build(deps): bump github.com/gabriel-vasile/mimetype (#1775)

Bumps [github.com/gabriel-vasile/mimetype](https://github.com/gabriel-vasile/mimetype) from 1.4.10 to 1.4.11. - [Release notes](https://github.com/gabriel-vasile/mimetype/releases) - [Commits](https://github.com/gabriel-vasile/mimetype/compare/v1.4.10...v1.4.11) --- updated-dependencies: - dependency-name: github.com/gabriel-vasile/mimetype dependency-version: 1.4.11 dependency-type: direct:production update-type: version-update:semver-patch ... Signed-off-by: dependabot[bot] <support@github.com> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
2025-11-04 16:02:02 +01:00
parent fcff855e16
commit 2da203613a
26 changed files with 1138 additions and 677 deletions
@@ -18,7 +18,7 @@ require (
 	github.com/davidbyttow/govips/v2 v2.16.0
 	github.com/dhowden/tag v0.0.0-20240417053706-3d75831295e8
 	github.com/dutchcoders/go-clamd v0.0.0-20170520113014-b970184f4d9e
-	github.com/gabriel-vasile/mimetype v1.4.10
+	github.com/gabriel-vasile/mimetype v1.4.11
 	github.com/ggwhite/go-masker v1.1.0
 	github.com/go-chi/chi/v5 v5.2.3
 	github.com/go-chi/render v1.0.3
@@ -349,8 +349,8 @@ github.com/fsnotify/fsnotify v1.4.7/go.mod h1:jwhsz4b93w/PPRr/qN1Yymfu8t87LnFCMo
 github.com/fsnotify/fsnotify v1.4.9/go.mod h1:znqG4EE+3YCdAaPaxE2ZRY/06pZUdp0tY4IgpuI1SZQ=
 github.com/fsnotify/fsnotify v1.9.0 h1:2Ml+OJNzbYCTzsxtv8vKSFD9PbJjmhYF14k/jKC7S9k=
 github.com/fsnotify/fsnotify v1.9.0/go.mod h1:8jBTzvmWwFyi3Pb8djgCCO5IBqzKJ/Jwo8TRcHyHii0=
-github.com/gabriel-vasile/mimetype v1.4.10 h1:zyueNbySn/z8mJZHLt6IPw0KoZsiQNszIpU+bX4+ZK0=
-github.com/gabriel-vasile/mimetype v1.4.10/go.mod h1:d+9Oxyo1wTzWdyVUPMmXFvp4F9tea18J8ufA774AB3s=
+github.com/gabriel-vasile/mimetype v1.4.11 h1:AQvxbp830wPhHTqc1u7nzoLT+ZFxGY7emj5DR5DYFik=
+github.com/gabriel-vasile/mimetype v1.4.11/go.mod h1:d+9Oxyo1wTzWdyVUPMmXFvp4F9tea18J8ufA774AB3s=
 github.com/gdexlab/go-render v1.0.1 h1:rxqB3vo5s4n1kF0ySmoNeSPRYkEsyHgln4jFIQY7v0U=
 github.com/gdexlab/go-render v1.0.1/go.mod h1:wRi5nW2qfjiGj4mPukH4UV0IknS1cHD4VgFTmJX5JzM=
 github.com/getkin/kin-openapi v0.13.0/go.mod h1:WGRs2ZMM1Q8LR1QBEwUxC6RJEfaBcD0s+pcEVXFuAjw=
@@ -0,0 +1,9 @@
+# Github is obeying this ignore file by default.
+# Run this command on local to ignore formatting commits in `git blame`
+# git config blame.ignoreRevsFile .git-blame-ignore-revs
+
+# Added a new column to supported_mimes.md
+# The supported_mimes.md file was a nice way to find when a file format was
+# introduced. However, when I changed to add a new column in the table, the
+# whole git blame got poisoned for the file.
+eb497f9bc5d31c6eab2929a112051218670137ba
@@ -97,7 +97,9 @@ or from a [file](https://pkg.go.dev/github.com/gabriel-vasile/mimetype#DetectFil
 </div>

 ## Contributing
-Contributions are unexpected but welcome. When submitting a PR for detection of
-a new file format, please make sure to add a record to the list of testcases
-from [mimetype_test.go](mimetype_test.go). For complex files a record can be added
-in the [testdata](testdata) directory.
+Contributions are never expected but very much welcome.
+[mimetype_tests](https://github.com/gabriel-vasile/mimetype_tests/actions/workflows/test.yml)
+shows which file formats are most often misidentified and can help prioritise.
+When submitting a PR for detection of a new file format, please make sure to
+add a record to the list of testcases in [mimetype_test.go](mimetype_test.go).
+For complex files a record can be added in the [testdata](testdata) directory.
@@ -2,6 +2,7 @@ package charset

 import (
 	"bytes"
+	"strings"
 	"unicode/utf8"

 	"github.com/gabriel-vasile/mimetype/internal/markup"
@@ -141,27 +142,25 @@ func FromXML(content []byte) string {
 	return FromPlain(content)
 }
 func fromXML(s scan.Bytes) string {
-	xml := []byte("<?XML")
+	xml := []byte("<?xml")
 	lxml := len(xml)
 	for {
-		if len(s) == 0 {
-			return ""
-		}
-		for scan.ByteIsWS(s.Peek()) {
-			s.Advance(1)
-		}
+		s.TrimLWS()
 		if len(s) <= lxml {
 			return ""
 		}
-		if !s.Match(xml, scan.IgnoreCase) {
-			s = s[1:] // safe to slice instead of s.Advance(1) because bounds are checked
-			continue
+
+		i, k := s.Search(xml, 0)
+		if i == -1 {
+			return ""
 		}
-		aName, aVal, hasMore := "", "", true
+		s.Advance(i + k)
+		var aName, aVal []byte
+		hasMore := true
 		for hasMore {
 			aName, aVal, hasMore = markup.GetAnAttribute(&s)
-			if aName == "encoding" && aVal != "" {
-				return aVal
+			if scan.Bytes(aName).Match([]byte("encoding"), 0) != -1 && len(aVal) != 0 {
+				return string(aVal)
 			}
 		}
 	}
@@ -198,10 +197,10 @@ func fromHTML(s scan.Bytes) string {
 			return ""
 		}
 		// Abort when <body is reached.
-		if s.Match(body, scan.IgnoreCase) {
+		if s.Match(body, scan.IgnoreCase) != -1 {
 			return ""
 		}
-		if !s.Match(meta, scan.IgnoreCase) {
+		if s.Match(meta, scan.IgnoreCase) == -1 {
 			s = s[1:] // safe to slice instead of s.Advance(1) because bounds are checked
 			continue
 		}
@@ -215,14 +214,16 @@ func fromHTML(s scan.Bytes) string {
 		needPragma := dontKnow

 		charset := ""
-		aName, aVal, hasMore := "", "", true
+		var aNameB, aValB []byte
+		hasMore := true
 		for hasMore {
-			aName, aVal, hasMore = markup.GetAnAttribute(&s)
+			aNameB, aValB, hasMore = markup.GetAnAttribute(&s)
+			aName := strings.ToLower(string(aNameB))
 			if attrList[aName] {
 				continue
 			}
 			// processing step
-			if len(aName) == 0 && len(aVal) == 0 {
+			if len(aName) == 0 && len(aValB) == 0 {
 				if needPragma == dontKnow {
 					continue
 				}
@@ -231,15 +232,18 @@ func fromHTML(s scan.Bytes) string {
 				}
 			}
 			attrList[aName] = true
-			if aName == "http-equiv" && scan.Bytes(aVal).Match([]byte("CONTENT-TYPE"), scan.IgnoreCase) {
-				gotPragma = true
-			} else if aName == "content" {
-				charset = string(extractCharsetFromMeta(scan.Bytes(aVal)))
+			switch aName {
+			case "http-equiv":
+				if scan.Bytes(aValB).Match([]byte("CONTENT-TYPE"), scan.IgnoreCase) != -1 {
+					gotPragma = true
+				}
+			case "content":
+				charset = string(extractCharsetFromMeta(scan.Bytes(aValB)))
 				if len(charset) != 0 {
 					needPragma = doNeedPragma
 				}
-			} else if aName == "charset" {
-				charset = aVal
+			case "charset":
+				charset = string(aValB)
 				needPragma = doNotNeedPragma
 			}
 		}
@@ -257,8 +257,11 @@ out:
 	return 0
 }

+// openArray is used instead of an inline []byte{'['} to avoid mem alllocs.
+var openArray = []byte{'['}
+
 func (p *parserState) consumeArray(b []byte, qs []query, lvl int) (n int) {
-	p.appendPath([]byte{'['}, qs)
+	p.appendPath(openArray, qs)
 	if len(b) == 0 {
 		return 0
 	}
@@ -5,43 +5,80 @@ import (
 	"encoding/binary"
 )

