don't rename certain zip/exe subtypes. fixes #1

CHANGELOG.md

@@ -4,6 +4,14 @@ Dates are given in YYYY-MM-DD format - for example, the 15th of October 2021 is
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/), and this project adheres to
 [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## Unreleased
+### Changed
+- Some extensions are considered to be always valid - these are:
+  - "filepart", "part", "crdownload": Partially downloaded files, renaming could break download
+  - "bak", "backup": Backup copies are a common idiom (e.g. "game.exe.bak") and should be respected
+### Fixed
+- Support for many file types that are subcategories of others (e.g., fif will no longer rename apk files to zip)
+
 ## v0.5.1 - 2022-04-12
 ### Added
 - When using the [`infer`] backend, fif is now able to detect [Mach-O](https://en.wikipedia.org/wiki/Mach-O) binaries

Cargo.lock

@@ -80,16 +80,16 @@ checksum = "baf1de4339761588bc0619e3cbc0120ee582ebb74b53b4efbf79117bd2da40fd"
 
 [[package]]
 name = "clap"
-version = "3.1.8"
+version = "3.1.14"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "71c47df61d9e16dc010b55dba1952a57d8c215dbb533fd13cdd13369aac73b1c"
+checksum = "535434c063ced786eb04aaf529308092c5ab60889e8fe24275d15de07b01fa97"
 dependencies = [
  "atty",
  "bitflags",
  "clap_derive",
+ "clap_lex",
  "indexmap",
  "lazy_static",
- "os_str_bytes",
  "termcolor",
  "terminal_size",
  "textwrap",
@@ -109,6 +109,15 @@ dependencies = [
  "syn",
 ]
 
+[[package]]
+name = "clap_lex"
+version = "0.2.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "a37c35f1112dad5e6e0b1adaff798507497a18fceeb30cceb3bae7d1427b9213"
+dependencies = [
+ "os_str_bytes",
+]
+
 [[package]]
 name = "crossbeam-channel"
 version = "0.5.4"
@@ -351,9 +360,9 @@ dependencies = [
 
 [[package]]
 name = "libc"
-version = "0.2.122"
+version = "0.2.125"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "ec647867e2bf0772e28c8bcde4f0d19a9216916e890543b5a03ed8ef27b8f259"
+checksum = "5916d2ae698f6de9bfb891ad7a8d65c09d232dc58cc4ac433c7da3b2fd84bc2b"
 
 [[package]]
 name = "lock_api"
@@ -382,9 +391,9 @@ checksum = "3e2e65a1a2e43cfcb47a895c4c8b10d1f4a61097f9f254f183aee60cad9c651d"
 
 [[package]]
 name = "memchr"
-version = "2.4.1"
+version = "2.5.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "308cc39be01b73d0d18f82a0e7b2a3df85245f84af96fdddc5d202d27e47b86a"
+checksum = "2dffe52ecf27772e601905b7522cb4ef790d2cc203488bbd0e2fe85fcb74566d"
 
 [[package]]
 name = "memoffset"
@@ -445,9 +454,6 @@ name = "os_str_bytes"
 version = "6.0.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "8e22443d1643a904602595ba1cd8f7d896afe56d26712531c5ff73a15b2fbf64"
-dependencies = [
- "memchr",
-]
 
 [[package]]
 name = "parking_lot"
@@ -461,9 +467,9 @@ dependencies = [
 
 [[package]]
 name = "parking_lot_core"
-version = "0.9.2"
+version = "0.9.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "995f667a6c822200b0433ac218e05582f0e2efa1b922a3fd2fbaadc5f87bab37"
+checksum = "09a279cbf25cb0757810394fbc1e359949b59e348145c643a939a525692e6929"
 dependencies = [
  "cfg-if",
  "libc",
@@ -618,9 +624,9 @@ dependencies = [
 
 [[package]]
 name = "rayon"
-version = "1.5.1"
+version = "1.5.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "c06aca804d41dbc8ba42dfd964f0d01334eceb64314b9ecf7c5fad5188a06d90"
+checksum = "fd249e82c21598a9a426a4e00dd7adc1d640b22445ec8545feef801d1a74c221"
 dependencies = [
  "autocfg",
  "crossbeam-deque",
@@ -630,14 +636,13 @@ dependencies = [
 
 [[package]]
 name = "rayon-core"
-version = "1.9.1"
+version = "1.9.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "d78120e2c850279833f1dd3582f730c4ab53ed95aeaaaa862a2a5c71b1656d8e"
+checksum = "9f51245e1e62e1f1629cbfec37b5793bbabcaeb90f30e94d2ba03564687353e4"
 dependencies = [
  "crossbeam-channel",
  "crossbeam-deque",
  "crossbeam-utils",
- "lazy_static",
  "num_cpus",
 ]
 
