fixed tests

- extension_from_path now returns Option<&OsStr>
- two new tests
- Findings uses a Path instead of a PathBuf, reducing allocations
- some unnecessary stuff removed, thanks clippy
- that is all

Cargo.lock

@@ -166,15 +166,25 @@ version = "1.1.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "de853764b47027c2e862a995c34978ffa63c1501f2e15f987ba11bd4f9bba193"
 
+[[package]]
+name = "fastrand"
+version = "1.4.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "ca5faf057445ce5c9d4329e382b2ce7ca38550ef3b73a5348362d5f24e0c7fe3"
+dependencies = [
+ "instant",
+]
+
 [[package]]
 name = "fif"
-version = "0.2.9"
+version = "0.2.10"
 dependencies = [
  "cached",
  "cfg-if",
  "clap",
  "env_logger",
  "exitcode",
+ "fastrand",
  "infer",
  "log",
  "mime_guess",
@@ -244,6 +254,15 @@ version = "0.3.6"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "0803735b9511d0956c68902a6513ca867819d6e43397adb6a5e903e2f09db734"
 
+[[package]]
+name = "instant"
+version = "0.1.9"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "61124eeebbd69b8190558df225adf7e4caafce0d743919e5d6b19652314ec5ec"
+dependencies = [
+ "cfg-if",
+]
+
 [[package]]
 name = "lazy_static"
 version = "1.4.0"
@@ -265,9 +284,9 @@ dependencies = [
 
 [[package]]
 name = "libc"
-version = "0.2.89"
+version = "0.2.91"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "538c092e5586f4cdd7dd8078c4a79220e3e168880218124dcbce860f0ea938c6"
+checksum = "8916b1f6ca17130ec6568feccee27c156ad12037880833a3b842a823236502e7"
 
 [[package]]
 name = "log"
@@ -650,9 +669,9 @@ checksum = "5fecdca9a5291cc2b8dcf7dc02453fee791a280f3743cb0905f8822ae463b3fe"
 
 [[package]]
 name = "walkdir"
-version = "2.3.1"
+version = "2.3.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "777182bc735b6424e1a57516d35ed72cb8019d85c8c9bf536dccb3445c1a2f7d"
+checksum = "808cf2735cd4b6866113f648b791c6adc5714537bc222d9347bb203386ffda56"
 dependencies = [
  "same-file",
  "winapi",

Cargo.toml

@@ -1,7 +1,7 @@
 [package]
 name = "fif"
 description = "A command-line tool for detecting and optionally correcting files with incorrect extensions."
-version = "0.2.9"
+version = "0.2.10"
 authors = ["Lynnesbian <lynne@bune.city>"]
 edition = "2018"
 license = "GPL-3.0-or-later"
@@ -58,6 +58,7 @@ default-features = false
 
 [dev-dependencies]
 tempfile = "3.2.0"
+fastrand = "1.4.0"
 
 [profile.release]
 lto = "thin"

clippy.sh

@@ -1,6 +1,6 @@
 #!/bin/bash
 fd -e rs -x touch {}
-cargo clippy -- \
+cargo clippy --tests -- \
  -W clippy::nursery \
  -W clippy::perf \
  -W clippy::pedantic \

src/findings.rs

@@ -1,4 +1,4 @@
-use std::path::PathBuf;
+use std::path::Path;
 
 use mime_guess::Mime;
 use smartstring::alias::String;
@@ -6,16 +6,16 @@ use smartstring::alias::String;
 use crate::inspectors::mime_extension_lookup;
 
 /// Information about a scanned file.
-pub struct Findings {
+pub struct Findings<'a> {
 	/// The location of the scanned file.
-	pub file: PathBuf, // TODO: replace with Path???? <'a> and all that
+	pub file: &'a Path,
 	/// Whether or not the file's extension is valid for its mimetype.
 	pub valid: bool,
 	/// The file's mimetype.
 	pub mime: Mime,
 }
 
-impl Findings {
+impl<'a> Findings<'a> {
 	pub fn recommended_extension(&self) -> Option<String> {
 		mime_extension_lookup(self.mime.clone()).map(|extensions| extensions[0].to_owned())
 	}

src/formats.rs

@@ -9,12 +9,13 @@ use snailquote::escape;
 
 use crate::scan_error::ScanError;
 use crate::{Findings, BACKEND};
+use std::ffi::OsStr;
 
