path: root/support/pkgcheck/src/utils.rs

use log::*;

use std::fs::read_dir;
use std::fs::read_link;
use std::os::unix::fs::PermissionsExt;
use std::path::Path;
use std::path::PathBuf;
use std::time::Duration;
use walkdir::DirEntry;

use once_cell::sync::Lazy; // 1.3.1
use regex::Regex;
use std::borrow::Cow;
use std::fs::File;
use std::io;
use std::process::Command;
use std::sync::atomic::Ordering;
use zip::result::ZipResult;

use std::fs;

#[derive(Debug, Clone, PartialEq, Eq)]
pub enum CheckType {
    Tds,
    Package,
}

pub struct Utils {
    kind: CheckType,
}

impl Utils {
    pub fn new(kind: CheckType) -> Utils {
        Utils { kind }
    }

    pub fn check_filesize(&self, fsize: u64, dir_entry_str: &str) -> bool {
        // if it is an empty file we don't need
        // to check further
        if fsize == 0 {
            e0005!(dir_entry_str);
            return false;
        }
        if fsize > 40 * 1024 * 1024 {
            if self.kind == CheckType::Package {
                w0005!(dir_entry_str, fsize / 1024 / 1024);
            } else {
                w0006!(dir_entry_str, fsize / 1024 / 1024);
            }
            return false;
        }

        true
    }

    // we check for the "good" characters, all other characters yield an
    // error message
    pub fn filename_has_bad_chars(&self, entry: &DirEntry, dir_entry_str: &str) {
        let s = entry.file_name().to_str().unwrap();
        for (i, c) in s.char_indices() {
            match c {
                'A'..='Z'
                | '0'..='9'
                | 'a'..='z'
                | '+'
                | ','
                | '-'
                | '.'
                | '/'
                | ':'
                | '='
                | '@'
                | '_' => (),
                _ => {
                    e0001!(c, dir_entry_str, i);
                }
            }
        }
    }

    pub fn is_unwanted_directory(&self, entry: &str, dir_entry_str: &str) {
        let names = ["__MACOSX"];
        for n in &names {
            if *n == entry {
                e0018!(dir_entry_str);
            };
        }
    }

    pub fn check_for_hidden_file(&self, entry: &str, dir_entry_str: &str) {
        if entry.starts_with('.') {
            match self.kind {
                CheckType::Package => {
                    e0007!(dir_entry_str);
                }
                CheckType::Tds => {
                    e0007t!(dir_entry_str);
                }
            }
        }
    }

    // fname is the base name without the directory part
    pub fn check_for_hidden_directory(&self, fname: &str, dir_entry_str: &str) {
        // We have to single out '.' and './' as they denote just the current directory
        // That case may happen if pkgcheck gets called like: `pkgcheck -d .` or `pkgcheck -d ./`
        if fname == "." || fname == "./" {
            return;
        };

        //if fname.map(|s| s.starts_with('.')).unwrap_or(false) {
        if fname.starts_with('.') {
            match self.kind {
                CheckType::Package => {
                    e0006!(dir_entry_str);
                }
                CheckType::Tds => {
                    e0006t!(dir_entry_str);
                }
            }
        }
    }

    pub fn check_for_temporary_file(&self, dir_entry_str: &str) {
        static RE: Lazy<Regex> = Lazy::new(regex_temporary_file_endings);

        if RE.is_match(dir_entry_str) {
            match self.kind {
                CheckType::Package => {
                    e0008!(dir_entry_str);
                }
                CheckType::Tds => {
                    e0008t!(dir_entry_str);
                }
            }
        }
    }

    pub fn check_for_empty_directory(&self, dir_entry_str: &str, dir_entry_display: &str) {
        match read_dir(dir_entry_str) {
            // try to read the directory specified
            Ok(contents) => {
                if contents.count() == 0 {
                    match self.kind {
                        CheckType::Package => {
                            e0004!(dir_entry_display);
                        }
                        CheckType::Tds => {
                            // Note: Karl Berry recommended to issue a warning only
                            //       as an empty directory in a TDS zip archive
                            //       - is automatically deleted before including
                            //         it in texlive
                            w0007!(dir_entry_display);
                        }
                    }
                }
            }
            Err(e) => {
                e0027!(dir_entry_display, e);
            }
        }
    }

