seer-core 0.49.1

//! Parser for .jp domains (JPRS format).
//!
//! JPRS uses a unique format where fields are prefixed with a letter code
//! followed by a Japanese label in brackets. Nameservers use the key
//! `p. [ネームサーバ]`.
//!
//! Example JPRS response:
//! ```text
//! Domain Information: [ドメイン情報]
//! a. [ドメイン名]                 TOKYO.JP
//! f. [組織名]                     東京ドメイン
//! g. [Organization]               Tokyo Domain
//! p. [ネームサーバ]               a.dns.jp
//! p. [ネームサーバ]               b.dns.jp
//! [状態]                          Reserved
//! [最終更新]                      2005/03/30 17:37:52 (JST)
//! ```

use chrono::{DateTime, NaiveDateTime, Utc};
use regex::Regex;

use super::{push_bounded, MAX_NAMESERVERS, MAX_STATUSES};
use crate::whois::parser::WhoisResponse;

static_regex! {
    /// Matches nameserver lines in both the legacy lettered format
    /// (`p. [ネームサーバ]   value`) and the current bracket-only format
    /// (`[Name Server]   value` / `[ネームサーバ]   value`).
    NS_PATTERN = r"(?m)^(?:p\.[ \t]+\[ネームサーバ\]|\[(?:Name Server|ネームサーバ)\])[ \t]+(.+)$";

    /// Matches organization (English): `g. [Organization]   value`
    ORG_PATTERN = r"(?m)^g\.[ \t]+\[Organization\][ \t]+(.+)$";

    /// Matches organization (Japanese): `f. [組織名]   value`
    ORG_JP_PATTERN = r"(?m)^f\.[ \t]+\[組織名\][ \t]+(.+)$";

    /// Matches state: `[状態]   value` / `[Status]   value` / `[State]   value`
    STATUS_PATTERN = r"(?m)^\[(?:状態|Status|State)\][ \t]+(.+)$";

    /// Matches last updated: `[最終更新]` / `[Last Updated]` (general-use .jp) /
    /// `[Last Update]` (attribute-type co.jp, or.jp, ne.jp, ac.jp, go.jp, …)
    ///   YYYY/MM/DD HH:MM:SS (JST)
    UPDATED_PATTERN = r"(?m)^\[(?:最終更新|Last Updated?)\][ \t]+(.+)$";

    /// Matches creation date: `[登録年月日]` / `[接続年月日]` / `[Created on]`
    /// (general-use .jp) and `[Registered Date]` / `[Connected Date]`
    /// (attribute-type co.jp etc.)   YYYY/MM/DD
    CREATED_PATTERN =
        r"(?m)^\[(?:登録年月日|接続年月日|Created on|Registered Date|Connected Date)\][ \t]+(.+)$";

    /// Matches expiration date: `[有効期限]` / `[Expires on]`   YYYY/MM/DD
    EXPIRES_PATTERN = r"(?m)^\[(?:有効期限|Expires on)\][ \t]+(.+)$";

    /// Matches DNSSEC signing key line, lettered `s. [署名鍵]` / `s. [Signing Key]`
    /// (the latter is what attribute-type co.jp etc. print under `/e`) or current
    /// `[Signing Key]` / `[署名鍵]`. The value portion is optional — an empty
    /// value means unsigned. Uses `[^\S\n]*` instead of `\s*` to avoid matching
    /// across newlines in multiline mode.
    SIGNING_KEY_LINE = r"(?m)^(?:s\.[ \t]+)?\[(?:Signing Key|署名鍵)\][^\S\n]*(.+)?$";

    /// Attribute-type domains (co.jp, or.jp, …) carry their expiry inside the
    /// state value: `[State]   Connected (2027/03/31)`.
    STATE_EXPIRY_PATTERN = r"\((\d{4}/\d{1,2}/\d{1,2})\)";

    /// Matches registrant lines in the current bracket format: prefer the English
    /// `[Registrant]` value; fall back to the Japanese `[登録者名]`.
    REGISTRANT_EN_PATTERN = r"(?m)^\[Registrant\][ \t]+(.+)$";
    REGISTRANT_JP_PATTERN = r"(?m)^\[登録者名\][ \t]+(.+)$";

