added text extension set, more test coverage, cargo update

CHANGELOG.md

@@ -7,6 +7,8 @@ Dates are given in YYYY-MM-DD format.
 - Better documentation for command line arguments
 - Added more stuff to test.py
 - PKGBUILD for Arch-based distros
+- Added Text extension set
+- More test coverage
 
 ### v0.2.11 (2021-04-04)
 #### Features

Cargo.lock

@@ -284,9 +284,9 @@ dependencies = [
 
 [[package]]
 name = "libc"
-version = "0.2.92"
+version = "0.2.93"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "56d855069fafbb9b344c0f962150cd2c1187975cb1c22c1522c240d8c4986714"
+checksum = "9385f66bf6105b241aa65a61cb923ef20efc665cb9f9bb50ac2f0c4b7f378d41"
 
 [[package]]
 name = "log"
@@ -549,9 +549,9 @@ checksum = "a2eb9349b6444b326872e140eb1cf5e7c522154d69e7a0ffb0fb81c06b37543f"
 
 [[package]]
 name = "syn"
-version = "1.0.68"
+version = "1.0.69"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "3ce15dd3ed8aa2f8eeac4716d6ef5ab58b6b9256db41d7e1a0224c2788e8fd87"
+checksum = "48fe99c6bd8b1cc636890bcc071842de909d902c81ac7dab53ba33c421ab8ffb"
 dependencies = [
  "proc-macro2",
  "quote",

Cargo.toml

@@ -67,6 +67,9 @@ fastrand = "1.4.0"
 [profile.release]
 lto = "thin"
 
+[profile.test]
+opt-level = 0
+
 # optimise dependencies, even when producing debug builds
 [profile.dev.package."*"]
 opt-level = 3

src/extension_set.rs

@@ -14,6 +14,8 @@ pub enum ExtensionSet {
 	Media,
 	/// Extensions used for document file formats, such as `pdf`, `odt`, `docx`, etc.
 	Documents,
+	/// Extensions used for text file formats, such as `txt`, `toml`, `html`, etc.
+	Text,
 	/// Extensions used for archive file formats, such as `zip`, `zst`, `gz`, etc.
 	Archives,
 }
@@ -34,9 +36,10 @@ impl ExtensionSet {
 			Self::Documents => vec![
 				"pdf", "doc", "docx", "ppt", "pptx", "xls", "xlsx", "csv", "tsv", "odt", "ods", "odp", "oda", "rtf", "ps",
 			],
+			Self::Text => mime_guess::get_mime_extensions_str("text/*").unwrap().to_vec(),
 			// many compressed file types follow the name scheme "application/x.+compressed.*" - maybe this can be used
 			// somehow to extract extensions for compressed files from mime_guess?
-			Self::Archives => vec!["zip", "tar", "gz", "zst", "xz", "rar", "7z", "bz", "bz2"],
+			Self::Archives => vec!["zip", "tar", "gz", "zst", "xz", "rar", "7z", "bz", "bz2", "tgz"],
 		}
 	}
 }

src/tests/mod.rs

@@ -8,9 +8,12 @@ use cfg_if::cfg_if;
 use mime_guess::mime::{APPLICATION_OCTET_STREAM, APPLICATION_PDF, IMAGE_JPEG, IMAGE_PNG};
 use mime_guess::Mime;
 
+use crate::findings::Findings;
+use crate::formats::{Format, Script};
 use std::borrow::Borrow;
 use std::collections::HashMap;
 use std::ffi::OsStr;
+use std::io::Read;
 use std::path::Path;
 
 const JPEG_BYTES: &[u8] = b"\xFF\xD8\xFF";
@@ -149,7 +152,10 @@ fn simple_directory() {
 				"pdf" => APPLICATION_PDF,
 				"zip" => application_zip(),
 				_ => APPLICATION_OCTET_STREAM, // general "fallback" type
-			}
+			},
+			"Incorrect MIME type detected - got {:?} for a {:?} file",
+			result.mime,
+			ext.unwrap()
 		);
 	}
 }
@@ -180,7 +186,8 @@ fn argument_parsing() {
 			hidden: false,
 			extensionless: false,
 			follow_symlinks: true
-		}
+		},
+		"ScanOpts are incorrect"
 	)
 }
 
@@ -224,3 +231,42 @@ fn identify_random_bytes() {
 	}
 	println!("No type found:\t{} counts", 500 - results.values().sum::<i32>())
 }
+
+#[test]
+fn outputs_move_commands() {
+	// create an example finding stating that "misnamed_file.png" has been identified as a jpeg file
+	let entries = vec![Ok(Findings {
+		file: Path::new("misnamed_file.png"),
+		valid: false,
+		mime: IMAGE_JPEG,
+	})];
+
+	let mut cursor = std::io::Cursor::new(Vec::new());
+	let mut contents = std::string::String::new();
+
+	Script::new()
+		.write_all(&entries, &mut cursor)
+		.expect("Failed to write to cursor");
+	cursor.set_position(0);
+	cursor
+		.read_to_string(&mut contents)
+		.expect("Failed to read from cursor to string");
+
+	// the output should contain a command like "mv -i misnamed_file.png misnamed_file.jpg"
+	assert!(
+		contents.contains("misnamed_file.jpg"),
+		"Output doesn't contain move command!"
+	)
+}
+
+#[test]
+fn media_contains_audio_video_images() {
+	use crate::extension_set::ExtensionSet::{Audio, Images, Media, Videos};
+	let media_exts = Media.extensions();
+
+	// assert every extension in the audio/video/image sets is contained in the media set
+	[Audio.extensions(), Videos.extensions(), Images.extensions()]
+		.concat()
+		.into_iter()
+		.for_each(|ext| assert!(media_exts.contains(&ext)));
+}