Cannot find document when document id contains non-ascii characters

[CorieW]

An issue was created for this in firebase-functions-python #220, but I believe this is an issue with this project, so I'm creating an issue here.

Issue

My document id:

H67iEkcHx6Nicg5eoZbmxARLyJo2_KoX4bc4jGmTKwSGLI9GBdF46Hbt2_中餐

The last two characters are Chinese characters.

Relevant code:

event.data.after.reference.update() (here)

Error message:

google.api_core.exceptions.NotFound: 404 No document to update: projects/{project_name}/databases/(default)/documents/{collection_name}/H67iEkcHx6Nicg5eoZbmxARLyJo2_KoX4bc4jGmTKwSGLI9GBdF46Hbt2_中�"

Notes

The issue could be that the string "H67iEkcHx6Nicg5eoZbmxARLyJo2_KoX4bc4jGmTKwSGLI9GBdF46Hbt2_中餐" is being converted into "H67iEkcHx6Nicg5eoZbmxARLyJo2_KoX4bc4jGmTKwSGLI9GBdF46Hbt2_\344\270\255\351\244\220" inside the event.data.after.reference.update function. Notably, the part "\344\270\255\351\244\220" is an octal representation of the UTF-8 encoding for 中餐. To convert it, each octal sequence (e.g. \344) can be converted into hexadecimal to get \xe4\xb8\xad\xe9\xa4\x90 which is the UTF-8 byte sequence for 中餐.

[daniel-sanche]

@CorieW were you able to reproduce this using the library alone, or just through firebase-functions?

I added a new system test that creates a new document using the relevant unicode string, modifies it, and reads it back, and I couldn't reproduce the issue. And everything looks as expected in the cloud console:

Maybe try running the test code in your environment to see if you can still reproduce it?

Do you know how the document was created? Does the id look normal in your cloud console?

[daniel-sanche]

I'm going to close this for now since I couldn't reproduce, but feel free to re-open if you still see it