    // just copied from zip-extensions-rs
    pub fn zip_extract(archive_file: &Path, target_dir: &Path) -> ZipResult<()> {
        let file = File::open(archive_file)?;
        let mut archive = zip::ZipArchive::new(file)?;
        archive.extract(target_dir)
    }

    pub fn unzip(&self, zip_archive: &str, out_dir: &str) -> ZipResult<()> {
        let archive_file = PathBuf::from(zip_archive);
        let target_dir = PathBuf::from(out_dir);

        Utils::zip_extract(&archive_file, &target_dir)
    }
}

pub fn temp_file_endings() -> Vec<(String, String)> {
    // https://github.com/github/gitignore/blob/master/TeX.gitignore
    // http://hopf.math.purdue.edu/doc/html/suffixes.html
    let v = vec![
        ("-blx.aux", "bibliography auxiliary file"),
        ("-blx.bib", "bibliography auxiliary file"),
        (".4ct", "htlatex related"),
        (".4tc", "htlatex related"),
        (".DS_Store", "Mac OS custom attribute file"),
        (".acn", "glossaries related"),
        (".acr", "glossaries related"),
        (".alg", ""),
        (".aux", "core latex auxiliary file"),
        (".backup", "backup file"),
        (".bak", "backup file"),
        (".bbl", "bibliography auxiliary file"),
        (".bcf", "bibliography auxiliary file"),
        (".blg", "bibliography log file"),
        (".brf", "hyperref related"),
        (".cb", "core latex auxiliary file"),
        (".cb2", "core latex auxiliary file"),
        (".cpt", "cprotect related"),
        (".dvi", "intermediate document"),
        (".ent", ""),
        (".fdb_latexmk", "latexmk related"),
        (".fff", "endfloat related"),
        (".fls", ""),
        (".fmt", "core latex auxiliary file"),
        (".fot", "core latex auxiliary file"),
        (".gaux", "generated by gregoriotex"),
        (".glg", "glossary related"),
        (".glo", "glossary related"),
        (".glog", "generated by gregoriotex"),
        (".gls", "glossary related"),
        (".glsdefs", "glossaries related"),
        (".gtex", "generated by gregoriotex"),
        (".hd", ""),
        (".idv", "htlatex related"),
        (".idx", "makeidx related"),
        (".ilg", "makeidx related"),
        (".ind", "makeidx related"),
        (".lg", "htlatex related"),
        (".loa", "core latex auxiliary file (list of algorithms)"),
        (".lod", "generated by easy-todo"),
        (".lof", "core latex auxiliary file (list of figures)"),
        (".log", "a log file for any flavor of TeX"),
        (".lol", "core latex auxiliary file (list of listings)"),
        (".los", "list of slides"),
        (".lot", "core latex auxiliary file"),
        (".lox", ""),
        (".lyx#", "LyX related autosave file"),
        (".maf", "generated by minitoc"),
        (".mlc", "generated by minitoc"),
        (".mlf", "generated by minitoc"),
        (".mlt", "generated by minitoc"),
        (".nav", "beamer related"),
        (".nlg", ""),
        (".nlo", ""),
        (".nls", ""),
        (".o", "C object file"),
        (".out", "Core latex auxiliary file"),
        (".pdfsync", "pdfsync related"),
        (".pre", "beamer related"),
        (".pyg", ""),
        (".run.xml", ""),
        (".sav", "used for saved data"),
        (".snm", "beamer related"),
        (".soc", ""),
        (".spl", "elsarticle related"),
        (".sta", "generated by standalone package"),
        (".swp", "vim swap file"),
        (".synctex", "synctex related"),
        (".synctex(busy)", "synctex related"),
        (".synctex.gz", "synctex related"),
        (".synctex.gz(busy)", "synctex related"),
        (".tpt", ""),
        (".tdo", "generated by todonotes (list of todos)"),
        (".thm", "amsthm related"),
        (".tmb", "generated by thumbs package"),
        (".tmp", "indicates a temporary file"),
        (".toc", "core latex auxiliary file (table of contents)"),
        (".trc", "htlatex related"),
        (".ttt", "endfloat related"),
        (".tuc", ""),
        (".upa", "generated by the soulpos package"),
        (".upb", "generated by the soulpos package"),
        (".url", "generated by jurabib"),
        (".vrb", "beamer related"),
        (".w18", "temporary file for the ifplatform package"),
        (".xdv", "intermediate document"),
        (".xref", "htlatex related"),
        (".xray", "dump of \\show output"),
        ("~", "a file name ending with ~ (tilde) is temporary anyway"),
        ("Thumbs.db", "thumbnails file in Windows"),
    ];