    // Also support the English-appended format (when /e is used)
    /// Matches `Name Server:   value` and the `p.`-prefixed form. For the `p.`
    /// branch an optional bracketed label (e.g. `[Name Server]`) is consumed but
    /// not captured, so only the trailing hostname lands in the capture group —
    /// otherwise a `p. [label] host` line yields `[label] host` as the nameserver.
    NS_EN_PATTERN = r"(?im)^(?:Name Server[ \t]*:?|p\.[ \t]*(?:\[[^\]]*\])?)[ \t]+(.+)$";
}

/// TLDs this parser handles.
pub(super) const TLDS: &[&str] = &["jp"];

/// Parses .jp domains using the JPRS format.
pub(super) fn parse(domain: &str, server: &str, raw: &str) -> WhoisResponse {
    let mut nameservers = Vec::new();
    let mut organization = None;
    let mut status = Vec::new();
    let mut has_signing_key = false;

    // Extract nameservers (Japanese format)
    for caps in NS_PATTERN.captures_iter(raw) {
        if let Some(m) = caps.get(1) {
            let ns = m.as_str().trim().to_lowercase();
            push_bounded(&mut nameservers, ns, MAX_NAMESERVERS);
        }
    }

    // If no Japanese-format NS found, try English format
    // (some .jp responses use "Name Server:" style after /e suffix)
    if nameservers.is_empty() {
        for caps in NS_EN_PATTERN.captures_iter(raw) {
            if let Some(m) = caps.get(1) {
                let ns = m.as_str().trim().to_lowercase();
                push_bounded(&mut nameservers, ns, MAX_NAMESERVERS);
            }
        }
    }

    // Extract organization (prefer English, fall back to Japanese)
    if let Some(caps) = ORG_PATTERN.captures(raw) {
        if let Some(m) = caps.get(1) {
            let org = m.as_str().trim().to_string();
            if !org.is_empty() {
                organization = Some(org);
            }
        }
    }
    if organization.is_none() {
        if let Some(caps) = ORG_JP_PATTERN.captures(raw) {
            if let Some(m) = caps.get(1) {
                let org = m.as_str().trim().to_string();
                if !org.is_empty() {
                    organization = Some(org);
                }
            }
        }
    }
    // Current bracket format: [Registrant] (English) / [登録者名] (Japanese)
    for pattern in [&REGISTRANT_EN_PATTERN, &REGISTRANT_JP_PATTERN] {
        if organization.is_some() {
            break;
        }
        if let Some(caps) = pattern.captures(raw) {
            if let Some(m) = caps.get(1) {
                let org = m.as_str().trim().to_string();
                if !org.is_empty() {
                    organization = Some(org);
                }
            }
        }
    }

    // Extract status
    if let Some(caps) = STATUS_PATTERN.captures(raw) {
        if let Some(m) = caps.get(1) {
            let s = m.as_str().trim().to_string();
            push_bounded(&mut status, s, MAX_STATUSES);
        }
    }

    // Extract dates (first occurrence of the label that parses, so an
    // empty `[Registered Date]` falls through to `[Connected Date]`).
    let first_date = |pattern: &Regex| {
        pattern
            .captures_iter(raw)
            .filter_map(|caps| caps.get(1))
            .find_map(|m| parse_jprs_date(m.as_str()))
    };
    let updated_date = first_date(&UPDATED_PATTERN);
    let creation_date = first_date(&CREATED_PATTERN);
    let expiration_date = first_date(&EXPIRES_PATTERN).or_else(|| {
        status.iter().find_map(|s| {
            STATE_EXPIRY_PATTERN
                .captures(s)
                .and_then(|caps| caps.get(1))
                .and_then(|m| parse_jprs_date(m.as_str()))
        })
    });