 /// The current version of fif, as defined in Cargo.toml.
 const VERSION: Option<&'static str> = option_env!("CARGO_PKG_VERSION");
 
 #[doc(hidden)]
-type Entries<'a> = [Result<Findings, ScanError<'a>>];
+type Entries<'a> = [Result<Findings<'a>, ScanError<'a>>];
 
 enum Writable<'a> {
 	String(&'a str),
@@ -36,6 +37,12 @@ impl<'a> From<&'a Path> for Writable<'a> {
 	}
 }
 
+impl<'a> From<&'a OsStr> for Writable<'a> {
+	fn from(p: &'a OsStr) -> Writable<'a> {
+		Writable::Path(p.as_ref())
+	}
+}
+
 fn smart_write<W: Write>(f: &mut W, writeables: &[Writable]) -> io::Result<()> {
 	// ehhhh
 	for writeable in writeables {
@@ -47,13 +54,15 @@ fn smart_write<W: Write>(f: &mut W, writeables: &[Writable]) -> io::Result<()> {
 				if let Some(string) = path.to_str() {
 					write!(f, "{}", escape(string))?
 				} else {
-					write!(f, "'''")?;
+					write!(f, "'")?;
 					#[cfg(unix)]
 					f.write_all(&*path.as_os_str().as_bytes())?;
-					#[cfg(windows)]
+					// TODO: implement bonked strings for windows
-					write!(f, "{}", path.as_os_str().to_string_lossy())?; // TODO: implement bonked strings for windows
+					// something like:
 					// f.write_all(&*path.as_os_str().encode_wide().collect::<Vec<u16>>())?;
-					write!(f, "'''")?
+					#[cfg(windows)]
+					write!(f, "{}", path.as_os_str().to_string_lossy())?;
+					write!(f, "'")?
 				}
 			}
 		}
@@ -78,9 +87,9 @@ pub trait Format {
 			match entry {
 				Ok(finding) => {
 					if let Some(ext) = finding.recommended_extension() {
-						self.rename(f, &finding.file, &finding.file.with_extension(ext.as_str()))?
+						self.rename(f, finding.file, &finding.file.with_extension(ext.as_str()))?
 					} else {
-						self.no_known_extension(f, &finding.file)?
+						self.no_known_extension(f, finding.file)?
 					}
 				}
 
@@ -100,7 +109,6 @@ pub trait Format {
 	}
 }
 
-// TODO: maybe make a batch script version for windows
 /// Bourne-Shell compatible script.
 pub struct Script {}
 
@@ -158,3 +166,66 @@ impl Format for Script {
 		writeln!(f, "\necho 'Done.'")
 	}
 }
+
+/// PowerShell script.
+pub struct PowerShell {}
+
+impl Format for PowerShell {
+	fn new() -> Self {
+		Self {}
+	}
+
+	fn rename<W: Write>(&self, f: &mut W, from: &Path, to: &Path) -> io::Result<()> {
+		// unfortunately there doesn't seem to be an equivalent of sh's `mv -i` -- passing the '-Confirm' flag will prompt
+		// the user to confirm every single rename, and using Move-Item -Force will always overwrite without prompting.
+		// there doesn't seem to be a way to rename the file, prompting only if the target already exists.
+		smart_write(
+			f,
+			&[
+				"Rename-Item -Path ".into(),
+				from.into(),
+				" -NewName ".into(),
+				to.file_name().unwrap().into(),
+				Writable::Newline,
+			],
+		)
+	}
+
+	fn no_known_extension<W: Write>(&self, f: &mut W, path: &Path) -> io::Result<()> {
+		smart_write(
+			f,
+			&[
+				"Write-Output @'\nNo known extension for ".into(),
+				path.into(),
+				"\n'@".into(),
+			],
+		)
+	}
+
+	fn unreadable<W: Write>(&self, f: &mut W, path: &Path) -> io::Result<()> {
+		smart_write(
+			f,
+			&["Write-Output @'\nFailed to read ".into(), path.into(), "\n'@".into()],
+		)
+	}
+
+	fn unknown_type<W: Write>(&self, f: &mut W, path: &Path) -> io::Result<()> {
+		smart_write(
+			f,
+			&["<# Failed to detect mime type for ".into(), path.into(), "#>".into()],
+		)
+	}
+
+	fn header<W: Write>(&self, _: &Entries, f: &mut W) -> io::Result<()> {
+		writeln!(
+			f,
+			"#!/usr/bin/env pwsh\n# Generated by fif {} ({} backend)",
+			VERSION.unwrap_or("???"),
+			BACKEND
+		)
+	}
+
+	fn footer<W: Write>(&self, _: &Entries, f: &mut W) -> io::Result<()> {
+		writeln!(f, "\nWrite-Output 'Done!'")
+	}
+}

src/inspectors.rs

@@ -17,10 +17,10 @@ use crate::mime_db::MimeDb;
 /// Rather than reading the entire file all at once into a [`BUF_SIZE`] buffer, it tends to be faster to read a small
 /// chunk of the file and trying to identify that, proceeding with the larger buffer if that fails. Many file formats
 /// can be identified with the first few dozen bytes, so the "happy path" will likely be taken in the majority of cases.
-const INITIAL_BUF_SIZE: usize = 128;
+pub const INITIAL_BUF_SIZE: usize = 128;
 
 /// The number of bytes to read if the file couldn't be identified from its first [`INITIAL_BUF_SIZE`] bytes.
-const BUF_SIZE: usize = 4096;
+pub const BUF_SIZE: usize = 4096;
 
 /// Tries to identify the mimetype of a file from a given path.
 pub fn mime_type<T: MimeDb>(db: &T, path: &Path) -> io::Result<Option<Mime>> {

src/main.rs

@@ -15,7 +15,7 @@
 // along with this program.  If not, see <https://www.gnu.org/licenses/>.
 
 use std::io::{stdout, BufWriter};
-use std::path::{Path, PathBuf};
+use std::path::Path;
 
 use cfg_if::cfg_if;
 use clap::Clap;
@@ -23,15 +23,15 @@ use log::{debug, error, info, trace, warn};
 use once_cell::sync::OnceCell;
 #[cfg(feature = "multi-threaded")]
 use rayon::prelude::*;
-use smartstring::alias::String;
 use walkdir::{DirEntry, WalkDir};
 
 use crate::findings::Findings;
-use crate::formats::{Format, Script};
+use crate::formats::{Format, PowerShell, Script};
 use crate::mime_db::MimeDb;
 use crate::parameters::{OutputFormat, ScanOpts};
 use crate::scan_error::ScanError;
 use env_logger::Env;
+use std::ffi::OsStr;
 use std::process::exit;
 
 mod extension_set;
@@ -125,16 +125,18 @@ fn main() {
 		exit(0);
 	}
 
-	match args.output_format {
+	let mut buffered_stdout = BufWriter::new(stdout());
-		OutputFormat::Script => {
+
-			let s = Script::new();
+	let result = match args.output_format {
-			if s.write_all(&results, &mut BufWriter::new(stdout().lock())).is_err() {
+		OutputFormat::Script => Script::new().write_all(&results, &mut buffered_stdout),
+		OutputFormat::PowerShell | OutputFormat::Powershell => PowerShell::new().write_all(&results, &mut buffered_stdout),
+		OutputFormat::Text => todo!(),
+	};
+
+	if result.is_err() {
 		error!("Failed to write to stdout.");
 		exit(exitcode::IOERR);
 	}
-		}
-		OutputFormat::Text => todo!(),
-	}
 
 	debug!("Done");
 }
@@ -186,17 +188,12 @@ fn wanted_file(entry: &DirEntry, exts: &[&str], scan_opts: &ScanOpts) -> bool {
 		return false;
 	}
 
-	exts.contains(&ext.unwrap().to_lowercase().as_str())
+	exts.contains(&ext.unwrap().to_string_lossy().to_lowercase().as_str())
 }
 
 /// Given a file path, returns its extension, using [`std::path::Path::extension`].
-///
+fn extension_from_path(path: &Path) -> Option<&OsStr> {
-/// The extension is currently [converted to a lossy string](std::ffi::OsStr::to_string_lossy), although it will
+	path.extension()
-/// (eventually) in future return an `OsStr` instead.
-// TODO: ↑
-fn extension_from_path(path: &Path) -> Option<String> {
-	path.extension(). // Get the path's extension
-		map(|e| String::from(e.to_string_lossy())) // Convert from OsStr to String
 }
 