-var (
-	// SevenZ matches a 7z archive.
-	SevenZ = prefix([]byte{0x37, 0x7A, 0xBC, 0xAF, 0x27, 0x1C})
-	// Gzip matches gzip files based on http://www.zlib.org/rfc-gzip.html#header-trailer.
-	Gzip = prefix([]byte{0x1f, 0x8b})
-	// Fits matches an Flexible Image Transport System file.
-	Fits = prefix([]byte{
+// SevenZ matches a 7z archive.
+func SevenZ(raw []byte, _ uint32) bool {
+	return bytes.HasPrefix(raw, []byte{0x37, 0x7A, 0xBC, 0xAF, 0x27, 0x1C})
+}
+
+// Gzip matches gzip files based on http://www.zlib.org/rfc-gzip.html#header-trailer.
+func Gzip(raw []byte, _ uint32) bool {
+	return bytes.HasPrefix(raw, []byte{0x1f, 0x8b})
+}
+
+// Fits matches an Flexible Image Transport System file.
+func Fits(raw []byte, _ uint32) bool {
+	return bytes.HasPrefix(raw, []byte{
 		0x53, 0x49, 0x4D, 0x50, 0x4C, 0x45, 0x20, 0x20, 0x3D, 0x20,
 		0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20,
 		0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x54,
 	})
-	// Xar matches an eXtensible ARchive format file.
-	Xar = prefix([]byte{0x78, 0x61, 0x72, 0x21})
-	// Bz2 matches a bzip2 file.
-	Bz2 = prefix([]byte{0x42, 0x5A, 0x68})
-	// Ar matches an ar (Unix) archive file.
-	Ar = prefix([]byte{0x21, 0x3C, 0x61, 0x72, 0x63, 0x68, 0x3E})
-	// Deb matches a Debian package file.
-	Deb = offset([]byte{
+}
+
+// Xar matches an eXtensible ARchive format file.
+func Xar(raw []byte, _ uint32) bool {
+	return bytes.HasPrefix(raw, []byte{0x78, 0x61, 0x72, 0x21})
+}
+
+// Bz2 matches a bzip2 file.
+func Bz2(raw []byte, _ uint32) bool {
+	return bytes.HasPrefix(raw, []byte{0x42, 0x5A, 0x68})
+}
+
+// Ar matches an ar (Unix) archive file.
+func Ar(raw []byte, _ uint32) bool {
+	return bytes.HasPrefix(raw, []byte{0x21, 0x3C, 0x61, 0x72, 0x63, 0x68, 0x3E})
+}
+
+// Deb matches a Debian package file.
+func Deb(raw []byte, _ uint32) bool {
+	return offset(raw, []byte{
 		0x64, 0x65, 0x62, 0x69, 0x61, 0x6E, 0x2D,
 		0x62, 0x69, 0x6E, 0x61, 0x72, 0x79,
 	}, 8)
-	// Warc matches a Web ARChive file.
-	Warc = prefix([]byte("WARC/1.0"), []byte("WARC/1.1"))
-	// Cab matches a Microsoft Cabinet archive file.
-	Cab = prefix([]byte("MSCF\x00\x00\x00\x00"))
-	// Xz matches an xz compressed stream based on https://tukaani.org/xz/xz-file-format.txt.
-	Xz = prefix([]byte{0xFD, 0x37, 0x7A, 0x58, 0x5A, 0x00})
-	// Lzip matches an Lzip compressed file.
-	Lzip = prefix([]byte{0x4c, 0x5a, 0x49, 0x50})
-	// RPM matches an RPM or Delta RPM package file.
-	RPM = prefix([]byte{0xed, 0xab, 0xee, 0xdb}, []byte("drpm"))
-	// Cpio matches a cpio archive file.
-	Cpio = prefix([]byte("070707"), []byte("070701"), []byte("070702"))
-	// RAR matches a RAR archive file.
-	RAR = prefix([]byte("Rar!\x1A\x07\x00"), []byte("Rar!\x1A\x07\x01\x00"))
-)
+}
+
+// Warc matches a Web ARChive file.
+func Warc(raw []byte, _ uint32) bool {
+	return bytes.HasPrefix(raw, []byte("WARC/1.0")) ||
+		bytes.HasPrefix(raw, []byte("WARC/1.1"))
+}
+
+// Cab matches a Microsoft Cabinet archive file.
+func Cab(raw []byte, _ uint32) bool {
+	return bytes.HasPrefix(raw, []byte("MSCF\x00\x00\x00\x00"))
+}
+
+// Xz matches an xz compressed stream based on https://tukaani.org/xz/xz-file-format.txt.
+func Xz(raw []byte, _ uint32) bool {
+	return bytes.HasPrefix(raw, []byte{0xFD, 0x37, 0x7A, 0x58, 0x5A, 0x00})
+}
+
+// Lzip matches an Lzip compressed file.
+func Lzip(raw []byte, _ uint32) bool {
+	return bytes.HasPrefix(raw, []byte{0x4c, 0x5a, 0x49, 0x50})
+}
+
+// RPM matches an RPM or Delta RPM package file.
+func RPM(raw []byte, _ uint32) bool {
+	return bytes.HasPrefix(raw, []byte{0xed, 0xab, 0xee, 0xdb}) ||
+		bytes.HasPrefix(raw, []byte("drpm"))
+}
+
+// RAR matches a RAR archive file.
+func RAR(raw []byte, _ uint32) bool {
+	return bytes.HasPrefix(raw, []byte("Rar!\x1A\x07\x00")) ||
+		bytes.HasPrefix(raw, []byte("Rar!\x1A\x07\x01\x00"))
+}

 // InstallShieldCab matches an InstallShield Cabinet archive file.
 func InstallShieldCab(raw []byte, _ uint32) bool {
@@ -78,6 +115,17 @@ func CRX(raw []byte, limit uint32) bool {
 	return Zip(raw[zipOffset:], limit)
 }

+// Cpio matches a cpio archive file.
+func Cpio(raw []byte, _ uint32) bool {
+	if len(raw) < 6 {
+		return false
+	}
+	return binary.LittleEndian.Uint16(raw) == 070707 || // binary cpio
+		bytes.HasPrefix(raw, []byte("070707")) || // portable ASCII cpios
+		bytes.HasPrefix(raw, []byte("070701")) ||
+		bytes.HasPrefix(raw, []byte("070702"))
+}
+
 // Tar matches a (t)ape (ar)chive file.
 // Tar files are divided into 512 bytes records. First record contains a 257
 // bytes header padded with NUL.
@@ -5,26 +5,50 @@ import (
 	"encoding/binary"
 )

