Decode an inbound email part in the charset it names, not always as UTF-8

[kevin9327]

What was wrong

extractEmailText (the SES → SNS email trigger) turned base64 and quoted-printable parts into bytes and then always read them as UTF-8, ignoring the charset on the part's Content-Type. A part sent as text/plain; charset=iso-8859-1 (or windows-1252, still common from Outlook and older mailers) saying Caf=E9 ferm=E9 reached the Bot as Caf� ferm�.

Change

decodeBody now takes the part's charset and decodes the bytes with TextDecoder(charset). A missing or unknown charset label still reads as UTF-8, so existing behaviour is unchanged for UTF-8 mail.

Test

New case in server/tests/responsibility-triggers.test.ts: an ISO-8859-1 quoted-printable part, a quoted windows-1252 base64 part, and a UTF-8 quoted-printable part.

• Before: Expected "Café fermé à minuit." Received "Caf� ferm� � minuit."
• After: bun test tests/responsibility-triggers.test.ts → 29 pass, 0 fail
• biome check clean on both files

[chatgpt-codex-connector]

You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard.