file-type 21.0.0 → 21.1.1

package/core.js

@@ -4,7 +4,7 @@ Primary entry point, Node.js specific entry point is index.js
 
 import * as Token from 'token-types';
 import * as strtok3 from 'strtok3/core';
-import {ZipHandler} from '@tokenizer/inflate';
+import {ZipHandler, GzipHandler} from '@tokenizer/inflate';
 import {getUintBE} from 'uint8array-extras';
 import {
 	stringToBytes,
@@ -234,11 +234,16 @@ export class FileTypeParser {
 	}
 
 	async fromBlob(blob) {
-		return this.fromStream(blob.stream());
+		const tokenizer = strtok3.fromBlob(blob, this.tokenizerOptions);
+		try {
+			return await this.fromTokenizer(tokenizer);
+		} finally {
+			await tokenizer.close();
+		}
 	}
 
 	async fromStream(stream) {
-		const tokenizer = await strtok3.fromWebStream(stream, this.tokenizerOptions);
+		const tokenizer = strtok3.fromWebStream(stream, this.tokenizerOptions);
 		try {
 			return await this.fromTokenizer(tokenizer);
 		} finally {
@@ -296,7 +301,7 @@ export class FileTypeParser {
 	}
 
 	checkString(header, options) {
-		return this.check(stringToBytes(header), options);
+		return this.check(stringToBytes(header, options?.encoding), options);
 	}
 
 	// Detections with a high degree of certainty in identifying the correct file type
@@ -310,7 +315,7 @@ export class FileTypeParser {
 
 		this.tokenizer = tokenizer;
 
-		await tokenizer.peekBuffer(this.buffer, {length: 12, mayBeLess: true});
+		await tokenizer.peekBuffer(this.buffer, {length: 32, mayBeLess: true});
 
 		// -- 2-byte signatures --
 
@@ -408,6 +413,30 @@ export class FileTypeParser {
 		}
 
 		if (this.check([0x1F, 0x8B, 0x8])) {
+			const gzipHandler = new GzipHandler(tokenizer);
+
+			const stream = gzipHandler.inflate();
+			let shouldCancelStream = true;
+			try {
+				let compressedFileType;
+				try {
+					compressedFileType = await this.fromStream(stream);
+				} catch {
+					shouldCancelStream = false;
+				}
+
+				if (compressedFileType && compressedFileType.ext === 'tar') {
+					return {
+						ext: 'tar.gz',
+						mime: 'application/gzip',
+					};
+				}
+			} finally {
+				if (shouldCancelStream) {
+					await stream.cancel();
+				}
+			}
+
 			return {
 				ext: 'gz',
 				mime: 'application/gzip',
@@ -577,6 +606,10 @@ export class FileTypeParser {
 
 						return {};
 				}
+			}).catch(error => {
+				if (!(error instanceof strtok3.EndOfStreamError)) {
+					throw error; // Re-throw non-EndOfStreamError
+				}
 			});
 
 			return fileType ?? {
@@ -932,6 +965,13 @@ export class FileTypeParser {
 			};
 		}
 
+		if (this.checkString('regf')) {
+			return {
+				ext: 'dat',
+				mime: 'application/x-ft-windows-registry-hive',
+			};
+		}
+
 		// -- 5-byte signatures --
 
 		if (this.check([0x4F, 0x54, 0x54, 0x4F, 0x00])) {
@@ -1283,6 +1323,15 @@ export class FileTypeParser {
 			}
 		}
 
+		// -- 10-byte signatures --
+
+		if (this.checkString('REGEDIT4\r\n')) {
+			return {
+				ext: 'reg',
+				mime: 'application/x-ms-regedit',
+			};
+		}
+
 		// -- 12-byte signatures --
 
 		// RIFF file format which might be AVI, WAV, QCP, etc
@@ -1443,8 +1492,8 @@ export class FileTypeParser {
 			};
 		}
 