-var (
-	// Flac matches a Free Lossless Audio Codec file.
-	Flac = prefix([]byte("\x66\x4C\x61\x43\x00\x00\x00\x22"))
-	// Midi matches a Musical Instrument Digital Interface file.
-	Midi = prefix([]byte("\x4D\x54\x68\x64"))
-	// Ape matches a Monkey's Audio file.
-	Ape = prefix([]byte("\x4D\x41\x43\x20\x96\x0F\x00\x00\x34\x00\x00\x00\x18\x00\x00\x00\x90\xE3"))
-	// MusePack matches a Musepack file.
-	MusePack = prefix([]byte("MPCK"))
-	// Au matches a Sun Microsystems au file.
-	Au = prefix([]byte("\x2E\x73\x6E\x64"))
-	// Amr matches an Adaptive Multi-Rate file.
-	Amr = prefix([]byte("\x23\x21\x41\x4D\x52"))
-	// Voc matches a Creative Voice file.
-	Voc = prefix([]byte("Creative Voice File"))
-	// M3u matches a Playlist file.
-	M3u = prefix([]byte("#EXTM3U"))
-	// AAC matches an Advanced Audio Coding file.
-	AAC = prefix([]byte{0xFF, 0xF1}, []byte{0xFF, 0xF9})
-)
+// Flac matches a Free Lossless Audio Codec file.
+func Flac(raw []byte, _ uint32) bool {
+	return bytes.HasPrefix(raw, []byte("\x66\x4C\x61\x43\x00\x00\x00\x22"))
+}
+
+// Midi matches a Musical Instrument Digital Interface file.
+func Midi(raw []byte, _ uint32) bool {
+	return bytes.HasPrefix(raw, []byte("\x4D\x54\x68\x64"))
+}
+
+// Ape matches a Monkey's Audio file.
+func Ape(raw []byte, _ uint32) bool {
+	return bytes.HasPrefix(raw, []byte("\x4D\x41\x43\x20\x96\x0F\x00\x00\x34\x00\x00\x00\x18\x00\x00\x00\x90\xE3"))
+}
+
+// MusePack matches a Musepack file.
+func MusePack(raw []byte, _ uint32) bool {
+	return bytes.HasPrefix(raw, []byte("MPCK"))
+}
+
+// Au matches a Sun Microsystems au file.
+func Au(raw []byte, _ uint32) bool {
+	return bytes.HasPrefix(raw, []byte("\x2E\x73\x6E\x64"))
+}
+
+// Amr matches an Adaptive Multi-Rate file.
+func Amr(raw []byte, _ uint32) bool {
+	return bytes.HasPrefix(raw, []byte("\x23\x21\x41\x4D\x52"))
+}
+
+// Voc matches a Creative Voice file.
+func Voc(raw []byte, _ uint32) bool {
+	return bytes.HasPrefix(raw, []byte("Creative Voice File"))
+}
+
+// M3u matches a Playlist file.
+func M3u(raw []byte, _ uint32) bool {
+	return bytes.HasPrefix(raw, []byte("#EXTM3U"))
+}
+
+// AAC matches an Advanced Audio Coding file.
+func AAC(raw []byte, _ uint32) bool {
+	return len(raw) > 1 && ((raw[0] == 0xFF && raw[1] == 0xF1) || (raw[0] == 0xFF && raw[1] == 0xF9))
+}

 // Mp3 matches an mp3 file.
 func Mp3(raw []byte, limit uint32) bool {
@@ -6,26 +6,52 @@ import (
 	"encoding/binary"
 )

-var (
-	// Lnk matches Microsoft lnk binary format.
-	Lnk = prefix([]byte{0x4C, 0x00, 0x00, 0x00, 0x01, 0x14, 0x02, 0x00})
-	// Wasm matches a web assembly File Format file.
-	Wasm = prefix([]byte{0x00, 0x61, 0x73, 0x6D})
-	// Exe matches a Windows/DOS executable file.
-	Exe = prefix([]byte{0x4D, 0x5A})
-	// Elf matches an Executable and Linkable Format file.
-	Elf = prefix([]byte{0x7F, 0x45, 0x4C, 0x46})
-	// Nes matches a Nintendo Entertainment system ROM file.
-	Nes = prefix([]byte{0x4E, 0x45, 0x53, 0x1A})
-	// SWF matches an Adobe Flash swf file.
-	SWF = prefix([]byte("CWS"), []byte("FWS"), []byte("ZWS"))
-	// Torrent has bencoded text in the beginning.
-	Torrent = prefix([]byte("d8:announce"))
-	// PAR1 matches a parquet file.
-	Par1 = prefix([]byte{0x50, 0x41, 0x52, 0x31})
-	// CBOR matches a Concise Binary Object Representation https://cbor.io/
-	CBOR = prefix([]byte{0xD9, 0xD9, 0xF7})
-)
+// Lnk matches Microsoft lnk binary format.
+func Lnk(raw []byte, _ uint32) bool {
+	return bytes.HasPrefix(raw, []byte{0x4C, 0x00, 0x00, 0x00, 0x01, 0x14, 0x02, 0x00})
+}
+
+// Wasm matches a web assembly File Format file.
+func Wasm(raw []byte, _ uint32) bool {
+	return bytes.HasPrefix(raw, []byte{0x00, 0x61, 0x73, 0x6D})
+}
+
+// Exe matches a Windows/DOS executable file.
+func Exe(raw []byte, _ uint32) bool {
+	return len(raw) > 1 && raw[0] == 0x4D && raw[1] == 0x5A
+}
+
+// Elf matches an Executable and Linkable Format file.
+func Elf(raw []byte, _ uint32) bool {
+	return bytes.HasPrefix(raw, []byte{0x7F, 0x45, 0x4C, 0x46})
+}
+
+// Nes matches a Nintendo Entertainment system ROM file.
+func Nes(raw []byte, _ uint32) bool {
+	return bytes.HasPrefix(raw, []byte{0x4E, 0x45, 0x53, 0x1A})
+}
+
+// SWF matches an Adobe Flash swf file.
+func SWF(raw []byte, _ uint32) bool {
+	return bytes.HasPrefix(raw, []byte("CWS")) ||
+		bytes.HasPrefix(raw, []byte("FWS")) ||
+		bytes.HasPrefix(raw, []byte("ZWS"))
+}
+
+// Torrent has bencoded text in the beginning.
+func Torrent(raw []byte, _ uint32) bool {
+	return bytes.HasPrefix(raw, []byte("d8:announce"))
+}
+
+// PAR1 matches a parquet file.
+func Par1(raw []byte, _ uint32) bool {
+	return bytes.HasPrefix(raw, []byte{0x50, 0x41, 0x52, 0x31})
+}
+
+// CBOR matches a Concise Binary Object Representation https://cbor.io/
+func CBOR(raw []byte, _ uint32) bool {
+	return bytes.HasPrefix(raw, []byte{0xD9, 0xD9, 0xF7})
+}

