asar_rust/

asar.rs

use crate::crawlfs::{crawl, FileMetadata, FileType};
use crate::disk::{
    read_archive_header_sync, read_file_sync, read_filesystem_sync, write_filesystem,
    ArchiveHeader, AsarError,
};
use crate::filesystem::{Filesystem, FilesystemEntry, ListOptions};
use crate::integrity::FileIntegrity;
use crate::path_validation::ensure_within;
use glob::Pattern;
use std::collections::HashMap;
use std::fs;
use std::io::{Read, Seek};
use std::path::{Path, PathBuf};
use std::sync::Arc;

/// Options for creating an ASAR archive.
///
/// Controls which files to include, file ordering,
/// and which files should be stored unpacked.
pub struct CreateOptions {
    /// Include dotfiles (files and directories starting with `.`).
    pub dot: bool,
    /// Path to a file specifying the ordering of files in the archive.
    pub ordering: Option<PathBuf>,
    /// Glob pattern for files to store unpacked alongside the archive.
    pub unpack: Option<String>,
    /// Glob pattern for directories whose contents should be stored unpacked.
    pub unpack_dir: Option<String>,
}

impl Default for CreateOptions {
    fn default() -> Self {
        CreateOptions {
            dot: true,
            ordering: None,
            unpack: None,
            unpack_dir: None,
        }
    }
}

/// An ASAR archive that can be read from or written to disk.
///
/// # Creating archives
///
/// ```no_run
/// use asar_rust::AsarArchive;
/// use std::path::Path;
///
/// let archive = AsarArchive::pack(Path::new("src"), Path::new("app.asar")).unwrap();
/// ```
///
/// # Reading archives
///
/// ```no_run
/// use asar_rust::AsarArchive;
/// use std::path::Path;
///
/// let archive = AsarArchive::open(Path::new("app.asar")).unwrap();
/// let files = archive.list().unwrap();
/// let data = archive.extract_file("main.js").unwrap();
/// ```
pub struct AsarArchive {
    filesystem: Arc<Filesystem>,
}

impl AsarArchive {
    /// Create a new ASAR archive from a directory.
    pub fn pack(src: &Path, dest: &Path) -> Result<Self, AsarError> {
        create_package(src, dest)?;
        Self::open(dest)
    }

    /// Create an ASAR archive with custom options.
    pub fn pack_with_options(src: &Path, dest: &Path, options: CreateOptions) -> Result<Self, AsarError> {
        create_package_with_options(src, dest, options)?;
        Self::open(dest)
    }

    /// Open an existing ASAR archive for reading.
    pub fn open(archive_path: &Path) -> Result<Self, AsarError> {
        let filesystem = read_filesystem_sync(archive_path)?;
        Ok(AsarArchive { filesystem })
    }

    /// Get the raw archive header.
    pub fn raw_header(&self) -> Result<ArchiveHeader, AsarError> {
        get_raw_header(self.filesystem.root_path())
    }

    /// List all files in the archive.
    pub fn list(&self) -> Result<Vec<String>, AsarError> {
        Ok(self.filesystem.list_files(None))
    }

    /// List files with pack/unpack state.
    pub fn list_with_flags(&self, opts: ListOptions) -> Result<Vec<String>, AsarError> {
        Ok(self.filesystem.list_files(Some(&opts)))
    }

    /// Get metadata for a file in the archive.
    pub fn stat(&self, filename: &str, follow_links: bool) -> Result<FilesystemEntry, AsarError> {
        self.filesystem.get_file(filename, follow_links).cloned()
    }

    /// Extract a single file's contents.
    pub fn extract_file(&self, filename: &str) -> Result<Vec<u8>, AsarError> {
        let info = self.filesystem.get_file(filename, true)?;
        match info {
            FilesystemEntry::File(file_entry) => read_file_sync(&self.filesystem, filename, file_entry),
            FilesystemEntry::Directory(_) => Err(AsarError::Other(format!("Expected file but found directory: {}", filename))),
            FilesystemEntry::Link(_) => Err(AsarError::Other(format!("Expected file but found link: {}", filename))),
        }
    }