-		if (this.check([0xFE, 0xFF])) { // UTF-16-BOM-LE
-			if (this.check([0, 60, 0, 63, 0, 120, 0, 109, 0, 108], {offset: 2})) {
+		if (this.check([0xFE, 0xFF])) { // UTF-16-BOM-BE
+			if (this.checkString('<?xml ', {offset: 2, encoding: 'utf-16be'})) {
 				return {
 					ext: 'xml',
 					mime: 'application/xml',
@@ -1462,7 +1511,7 @@ export class FileTypeParser {
 			};
 		}
 
-		// Increase sample size from 12 to 256.
+		// Increase sample size from 32 to 256.
 		await tokenizer.peekBuffer(this.buffer, {length: Math.min(256, tokenizer.fileInfo.size), mayBeLess: true});
 
 		if (this.check([0x61, 0x63, 0x73, 0x70], {offset: 36})) {
@@ -1636,21 +1685,29 @@ export class FileTypeParser {
 			};
 		}
 
-		if (this.check([0xFF, 0xFE])) { // UTF-16-BOM-BE
-			if (this.check([60, 0, 63, 0, 120, 0, 109, 0, 108, 0], {offset: 2})) {
+		if (this.check([0xFF, 0xFE])) { // UTF-16-BOM-LE
+			const encoding = 'utf-16le';
+			if (this.checkString('<?xml ', {offset: 2, encoding})) {
 				return {
 					ext: 'xml',
 					mime: 'application/xml',
 				};
 			}
 
-			if (this.check([0xFF, 0x0E, 0x53, 0x00, 0x6B, 0x00, 0x65, 0x00, 0x74, 0x00, 0x63, 0x00, 0x68, 0x00, 0x55, 0x00, 0x70, 0x00, 0x20, 0x00, 0x4D, 0x00, 0x6F, 0x00, 0x64, 0x00, 0x65, 0x00, 0x6C, 0x00], {offset: 2})) {
+			if (this.check([0xFF, 0x0E], {offset: 2}) && this.checkString('SketchUp Model', {offset: 4, encoding})) {
 				return {
 					ext: 'skp',
 					mime: 'application/vnd.sketchup.skp',
 				};
 			}
 
+			if (this.checkString('Windows Registry Editor Version 5.00\r\n', {offset: 2, encoding})) {
+				return {
+					ext: 'reg',
+					mime: 'application/x-ms-regedit',
+				};
+			}
+
 			return undefined; // Some text based format
 		}
 

package/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "file-type",
-	"version": "21.0.0",
+	"version": "21.1.1",
 	"description": "Detect the file type of a file, stream, or data",
 	"license": "MIT",
 	"repository": "sindresorhus/file-type",
@@ -243,12 +243,15 @@
 		"jar",
 		"rm",
 		"ppsm",
-		"ppsx"
+		"ppsx",
+		"tar.gz",
+		"reg",
+		"dat"
 	],
 	"dependencies": {
-		"@tokenizer/inflate": "^0.2.7",
-		"strtok3": "^10.2.2",
-		"token-types": "^6.0.0",
+		"@tokenizer/inflate": "^0.4.1",
+		"strtok3": "^10.3.4",
+		"token-types": "^6.1.1",
 		"uint8array-extras": "^1.4.0"
 	},
 	"devDependencies": {

package/readme.md

@@ -394,6 +394,7 @@ console.log(fileType);
 
 ### Available third-party file-type detectors
 
+- [@file-type/av](https://github.com/Borewit/file-type-av): Improves detection of audio and video file formats, with accurate differentiation between the two
 - [@file-type/xml](https://github.com/Borewit/file-type-xml): Detects common XML file types, such as GLM, KML, MusicXML, RSS, SVG, and XHTML
 
 ### Detector execution flow
@@ -494,6 +495,7 @@ abortController.abort(); // Abort file-type reading from the Blob stream.
 - [`cr3`](https://fileinfo.com/extension/cr3) - Canon Raw image file (v3)
 - [`crx`](https://developer.chrome.com/extensions/crx) - Google Chrome extension
 - [`cur`](https://en.wikipedia.org/wiki/ICO_(file_format)) - Icon file
+- [`dat`](https://en.wikipedia.org/wiki/Windows_Registry) - Windows registry hive file
 - [`dcm`](https://en.wikipedia.org/wiki/DICOM#Data_format) - DICOM Image File
 - [`deb`](https://en.wikipedia.org/wiki/Deb_(file_format)) - Debian package
 - [`dmg`](https://en.wikipedia.org/wiki/Apple_Disk_Image) - Apple Disk Image
@@ -596,6 +598,7 @@ abortController.abort(); // Abort file-type reading from the Blob stream.
 - [`qcp`](https://en.wikipedia.org/wiki/QCP) - Tagged and chunked data
 - [`raf`](https://en.wikipedia.org/wiki/Raw_image_format) - Fujifilm RAW image file
 - [`rar`](https://en.wikipedia.org/wiki/RAR_(file_format)) - Archive file
+- [`reg`](https://en.wikipedia.org/wiki/Windows_Registry) - Windows registry (entries) file format
 - [`rm`](https://en.wikipedia.org/wiki/RealMedia) - RealMedia
 - [`rpm`](https://fileinfo.com/extension/rpm) - Red Hat Package Manager file
 - [`rtf`](https://en.wikipedia.org/wiki/Rich_Text_Format) - Rich Text Format
@@ -607,7 +610,8 @@ abortController.abort(); // Abort file-type reading from the Blob stream.
 - [`sqlite`](https://www.sqlite.org/fileformat2.html) - SQLite file
 - [`stl`](https://en.wikipedia.org/wiki/STL_(file_format)) - Standard Tesselated Geometry File Format (ASCII only)
 - [`swf`](https://en.wikipedia.org/wiki/SWF) - Adobe Flash Player file
-- [`tar`](https://en.wikipedia.org/wiki/Tar_(computing)#File_format) - Tarball archive file
+- [`tar`](https://en.wikipedia.org/wiki/Tar_(computing)#File_format) - Tape archive or tarball
+- [`tar.gz`](https://en.wikipedia.org/wiki/Gzip) - Gzipped tape archive (tarball)
 - [`tif`](https://en.wikipedia.org/wiki/Tagged_Image_File_Format) - Tagged Image file
 - [`ttc`](https://en.wikipedia.org/wiki/TrueType#TrueType_Collection) - TrueType Collection font
 - [`ttf`](https://en.wikipedia.org/wiki/TrueType) - TrueType font

package/supported.js

@@ -174,6 +174,9 @@ export const extensions = [
 	'rm',
 	'ppsm',
 	'ppsx',
+	'tar.gz',
+	'reg',
+	'dat',
 ];
 
 export const mimeTypes = [
@@ -348,4 +351,6 @@ export const mimeTypes = [
 	'application/vnd.ms-powerpoint.presentation.macroenabled.12',
 	'application/java-archive',
 	'application/vnd.rn-realmedia',
+	'application/x-ms-regedit',
+	'application/x-ft-windows-registry-hive',
 ];

package/util.js

@@ -1,6 +1,26 @@
 import {StringType} from 'token-types';
 
-export function stringToBytes(string) {
+export function stringToBytes(string, encoding) {
+	if (encoding === 'utf-16le') {
+		const bytes = [];
+		for (let index = 0; index < string.length; index++) {
+			const code = string.charCodeAt(index); // eslint-disable-line unicorn/prefer-code-point
+			bytes.push(code & 0xFF, (code >> 8) & 0xFF); // High byte
+		}
+
+		return bytes;
+	}
+
+	if (encoding === 'utf-16be') {
+		const bytes = [];
+		for (let index = 0; index < string.length; index++) {
+			const code = string.charCodeAt(index); // eslint-disable-line unicorn/prefer-code-point
+			bytes.push((code >> 8) & 0xFF, code & 0xFF); // Low byte
+		}
+
+		return bytes;
+	}
+
 	return [...string].map(character => character.charCodeAt(0)); // eslint-disable-line unicorn/prefer-code-point
 }