[wleeaf]

The zone ID pattern only allows letters and digits, so link-local addresses on common interfaces are rejected:

ipaddr.IPv6.isValid('fe80::1%eth0');      // true
ipaddr.IPv6.isValid('fe80::1%eth0.100');  // false (VLAN interface)
ipaddr.IPv6.parse('fe80::1%br-lan');      // throws: string is not formatted like an IPv6 Address (bridge, e.g. OpenWrt)
ipaddr.IPv6.parse('fe80::1%en_0');        // throws

RFC 6874 (section 2) defines ZoneID = 1*( unreserved / pct-encoded ), and unreserved is ALPHA / DIGIT / "-" / "." / "_" / "~". Python's ipaddress accepts them too (ipaddress.IPv6Address('fe80::1%br-lan')).

The zoneIndex pattern is now %[0-9a-z._~-]{1,}. It is shared by the native and transitional IPv6 patterns and by the zone extraction in expandIPv6, so parse, isValid, parseCIDR and IPv4-mapped addresses with a zone all accept the new characters. A zone ID still can't be empty or contain spaces, % or /.

The new test accepts zone IDs with the other RFC 6874 unreserved characters covers ., -, _ and ~, an IPv4-mapped address, parseCIDR, and two invalid zone IDs. It fails without the change. npm test (71 tests) and npm run lint pass.

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has insufficient remaining quota.