 /// Inspects the given entry, returning a [`Findings`] on success and a [`ScanError`] on failure.
@@ -230,13 +227,13 @@ fn scan_file(entry: &DirEntry) -> Result<Findings, ScanError> {
 
 	let valid = match known_exts {
 		// there is a known set of extensions for this mimetype, and the file has an extension
-		Some(e) if entry_ext.is_some() => e.contains(&entry_ext.unwrap().to_lowercase().into()),
+		Some(e) if entry_ext.is_some() => e.contains(&entry_ext.unwrap().to_string_lossy().to_lowercase().into()),
 		// either this file has no extension, or there is no known set of extensions for this mimetype :(
 		Some(_) | None => false,
 	};
 
 	Ok(Findings {
-		file: entry.path().to_path_buf(),
+		file: entry.path(),
 		valid,
 		mime: result,
 	})
@@ -266,7 +263,7 @@ fn scan_from_walkdir(entries: &[DirEntry]) -> Vec<Result<Findings, ScanError>> {
 
 /// Scans a given directory with [`WalkDir`], filters with [`wanted_file`], checks for errors, and returns a vector of
 /// [DirEntry]s.
-fn scan_directory(dirs: &PathBuf, exts: &[&str], scan_opts: &ScanOpts) -> Option<Vec<DirEntry>> {
+fn scan_directory(dirs: &Path, exts: &[&str], scan_opts: &ScanOpts) -> Option<Vec<DirEntry>> {
 	let stepper = WalkDir::new(dirs).into_iter();
 	let mut probably_fatal_error = false;
 	let entries: Vec<DirEntry> = stepper

src/parameters.rs

@@ -3,13 +3,26 @@
 use std::path::PathBuf;
 
 use crate::extension_set::ExtensionSet;
+use cfg_if::cfg_if;
 use clap::{AppSettings, Clap};
 use smartstring::{LazyCompact, SmartString};
 
+cfg_if! {
+	if #[cfg(windows)] {
+		const DEFAULT_FORMAT: &str = "powershell";
+	} else {
+		const DEFAULT_FORMAT: &str = "script";
+	}
+}
+
 #[derive(Clap, PartialEq, Debug)]
 pub enum OutputFormat {
 	/// A Bourne shell compatible script.
 	Script,
+	/// A PowerShell script.
+	PowerShell,
+	/// Also a PowerShell script, with different casing to allow for `fif -o powershell`.
+	Powershell,
 	/// Plain text.
 	Text,
 }
@@ -52,7 +65,7 @@ pub struct Parameters {
 	pub scan_extensionless: bool,
 
 	/// Output format to use
-	#[clap(short, long, default_value = "script", arg_enum)]
+	#[clap(short, long, default_value = DEFAULT_FORMAT, arg_enum)]
 	pub output_format: OutputFormat,
 
 	/// Directory to process

src/tests/mod.rs

@@ -1,13 +1,15 @@
-use crate::inspectors::mime_extension_lookup;
+use crate::inspectors::{mime_extension_lookup, BUF_SIZE};
-use crate::mime_db::*;
 use crate::{extension_from_path, init_db, scan_directory, scan_from_walkdir};
 
 use crate::parameters::{Parameters, ScanOpts};
+use crate::mime_db::MimeDb;
 use cfg_if::cfg_if;
 use mime_guess::mime::{APPLICATION_OCTET_STREAM, APPLICATION_PDF, IMAGE_JPEG, IMAGE_PNG};
 use mime_guess::Mime;
 use smartstring::alias::String;
+use std::borrow::Borrow;
 use std::collections::HashMap;
+use std::ffi::OsStr;
 use std::path::Path;
 
 const JPEG_BYTES: &[u8] = b"\xFF\xD8\xFF";
@@ -17,12 +19,12 @@ const ZIP_BYTES: &[u8] = b"PK\x03\x04";
 
 cfg_if! {
 	if #[cfg(any(all(unix, feature = "infer-backend"), all(not(unix), not(feature = "xdg-mime-backend"))))] {
-		fn get_mime_db() -> InferDb {
+		fn get_mime_db() -> crate::mime_db::InferDb {
-			InferDb::init()
+			crate::mime_db::InferDb::init()
 		}
 	} else {
-		fn get_mime_db() -> XdgDb {
+		fn get_mime_db() -> crate::mime_db::XdgDb {
-			XdgDb::init()
+			crate::mime_db::XdgDb::init()
 		}
 	}
 }
@@ -34,11 +36,11 @@ fn application_zip() -> Mime {
 