    /// Extract all files from the archive to a destination directory.
    pub fn extract_all(&self, dest: &Path) -> Result<(), AsarError> {
        extract_all_from_fs(&self.filesystem, dest)
    }
}

/// Create a new ASAR archive from a directory using default options.
pub fn create_package(src: &Path, dest: &Path) -> Result<(), AsarError> {
    create_package_with_options(src, dest, CreateOptions::default())
}

/// Create a new ASAR archive from a directory with custom options.
pub fn create_package_with_options(
    src: &Path,
    dest: &Path,
    options: CreateOptions,
) -> Result<(), AsarError> {
    let pattern = format!("{}/**/*", src.display());
    let (filenames, metadata) = crawl(&pattern)?;
    let (filenames, metadata) = if options.dot {
        (filenames, metadata)
    } else {
        let filtered: Vec<_> = filenames.into_iter().filter(|p| {
            let name = p.file_name().and_then(|n| n.to_str()).unwrap_or("");
            !name.starts_with('.')
        }).collect();
        let filtered_set: std::collections::HashSet<_> = filtered.iter().cloned().collect();
        let filtered_meta: HashMap<_, _> = metadata.into_iter()
            .filter(|(p, _)| filtered_set.contains(p))
            .collect();
        (filtered, filtered_meta)
    };
    create_package_from_files(src, dest, &filenames, metadata, options)
}

fn canonicalize_stripped(path: &Path) -> PathBuf {
    let canonical = std::fs::canonicalize(path).unwrap_or_else(|_| path.to_path_buf());
    let s = canonical.to_string_lossy();
    if cfg!(windows) && s.starts_with(r"\\?\") {
        PathBuf::from(&s[4..])
    } else {
        canonical
    }
}

/// Create an ASAR archive from a pre-crawled list of files.
///
/// This is the lower-level entry point used by `create_package` and
/// `create_package_with_options`. It takes a pre-computed list of
/// filenames and their metadata.
pub fn create_package_from_files(
    src: &Path,
    dest: &Path,
    filenames: &[PathBuf],
    metadata: HashMap<PathBuf, FileMetadata>,
    options: CreateOptions,
) -> Result<(), AsarError> {
    let canonical_src = canonicalize_stripped(src);
    let dest = dest.to_path_buf();

    let mut filesystem = Filesystem::new(&canonical_src);
    let mut file_entries: Vec<(PathBuf, bool)> = Vec::new();

    let mut filenames_sorted: Vec<&PathBuf> = filenames.iter().collect();
    filenames_sorted.sort_unstable();

    if let Some(ref ordering_path) = options.ordering
        && let Ok(content) = fs::read_to_string(ordering_path)
    {
        let ordering_files: Vec<String> = content
            .lines()
            .map(|line| {
                let line = if line.contains(':') {
                    line.split(':').next_back().unwrap_or(line)
                } else {
                    line
                };
                line.trim().trim_start_matches('/').to_string()
            })
            .collect();

        let mut path_index: std::collections::HashMap<&str, Vec<&PathBuf>> = std::collections::HashMap::new();
        for filename in filenames {
            let fname = filename.to_str().unwrap_or("");
            path_index.entry(fname).or_default().push(filename);
        }

        let mut sorted = Vec::new();
        let mut seen: std::collections::HashSet<PathBuf> = std::collections::HashSet::new();

        for ordering_file in &ordering_files {
            if let Some(matches) = path_index.get(ordering_file.as_str()) {
                for filename in matches {
                    if seen.insert((*filename).clone()) {
                        sorted.push(*filename);
                    }
                }
            }
        }

        for filename in filenames {
            if seen.insert(filename.clone()) {
                sorted.push(filename);
            }
        }

        filenames_sorted = sorted;
    }

    for filename in &filenames_sorted {
        let file_meta = metadata.get(*filename);
        let file_type = file_meta.map(|m| m.file_type.clone());
        let abs_filename = canonicalize_stripped(filename);
        let archive_path = abs_filename.strip_prefix(&canonical_src).unwrap_or(&abs_filename);