 // Java bytecode and Mach-O binaries share the same magic number.
 // More info here https://github.com/threatstack/libmagic/blob/master/magic/Magdir/cafebabe
@@ -168,8 +194,10 @@ func Marc(raw []byte, limit uint32) bool {
 //
 // [glTF specification]: https://registry.khronos.org/glTF/specs/2.0/glTF-2.0.html
 // [IANA glTF entry]: https://www.iana.org/assignments/media-types/model/gltf-binary
-var GLB = prefix([]byte("\x67\x6C\x54\x46\x02\x00\x00\x00"),
-	[]byte("\x67\x6C\x54\x46\x01\x00\x00\x00"))
+func GLB(raw []byte, _ uint32) bool {
+	return bytes.HasPrefix(raw, []byte("\x67\x6C\x54\x46\x02\x00\x00\x00")) ||
+		bytes.HasPrefix(raw, []byte("\x67\x6C\x54\x46\x01\x00\x00\x00"))
+}

 // TzIf matches a Time Zone Information Format (TZif) file.
 // See more: https://tools.ietf.org/id/draft-murchison-tzdist-tzif-00.html#rfc.section.3
@@ -1,13 +1,21 @@
 package magic

-var (
-	// Sqlite matches an SQLite database file.
-	Sqlite = prefix([]byte{
+import "bytes"
+
+// Sqlite matches an SQLite database file.
+func Sqlite(raw []byte, _ uint32) bool {
+	return bytes.HasPrefix(raw, []byte{
 		0x53, 0x51, 0x4c, 0x69, 0x74, 0x65, 0x20, 0x66,
 		0x6f, 0x72, 0x6d, 0x61, 0x74, 0x20, 0x33, 0x00,
 	})
-	// MsAccessAce matches Microsoft Access dababase file.
-	MsAccessAce = offset([]byte("Standard ACE DB"), 4)
-	// MsAccessMdb matches legacy Microsoft Access database file (JET, 2003 and earlier).
-	MsAccessMdb = offset([]byte("Standard Jet DB"), 4)
-)
+}
+
+// MsAccessAce matches Microsoft Access dababase file.
+func MsAccessAce(raw []byte, _ uint32) bool {
+	return offset(raw, []byte("Standard ACE DB"), 4)
+}
+
+// MsAccessMdb matches legacy Microsoft Access database file (JET, 2003 and earlier).
+func MsAccessMdb(raw []byte, _ uint32) bool {
+	return offset(raw, []byte("Standard Jet DB"), 4)
+}
@@ -5,14 +5,31 @@ import (
 	"encoding/binary"
 )

-var (
-	// Fdf matches a Forms Data Format file.
-	Fdf = prefix([]byte("%FDF"))
-	// Mobi matches a Mobi file.
-	Mobi = offset([]byte("BOOKMOBI"), 60)
-	// Lit matches a Microsoft Lit file.
-	Lit = prefix([]byte("ITOLITLS"))
-)
+// Pdf matches a Portable Document Format file.
+// https://github.com/file/file/blob/11010cc805546a3e35597e67e1129a481aed40e8/magic/Magdir/pdf
+func Pdf(raw []byte, _ uint32) bool {
+	// usual pdf signature
+	return bytes.HasPrefix(raw, []byte("%PDF-")) ||
+		// new-line prefixed signature
+		bytes.HasPrefix(raw, []byte("\012%PDF-")) ||
+		// UTF-8 BOM prefixed signature
+		bytes.HasPrefix(raw, []byte("\xef\xbb\xbf%PDF-"))
+}
+
+// Fdf matches a Forms Data Format file.
+func Fdf(raw []byte, _ uint32) bool {
+	return bytes.HasPrefix(raw, []byte("%FDF"))
+}
+
+// Mobi matches a Mobi file.
+func Mobi(raw []byte, _ uint32) bool {
+	return offset(raw, []byte("BOOKMOBI"), 60)
+}
+
+// Lit matches a Microsoft Lit file.
+func Lit(raw []byte, _ uint32) bool {
+	return bytes.HasPrefix(raw, []byte("ITOLITLS"))
+}

 // PDF matches a Portable Document Format file.
 // The %PDF- header should be the first thing inside the file but many
@@ -4,14 +4,20 @@ import (
 	"bytes"
 )

-var (
-	// Woff matches a Web Open Font Format file.
-	Woff = prefix([]byte("wOFF"))
-	// Woff2 matches a Web Open Font Format version 2 file.
-	Woff2 = prefix([]byte("wOF2"))
-	// Otf matches an OpenType font file.
-	Otf = prefix([]byte{0x4F, 0x54, 0x54, 0x4F, 0x00})
-)
+// Woff matches a Web Open Font Format file.
+func Woff(raw []byte, _ uint32) bool {
+	return bytes.HasPrefix(raw, []byte("wOFF"))
+}
+
+// Woff2 matches a Web Open Font Format version 2 file.
+func Woff2(raw []byte, _ uint32) bool {
+	return bytes.HasPrefix(raw, []byte("wOF2"))
+}
+
+// Otf matches an OpenType font file.
+func Otf(raw []byte, _ uint32) bool {
+	return bytes.HasPrefix(raw, []byte{0x4F, 0x54, 0x54, 0x4F, 0x00})
+}

 // Ttf matches a TrueType font file.
 func Ttf(raw []byte, limit uint32) bool {
@@ -4,24 +4,33 @@ import (
 	"bytes"
 )