@@ -714,18 +719,18 @@ checksum = "d29ab0c6d3fc0ee92fe66e2d99f700eab17a8d57d1c1d3b748380fb20baa78cd"
 
 [[package]]
 name = "serde"
-version = "1.0.136"
+version = "1.0.137"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "ce31e24b01e1e524df96f1c2fdd054405f8d7376249a5110886fb4b658484789"
+checksum = "61ea8d54c77f8315140a05f4c7237403bf38b72704d031543aa1d16abbf517d1"
 dependencies = [
  "serde_derive",
 ]
 
 [[package]]
 name = "serde_derive"
-version = "1.0.136"
+version = "1.0.137"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "08597e7152fcd306f41838ed3e37be9eaeed2b61c42e2117266a554fab4662f9"
+checksum = "1f26faba0c3959972377d3b2d306ee9f71faee9714294e41bb777f83f88578be"
 dependencies = [
  "proc-macro2",
  "quote",
@@ -734,9 +739,9 @@ dependencies = [
 
 [[package]]
 name = "serde_json"
-version = "1.0.79"
+version = "1.0.80"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "8e8d9fa5c3b304765ce1fd9c4c8a3de2c8db365a5b91be52f186efc675681d95"
+checksum = "f972498cf015f7c0746cac89ebe1d6ef10c293b94175a243a2d9442c163d9944"
 dependencies = [
  "itoa",
  "ryu",
@@ -782,9 +787,9 @@ checksum = "a2eb9349b6444b326872e140eb1cf5e7c522154d69e7a0ffb0fb81c06b37543f"
 
 [[package]]
 name = "syn"
-version = "1.0.91"
+version = "1.0.92"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "b683b2b825c8eef438b77c36a06dc262294da3d5a5813fac20da149241dcd44d"
+checksum = "7ff7c592601f11445996a06f8ad0c27f094a58857c2f89e97974ab9235b92c52"
 dependencies = [
  "proc-macro2",
  "quote",
@@ -842,18 +847,18 @@ dependencies = [
 
 [[package]]
 name = "thiserror"
-version = "1.0.30"
+version = "1.0.31"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "854babe52e4df1653706b98fcfc05843010039b406875930a70e4d9644e5c417"
+checksum = "bd829fe32373d27f76265620b5309d0340cb8550f523c1dda251d6298069069a"
 dependencies = [
  "thiserror-impl",
 ]
 
 [[package]]
 name = "thiserror-impl"
-version = "1.0.30"
+version = "1.0.31"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "aa32fd3f627f367fe16f893e2597ae3c05020f8bba2666a4e6ea73d377e5714b"
+checksum = "0396bc89e626244658bef819e22d0cc459e795a5ebe878e6ec336d1674a8d79a"
 dependencies = [
  "proc-macro2",
  "quote",
@@ -958,9 +963,9 @@ checksum = "712e227841d057c1ee1cd2fb22fa7e5a5461ae8e48fa2ca79ec42cfc1931183f"
 
 [[package]]
 name = "windows-sys"
-version = "0.34.0"
+version = "0.36.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "5acdd78cb4ba54c0045ac14f62d8f94a03d10047904ae2a40afa1e99d8f70825"
+checksum = "ea04155a16a59f9eab786fe12a4a450e75cdb175f9e0d80da1e17db09f55b8d2"
 dependencies = [
  "windows_aarch64_msvc",
  "windows_i686_gnu",
@@ -971,33 +976,33 @@ dependencies = [
 
 [[package]]
 name = "windows_aarch64_msvc"
-version = "0.34.0"
+version = "0.36.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "17cffbe740121affb56fad0fc0e421804adf0ae00891205213b5cecd30db881d"
+checksum = "9bb8c3fd39ade2d67e9874ac4f3db21f0d710bee00fe7cab16949ec184eeaa47"
 
 [[package]]
 name = "windows_i686_gnu"
-version = "0.34.0"
+version = "0.36.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "2564fde759adb79129d9b4f54be42b32c89970c18ebf93124ca8870a498688ed"
+checksum = "180e6ccf01daf4c426b846dfc66db1fc518f074baa793aa7d9b9aaeffad6a3b6"
 
 [[package]]
 name = "windows_i686_msvc"
-version = "0.34.0"
+version = "0.36.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "9cd9d32ba70453522332c14d38814bceeb747d80b3958676007acadd7e166956"
+checksum = "e2e7917148b2812d1eeafaeb22a97e4813dfa60a3f8f78ebe204bcc88f12f024"
 
 [[package]]
 name = "windows_x86_64_gnu"
-version = "0.34.0"
+version = "0.36.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "cfce6deae227ee8d356d19effc141a509cc503dfd1f850622ec4b0f84428e1f4"
+checksum = "4dcd171b8776c41b97521e5da127a2d86ad280114807d0b2ab1e462bc764d9e1"
 
