dates-le 0.2.2

//! The terminal surface.
//!
//! stdout is always protocol — one JSON report per line, one line per
//! file. stderr is always for the human, and is a projection of the same
//! reports rather than parallel prose.

use std::io::{Read, Write};
use std::path::PathBuf;
use std::process::ExitCode;

use crate::extract::{extended, resolve_format};
use crate::scan::{self, FileReport, ScanOptions};
use crate::walk::{self, WalkOptions};

const USAGE: &str = "usage: dates-le [options] <file|dir>...
       dates-le [options] --stdin [--format <format>]
       dates-le mcp
       dates-le --version | --help

Finds every date and timestamp in a tree and puts them where a person
can read them: ISO 8601 in every form it is written — extended, basic,
week and ordinal — RFC 2822, Unix epochs from seconds to nanoseconds,
US notations, log and syslog lines, Apache access logs, and the strings
inside date constructors that nothing else would recognise as dates.

Every file is read. A name that matches no format is scanned with the
patterns every format shares, so a .py, .go, .toml or .md file yields
its dates rather than being skipped.

Each one carries the instant it actually resolves to, so `2024-01-15`,
`1705276800` and `Mon, 15 Jan 2024` can be compared rather than read.

Options:
  --after <date>       keep dates at or after this instant
  --before <date>      keep dates strictly before this instant
  --sort               order by instant rather than by position
  --dedupe             collapse repeated dates to their first occurrence
  --iso                add each instant as a UTC ISO 8601 string
  --tz <zone>          resolve dates that carry no timezone in this
                       IANA zone, e.g. UTC or America/New_York, instead
                       of this machine's
  --format <format>    force a format instead of inferring it from the
                       file name; a name nothing recognises falls back
                       to the shared patterns rather than failing
  --year <year>        the year a syslog line is assumed to be in,
                       since the line does not carry one. Defaults to
                       this one, which makes that answer move
  --values             print only the dates, one per line, for piping
  --strict             exit 2 if any file could not be read, rather than
                       reporting it and carrying on
  --stdin              read one document from stdin
  --hidden             walk hidden files and directories too
  --no-ignore          walk files that .gitignore excludes

--after and --before accept anything this tool can read, so
`--after 2024-01-15` and `--after 'March 5, 2024'` both work.

A date with no timezone resolves against this machine's. Use --tz to
name one instead; the answer genuinely differs by zone, exactly as it
does for the code being read.

A file that is not text — a PNG, a zip — is not read and not reported;
it was never a candidate. It is counted in the summary so the coverage
is still stated. A file that IS text and could not be read, or is not
UTF-8, is named on stderr and carried in the report, and does not by
itself fail the run. --strict turns those back into a failure.

Exit codes follow grep: 0 dates found · 1 none found · 2 malformed
question. Finding none is an answer, not an error.";

/// Every flag the parser accepts. Held equal to the flags named in
/// USAGE by a test, and consulted at runtime so the list is what the
/// parser actually honours.
const FLAGS: [&str; 14] = [
    "--strict",
    "--tz",
    "--after",
    "--before",
    "--sort",
    "--dedupe",
    "--iso",
    "--format",
    "--year",
    "--values",
    "--stdin",
    "--hidden",
    "--no-ignore",
    "--help",
];

#[derive(Debug)]
struct Invocation {
    scan: ScanOptions,
    walk: WalkOptions,
    values: bool,
    strict: bool,
    stdin: bool,
    roots: Vec<PathBuf>,
}

pub fn run(arguments: &[String]) -> ExitCode {
    if arguments.first().map(String::as_str) == Some("mcp") {
        return crate::mcp::serve();
    }
    if arguments.iter().any(|argument| argument == "--version") {
        println!("dates-le {}", env!("CARGO_PKG_VERSION"));
        return ExitCode::SUCCESS;
    }
    if arguments.is_empty() || arguments.iter().any(|argument| argument == "--help") {
        println!("{USAGE}");
        return ExitCode::SUCCESS;
    }

    let invocation = match parse_arguments(arguments) {
        Ok(invocation) => invocation,
        Err(message) => return refuse(&message),
    };

    let scanned = match gather(&invocation) {
        Ok(scanned) => scanned,
        Err(message) => return refuse(&message),
    };

    report(&scanned, invocation.values);
    scan::exit_code(&scanned.reports, invocation.strict)
}

fn refuse(message: &str) -> ExitCode {
    eprintln!("dates-le: {message}");
    eprintln!("try `dates-le --help`");
    ExitCode::from(2)
}