    // DNSSEC — check if signing key has an actual value
    if let Some(caps) = SIGNING_KEY_LINE.captures(raw) {
        if let Some(m) = caps.get(1) {
            if !m.as_str().trim().is_empty() {
                has_signing_key = true;
            }
        }
    }

    // Only claim JPRS identity fields when the response actually carried
    // registration data. Synthesizing them onto a "No match!!" response
    // would trip is_available()'s registered-data guard, making
    // unregistered .jp domains impossible to report as available.
    let has_data = organization.is_some()
        || !nameservers.is_empty()
        || creation_date.is_some()
        || expiration_date.is_some()
        || !status.is_empty();

    WhoisResponse {
        domain: domain.to_string(),
        registrar: has_data.then(|| "JPRS".to_string()),
        registrant: organization.clone(),
        organization,
        registrant_country: has_data.then(|| "JP".to_string()),
        creation_date,
        expiration_date,
        updated_date,
        nameservers,
        status,
        dnssec: if has_signing_key {
            Some("signedDelegation".to_string())
        } else {
            None
        },
        whois_server: server.to_string(),
        raw_response: raw.to_string(),
        ..Default::default()
    }
}

fn parse_jprs_date(date_str: &str) -> Option<DateTime<Utc>> {
    let cleaned = date_str.trim().trim_end_matches("(JST)").trim();

    // JPRS uses "YYYY/MM/DD HH:MM:SS" or "YYYY/MM/DD"
    if let Ok(dt) = NaiveDateTime::parse_from_str(cleaned, "%Y/%m/%d %H:%M:%S") {
        // JST is UTC+9. Checked: the body is untrusted (port 43 is
        // plaintext), and a year at chrono's lower bound would make the
        // plain `-` operator panic on underflow.
        return dt.and_utc().checked_sub_signed(chrono::Duration::hours(9));
    }
    if let Ok(dt) =
        NaiveDateTime::parse_from_str(&format!("{} 00:00:00", cleaned), "%Y/%m/%d %H:%M:%S")
    {
        return Some(dt.and_utc());
    }

    None
}

#[cfg(test)]
mod tests {
    use super::*;

    const SAMPLE_JPRS_RESPONSE: &str = r#"[ JPRS database provides information on network administration. Its use is    ]
[ restricted to network administration purposes. For further information,     ]
[ use 'whois -h whois.jprs.jp help'. To suppress Japanese output, add'/e'     ]
[ at the end of command, e.g. 'whois -h whois.jprs.jp xxx/e'.                 ]
Domain Information: [ドメイン情報]
a. [ドメイン名]                 TOKYO.JP
e. [そしきめい]
f. [組織名]                     東京ドメイン
g. [Organization]               Tokyo Domain
k. [組織種別]
l. [Organization Type]
m. [登録担当者]
n. [技術連絡担当者]
p. [ネームサーバ]               a.dns.jp
p. [ネームサーバ]               b.dns.jp
p. [ネームサーバ]               c.dns.jp
p. [ネームサーバ]               d.dns.jp
p. [ネームサーバ]               e.dns.jp
p. [ネームサーバ]               f.dns.jp
p. [ネームサーバ]               g.dns.jp
p. [ネームサーバ]               h.dns.jp
s. [署名鍵]
[状態]                          Reserved
[登録年月日]
[接続年月日]
[最終更新]                      2005/03/30 17:37:52 (JST)"#;

    const SAMPLE_JPRS_DOMAIN: &str = r#"[ JPRS database provides information on network administration. Its use is    ]
Domain Information: [ドメイン情報]
a. [ドメイン名]                 EXAMPLE.JP
f. [組織名]                     株式会社エグザンプル
g. [Organization]               Example Inc.
p. [ネームサーバ]               ns1.example.jp
p. [ネームサーバ]               ns2.example.jp
s. [署名鍵]
[状態]                          Active
[登録年月日]                    2001/03/22
[接続年月日]                    2001/03/22
[最終更新]                      2024/04/01 01:05:10 (JST)"#;