        let should_unpack = {
            let fname = archive_path.to_str().unwrap_or("");
            let mut unpack = false;
            if let Some(ref unpack_pattern) = options.unpack {
                unpack = Pattern::new(unpack_pattern)
                    .map(|p| p.matches(fname))
                    .unwrap_or(false);
            }
            if !unpack
                && let Some(ref unpack_dir_pattern) = options.unpack_dir
            {
                unpack = Pattern::new(unpack_dir_pattern)
                    .map(|p| p.matches(fname))
                    .unwrap_or(false);
            }
            unpack
        };

        match file_type {
            Some(FileType::Directory) => {
                filesystem.insert_directory(archive_path, should_unpack)?;
            }
            Some(FileType::File) => {
                let size = file_meta.unwrap().size;
                let executable = false;
                let integrity = compute_file_integrity(&abs_filename);
                filesystem.insert_file(archive_path, size, executable, should_unpack, integrity)?;
                file_entries.push((abs_filename, should_unpack));
            }
            Some(FileType::Link) => {
                let link = fs::read_link(filename)
                    .map_err(AsarError::Io)?
                    .to_str()
                    .unwrap_or("")
                    .to_string();
                let resolved = resolve_link(&canonical_src, archive_path, &link);
                if Path::new(&resolved).is_absolute() || resolved.starts_with("..") {
                    return Err(AsarError::Other(format!(
                        "{}: file \"{}\" links out of the package",
                        filename.display(),
                        resolved
                    )));
                }
                filesystem.insert_link(archive_path, resolved, should_unpack)?;
                file_entries.push((abs_filename, should_unpack));
            }
            None => {
                return Err(AsarError::Other(format!(
                    "Unknown file type: {}",
                    filename.display()
                )));
            }
        }
    }

    write_filesystem(&dest, &filesystem, &file_entries, &metadata)
}

fn compute_file_integrity(path: &Path) -> Option<FileIntegrity> {
    if let Ok(mut file) = fs::File::open(path) {
        FileIntegrity::from_reader(&mut file).ok()
    } else {
        None
    }
}

fn resolve_link(src: &Path, parent_path: &Path, symlink: &str) -> String {
    let parent = parent_path.parent().unwrap_or(Path::new("."));
    let target = parent.join(symlink);
    target
        .strip_prefix(src)
        .unwrap_or(&target)
        .to_str()
        .unwrap_or("")
        .to_string()
}

/// Get metadata for a file in the archive.
pub fn stat_file(
    archive_path: &Path,
    filename: &str,
    follow_links: bool,
) -> Result<FilesystemEntry, AsarError> {
    let filesystem = read_filesystem_sync(archive_path)?;
    filesystem
        .get_file(filename, follow_links)
        .cloned()
}

/// Read the raw archive header without caching the filesystem.
pub fn get_raw_header(archive_path: &Path) -> Result<ArchiveHeader, AsarError> {
    read_archive_header_sync(archive_path)
}

/// List all files in the archive.
pub fn list_package(
    archive_path: &Path,
    options: Option<ListOptions>,
) -> Result<Vec<String>, AsarError> {
    let filesystem = read_filesystem_sync(archive_path)?;
    Ok(filesystem.list_files(options.as_ref()))
}

/// Extract a single file's contents from the archive.
pub fn extract_file(
    archive_path: &Path,
    filename: &str,
    follow_links: bool,
) -> Result<Vec<u8>, AsarError> {
    let filesystem = read_filesystem_sync(archive_path)?;
    let info = filesystem
        .get_file(filename, follow_links)?;

    match info {
        FilesystemEntry::File(file_entry) => read_file_sync(&filesystem, filename, file_entry),
        FilesystemEntry::Directory(_) => Err(AsarError::Other(format!(
            "Expected file but found directory: {}",
            filename
        ))),
        FilesystemEntry::Link(_) => Err(AsarError::Other(format!(
            "Expected file but found link: {}",
            filename
        ))),
    }
}