/// `--tz` is honoured before anything else is read, because `--after`
/// and `--before` parse dates too and must land in the same zone as the
/// documents. Order on the command line should not change an answer.
fn apply_zone_first(arguments: &[String]) -> Result<(), String> {
    let Some(index) = arguments.iter().position(|argument| argument == "--tz") else {
        return Ok(());
    };
    let raw = arguments.get(index + 1).ok_or("--tz needs a value")?;
    let zone = crate::extract::time::zone_by_name(raw)
        .ok_or_else(|| format!("{raw:?} is not an IANA timezone name"))?;
    crate::extract::time::set_zone(Some(zone));
    Ok(())
}

fn parse_arguments(arguments: &[String]) -> Result<Invocation, String> {
    apply_zone_first(arguments)?;
    let mut invocation = Invocation {
        scan: ScanOptions::default(),
        walk: WalkOptions::default(),
        values: false,
        strict: false,
        stdin: false,
        roots: Vec::new(),
    };

    let mut index = 0;
    while index < arguments.len() {
        let argument = arguments[index].as_str();
        let value = |name: &str| -> Result<String, String> {
            arguments
                .get(index + 1)
                .cloned()
                .ok_or_else(|| format!("{name} needs a value"))
        };

        match argument {
            "--after" => {
                let raw = value("--after")?;
                invocation.scan.after = Some(instant(&raw)?);
                index += 1;
            }
            "--before" => {
                let raw = value("--before")?;
                invocation.scan.before = Some(instant(&raw)?);
                index += 1;
            }
            "--format" => {
                invocation.scan.format = Some(value("--format")?);
                index += 1;
            }
            // Already applied by `apply_zone_first`; skip its value.
            "--tz" => index += 1,
            "--year" => {
                let raw = value("--year")?;
                invocation.scan.year = raw
                    .parse()
                    .map_err(|_| format!("--year needs a year, not {raw:?}"))?;
                index += 1;
            }
            "--sort" => invocation.scan.sort = true,
            "--dedupe" => invocation.scan.dedupe = true,
            "--iso" => invocation.scan.iso = true,
            "--values" => invocation.values = true,
            "--strict" => invocation.strict = true,
            "--stdin" => invocation.stdin = true,
            "--hidden" => invocation.walk.hidden = true,
            "--no-ignore" => invocation.walk.respect_ignore = false,
            other if other.starts_with("--") => {
                return Err(format!(
                    "unknown option {other:?} — one of: {}",
                    FLAGS.join(", ")
                ));
            }
            path => invocation.roots.push(PathBuf::from(path)),
        }
        index += 1;
    }

    if !invocation.stdin && invocation.roots.is_empty() {
        return Err("name a file or directory, or pass --stdin".into());
    }
    Ok(invocation)
}

/// A boundary for `--after` / `--before`, read by the same parser that
/// reads the documents. A tool that could not read its own output as
/// input would be a strange one.
fn instant(raw: &str) -> Result<i64, String> {
    extended::instant(raw).ok_or_else(|| format!("{raw:?} is not a date this can read"))
}

/// What a run read: a report per text file, and a count of the files
/// that were never text candidates.
///
/// The count is carried rather than dropped because a walk that covered
/// less than the tree has to say so — but a PNG is not a file that
/// failed to be read, so it gets a number in the summary rather than a
/// line in the report and a hold over the exit code.
struct Scanned {
    reports: Vec<FileReport>,
    binary: usize,
}