    #[test]
    fn test_jprs_nameservers() {
        let result = parse("tokyo.jp", "whois.jprs.jp", SAMPLE_JPRS_RESPONSE);

        assert_eq!(result.nameservers.len(), 8);
        assert!(result.nameservers.contains(&"a.dns.jp".to_string()));
        assert!(result.nameservers.contains(&"h.dns.jp".to_string()));
    }

    #[test]
    fn test_jprs_organization() {
        let result = parse("tokyo.jp", "whois.jprs.jp", SAMPLE_JPRS_RESPONSE);

        assert_eq!(result.organization, Some("Tokyo Domain".to_string()));
    }

    #[test]
    fn test_jprs_status() {
        let result = parse("tokyo.jp", "whois.jprs.jp", SAMPLE_JPRS_RESPONSE);

        assert!(result.status.contains(&"Reserved".to_string()));
    }

    #[test]
    fn test_jprs_updated_date() {
        let result = parse("tokyo.jp", "whois.jprs.jp", SAMPLE_JPRS_RESPONSE);

        assert!(result.updated_date.is_some());
    }

    #[test]
    fn test_jprs_domain_with_dates() {
        let result = parse("example.jp", "whois.jprs.jp", SAMPLE_JPRS_DOMAIN);

        assert_eq!(result.nameservers.len(), 2);
        assert!(result.nameservers.contains(&"ns1.example.jp".to_string()));
        assert!(result.nameservers.contains(&"ns2.example.jp".to_string()));
        assert_eq!(result.organization, Some("Example Inc.".to_string()));
        assert!(result.status.contains(&"Active".to_string()));
        assert!(result.creation_date.is_some());
        assert!(result.updated_date.is_some());
    }

    #[test]
    fn test_jprs_no_signing_key() {
        let result = parse("tokyo.jp", "whois.jprs.jp", SAMPLE_JPRS_RESPONSE);

        // s. [署名鍵] is empty, so no DNSSEC
        assert!(result.dnssec.is_none());
    }

    /// English-appended (/e) variant whose NS lines carry a `p.` prefix and a
    /// bracketed label that is NOT the exact Japanese `[ネームサーバ]`. The strict
    /// Japanese NS_PATTERN doesn't match, so parsing falls back to NS_EN_PATTERN.
    /// That fallback must extract only the hostname, not swallow the bracket
    /// label into the nameserver value.
    const SAMPLE_JPRS_EN_BRACKET_NS: &str = r#"Domain Information:
a. [Domain Name]                EXAMPLE.JP
p.        [Name Server]              a.dns.jp
p.        [Name Server]              b.dns.jp
[State]                         Active"#;

    #[test]
    fn test_jprs_en_fallback_strips_bracket_label() {
        let result = parse("example.jp", "whois.jprs.jp", SAMPLE_JPRS_EN_BRACKET_NS);

        assert_eq!(
            result.nameservers,
            vec!["a.dns.jp".to_string(), "b.dns.jp".to_string()],
            "fallback must capture only the hostname, not the bracket label"
        );
    }

    /// Current JPRS format (captured live 2026-07): bracket labels with NO
    /// letter prefixes — `[Name Server]`, Japanese date labels, and an expiry
    /// field `[有効期限]` that JPRS now publishes.
    const SAMPLE_JPRS_CURRENT: &str = r#"[ JPRS database provides information on network administration. Its use is    ]
Domain Information: [ドメイン情報]
[Domain Name]                   JPRS.JP

[登録者名]                      株式会社日本レジストリサービス
[Registrant]                    Japan Registry Services Co.,Ltd.

[Name Server]                   ns1.jprs.jp
[Name Server]                   ns2.jprs.jp
[Name Server]                   ns3.jprs.jp
[Name Server]                   ns4.jprs.jp
[Signing Key]                   7240 8 2 (
                                E147A85589E24FE0DBB5980C73501B5D
                                D656BE5550714F150BE574AE8777B77D )

[登録年月日]                    2001/02/02
[有効期限]                      2027/02/28
[状態]                          Active
[最終更新]                      2026/03/01 01:05:03 (JST)"#;

    /// Same record queried with the `/e` suffix: English-only labels.
    const SAMPLE_JPRS_CURRENT_EN: &str = r#"Domain Information:
[Domain Name]                   JPRS.JP

[Registrant]                    Japan Registry Services Co.,Ltd.