-var (
-	// AVIF matches an AV1 Image File Format still or animated.
-	// Wikipedia page seems outdated listing image/avif-sequence for animations.
-	// https://github.com/AOMediaCodec/av1-avif/issues/59
-	AVIF = ftyp([]byte("avif"), []byte("avis"))
-	// ThreeGP matches a 3GPP file.
-	ThreeGP = ftyp(
+// AVIF matches an AV1 Image File Format still or animated.
+// Wikipedia page seems outdated listing image/avif-sequence for animations.
+// https://github.com/AOMediaCodec/av1-avif/issues/59
+func AVIF(raw []byte, _ uint32) bool {
+	return ftyp(raw, []byte("avif"), []byte("avis"))
+}
+
+// ThreeGP matches a 3GPP file.
+func ThreeGP(raw []byte, _ uint32) bool {
+	return ftyp(raw,
 		[]byte("3gp1"), []byte("3gp2"), []byte("3gp3"), []byte("3gp4"),
 		[]byte("3gp5"), []byte("3gp6"), []byte("3gp7"), []byte("3gs7"),
 		[]byte("3ge6"), []byte("3ge7"), []byte("3gg6"),
 	)
-	// ThreeG2 matches a 3GPP2 file.
-	ThreeG2 = ftyp(
+}
+
+// ThreeG2 matches a 3GPP2 file.
+func ThreeG2(raw []byte, _ uint32) bool {
+	return ftyp(raw,
 		[]byte("3g24"), []byte("3g25"), []byte("3g26"), []byte("3g2a"),
 		[]byte("3g2b"), []byte("3g2c"), []byte("KDDI"),
 	)
-	// AMp4 matches an audio MP4 file.
-	AMp4 = ftyp(
+}
+
+// AMp4 matches an audio MP4 file.
+func AMp4(raw []byte, _ uint32) bool {
+	return ftyp(raw,
 		// audio for Adobe Flash Player 9+
 		[]byte("F4A "), []byte("F4B "),
 		// Apple iTunes AAC-LC (.M4A) Audio
@@ -31,33 +40,61 @@ var (
 		// Nero Digital AAC Audio
 		[]byte("NDAS"),
 	)
-	// Mqv matches a Sony / Mobile QuickTime  file.
-	Mqv = ftyp([]byte("mqt "))
-	// M4a matches an audio M4A file.
-	M4a = ftyp([]byte("M4A "))
-	// M4v matches an Appl4 M4V video file.
-	M4v = ftyp([]byte("M4V "), []byte("M4VH"), []byte("M4VP"))
-	// Heic matches a High Efficiency Image Coding (HEIC) file.
-	Heic = ftyp([]byte("heic"), []byte("heix"))
-	// HeicSequence matches a High Efficiency Image Coding (HEIC) file sequence.
-	HeicSequence = ftyp([]byte("hevc"), []byte("hevx"))
-	// Heif matches a High Efficiency Image File Format (HEIF) file.
-	Heif = ftyp([]byte("mif1"), []byte("heim"), []byte("heis"), []byte("avic"))
-	// HeifSequence matches a High Efficiency Image File Format (HEIF) file sequence.
-	HeifSequence = ftyp([]byte("msf1"), []byte("hevm"), []byte("hevs"), []byte("avcs"))
-	// Mj2 matches a Motion JPEG 2000 file: https://en.wikipedia.org/wiki/Motion_JPEG_2000.
-	Mj2 = ftyp([]byte("mj2s"), []byte("mjp2"), []byte("MFSM"), []byte("MGSV"))
-	// Dvb matches a Digital Video Broadcasting file: https://dvb.org.
-	// https://cconcolato.github.io/mp4ra/filetype.html
-	// https://github.com/file/file/blob/512840337ead1076519332d24fefcaa8fac36e06/magic/Magdir/animation#L135-L154
-	Dvb = ftyp(
+}
+
+// Mqv matches a Sony / Mobile QuickTime  file.
+func Mqv(raw []byte, _ uint32) bool {
+	return ftyp(raw, []byte("mqt "))
+}
+
+// M4a matches an audio M4A file.
+func M4a(raw []byte, _ uint32) bool {
+	return ftyp(raw, []byte("M4A "))
+}
+
+// M4v matches an Appl4 M4V video file.
+func M4v(raw []byte, _ uint32) bool {
+	return ftyp(raw, []byte("M4V "), []byte("M4VH"), []byte("M4VP"))
+}
+
+// Heic matches a High Efficiency Image Coding (HEIC) file.
+func Heic(raw []byte, _ uint32) bool {
+	return ftyp(raw, []byte("heic"), []byte("heix"))
+}
+
+// HeicSequence matches a High Efficiency Image Coding (HEIC) file sequence.
+func HeicSequence(raw []byte, _ uint32) bool {
+	return ftyp(raw, []byte("hevc"), []byte("hevx"))
+}
+
+// Heif matches a High Efficiency Image File Format (HEIF) file.
+func Heif(raw []byte, _ uint32) bool {
+	return ftyp(raw, []byte("mif1"), []byte("heim"), []byte("heis"), []byte("avic"))
+}
+
+// HeifSequence matches a High Efficiency Image File Format (HEIF) file sequence.
+func HeifSequence(raw []byte, _ uint32) bool {
+	return ftyp(raw, []byte("msf1"), []byte("hevm"), []byte("hevs"), []byte("avcs"))
+}
+
+// Mj2 matches a Motion JPEG 2000 file: https://en.wikipedia.org/wiki/Motion_JPEG_2000.
+func Mj2(raw []byte, _ uint32) bool {
+	return ftyp(raw, []byte("mj2s"), []byte("mjp2"), []byte("MFSM"), []byte("MGSV"))
+}
+
+// Dvb matches a Digital Video Broadcasting file: https://dvb.org.
+// https://cconcolato.github.io/mp4ra/filetype.html
+// https://github.com/file/file/blob/512840337ead1076519332d24fefcaa8fac36e06/magic/Magdir/animation#L135-L154
+func Dvb(raw []byte, _ uint32) bool {
+	return ftyp(raw,
 		[]byte("dby1"), []byte("dsms"), []byte("dts1"), []byte("dts2"),
 		[]byte("dts3"), []byte("dxo "), []byte("dmb1"), []byte("dmpf"),
 		[]byte("drc1"), []byte("dv1a"), []byte("dv1b"), []byte("dv2a"),
 		[]byte("dv2b"), []byte("dv3a"), []byte("dv3b"), []byte("dvr1"),
 		[]byte("dvt1"), []byte("emsg"))
-	// TODO: add support for remaining video formats at ftyps.com.
-)
+}
+
+// TODO: add support for remaining video formats at ftyps.com.