fn gather(invocation: &Invocation) -> Result<Scanned, String> {
    if invocation.stdin {
        // No file name to infer from, so an unnamed format falls back —
        // the same answer a `.py` file gets, and no special case.
        let language = resolve_format(invocation.scan.format.as_deref(), None);
        let mut content = String::new();
        std::io::stdin()
            .read_to_string(&mut content)
            .map_err(|error| format!("could not read stdin: {error}"))?;
        return Ok(Scanned {
            reports: vec![scan::scan_text(
                "<stdin>",
                scan::without_bom(&content),
                language,
                &invocation.scan,
            )],
            binary: 0,
        });
    }

    for root in &invocation.roots {
        if !root.exists() {
            return Err(format!("{} does not exist", root.display()));
        }
    }
    let files = walk::collect(&invocation.roots, invocation.walk);
    let reports: Vec<FileReport> = files
        .iter()
        .filter_map(|path| scan::scan_file(path, &invocation.scan))
        .collect();
    Ok(Scanned {
        binary: files.len() - reports.len(),
        reports,
    })
}

fn report(scanned: &Scanned, values_only: bool) {
    let reports = &scanned.reports;
    let stdout = std::io::stdout();
    let mut out = stdout.lock();

    if values_only {
        for report in reports {
            for date in &report.dates {
                let _ = writeln!(out, "{}", date.value);
            }
        }
        return;
    }

    for report in reports {
        if let Ok(line) = serde_json::to_string(report) {
            let _ = writeln!(out, "{line}");
        }
    }

    let total: usize = reports.iter().map(|report| report.dates.len()).sum();
    let skipped = reports
        .iter()
        .filter(|report| report.skipped.is_some())
        .count();
    let files = reports.len() - skipped;
    eprintln!(
        "{total} date{} in {files} file{}{}{}",
        plural(total),
        plural(files),
        if skipped == 0 {
            String::new()
        } else {
            format!(", {skipped} skipped")
        },
        // Counted, never named: naming fourteen PNGs is the noise that
        // stops anyone reading the line that matters.
        if scanned.binary == 0 {
            String::new()
        } else {
            format!(
                ", {} binary file{} skipped",
                scanned.binary,
                plural(scanned.binary)
            )
        }
    );
    // Named, every one of them. A tool that quietly reads fewer files
    // than it was pointed at is worse than one that fails.
    for report in reports.iter().filter(|report| report.skipped.is_some()) {
        eprintln!(
            "  skipped {}: {}",
            report.file,
            report.skipped.as_deref().unwrap_or_default()
        );
    }
}

fn plural(count: usize) -> &'static str {
    if count == 1 { "" } else { "s" }
}

#[cfg(test)]
mod tests {
    use super::*;
    use crate::extract::SUPPORTED_FORMATS;
    use crate::extract::format::FALLBACK_FORMAT;

    fn parse(arguments: &[&str]) -> Result<Invocation, String> {
        parse_arguments(
            &arguments
                .iter()
                .map(|argument| (*argument).to_string())
                .collect::<Vec<_>>(),
        )
    }

    #[test]
    fn every_flag_the_parser_honours_is_documented() {
        for flag in FLAGS {
            assert!(USAGE.contains(flag), "{flag} is not in the usage text");
        }
    }

    #[test]
    fn every_flag_the_usage_names_is_honoured() {
        for word in USAGE.split_whitespace() {
            let flag = word.trim_end_matches([',', '.', '·']);
            if flag.starts_with("--") && flag.len() > 2 {
                assert!(
                    FLAGS.contains(&flag) || flag == "--version",
                    "{flag} is documented and not honoured"
                );
            }
        }
    }

    #[test]
    fn every_exit_code_is_documented() {
        for code in ["0", "1", "2"] {
            assert!(USAGE.contains(code), "exit code {code} is undocumented");
        }
    }

    #[test]
    fn a_boundary_is_read_by_the_same_parser_as_a_document() {
        let invocation = parse(&["--after", "2024-01-15", "."]).expect("parses");
        assert_eq!(invocation.scan.after, Some(1_705_276_800_000));
        let written = parse(&["--after", "March 5, 2024", "."]).expect("parses");
        assert!(written.scan.after.is_some());
    }

    #[test]
    fn a_boundary_that_is_not_a_date_is_refused() {
        let error = parse(&["--after", "soon", "."]).expect_err("refuses");
        assert!(error.contains("not a date"), "{error}");
    }

