outcome != OutcomeType.Successful)
	{
		// retry policy comes here as expected
		Assert.Pass();
	}
	if (result.Result == null)
	{
		// timeout policy comes here
		// why is Outcome not OutcomeType.Failure just like retry policy?
	}
	Assert.Fail();
}

[reisenberger]

Closing this as I think we’re done here, and we’ve updated the wiki to cover this more precisely. However, if you do need anything else, please re-open the issue or start a fresh one. Also updated the title of the issue for future discoverability. Thanks for your great question!

[reisenberger]

@rmandvikar Thanks again for your great example which highlights these important points! We’ve updated the documentation for both PolicyWrap and Fallback to make this clearer.

Let us know if there is anything else we can help with.