Summary

Fixes UnicodeDecodeError by allowing explicit encoding selection when reading .env files.

Resolves #2016

Description

Modified the _read_file method in starlette/config.py to accept an encoding parameter.
This ensures .env files can be correctly decoded regardless of the system's default locale encoding.

Key changes:

• Added encoding parameter (default: "utf-8") to _read_file
• Passed encoding explicitly to open() when reading .env files
• Retained existing parsing logic (key/value handling unchanged)

Motivation

The current implementation uses open(file_name) without specifying an encoding, which defaults to the system’s locale (e.g., GBK on Windows in Chinese locales).
This leads to UnicodeDecodeError when the .env file is encoded in UTF-8 (a common format).

Expected Outcome

After this fix:

• ✅ UTF-8 .env files can be read on Windows systems using GBK/CP949 locales
• ✅ Developers have flexibility to override encoding if needed
• ✅ Backward compatibility preserved (utf-8 is default)
• ✅ No additional runtime cost introduced

Checklist

• I understand that this PR may be closed in case there was no previous discussion. (This doesn't apply to typos!)
• I've added a test for each change that was introduced, and I tried as much as possible to make a single atomic change.
• I've updated the documentation accordingly.