    /// A format nothing recognises is read with the shared patterns,
    /// the same as a file whose name says nothing. Refusing it was the
    /// tool declining to read most of a repository.
    #[test]
    fn an_unknown_format_falls_back_rather_than_failing() {
        let invocation = parse(&["--format", "rust", "."]).expect("parses");
        assert_eq!(invocation.scan.format.as_deref(), Some("rust"));
        assert_eq!(resolve_format(Some("rust"), None), FALLBACK_FORMAT);
    }

    #[test]
    fn every_advertised_format_is_accepted_by_name() {
        for name in SUPPORTED_FORMATS {
            assert!(parse(&["--format", name, "."]).is_ok(), "{name}");
        }
    }

    #[test]
    fn stdin_without_a_format_reads_the_document_anyway() {
        assert!(parse(&["--stdin"]).is_ok());
    }

    #[test]
    fn naming_nothing_at_all_is_refused() {
        let error = parse(&["--sort"]).expect_err("refuses");
        assert!(error.contains("file or directory"), "{error}");
    }

    #[test]
    fn an_unknown_option_is_refused_rather_than_read_as_a_path() {
        let error = parse(&["--nope", "."]).expect_err("refuses");
        assert!(error.contains("--nope"), "{error}");
    }

    #[test]
    fn a_flag_missing_its_value_says_which() {
        assert!(
            parse(&["--after"])
                .expect_err("refuses")
                .contains("--after")
        );
        assert!(parse(&["--year"]).expect_err("refuses").contains("--year"));
    }

    #[test]
    fn a_year_that_is_not_a_number_is_refused() {
        let error = parse(&["--year", "soon", "."]).expect_err("refuses");
        assert!(error.contains("--year"), "{error}");
    }

    #[test]
    fn a_named_zone_is_accepted_and_a_made_up_one_is_not() {
        assert!(parse(&["--tz", "UTC", "."]).is_ok());
        assert!(parse(&["--tz", "America/New_York", "."]).is_ok());
        let error = parse(&["--tz", "Mars/Olympus", "."]).expect_err("refuses");
        assert!(error.contains("IANA"), "{error}");
        crate::extract::time::set_zone(None);
    }

    /// The zone has to apply to the boundaries too, or `--after` would
    /// mean something different depending on where it was typed.
    #[test]
    fn the_zone_applies_before_a_boundary_is_read() {
        let utc = parse(&["--tz", "UTC", "--after", "2024-01-15 00:00:00", "."])
            .expect("parses")
            .scan
            .after;
        let east = parse(&[
            "--tz",
            "America/New_York",
            "--after",
            "2024-01-15 00:00:00",
            ".",
        ])
        .expect("parses")
        .scan
        .after;
        crate::extract::time::set_zone(None);
        assert_ne!(utc, east, "a zone-less boundary is not zone-independent");
        assert_eq!(east.unwrap() - utc.unwrap(), 5 * 3_600_000);
    }

    /// Written after `--tz` was read in argument order and quietly gave
    /// a different answer depending on which flag came first.
    #[test]
    fn the_zone_applies_wherever_it_appears() {
        let first = parse(&["--tz", "UTC", "--after", "2024-01-15 00:00:00", "."])
            .expect("parses")
            .scan
            .after;
        let last = parse(&["--after", "2024-01-15 00:00:00", "--tz", "UTC", "."])
            .expect("parses")
            .scan
            .after;
        crate::extract::time::set_zone(None);
        assert_eq!(first, last);
    }

    #[test]
    fn paths_accumulate() {
        let invocation = parse(&["a.json", "b.log"]).expect("parses");
        assert_eq!(invocation.roots.len(), 2);
    }

    #[test]
    fn the_walk_flags_invert_the_defaults() {
        let invocation = parse(&["--hidden", "--no-ignore", "."]).expect("parses");
        assert!(invocation.walk.hidden);
        assert!(!invocation.walk.respect_ignore);
    }

    #[test]
    fn the_default_year_is_this_one() {
        assert_eq!(
            ScanOptions::default().year,
            crate::extract::time::current_year()
        );
    }
}