retex-engine 1.0.0-rc.2

//! The job's start time: `\time`, `\day`, `\month`, `\year`, the log
//! banner and the PDF creation date.
//!
//! LuaTeX takes the time once, at startup: from `SOURCE_DATE_EPOCH` when it
//! is set (with or without `FORCE_SOURCE_DATE`), otherwise from the clock.
//! It converts it to local time, or to UTC with the `-utc` option. Dates in
//! PDF format carry the UTC offset (`D:19700101010000+01'00'` for epoch 0 in
//! Central Europe, `D:19700101000000Z` with `TZ=UTC` or `-utc`).
//!
//! The local offset is found the way the C library finds it: from `$TZ` (a
//! zone name, optionally prefixed with `:`, or a POSIX TZ string), else from
//! `/etc/localtime`, reading the binary TZif files (RFC 8536) and their POSIX
//! TZ footer rules (POSIX.1, *TZ* variable) for times after the last
//! transition. Anything unreadable falls back to UTC.

/// Where zone files live unless `$TZDIR` says otherwise.
const ZONEINFO: &str = "/usr/share/zoneinfo";

/// An instant together with the UTC offset in effect there.
#[derive(Clone, Copy, Debug, Default, PartialEq, Eq)]
pub struct StartTime {
    /// Seconds since 1970-01-01 00:00:00 UTC.
    pub unix: i64,
    /// Offset of the time zone used, in seconds east of UTC.
    pub offset: i64,
}

/// A broken-down date and time.
#[derive(Clone, Copy, Debug, PartialEq, Eq)]
pub struct Civil {
    /// The year.
    pub year: i64,
    /// The month, 1–12.
    pub month: i64,
    /// The day of the month, 1–31.
    pub day: i64,
    /// The hour, 0–23.
    pub hour: i64,
    /// The minute, 0–59.
    pub minute: i64,
    /// The second, 0–59.
    pub second: i64,
}

impl StartTime {
    /// The start time of a job: `epoch` (from `SOURCE_DATE_EPOCH`) or the
    /// current time, in the local time zone or, if `utc`, in UTC.
    pub fn new(epoch: Option<i64>, utc: bool) -> StartTime {
        let unix = epoch.unwrap_or_else(now);
        StartTime {
            unix,
            offset: if utc { 0 } else { local_offset(unix) },
        }
    }

    /// The local date and time.
    pub fn civil(&self) -> Civil {
        let local = self.unix.saturating_add(self.offset);
        let (year, month, day) = civil_from_days(local.div_euclid(86400));
        let secs = local.rem_euclid(86400);
        Civil {
            year,
            month,
            day,
            hour: secs / 3600,
            minute: secs / 60 % 60,
            second: secs % 60,
        }
    }

    /// The date in PDF format, as `pdf.getcreationdate()` and
    /// `\pdffeedback creationdate` return it and as the Info dictionary
    /// contains it: `D:YYYYMMDDHHmmSS` followed by `Z` or the offset as
    /// `+HH'mm'`.
    pub fn pdf_date(&self) -> String {
        let c = self.civil();
        let mut s = format!(
            "D:{:04}{:02}{:02}{:02}{:02}{:02}",
            c.year, c.month, c.day, c.hour, c.minute, c.second
        );
        if self.offset == 0 {
            s.push('Z');
        } else {
            let sign = if self.offset < 0 { '-' } else { '+' };
            let o = self.offset.abs() / 60;
            s.push_str(&format!("{sign}{:02}'{:02}'", o / 60, o % 60));
        }
        s
    }
}

/// The random seed of a job that does not set one.
///
/// LuaTeX 1.22 takes the clock's microseconds times 1000 plus its seconds
/// modulo 1000 (`\randomseed` is 720667377 at 1792245377.720667 s), also
/// when `SOURCE_DATE_EPOCH` is set, so the default sequence differs from
/// run to run.
pub fn startup_random_seed() -> i32 {
    let d = std::time::SystemTime::now()
        .duration_since(std::time::UNIX_EPOCH)
        .unwrap_or_default();
    (i64::from(d.subsec_micros()) * 1000 + (d.as_secs() % 1000) as i64) as i32
}

/// The current time in seconds since the epoch.
pub fn now() -> i64 {
    std::time::SystemTime::now()
        .duration_since(std::time::UNIX_EPOCH)
        .map_or(0, |d| d.as_secs() as i64)
}

/// `SOURCE_DATE_EPOCH` from the environment, interpreted like LuaTeX does
/// (see [`parse_source_date_epoch`]).
pub fn source_date_epoch() -> Option<i64> {
    let v = std::env::var_os("SOURCE_DATE_EPOCH")?;
    parse_source_date_epoch(&v.to_string_lossy())
}