 [[package]]
 name = "windows_x86_64_msvc"
-version = "0.34.0"
+version = "0.36.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "d19538ccc21819d01deaf88d6a17eae6596a12e9aafdbb97916fb49896d89de9"
+checksum = "c811ca4a8c853ef420abd8592ba53ddbbac90410fab6903b3e79972a631f7680"
 
 [[package]]
 name = "xdg-mime"

src/files.rs

@@ -127,16 +127,39 @@ pub fn scan_file(entry: &DirEntry, canonical_paths: bool) -> Result<Findings, Sc
 		Ok(Some(result)) => result,
 	};
 
-	// set of known extensions for the given MIME type
-	let known_exts = mime_extension_lookup(result.essence_str().into());
-	// file extension for this particular file
-	let entry_ext = path.extension();
 
-	let valid = match known_exts {
+	// determine whether or not the file's current extension is valid
-		// there is a known set of extensions for this MIME type, and the file has an extension
+	let valid = if let Some(entry_ext) = path.extension() {
-		Some(e) if entry_ext.is_some() => e.contains(&entry_ext.unwrap().to_string_lossy().to_lowercase().into()),
+		// discard invalid UTF-8 and convert to lowercase. all extensions in both backend's databases are lowercase
-		// either this file has no extension, or there is no known set of extensions for this MIME type :(
+		// ascii, so this assumption is fine.
-		Some(_) | None => false,
+		let entry_ext = entry_ext.to_string_lossy().to_lowercase();
+
+		// if the file has any of these extensions, it is probably either:
+		// - a copy of another file renamed for backup purposes (e.g. a word processor might save by renaming "my.doc" to
+		//   "my.doc.bak", then creating "my.doc", leaving the backup for safekeeping), which shouldn't be renamed so as
+		//   not to break the backup program
+		// - a partially downloaded file, which shouldn't be renamed to avoid corrupting it and blocking the downloader
+		//   from resuming
+		if ["bak", "backup", "filepart", "part", "crdownload"]
+			.iter()
+			.any(|ext| ext == &entry_ext)
+		{
+			true
+		} else {
+			// otherwise, check to see whether there's a known extension for this file type
+
+			// retrieve set of known extensions for the given MIME type
+			let known_exts = mime_extension_lookup(result.essence_str().into());
+			match known_exts {
+				// there is a known set of extensions for this MIME type - is entry_ext in the given set?
+				Some(e) => e.contains(&entry_ext.into()),
+				// there is no known set of extensions for this MIME type :(
+				None => false,
+			}
+		}
+	} else {
+		// this file has no extension
+		false
 	};
 
 	let path = if canonical_paths {
@@ -350,16 +373,47 @@ pub fn mime_extension_lookup(essence: String) -> Option<Vec<String>> {
 				// classic office files considered harmful
 				vec![String::from("doc"), String::from("xls"), String::from("ppt")]
 			} else if essence == "application/zip" {
-				// neither xdg-mime nor infer seem to be able to detect office XML files properly...
+				// both backends seem to be unable to consistently detect OOXML files, so they should be considered valid
+				// extensions for zip files to prevent them being erroneously renamed.
+				// additionally, there are various popular formats that are just renamed zip files, such as android's apk
+				// format, that also shouldn't be renamed.
 				[
-					vec![String::from("zip"), String::from("docx"), String::from("xlsx"), String::from("pptx")],
+					vec![
+						String::from("zip"),
+						String::from("docx"),
+						String::from("xlsx"),
+						String::from("pptx"),
+						String::from("apk"),
+						String::from("ipa"),
+						String::from("docbook"),
+						String::from("kdenlive"),
+						String::from("vcpkg"),
+						String::from("nupkg"),
+						String::from("whl"),
+						String::from("xpi"),
+					],
 					possible_exts,
 				]
 				.concat()
 			} else if essence == "application/x-ms-dos-executable" {
-				// .dll, .exe, and .scr files are given the same MIME type... but you definitely don't want to rename one to the
+				// .dll, .exe, .scr, etc. files are given the same MIME type, and aren't really distinguishable from each other
-				// other!
+				// ... but you definitely don't want to rename one to the other!
-				[vec![String::from("dll"), String::from("exe"), String::from("scr")], possible_exts].concat()
+				[
+					vec![
+						String::from("exe"),
+						String::from("dll"),
+						String::from("scr"),
+						String::from("com"),
+						String::from("dll16"),
+						String::from("drv"),
+						String::from("drv16"),
+						String::from("cpl"),
+						String::from("msstyles"),
+						String::from("sys"),
+					],
+					possible_exts,
+				]
+				.concat()
 			} else {
 				possible_exts
 			})