[Name Server]                   ns1.jprs.jp
[Name Server]                   ns2.jprs.jp
[Signing Key]                   7240 8 2 (
                                E147A85589E24FE0DBB5980C73501B5D )

[Created on]                    2001/02/02
[Expires on]                    2027/02/28
[Status]                        Active
[Last Updated]                  2026/03/01 01:05:03 (JST)"#;

    #[test]
    fn test_jprs_current_format_nameservers_and_dates() {
        let result = parse("jprs.jp", "whois.jprs.jp", SAMPLE_JPRS_CURRENT);

        assert_eq!(
            result.nameservers,
            vec!["ns1.jprs.jp", "ns2.jprs.jp", "ns3.jprs.jp", "ns4.jprs.jp"]
        );
        assert!(result.creation_date.is_some(), "created from [登録年月日]");
        assert!(result.expiration_date.is_some(), "expiry from [有効期限]");
        assert!(result.status.contains(&"Active".to_string()));
        assert_eq!(
            result.organization.as_deref(),
            Some("Japan Registry Services Co.,Ltd.")
        );
        assert_eq!(result.dnssec.as_deref(), Some("signedDelegation"));
    }

    /// Unregistered .jp: JPRS answers "No match!!". The parser must not
    /// hardcode registrar/country onto a no-match response — a synthesized
    /// registrar makes `is_available()` bail via its registered-data guard,
    /// so unregistered .jp domains could never be reported available.
    #[test]
    fn test_jprs_no_match_is_available() {
        let raw = "[ JPRS database provides information on network administration. Its use is    ]\n\
                   [ restricted to network administration purposes. For further information,     ]\n\
                   No match!!\n";
        let result = parse("nosuchdomain.jp", "whois.jprs.jp", raw);
        assert!(
            result.registrar.is_none(),
            "no-match must not synthesize a registrar"
        );
        assert!(
            result.is_available(),
            "No match!! must mark the domain available"
        );
    }

    #[test]
    fn test_jprs_current_format_english_labels() {
        let result = parse("jprs.jp", "whois.jprs.jp", SAMPLE_JPRS_CURRENT_EN);

        assert_eq!(result.nameservers, vec!["ns1.jprs.jp", "ns2.jprs.jp"]);
        assert!(result.creation_date.is_some(), "created from [Created on]");
        assert!(result.expiration_date.is_some(), "expiry from [Expires on]");
        assert!(result.status.contains(&"Active".to_string()));
        assert_eq!(result.dnssec.as_deref(), Some("signedDelegation"));
    }

    /// An EMPTY bracket field must not capture the next line: `\s+` after
    /// the label used to match the line break, turning an empty
    /// `p. [Name Server]` into the nameserver `s. [signing key]`.
    #[test]
    fn test_jprs_empty_fields_do_not_capture_next_line() {
        let raw = "Domain Information:\n\
                   a. [Domain Name]                EXAMPLE.CO.JP\n\
                   g. [Organization]               \n\
                   l. [Organization Type]          Company\n\
                   p. [Name Server]                \n\
                   s. [Signing Key]                \n\
                   [State]                         Connected (2027/03/31)\n";
        let result = parse("example.co.jp", "whois.jprs.jp", raw);
        assert!(
            result.nameservers.is_empty(),
            "empty NS line must not yield the next line: {:?}",
            result.nameservers
        );
        assert_eq!(
            result.organization, None,
            "empty [Organization] must not capture [Organization Type]"
        );
    }

    /// Attribute-type .jp (co.jp, or.jp, ne.jp, ac.jp, go.jp …) queried with
    /// `/e`: dates are `[Registered Date]` / `[Connected Date]` /
    /// `[Last Update]`, the signing key is `s. [Signing Key]`, and the expiry
    /// is carried in the state value `Connected (YYYY/MM/DD)`.
    const SAMPLE_JPRS_CO_JP_EN: &str = "\
[ JPRS database provides information on network administration. Its use is    ]
[ restricted to network administration purposes. For further information,     ]
[ use 'whois -h whois.jprs.jp help'. To suppress Japanese output, add'/e'     ]
[ at the end of command, e.g. 'whois -h whois.jprs.jp xxx/e'.                 ]