 // QuickTime matches a QuickTime File Format file.
 // https://www.loc.gov/preservation/digital/formats/fdd/fdd000052.shtml
@@ -2,66 +2,127 @@ package magic

 import "bytes"

-var (
-	// Png matches a Portable Network Graphics file.
-	// https://www.w3.org/TR/PNG/
-	Png = prefix([]byte{0x89, 0x50, 0x4E, 0x47, 0x0D, 0x0A, 0x1A, 0x0A})
-	// Apng matches an Animated Portable Network Graphics file.
-	// https://wiki.mozilla.org/APNG_Specification
-	Apng = offset([]byte("acTL"), 37)
-	// Jpg matches a Joint Photographic Experts Group file.
-	Jpg = prefix([]byte{0xFF, 0xD8, 0xFF})
-	// Jp2 matches a JPEG 2000 Image file (ISO 15444-1).
-	Jp2 = jpeg2k([]byte{0x6a, 0x70, 0x32, 0x20})
-	// Jpx matches a JPEG 2000 Image file (ISO 15444-2).
-	Jpx = jpeg2k([]byte{0x6a, 0x70, 0x78, 0x20})
-	// Jpm matches a JPEG 2000 Image file (ISO 15444-6).
-	Jpm = jpeg2k([]byte{0x6a, 0x70, 0x6D, 0x20})
-	// Gif matches a Graphics Interchange Format file.
-	Gif = prefix([]byte("GIF87a"), []byte("GIF89a"))
-	// Bmp matches a bitmap image file.
-	Bmp = prefix([]byte{0x42, 0x4D})
-	// Ps matches a PostScript file.
-	Ps = prefix([]byte("%!PS-Adobe-"))
-	// Psd matches a Photoshop Document file.
-	Psd = prefix([]byte("8BPS"))
-	// Ico matches an ICO file.
-	Ico = prefix([]byte{0x00, 0x00, 0x01, 0x00}, []byte{0x00, 0x00, 0x02, 0x00})
-	// Icns matches an ICNS (Apple Icon Image format) file.
-	Icns = prefix([]byte("icns"))
-	// Tiff matches a Tagged Image File Format file.
-	Tiff = prefix([]byte{0x49, 0x49, 0x2A, 0x00}, []byte{0x4D, 0x4D, 0x00, 0x2A})
-	// Bpg matches a Better Portable Graphics file.
-	Bpg = prefix([]byte{0x42, 0x50, 0x47, 0xFB})
-	// Xcf matches GIMP image data.
-	Xcf = prefix([]byte("gimp xcf"))
-	// Pat matches GIMP pattern data.
-	Pat = offset([]byte("GPAT"), 20)
-	// Gbr matches GIMP brush data.
-	Gbr = offset([]byte("GIMP"), 20)
-	// Hdr matches Radiance HDR image.
-	// https://web.archive.org/web/20060913152809/http://local.wasp.uwa.edu.au/~pbourke/dataformats/pic/
-	Hdr = prefix([]byte("#?RADIANCE\n"))
-	// Xpm matches X PixMap image data.
-	Xpm = prefix([]byte{0x2F, 0x2A, 0x20, 0x58, 0x50, 0x4D, 0x20, 0x2A, 0x2F})
-	// Jxs matches a JPEG XS coded image file (ISO/IEC 21122-3).
-	Jxs = prefix([]byte{0x00, 0x00, 0x00, 0x0C, 0x4A, 0x58, 0x53, 0x20, 0x0D, 0x0A, 0x87, 0x0A})
-	// Jxr matches Microsoft HD JXR photo file.
-	Jxr = prefix([]byte{0x49, 0x49, 0xBC, 0x01})
-)
+// Png matches a Portable Network Graphics file.
+// https://www.w3.org/TR/PNG/
+func Png(raw []byte, _ uint32) bool {
+	return bytes.HasPrefix(raw, []byte{0x89, 0x50, 0x4E, 0x47, 0x0D, 0x0A, 0x1A, 0x0A})
+}

-func jpeg2k(sig []byte) Detector {
-	return func(raw []byte, _ uint32) bool {
-		if len(raw) < 24 {
-			return false
-		}
+// Apng matches an Animated Portable Network Graphics file.
+// https://wiki.mozilla.org/APNG_Specification
+func Apng(raw []byte, _ uint32) bool {
+	return offset(raw, []byte("acTL"), 37)
+}

