use std::io;
use std::path::Path;
use std::process::Command;
use std::str;
use std::str::Utf8Error;

use regex::Regex;

#[derive(Debug, PartialEq)]
pub struct Blob<'a> {
    pub object: &'a str,
    pub path: &'a Path,
}

pub fn crawl_git_tree<P: AsRef<Path>>(path: P) -> io::Result<Vec<u8>> {
    let output = Command::new("git")
        .current_dir(path)
        .args(&["ls-tree", "-zr", "HEAD"])
        .output()?;

    if !output.status.success() {
        return Err(io::Error::new(
            io::ErrorKind::Other,
            "git ls-tree did not run successfully",
        ));
    }

    Ok(output.stdout)
}

pub fn parse_ls_tree_output<'a>(output: &'a [u8]) -> Result<Vec<Blob<'a>>, Utf8Error> {
    let re = Regex::new(r"^[^ ]+ [^ ]+ ([^\t]+)\t(.+)$").unwrap();

    let mut blobs = Vec::new();
    for line in output.split(|&byte| byte == 0) {
        let line = str::from_utf8(line)?;

        if let Some(captures) = re.captures(line) {
            blobs.push(Blob {
                object: captures.get(1).unwrap().as_str(),
                path: Path::new(captures.get(2).unwrap().as_str()),
            })
        }
    }

    Ok(blobs)
}

#[cfg(test)]
mod tests {
    use super::*;
    use crate::git::{parse_ls_tree_output, Blob};

    const TEST_OUTPUT: &[u8] = b"100644 blob 424965736fa85c814e0ecb12d42ec81021304ce9\tREADME.md\0invalid\0100644 blob 5ec8f34bce9b5b53f3d809591914fa1601534a8a	src/main.rs\0";

    #[test]
    fn test_parse_ls_tree_output() {
        let expected = vec![
            Blob {
                object: "424965736fa85c814e0ecb12d42ec81021304ce9",
                path: Path::new("README.md"),
            },
            Blob {
                object: "5ec8f34bce9b5b53f3d809591914fa1601534a8a",
                path: Path::new("src/main.rs"),
            },
        ];

        assert_eq!(parse_ls_tree_output(TEST_OUTPUT).unwrap(), expected);
    }
}