/// Interprets a `SOURCE_DATE_EPOCH` value; `None` means "use the clock".
///
/// Observed behaviour of LuaTeX 1.22 (the value is honoured whether or not
/// `FORCE_SOURCE_DATE` is set):
///
/// * an empty value is ignored;
/// * leading white space and a sign are accepted, then decimal digits that
///   must reach the end of the value; anything else (`abc`, `5x`, `5 `,
///   `0x10`, blanks only) and values that overflow 64 bits count as 0;
/// * a minus sign negates modulo 2⁶⁴; the result is then truncated to
///   32 bits, and negative results are ignored (`-5`, `4294967295`), so
///   `99999999999` means 1215752191 and `4294967301` means 5.
pub fn parse_source_date_epoch(v: &str) -> Option<i64> {
    if v.is_empty() {
        return None;
    }
    let s = v.trim_start_matches([' ', '\t', '\n', '\x0b', '\x0c', '\r']);
    let (neg, digits) = match s.as_bytes().first() {
        Some(b'-') => (true, &s[1..]),
        Some(b'+') => (false, &s[1..]),
        _ => (false, s),
    };
    let value = if !digits.is_empty() && digits.bytes().all(|b| b.is_ascii_digit()) {
        digits.parse::<u64>().unwrap_or(0)
    } else {
        0
    };
    let value = if neg { value.wrapping_neg() } else { value };
    let t = value as u32 as i32;
    (t >= 0).then_some(i64::from(t))
}

/// Returns the local UTC offset (seconds east of UTC) at Unix time `t`.
pub fn local_offset(t: i64) -> i64 {
    match std::env::var("TZ") {
        Ok(tz) => offset_for_tz(&tz, t).unwrap_or(0),
        Err(_) => std::fs::read("/etc/localtime")
            .ok()
            .and_then(|d| tzif_offset(&d, t))
            .unwrap_or(0),
    }
}

/// The offset at `t` for a `TZ` value.
pub fn offset_for_tz(tz: &str, t: i64) -> Option<i64> {
    let tz = tz.strip_prefix(':').unwrap_or(tz);
    if tz.is_empty() {
        return Some(0);
    }
    let path = if tz.starts_with('/') {
        Some(std::path::PathBuf::from(tz))
    } else if !tz.contains("..") {
        let dir = std::env::var_os("TZDIR").unwrap_or_else(|| ZONEINFO.into());
        Some(std::path::Path::new(&dir).join(tz))
    } else {
        None
    };
    if let Some(data) = path.and_then(|p| std::fs::read(p).ok())
        && let Some(off) = tzif_offset(&data, t)
    {
        return Some(off);
    }
    PosixTz::parse(tz).map(|r| r.offset_at(t))
}

fn be32(d: &[u8], p: usize) -> Option<i64> {
    d.get(p..p + 4)
        .map(|b| i32::from_be_bytes([b[0], b[1], b[2], b[3]]) as i64)
}

fn be64(d: &[u8], p: usize) -> Option<i64> {
    d.get(p..p + 8).map(|b| {
        let mut a = [0u8; 8];
        a.copy_from_slice(b);
        i64::from_be_bytes(a)
    })
}

/// The UTC offset at `t` according to a TZif file.
pub fn tzif_offset(d: &[u8], t: i64) -> Option<i64> {
    if d.get(0..4)? != b"TZif" {
        return None;
    }
    let version = *d.get(4)?;
    let counts = |base: usize| -> Option<[usize; 6]> {
        let mut c = [0usize; 6];
        for (i, v) in c.iter_mut().enumerate() {
            *v = usize::try_from(be32(d, base + 20 + 4 * i)?).ok()?;
        }
        Some(c)
    };
    let [isut, isstd, leap, timecnt, typecnt, charcnt] = counts(0)?;
    let v1_len = 44 + timecnt * 5 + typecnt * 6 + charcnt + leap * 8 + isstd + isut;
    // Use the 64-bit data block of version 2+ files.
    let (base, tsize, c) = if version >= b'2' {
        (v1_len, 8, counts(v1_len)?)
    } else {
        (0, 4, counts(0)?)
    };
    let [isut, isstd, leap, timecnt, typecnt, charcnt] = c;
    if typecnt == 0 {
        return None;
    }
    let times = base + 44;
    let idx = times + timecnt * tsize;
    let types = idx + timecnt;
    let read_t = |i: usize| {
        if tsize == 8 {
            be64(d, times + 8 * i)
        } else {
            be32(d, times + 4 * i)
        }
    };
    let ttinfo = |i: usize| -> Option<(i64, bool)> {
        let p = types + 6 * i;
        Some((be32(d, p)?, *d.get(p + 4)? != 0))
    };
    let footer = types + typecnt * 6 + charcnt + leap * (tsize + 4) + isstd + isut;