/// Extract all files from the archive to a destination directory.
pub fn extract_all(archive_path: &Path, dest: &Path) -> Result<(), AsarError> {
    let filesystem = read_filesystem_sync(archive_path)?;
    extract_all_from_fs(&filesystem, dest)
}

fn extract_all_from_fs(filesystem: &Arc<Filesystem>, dest: &Path) -> Result<(), AsarError> {
    let file_list = filesystem.list_files(None);

    fs::create_dir_all(dest)?;

    // Pre-read entire data section for efficient extraction
    let archive_path = filesystem.root_path();
    let header_size = filesystem.header_size() as u64;
    let archive_size = fs::metadata(archive_path)?.len();
    let data_start = 8 + header_size;
    let data_size = archive_size.saturating_sub(data_start);
    let mut data_buf = Vec::new();
    if data_size > 0 {
        let mut file = fs::File::open(archive_path)?;
        file.seek(std::io::SeekFrom::Start(data_start))?;
        let ds = usize::try_from(data_size).map_err(|_| AsarError::Other("data size overflow".into()))?;
        data_buf.resize(ds, 0);
        file.read_exact(&mut data_buf)?;
    }

    let mut extraction_errors: Vec<String> = Vec::new();

    for full_path in &file_list {
        let filename = &full_path[1..];
        let dest_filename = ensure_within(dest, filename)?;

        let file_entry = filesystem.get_file(filename, cfg!(windows))?;

        match file_entry {
            FilesystemEntry::Directory(_) => {
                fs::create_dir_all(&dest_filename)?;
            }
            FilesystemEntry::Link(link_entry) => {
                let link_path = dest.join(&link_entry.link);
                let dest_dir = dest_filename.parent().unwrap_or(Path::new("."));
                let _relative_path = pathdiff::diff_paths(&link_path, dest_dir)
                    .unwrap_or_else(|| PathBuf::from(&link_entry.link));
                let _ = fs::remove_file(&dest_filename);
                #[cfg(unix)]
                {
                    std::os::unix::fs::symlink(&_relative_path, &dest_filename)?;
                }
                #[cfg(windows)]
                {
                    if let Some(parent) = dest_filename.parent() {
                        fs::create_dir_all(parent)?;
                    }
                }
            }
            FilesystemEntry::File(file_info) => {
                let result: Result<Vec<u8>, AsarError> = if file_info.unpacked {
                    let unpacked_dir = format!("{}.unpacked", filesystem.root_path().display());
                        fs::read(ensure_within(Path::new(&unpacked_dir), filename)?)
                            .map_err(AsarError::Io)
                } else if file_info.size == 0 {
                    Ok(Vec::new())
                } else {
                    let offset: u64 = file_info.offset.parse().map_err(|_| AsarError::Other("Invalid offset".to_string()))?;
                    let start = usize::try_from(offset).map_err(|_| AsarError::Other("offset overflow".into()))?;
                    let end = start + usize::try_from(file_info.size).map_err(|_| AsarError::Other("size overflow".into()))?;
                    if end <= data_buf.len() {
                        Ok(data_buf[start..end].to_vec())
                    } else {
                        Err(AsarError::Other("Data out of bounds".to_string()))
                    }
                };

                match result {
                    Ok(content) => {
                        if let Some(parent) = dest_filename.parent() {
                            fs::create_dir_all(parent)?;
                        }
                        fs::write(&dest_filename, content)?;
                        #[cfg(unix)]
                        if file_info.executable {
                            use std::os::unix::fs::PermissionsExt;
                            if let Ok(meta) = fs::metadata(&dest_filename) {
                                let mut perms = meta.permissions();
                                perms.set_mode(0o755);
                                let _ = fs::set_permissions(&dest_filename, perms);
                            }
                        }
                    }
                    Err(e) => {
                        extraction_errors.push(format!("{}: {}", full_path, e));
                    }
                }
            }
        }
    }

    if !extraction_errors.is_empty() {
        return Err(AsarError::Other(format!(
            "Unable to extract some files:\n\n{}",
            extraction_errors.join("\n\n")
        )));
    }

    Ok(())
}