 #[test]
 fn get_ext() {
-	let mut ext_checks = HashMap::new();
+	let mut ext_checks: HashMap<_, Option<&OsStr>> = HashMap::new();
-	ext_checks.insert(Path::new("test.txt"), Some(String::from("txt")));
+	ext_checks.insert(Path::new("test.txt"), Some(OsStr::new("txt")));
-	ext_checks.insert(Path::new("test.zip"), Some(String::from("zip")));
+	ext_checks.insert(Path::new("test.zip"), Some(OsStr::new("zip")));
-	ext_checks.insert(Path::new("test.tar.gz"), Some(String::from("gz")));
+	ext_checks.insert(Path::new("test.tar.gz"), Some(OsStr::new("gz")));
-	ext_checks.insert(Path::new("test."), Some(String::from("")));
+	ext_checks.insert(Path::new("test."), Some(OsStr::new("")));
 	ext_checks.insert(Path::new("test"), None);
 	ext_checks.insert(Path::new(".hidden"), None);
 
@@ -102,7 +104,7 @@ fn simple_directory() {
 
 	let entries = scan_directory(
 		&dir.path().to_path_buf(),
-		&vec!["jpg", "jpeg", "png", "pdf", "zip"],
+		&["jpg", "jpeg", "png", "pdf", "zip"],
 		&scan_opts,
 	)
 	.expect("Directory scan failed.");
@@ -119,30 +121,32 @@ fn simple_directory() {
 		if !result.valid {
 			// this should be "wrong.jpg", which is a misnamed png file
 			// 1. ensure extension is "png"
-			assert_eq!(extension_from_path(&*result.file).unwrap(), String::from("jpg"));
+			assert_eq!(extension_from_path(&*result.file).unwrap(), OsStr::new("jpg"));
 			// 2. ensure mime type detected is IMAGE_PNG
 			assert_eq!(result.mime, IMAGE_PNG);
 			// 3. ensure recommended extension is in the list of known extensions for PNG files
 			assert!(mime_extension_lookup(IMAGE_PNG)
 				.unwrap()
-				.contains(&result.recommended_extension().unwrap().into()));
+				.contains(&result.recommended_extension().unwrap()));
 			continue;
 		}
 
 		// check if the recommended extension for this file is in the list of known extensions for its mimetype
 		assert!(mime_extension_lookup(result.mime.clone())
 			.unwrap()
-			.contains(&result.recommended_extension().unwrap().into()));
+			.contains(&result.recommended_extension().unwrap()));
 
 		// make sure the guessed mimetype is correct based on the extension of the scanned file
+		let ext = extension_from_path(result.file);
+		assert!(ext.is_some());
 		assert_eq!(
 			result.mime,
-			match extension_from_path(&*result.file).as_deref() {
+			match ext.unwrap().to_string_lossy().borrow() {
-				Some("jpg") | Some("jpeg") => IMAGE_JPEG,
+				"jpg" | "jpeg" => IMAGE_JPEG,
-				Some("png") => IMAGE_PNG,
+				"png" => IMAGE_PNG,
-				Some("pdf") => APPLICATION_PDF,
+				"pdf" => APPLICATION_PDF,
-				Some("zip") => application_zip(),
+				"zip" => application_zip(),
-				Some(_) | None => APPLICATION_OCTET_STREAM, // general "fallback" type
+				_ => APPLICATION_OCTET_STREAM, // general "fallback" type
 			}
 		);
 	}
@@ -162,3 +166,29 @@ fn argument_parsing() {
 	// exts should be none
 	assert!(args.exts.is_none());
 }
+
+#[test]
+fn rejects_bad_args() {
+	use clap::Clap;
+	assert!(Parameters::try_parse_from(vec!["fif", "-abcdefg", "-E", "-e"]).is_err());
+}
+
+#[test]
+fn identify_random_bytes() {
+	let db = get_mime_db();
+	let rng = fastrand::Rng::new();
+	let mut bytes: Vec<u8>;
+	let mut results: HashMap<Mime, i32> = HashMap::new();
+
+	for _ in 1..500 {
+		bytes = std::iter::repeat_with(|| rng.u8(..)).take(BUF_SIZE * 2).collect();
+		if let Some(detected_type) = db.get_type(&*bytes) {
+			*results.entry(detected_type).or_insert(0) += 1;
+		}
+	}
+
+	for (mime, count) in &results {
+		println!("{}:\t{} counts", mime, count);
+	}
+	println!("No type found:\t{} counts", 500 - results.values().sum::<i32>())
+}