    // Find the last transition at or before t.
    let mut last: Option<usize> = None;
    for i in 0..timecnt {
        if read_t(i)? <= t {
            last = Some(i);
        } else {
            break;
        }
    }
    match last {
        Some(i) if i + 1 == timecnt && version >= b'2' => {
            // After the last transition, the footer rule applies.
            if let Some(rule) = footer_rule(d, footer) {
                return Some(rule.offset_at(t));
            }
            let ty = *d.get(idx + i)? as usize;
            Some(ttinfo(ty)?.0)
        }
        Some(i) => {
            let ty = *d.get(idx + i)? as usize;
            Some(ttinfo(ty)?.0)
        }
        None => {
            if timecnt == 0
                && version >= b'2'
                && let Some(rule) = footer_rule(d, footer)
            {
                return Some(rule.offset_at(t));
            }
            // Before the first transition: time type 0 (RFC 8536, 3.2).
            Some(ttinfo(0)?.0)
        }
    }
}

fn footer_rule(d: &[u8], at: usize) -> Option<PosixTz> {
    let rest = d.get(at..)?;
    let rest = rest.strip_prefix(b"\n")?;
    let end = rest.iter().position(|&b| b == b'\n')?;
    let s = std::str::from_utf8(&rest[..end]).ok()?;
    if s.is_empty() {
        return None;
    }
    PosixTz::parse(s)
}

/// A DST transition rule date.
#[derive(Clone, Copy, Debug, PartialEq)]
enum RuleDate {
    /// `Jn`: day 1–365, February 29 never counted.
    Julian1(i64),
    /// `n`: day 0–365, counting February 29.
    Julian0(i64),
    /// `Mm.w.d`: day `d` (0 = Sunday) of week `w` (5 = last) of month `m`.
    MonthWeekDay(i64, i64, i64),
}

/// A parsed POSIX TZ string.
#[derive(Clone, Copy, Debug, PartialEq)]
struct PosixTz {
    /// Standard offset, seconds east of UTC.
    std: i64,
    /// DST offset and rules: (offset, start, start time, end, end time).
    dst: Option<(i64, RuleDate, i64, RuleDate, i64)>,
}

struct Cursor<'a> {
    s: &'a [u8],
    p: usize,
}

impl Cursor<'_> {
    fn peek(&self) -> Option<u8> {
        self.s.get(self.p).copied()
    }

    fn eat(&mut self, c: u8) -> bool {
        if self.peek() == Some(c) {
            self.p += 1;
            true
        } else {
            false
        }
    }

    fn name(&mut self) -> Option<()> {
        if self.eat(b'<') {
            while self.peek()? != b'>' {
                self.p += 1;
            }
            self.p += 1;
            return Some(());
        }
        let start = self.p;
        while self.peek().is_some_and(|c| c.is_ascii_alphabetic()) {
            self.p += 1;
        }
        (self.p - start >= 3).then_some(())
    }

    fn number(&mut self) -> Option<i64> {
        let start = self.p;
        while self.peek().is_some_and(|c| c.is_ascii_digit()) {
            self.p += 1;
        }
        std::str::from_utf8(&self.s[start..self.p])
            .ok()?
            .parse()
            .ok()
    }

    /// `[+-]hh[:mm[:ss]]` in seconds.
    fn hms(&mut self) -> Option<i64> {
        let neg = if self.eat(b'-') {
            true
        } else {
            self.eat(b'+');
            false
        };
        let mut v = self.number()? * 3600;
        if self.eat(b':') {
            v += self.number()? * 60;
            if self.eat(b':') {
                v += self.number()?;
            }
        }
        Some(if neg { -v } else { v })
    }

    fn date(&mut self) -> Option<RuleDate> {
        if self.eat(b'J') {
            Some(RuleDate::Julian1(self.number()?))
        } else if self.eat(b'M') {
            let m = self.number()?;
            self.eat(b'.').then_some(())?;
            let w = self.number()?;
            self.eat(b'.').then_some(())?;
            let d = self.number()?;
            Some(RuleDate::MonthWeekDay(m, w, d))
        } else {
            Some(RuleDate::Julian0(self.number()?))
        }
    }
}

