fix(email): decode IMAP folder names from modified UTF-7 (#3514)

[maxmilian]

Summary

IMAP mailbox names are encoded in modified UTF-7 (RFC 3501), but nothing decoded them, so folders with non-Latin names rendered as their raw wire form (e.g. &BB4EQgQ,BEAEMAQyBDsENQQ9BD0ESwQ1- instead of Cyrillic). imaplib does no decoding and Python ships no imap4-utf-7 codec.

This adds a stdlib-only src/imap_utf7 codec and establishes one invariant: folder names are kept decoded everywhere in the Python layer, and _q() — the single mailbox-quoting chokepoint for SELECT/COPY/MOVE/APPEND — re-encodes to modified UTF-7 on the way to the server. Decoding is applied where names are read from the LIST response (_folder_name_from_list_line, the /email/folders endpoint, and the \Sent/\Drafts/\Junk detectors), so non-ASCII folders both display correctly and stay selectable (no display-only regression). ASCII names are unchanged by encode/decode.

Files: new src/imap_utf7.py; routes/email_routes.py, routes/email_helpers.py, mcp_servers/email_server.py (decode on read, encode in _q).

Note: the report's expected/actual labels were swapped (&BB4EQgQ,…- actually decodes to Отправленные/“Sent”, not Входящие), but the underlying bug — no MUTF-7 decoding — is exactly as described.

Target branch

• This PR targets dev, not main.

Linked Issue

Closes #3514

Type of Change

• Bug fix
• Feature
• Refactor
• Documentation

Checklist

• I searched open issues and open PRs — this is not a duplicate.
• This PR targets dev, not main.
• The change is limited in scope to the linked issue.
• I ran the app end-to-end against a live IMAP account with non-ASCII folders. (No server with Cyrillic folders on hand; covered by unit tests using the real RFC 3501 wire form instead.)

How to Test

Backend-only, no UI changes. Run the codec + IMAP quoting tests:

python -m pytest tests/test_imap_utf7.py tests/test_imap_mailbox_quoting.py

Or check the round-trip directly:

from src.imap_utf7 import decode_imap_utf7, encode_imap_utf7
decode_imap_utf7("&BB4EQgQ,BEAEMAQyBDsENQQ9BD0ESwQ1-")   # -> 'Отправленные'
encode_imap_utf7("Отправленные")                          # -> '&BB4EQgQ,BEAEMAQyBDsENQQ9BD0ESwQ1-'

To verify against a real account: add an IMAP account that has a folder with a non-Latin name and open the folder list — it now shows the decoded Unicode name, and selecting/reading that folder still works (the name is re-encoded before it hits the server).

Testing

• python -m pytest tests/test_imap_utf7.py tests/test_imap_mailbox_quoting.py → 17 passed (9 new codec tests + existing quoting tests, which still pass since ASCII names are unchanged).
• python -m py_compile on all four changed files → clean.

[maxmilian]

Re-verified on a rebase onto today's dev (c80462e) — still cleanly applicable and, as far as I can tell, still the only Modified-UTF-7 handling in the tree (no decoder landed upstream since this was opened; #3514 remains open).

• tests/test_imap_utf7.py + tests/test_imap_mailbox_quoting.py → 17 passed
• full email/IMAP-related selection (-k "email or imap") → 213 passed

Happy to adjust scope or squash if the maintainers prefer a different shape for this.