Domain Information:
a. [Domain Name]                GOOGLE.CO.JP
g. [Organization]               Google Japan G.K.
l. [Organization Type]          Limited Liability Company
m. [Administrative Contact]     DL152JP
n. [Technical Contact]          TW124137JP
p. [Name Server]                ns1.google.com
p. [Name Server]                ns2.google.com
p. [Name Server]                ns3.google.com
p. [Name Server]                ns4.google.com
s. [Signing Key]                \n\
[State]                         Connected (2027/03/31)
[Lock Status]                   DomainTransferLocked
[Lock Status]                   AgentChangeLocked
[Registered Date]               2001/03/22
[Connected Date]                2001/03/22
[Last Update]                   2026/04/01 01:05:10 (JST)
";

    #[test]
    fn test_jprs_attribute_type_co_jp_labels() {
        use chrono::{Datelike, Timelike};
        let result = parse("google.co.jp", "whois.jprs.jp", SAMPLE_JPRS_CO_JP_EN);

        assert_eq!(
            result.nameservers,
            vec![
                "ns1.google.com",
                "ns2.google.com",
                "ns3.google.com",
                "ns4.google.com"
            ]
        );
        assert_eq!(result.organization.as_deref(), Some("Google Japan G.K."));
        let created = result
            .creation_date
            .expect("created from [Registered Date]");
        assert_eq!(
            (created.year(), created.month(), created.day()),
            (2001, 3, 22)
        );
        let updated = result.updated_date.expect("updated from [Last Update]");
        // 01:05:10 JST == 16:05:10 UTC the previous day.
        assert_eq!(
            (updated.month(), updated.day(), updated.hour()),
            (3, 31, 16),
            "JST must be converted to UTC"
        );
        let expires = result
            .expiration_date
            .expect("expiry from the Connected (YYYY/MM/DD) state");
        assert_eq!(
            (expires.year(), expires.month(), expires.day()),
            (2027, 3, 31)
        );
        assert_eq!(
            result.dnssec, None,
            "an empty s. [Signing Key] means unsigned"
        );

        // A populated `s. [Signing Key]` line marks the zone as signed.
        let signed = SAMPLE_JPRS_CO_JP_EN.replace(
            "s. [Signing Key]                \n",
            "s. [Signing Key]                12345 8 2 ABCDEF0123456789\n",
        );
        assert_ne!(signed, SAMPLE_JPRS_CO_JP_EN, "fixture replacement applied");
        let result = parse("google.co.jp", "whois.jprs.jp", &signed);
        assert_eq!(result.dnssec.as_deref(), Some("signedDelegation"));
    }

    /// A hostile / on-path server (port 43 is plaintext) can send a date at
    /// chrono's lower bound; subtracting the JST offset must not underflow
    /// and panic — the date is simply dropped.
    #[test]
    fn test_jprs_date_at_chrono_min_does_not_panic() {
        let min = NaiveDateTime::MIN.format("%Y/%m/%d %H:%M:%S").to_string();
        assert!(
            NaiveDateTime::parse_from_str(&min, "%Y/%m/%d %H:%M:%S").is_ok(),
            "fixture must reach the JST-offset subtraction: {min}"
        );
        assert_eq!(parse_jprs_date(&format!("{min} (JST)")), None);

        let raw = format!(
            "Domain Information:\n\
             [Domain Name]                   EXAMPLE.JP\n\
             [Name Server]                   ns1.example.jp\n\
             [Last Updated]                  {min} (JST)\n"
        );
        let result = parse("example.jp", "whois.jprs.jp", &raw);
        assert_eq!(result.updated_date, None);
        assert_eq!(result.nameservers, vec!["ns1.example.jp"]);
    }
}