    v.into_iter()
        .map(|(i, j)| (i.to_string(), j.to_string()))
        .collect()
}

pub fn regex_temporary_file_endings() -> Regex {
    let mut rv = String::new();
    let mut first_time = true;
    for (p, _) in temp_file_endings() {
        if first_time {
            rv.push('(');
            first_time = false;
        } else {
            rv.push('|');
        }
        let px = str::replace(&p, ".", "\\.");

        rv.push_str(&px);
    }
    rv.push_str(")$");
    Regex::new(&rv).unwrap()
}

pub fn get_symlink(entry: &DirEntry) -> Result<Option<PathBuf>, io::Error> {
    let r = entry.path().to_str().unwrap();

    match read_link(r) {
        Ok(o) => {
            let full_path = if o.is_absolute() {
                o
            } else {
                // make the relative path absolute
                let p = entry.path().parent().unwrap();
                p.join(&o)
            };
            if full_path.exists() {
                Ok(Some(full_path))
            } else {
                Ok(None)
            }
        }
        Err(e) => Err(e),
    }
}

pub fn _is_symlink_broken(entry: &DirEntry) -> Result<bool, io::Error> {
    let r = entry.path().to_str().unwrap();

    match read_link(r) {
        Ok(o) => {
            let full_path = if o.is_absolute() {
                o
            } else {
                // make the relative path absolute
                let p = entry.path().parent().unwrap();
                p.join(&o)
            };
            Ok(!full_path.exists())
        }
        Err(e) => Err(e),
    }
}

// Runs `pdfinfo` to check a PDF document. If `pdfinfo`
// returns non zero we assume that the PDF document is
// corrupted.
pub fn is_pdf_ok(fname: &str) -> CmdReturn {
    run_cmd("pdfinfo", &[fname])
}

pub fn get_perms(path: &Path) -> u32 {
    match path.metadata() {
        Ok(p) => p.permissions().mode(),
        Err(_e) => 0,
    }
}

pub fn others_match(p: u32, m: u32) -> bool {
    p == m
}

#[test]
fn test_others_have() {
    assert_eq!(others_have(0o600, 4), false);
    assert_eq!(others_have(0o601, 4), false);
    assert_eq!(others_have(0o602, 4), false);
    assert_eq!(others_have(0o603, 4), false);
    assert_eq!(others_have(0o604, 4), true);
    assert_eq!(others_have(0o605, 4), true);
    assert_eq!(others_have(0o606, 4), true);
    assert_eq!(others_have(0o607, 4), true);
}

// It checks if a permission `p` has the bits
// given in `m` set for others
pub fn others_have(p: u32, m: u32) -> bool {
    let p1 = p & 0o0007;
    p1 & m == m
}

#[test]
fn test_owner_has() {
    assert_eq!(owner_has(0o000, 4), false);
    assert_eq!(owner_has(0o100, 4), false);
    assert_eq!(owner_has(0o200, 4), false);
    assert_eq!(owner_has(0o300, 4), false);
    assert_eq!(owner_has(0o400, 4), true);
    assert_eq!(owner_has(0o505, 4), true);
    assert_eq!(owner_has(0o605, 4), true);
    assert_eq!(owner_has(0o705, 4), true);
}

// It checks if a permission `p` has the bits
// given in `m` set for the owner.
pub fn owner_has(p: u32, m: u32) -> bool {
    let p1 = p & 0o7777;
    let m1 = m << 6;
    p1 & m1 == m1
}

