fix: strip BOM and whitespace from authorization token

[sentry]

This PR addresses a UnicodeEncodeError that occurred when the Codecov CLI attempted to send an Authorization header containing a Byte Order Mark (BOM) character (\ufeff).

The error, "UnicodeEncodeError: 'latin-1' codec can't encode character '\ufeff'", was triggered because urllib3 (used by requests) attempts to encode HTTP headers using latin-1, which does not support the \ufeff character.

The root cause was identified as the token string, often provided via CLI arguments or environment variables, sometimes containing a leading \ufeff character (e.g., from copy-pasting from certain sources or specific file encodings). The CLI was using this token verbatim when constructing the Authorization header.

To fix this, the get_token_header and get_token_header_or_fail functions in codecov_cli/helpers/request.py have been updated to strip the \ufeff character and any surrounding whitespace from the token string before it is used to form the Authorization header. This ensures that only valid, encodable characters are passed to the HTTP client.

Fixes CLI-EP