Behavior of break in the outermost for different from the documentation

[lin-hitonami]

Describe the bug
According to the documentation, only the for loop at the outermost scope is parallelized, and the outermost for loop not at the outermost scope is not parallelized, and parallelized for loops cannot be broken. However, according to my test, error raises when breaking the outermost for loop that is not at the outermost scope.

To Reproduce

import taichi as ti
ti.init()

@ti.kernel
def foo(k: ti.i32):
    if k > 42:
        for i in range(10):
            break

foo(1)

Log/Screenshots
Please post the full log of the program (instead of just a few lines around the error message, unless the log is > 1000 lines). This will help us diagnose what's happening. For example:

Traceback (most recent call last):
  File "test.py", line 10, in <module>
    foo(1)
  File "/Users/lin/taichi/python/taichi/lang/kernel_impl.py", line 736, in wrapped
    return primal(*args, **kwargs)
  File "/Users/lin/taichi/python/taichi/lang/kernel_impl.py", line 667, in __call__
    return self.compiled_functions[key](*args)
  File "/Users/lin/taichi/python/taichi/lang/kernel_impl.py", line 619, in func__
    t_kernel(launch_ctx)
RuntimeError: [ir.h:as@225] Assertion failure: is<T>()