refactor: use semantically correct ICRC-3 types for attribute values

[sea-snake]

Motivation

The ICRC-3 attribute message encoded all values as Blob, even for human-readable strings and numeric timestamps. Using the semantically correct ICRC-3 Value types makes the encoded message more self-describing and easier to decode on the client side.

Changes

Backend (attributes.rs)

Updated prepare_icrc3_attributes to use the appropriate ICRC-3 Value type for each entry:

Key	Before	After	Reason
Attribute values (email, name, verified_email)	Blob	Text	Human-readable strings
implicit:origin	Blob	Text	URL string
implicit:issued_at_timestamp_ns	Blob (string bytes)	Nat	Numeric value
implicit:nonce	Blob	Blob (unchanged)	Raw 32 bytes

Test app (auth.ts, index.html, index.tsx)

• Added optional icrc3Nonce parameter (base64 input field) so an app-supplied nonce can be used instead of a random one.

Tests

• Updated unit tests in icrc3_attribute_message_tests to use Icrc3Value::Text and verify decoded values.
• Updated integration test assertions for implicit:origin (Text) and implicit:issued_at_timestamp_ns (Nat).
• Updated e2e test helper from decodeIcrc3BlobEntries to decodeIcrc3TextEntries to extract Text variant entries.
• Added e2e verification of all implicit entries: implicit:origin is Text, implicit:nonce is a 32-byte Blob, and implicit:issued_at_timestamp_ns is a Nat > 0.
• Added "with app-supplied nonce" e2e test that passes a known nonce and verifies it appears byte-for-byte in the decoded ICRC-3 map.

---

← Previous: #3768

Next: #3771 →