impl PosixTz {
    fn parse(s: &str) -> Option<PosixTz> {
        let mut c = Cursor {
            s: s.as_bytes(),
            p: 0,
        };
        c.name()?;
        let std = -c.hms()?;
        if c.peek().is_none() {
            return Some(PosixTz { std, dst: None });
        }
        c.name()?;
        let dst_off = if matches!(c.peek(), Some(b',') | None) {
            std + 3600
        } else {
            -c.hms()?
        };
        let (start, st, end, et) = if c.eat(b',') {
            let start = c.date()?;
            let st = if c.eat(b'/') { c.hms()? } else { 7200 };
            c.eat(b',').then_some(())?;
            let end = c.date()?;
            let et = if c.eat(b'/') { c.hms()? } else { 7200 };
            (start, st, end, et)
        } else {
            // POSIX leaves the default rules implementation-defined; use the
            // US rules like glibc.
            (
                RuleDate::MonthWeekDay(3, 2, 0),
                7200,
                RuleDate::MonthWeekDay(11, 1, 0),
                7200,
            )
        };
        Some(PosixTz {
            std,
            dst: Some((dst_off, start, st, end, et)),
        })
    }

    fn offset_at(&self, t: i64) -> i64 {
        let Some((dst, start, st, end, et)) = self.dst else {
            return self.std;
        };
        let year = civil_from_days((t + self.std).div_euclid(86400)).0;
        // Transition instants in UTC: the start is given in standard local
        // time, the end in daylight local time.
        let s = days_from_civil(year, 1, 1) * 86400 + rule_day_offset(start, year) * 86400 + st
            - self.std;
        let e = days_from_civil(year, 1, 1) * 86400 + rule_day_offset(end, year) * 86400 + et - dst;
        let in_dst = if s < e {
            t >= s && t < e
        } else {
            !(t >= e && t < s)
        };
        if in_dst { dst } else { self.std }
    }
}

fn is_leap(y: i64) -> bool {
    (y % 4 == 0 && y % 100 != 0) || y % 400 == 0
}

/// Days since January 1 of `year` for a rule date.
fn rule_day_offset(r: RuleDate, year: i64) -> i64 {
    match r {
        RuleDate::Julian1(n) => {
            let n = n.clamp(1, 365);
            if is_leap(year) && n >= 60 { n } else { n - 1 }
        }
        RuleDate::Julian0(n) => n,
        RuleDate::MonthWeekDay(m, w, d) => {
            let m = m.clamp(1, 12);
            let first = days_from_civil(year, m, 1);
            let jan1 = days_from_civil(year, 1, 1);
            // 1970-01-01 was a Thursday (4).
            let wd_first = (first + 4).rem_euclid(7);
            let mut day = first + (d - wd_first).rem_euclid(7) + 7 * (w.clamp(1, 5) - 1);
            let (ny, nm) = if m == 12 {
                (year + 1, 1)
            } else {
                (year, m + 1)
            };
            let next = days_from_civil(ny, nm, 1);
            while day >= next {
                day -= 7;
            }
            day - jan1
        }
    }
}

/// Days since 1970-01-01 of a proleptic Gregorian date.
pub fn days_from_civil(y: i64, m: i64, d: i64) -> i64 {
    let y = if m <= 2 { y - 1 } else { y };
    let era = y.div_euclid(400);
    let yoe = y - era * 400;
    let mp = (m + 9) % 12;
    let doy = (153 * mp + 2) / 5 + d - 1;
    let doe = yoe * 365 + yoe / 4 - yoe / 100 + doy;
    era * 146097 + doe - 719468
}

/// The proleptic Gregorian date (year, month, day) of a day number.
pub fn civil_from_days(z: i64) -> (i64, i64, i64) {
    let z = z + 719468;
    let era = z.div_euclid(146097);
    let doe = z - era * 146097;
    let yoe = (doe - doe / 1460 + doe / 36524 - doe / 146096) / 365;
    let y = yoe + era * 400;
    let doy = doe - (365 * yoe + yoe / 4 - yoe / 100);
    let mp = (5 * doy + 2) / 153;
    let d = doy - (153 * mp + 2) / 5 + 1;
    let m = if mp < 10 { mp + 3 } else { mp - 9 };
    (if m <= 2 { y + 1 } else { y }, m, d)
}

