Grid tables with double-width characters crash rst2myst: “null bytes should be removed by now” (Docutils NUL \x00 leaks into tokens)

[coxxny]

Describe the bug

context
When I run rst2myst on an .rst file (UTF-8) that contains a grid table with double-width characters (e.g., Japanese):

+-------+-------+
|  列A  |  列B  |
+-------+-------+
|  あい |  かき |
+-------+-------+

expectation
I expected the table to convert to MyST/Markdown (or safely fall back) without errors—just like Docutils' rst2html.py renders the same file correctly.

bug
But instead rst2myst crashes with a null-byte error.
Here’s an error message I ran into:

$ rst2myst convert test.rst
test.rst -> test.md
FAILED:
null bytes should be removed by now

FINISHED ALL! (extensions: [])

problem
This is a problem for people converting RST docs that include fullwidth text—e.g., multilingual projects using rst-to-myst—because the build fails on valid tables. Internally, Docutils pads double-width characters with a NUL (\x00) during table parsing; that sentinel appears to leak into token text in the RST→MyST path and triggers the renderer’s null-byte error, breaking documentation builds.

Reproduce the bug

1. Create test.rst (UTF-8, no BOM) with a grid table that includes double-width characters.
2. run rst2myst convert test.rst.

List your environment

No response