setting dtype of array in ti.func causes TypeError: Expect TensorType element for Ndarray with element_dim: 1 > 0

[99991]

Describe the bug

It does not seem to be possible to set the dtype of ti.types.ndarray(dtype=vec3, ndim=1) in the signature of a ti.func to a vector type.

To Reproduce

import taichi as ti
ti.init()

vec3 = ti.types.vector(3, ti.f32)

@ti.func #                         vvvvvvvvvv-------------- Works without this part.
def child_func(a: ti.types.ndarray(dtype=vec3, ndim=1)):
    pass

@ti.kernel
def parent_func(x: ti.types.ndarray(dtype=vec3, ndim=1)):
    child_func(x)

arr = ti.ndarray(vec3, shape=(4))
parent_func(arr)

Log/Screenshots

$ python my_sample_code.py
[Taichi] version 1.4.1, llvm 15.0.4, commit e67c674e, linux, python 3.7.4
[Taichi] Starting on arch=x64
Traceback (most recent call last):
  File "test.py", line 15, in <module>
    parent_func(arr)
  File "/home/myusername/miniconda3/lib/python3.7/site-packages/taichi/lang/kernel_impl.py", line 976, in wrapped
    raise type(e)('\n' + str(e)) from None
taichi.lang.exception.TaichiCompilationError: 
File "test.py", line 12, in parent_func:
    child_func(x)
File "test.py", line 6, in child_func:
@ti.func
Traceback (most recent call last):
  File "/home/myusername/miniconda3/lib/python3.7/site-packages/taichi/lang/ast/ast_transformer_utils.py", line 25, in __call__
    return method(ctx, node)
  File "/home/myusername/miniconda3/lib/python3.7/site-packages/taichi/lang/ast/ast_transformer.py", line 652, in build_FunctionDef
    data.get_type())
  File "/home/myusername/miniconda3/lib/python3.7/site-packages/taichi/types/ndarray_type.py", line 112, in check_matched
    f"Expect TensorType element for Ndarray with element_dim: {self.dtype.ndim} > 0"
TypeError: Expect TensorType element for Ndarray with element_dim: 1 > 0

Additional comments

• Scalar types instead of vec3 work, e.g. by replacing vec3 = ti.types.vector(3, ti.f32) with vec3 = ti.f32.
• Not giving type information at all works, too.

[ailzhang]

@99991 I've confirmed that this error exists in v1.4.1 and is fixed in master, would you minding giving nightly a try?

pip install -i https://pypi.taichi.graphics/simple/ taichi-nightly

[99991]

Works for me.