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

    #[test]
    fn civil_round_trip() {
        assert_eq!(days_from_civil(1970, 1, 1), 0);
        assert_eq!(civil_from_days(0), (1970, 1, 1));
        assert_eq!(civil_from_days(-1), (1969, 12, 31));
        assert_eq!(days_from_civil(2000, 3, 1), 11017);
        for z in [-800000, -1, 0, 59, 10957, 20000, 2932896] {
            let (y, m, d) = civil_from_days(z);
            assert_eq!(days_from_civil(y, m, d), z);
        }
    }

    #[test]
    fn posix_rules() {
        let cet = PosixTz::parse("CET-1CEST,M3.5.0,M10.5.0/3").unwrap();
        assert_eq!(cet.offset_at(0), 3600);
        // 2026-07-01 12:00 UTC
        assert_eq!(
            cet.offset_at(days_from_civil(2026, 7, 1) * 86400 + 43200),
            7200
        );
        // 2026-03-29 00:59:59 UTC is still CET, 01:00 UTC is CEST.
        let t = days_from_civil(2026, 3, 29) * 86400 + 3600;
        assert_eq!(cet.offset_at(t - 1), 3600);
        assert_eq!(cet.offset_at(t), 7200);
        let ny = PosixTz::parse("EST5EDT,M3.2.0,M11.1.0").unwrap();
        assert_eq!(ny.offset_at(0), -18000);
        let india = PosixTz::parse("IST-5:30").unwrap();
        assert_eq!(india.offset_at(0), 19800);
        let quoted = PosixTz::parse("<+0330>-3:30").unwrap();
        assert_eq!(quoted.offset_at(0), 12600);
        assert!(PosixTz::parse("X").is_none());
    }

    #[test]
    fn zone_files_if_present() {
        // These match lualatex's dates for SOURCE_DATE_EPOCH=0.
        if std::path::Path::new("/usr/share/zoneinfo/America/New_York").exists() {
            assert_eq!(offset_for_tz("America/New_York", 0), Some(-18000));
            assert_eq!(offset_for_tz("Asia/Kolkata", 0), Some(19800));
            assert_eq!(offset_for_tz("Europe/Paris", 0), Some(3600));
            assert_eq!(offset_for_tz("Europe/Paris", 1_790_000_000), Some(7200));
        }
        assert_eq!(offset_for_tz("UTC0", 0), Some(0));
        assert_eq!(offset_for_tz("", 0), Some(0));
    }

    #[test]
    fn source_date_epoch_like_luatex() {
        // Values observed with LuaTeX 1.22 (\time, \day, … and the banner).
        let cases: &[(&str, Option<i64>)] = &[
            ("", None),
            ("0", Some(0)),
            ("abc", Some(0)),
            ("-5", None),
            (" 5", Some(5)),
            ("5x", Some(0)),
            ("5 ", Some(0)),
            ("  ", Some(0)),
            ("+7", Some(7)),
            ("-0", Some(0)),
            ("0x10", Some(0)),
            ("99999999999", Some(1_215_752_191)),
            ("18446744073709551616", Some(0)),
            ("4294967295", None),
            ("2147483648", None),
            ("2147483647", Some(2_147_483_647)),
            ("4294967296", Some(0)),
            ("4294967301", Some(5)),
            ("-4294967291", Some(5)),
        ];
        for &(v, want) in cases {
            assert_eq!(parse_source_date_epoch(v), want, "{v:?}");
        }
    }

    #[test]
    fn dates_like_luatex() {
        let d = |unix, offset| StartTime { unix, offset }.pdf_date();
        assert_eq!(d(0, 0), "D:19700101000000Z");
        assert_eq!(d(0, 3600), "D:19700101010000+01'00'");
        assert_eq!(d(0, -18000), "D:19691231190000-05'00'");
        assert_eq!(d(0, 19800), "D:19700101053000+05'30'");
        assert_eq!(d(1_790_000_000, 7200), "D:20260921161320+02'00'");
        assert_eq!(d(1_790_000_000, 0), "D:20260921141320Z");
        let c = StartTime {
            unix: 1_790_000_000,
            offset: 7200,
        }
        .civil();
        assert_eq!(
            (c.year, c.month, c.day, c.hour, c.minute, c.second),
            (2026, 9, 21, 16, 13, 20)
        );
        assert_eq!(
            StartTime::new(Some(5), true),
            StartTime { unix: 5, offset: 0 }
        );
    }

    #[test]
    fn garbage_is_not_a_panic() {
        assert_eq!(tzif_offset(b"TZif2", 0), None);
        assert_eq!(tzif_offset(b"nope", 0), None);
        let mut d = b"TZif2".to_vec();
        d.resize(44, 0);
        d[39] = 1; // typecnt = 1 but no data
        assert_eq!(tzif_offset(&d, 0), None);
    }
}