[Bug] Field with offset cannot print or to_numpy

[chunleili]

Describe the bug
Field with offset cannot print or to_numpy

To Reproduce

import taichi as ti
ti.init(debug=True)
shape = (3,3)
new_shape = (shape[0]+2, shape[1]+2)
val =  ti.field(dtype=ti.f32, shape=(new_shape), offset=(-1, -1))
val.fill(1.0)
print(val)

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:

Output:
[Taichi] version 1.6.0, llvm 15.0.1, commit 4c3e7cd3, win, python 3.10.9[Taichi] Starting on arch=x64
Traceback (most recent call last):
  File "d:\Dev\tiPBD\misc\code_bak\offset.py", line 7, in <module>      
    print(val)
  File "C:\Users\GRBJ200045\.conda\envs\ti-build\lib\site-packages\taichi\lang\field.py", line 240, in __str__
    return str(self.to_numpy())
  File "C:\Users\GRBJ200045\.conda\envs\ti-build\lib\site-packages\taichi\lang\util.py", line 311, in wrapped
    return func(*args, **kwargs)
  File "C:\Users\GRBJ200045\.conda\envs\ti-build\lib\site-packages\taichi\lang\field.py", line 302, in to_numpy
    tensor_to_ext_arr(self, arr)
  File "C:\Users\GRBJ200045\.conda\envs\ti-build\lib\site-packages\taichi\lang\kernel_impl.py", line 1033, in wrapped
    raise type(e)('\n' + str(e)) from None
taichi.lang.exception.TaichiAssertionError:
[kernel=tensor_to_ext_arr_c6_0] Out of bound access to ndarray at arg 0 
with indices [-1, -1]
File "C:\Users\GRBJ200045\.conda\envs\ti-build\lib\site-packages\taichi\_kernels.py", line 42, in tensor_to_ext_arr:
        arr[I] = tensor[I]

[ailzhang]

I believe this bug has been there for a while.... @chunleili would you mind help verifying if this was true for old taichi versions?

Btw the fix for it is kinda simple but we probably need to spread this in many places... https://github.com/taichi-dev/taichi/compare/master...ailzhang:example_fix?expand=1