#[allow(dead_code)]
fn owner_match(p: u32, m: u32) -> bool {
    let p1 = p & 0o0700;

    let m1 = m << 6;
    p1 == m1
}

// Formats a permission value to octal for output
pub fn perms_to_string(p: u32) -> Cow<'static, str> {
    format!("{:04o}", p & 0o7777).into()
}

pub struct CmdReturn {
    pub status: bool,
    pub output: Option<String>,
}

// Runs a command in a shell
// If the command returns 0 stdout gets captured. Otherwise
// stderr gets captured and returned.
pub fn run_cmd(cmd: &str, argument: &[&str]) -> CmdReturn {
    let output = Command::new(cmd)
        .args(argument.iter())
        .output()
        .unwrap_or_else(|_| panic!("Failed to execute process `{}`", cmd));

    if output.status.success() {
        CmdReturn {
            status: true,
            output: Some(String::from_utf8_lossy(&output.stdout).to_string()),
        }
    } else {
        CmdReturn {
            status: false,
            output: Some(String::from_utf8_lossy(&output.stderr).to_string()),
        }
    }
}

// returns true if file is a directory and does exist
// returns false otherwise
pub fn exists_dir(file: &str) -> bool {
    match fs::metadata(file) {
        Ok(attr) => attr.is_dir(),
        Err(_) => false,
    }
}

// returns true if file is a regular file and does exist
// returns false otherwise
pub fn exists_file(file: &str) -> bool {
    match fs::metadata(file) {
        Ok(attr) => attr.is_file(),
        Err(_) => false,
    }
}

#[test]
fn test_dirname() {
    assert!(dirname("/etc/fstab") == Some("/etc"));
    assert!(dirname("/etc/") == Some("/etc"));
    assert!(dirname("/") == Some("/"));
}

#[allow(dead_code)]
pub fn dirname(entry: &str) -> Option<&str> {
    if let Some(stripped) = entry.strip_suffix('/') {
        if entry.len() == 1 {
            return Some(entry);
        }
        return Some(stripped);
    }

    let pos = entry.rfind('/');
    match pos {
        None => None,
        Some(pos) => Some(&entry[..pos]),
    }
}

#[test]
fn test_format_duration() {
    assert!(format_duration(&Duration::new(5, 0)) == String::from("5sec"));
    assert!(format_duration(&Duration::new(105, 0)) == String::from("1min 45sec"));
    assert!(format_duration(&Duration::new(3801, 0)) == String::from("1h 3min 21sec"));
    assert!(format_duration(&Duration::new(25449, 0)) == String::from("7h 4min 9sec"));
    assert!(format_duration(&Duration::new(108245, 0)) == String::from("1d 6h 4min 5sec"));
    assert!(format_duration(&Duration::new(0, 0)) == String::from("0sec"));
}

#[test]
fn test_filename() {
    assert!(filename("/etc/fstab") == Some("fstab"));
    assert!(filename("fstab") == Some("fstab"));
    assert!(filename("../pkgcheck.rs/testdirs/fira.tds.zip") == Some("fira.tds.zip"));
    assert!(filename("/etc/") == None);
    assert!(filename("/") == None);
}

// We return the right part of a path name if it does not end with a `/`
pub fn filename(entry: &str) -> Option<&str> {
    if entry.ends_with('/') {
        return None;
    }

    let pos = entry.rfind('/');
    match pos {
        None => Some(entry),
        Some(pos) => Some(&entry[pos + 1..]),
    }
}

// Found here: https://codereview.stackexchange.com/questions/98536/extracting-the-last-component-basename-of-a-filesystem-path
pub fn basename(path: &str) -> Cow<str> {
    let mut pieces = path.rsplitn(2, |c| c == '/' || c == '\\');
    match pieces.next() {
        Some(p) => p.into(),
        None => path.into(),
    }
}

pub fn format_duration(duration: &Duration) -> String {
    let seconds = duration.as_secs();
    let days = seconds / 86400;
    let hours = (seconds % 86400) / 3600;
    let minutes = (seconds % 3600) / 60;
    let seconds = seconds % 60;

    let mut result = String::new();
    if days > 0 {
        result.push_str(&format!("{}d ", days));
    }

    if hours > 0 {
        result.push_str(&format!("{}h ", hours));
    }

    if minutes > 0 {
        result.push_str(&format!("{}min ", minutes));
    }

    if seconds > 0 || result.is_empty() {
        result.push_str(&format!("{}sec", seconds));
    }

    return result.trim().to_string();
}