Flyte enables production-grade orchestration for machine learning workflows and data processing created to accelerate local workflows to production.

Flyte

Wonder if something changed with chaining tasks using `>>`? Seeing this error when trying to chain tasks inside a dynamic task

```File "/usr/local/lib/python3.9/site-packages/flytekit/core/promise.py", line 297, in __rshift__
        if not self.is_ready and other.ref:

Message:

    'Output' object has no attribute 'ref'```

but I can't reproduce it in a basic example.. :confused:

can you paste your current example redacted?

that code hasn’t really changed in a while, but maybe something touched it in an edge case?

let me try. It's something with the 1.9.0/1 bc when I rolled back my flytekit docker to 1.8.1; it was ok.

yea...my minimal example started working instead of breaking when I put it all in one file :lolcry: instead of importing from diff files

:open_mouth: its when the output is of type Tuple.

```
from flytekit import dynamic, task, workflow
from typing import Tuple

@task
def task_2(
    condition: bool
) -> str:
    return str(condition)

@task
def task_3(
    condition: bool
) -> Tuple[str, str]:
    return str(condition), str(not condition)
    

@dynamic
def dynamic_1(
    condition: bool
):
    job3 = task_2(
        condition=condition
    )
    job4 = task_3(
        condition=condition
    )
    job3 >> job4


@workflow
def wf_2(
    condition: bool
):
    dynamic_1(
        condition=condition
    )```

but it works presumably if you do
```job4, job5 = task_3(
            condition=condition
        )```


ok i pruned a little bit; since the other tasks weren't relevant to this

```job4, job5 = task_3(
        condition=condition
    )
    job3 &gt;&gt; job4```

i think that’s the right way to call it though

The same happens with a NamedTuple, what is the suggested way here?

```@task
def task1() -&gt; None:
    print("Task1")

Task2Out = NamedTuple("Task2Out", out1=int, out2=int)

@task
def task2() -&gt; Task2Out:
    return Task2Out(1, 2)


@workflow
def demo_wf() -&gt; None:
    out1 = task1()
    out2 = task2()

    out1 &gt;&gt; out2```