-		if !bytes.Equal(raw[4:8], []byte{0x6A, 0x50, 0x20, 0x20}) &&
-			!bytes.Equal(raw[4:8], []byte{0x6A, 0x50, 0x32, 0x20}) {
-			return false
-		}
-		return bytes.Equal(raw[20:24], sig)
+// Jpg matches a Joint Photographic Experts Group file.
+func Jpg(raw []byte, _ uint32) bool {
+	return bytes.HasPrefix(raw, []byte{0xFF, 0xD8, 0xFF})
+}
+
+// Jp2 matches a JPEG 2000 Image file (ISO 15444-1).
+func Jp2(raw []byte, _ uint32) bool {
+	return jpeg2k(raw, []byte{0x6a, 0x70, 0x32, 0x20})
+}
+
+// Jpx matches a JPEG 2000 Image file (ISO 15444-2).
+func Jpx(raw []byte, _ uint32) bool {
+	return jpeg2k(raw, []byte{0x6a, 0x70, 0x78, 0x20})
+}
+
+// Jpm matches a JPEG 2000 Image file (ISO 15444-6).
+func Jpm(raw []byte, _ uint32) bool {
+	return jpeg2k(raw, []byte{0x6a, 0x70, 0x6D, 0x20})
+}
+
+// Gif matches a Graphics Interchange Format file.
+func Gif(raw []byte, _ uint32) bool {
+	return bytes.HasPrefix(raw, []byte("GIF87a")) ||
+		bytes.HasPrefix(raw, []byte("GIF89a"))
+}
+
+// Bmp matches a bitmap image file.
+func Bmp(raw []byte, _ uint32) bool {
+	return bytes.HasPrefix(raw, []byte{0x42, 0x4D})
+}
+
+// Ps matches a PostScript file.
+func Ps(raw []byte, _ uint32) bool {
+	return bytes.HasPrefix(raw, []byte("%!PS-Adobe-"))
+}
+
+// Psd matches a Photoshop Document file.
+func Psd(raw []byte, _ uint32) bool {
+	return bytes.HasPrefix(raw, []byte("8BPS"))
+}
+
+// Ico matches an ICO file.
+func Ico(raw []byte, _ uint32) bool {
+	return bytes.HasPrefix(raw, []byte{0x00, 0x00, 0x01, 0x00}) ||
+		bytes.HasPrefix(raw, []byte{0x00, 0x00, 0x02, 0x00})
+}
+
+// Icns matches an ICNS (Apple Icon Image format) file.
+func Icns(raw []byte, _ uint32) bool {
+	return bytes.HasPrefix(raw, []byte("icns"))
+}
+
+// Tiff matches a Tagged Image File Format file.
+func Tiff(raw []byte, _ uint32) bool {
+	return bytes.HasPrefix(raw, []byte{0x49, 0x49, 0x2A, 0x00}) ||
+		bytes.HasPrefix(raw, []byte{0x4D, 0x4D, 0x00, 0x2A})
+}
+
+// Bpg matches a Better Portable Graphics file.
+func Bpg(raw []byte, _ uint32) bool {
+	return bytes.HasPrefix(raw, []byte{0x42, 0x50, 0x47, 0xFB})
+}
+
+// Xcf matches GIMP image data.
+func Xcf(raw []byte, _ uint32) bool {
+	return bytes.HasPrefix(raw, []byte("gimp xcf"))
+}
+
+// Pat matches GIMP pattern data.
+func Pat(raw []byte, _ uint32) bool {
+	return offset(raw, []byte("GPAT"), 20)
+}
+
+// Gbr matches GIMP brush data.
+func Gbr(raw []byte, _ uint32) bool {
+	return offset(raw, []byte("GIMP"), 20)
+}
+
+// Hdr matches Radiance HDR image.
+// https://web.archive.org/web/20060913152809/http://local.wasp.uwa.edu.au/~pbourke/dataformats/pic/
+func Hdr(raw []byte, _ uint32) bool {
+	return bytes.HasPrefix(raw, []byte("#?RADIANCE\n"))
+}
+
+// Xpm matches X PixMap image data.
+func Xpm(raw []byte, _ uint32) bool {
+	return bytes.HasPrefix(raw, []byte{0x2F, 0x2A, 0x20, 0x58, 0x50, 0x4D, 0x20, 0x2A, 0x2F})
+}
+
+// Jxs matches a JPEG XS coded image file (ISO/IEC 21122-3).
+func Jxs(raw []byte, _ uint32) bool {
+	return bytes.HasPrefix(raw, []byte{0x00, 0x00, 0x00, 0x0C, 0x4A, 0x58, 0x53, 0x20, 0x0D, 0x0A, 0x87, 0x0A})
+}
+
+// Jxr matches Microsoft HD JXR photo file.
+func Jxr(raw []byte, _ uint32) bool {
+	return bytes.HasPrefix(raw, []byte{0x49, 0x49, 0xBC, 0x01})
+}
+
+func jpeg2k(raw []byte, sig []byte) bool {
+	if len(raw) < 24 {
+		return false
 	}
+
+	if !bytes.Equal(raw[4:8], []byte{0x6A, 0x50, 0x20, 0x20}) &&
+		!bytes.Equal(raw[4:8], []byte{0x6A, 0x50, 0x32, 0x20}) {
+		return false
+	}
+	return bytes.Equal(raw[20:24], sig)
 }

 // Webp matches a WebP file.
@@ -108,3 +169,20 @@ func Jxl(raw []byte, _ uint32) bool {
 	return bytes.HasPrefix(raw, []byte{0xFF, 0x0A}) ||
 		bytes.HasPrefix(raw, []byte("\x00\x00\x00\x0cJXL\x20\x0d\x0a\x87\x0a"))
 }
+
+// DXF matches Drawing Exchange Format AutoCAD file.
+// There does not seem to be a clear specification and the files in the wild
+// differ wildly.
+// https://images.autodesk.com/adsk/files/autocad_2012_pdf_dxf-reference_enu.pdf
+//
+// I collected these signatures by downloading a few dozen files from
+// http://cd.textfiles.com/amigaenv/DXF/OBJEKTE/ and
+// https://sembiance.com/fileFormatSamples/poly/dxf/ and then
+// xxd -l 16 {} | sort | uniq.
+// These signatures are only for the ASCII version of DXF. There is a binary version too.
+func DXF(raw []byte, _ uint32) bool {
+	return bytes.HasPrefix(raw, []byte("  0\x0ASECTION\x0A")) ||
+		bytes.HasPrefix(raw, []byte("  0\x0D\x0ASECTION\x0D\x0A")) ||
+		bytes.HasPrefix(raw, []byte("0\x0ASECTION\x0A")) ||
+		bytes.HasPrefix(raw, []byte("0\x0D\x0ASECTION\x0D\x0A"))
+}
@@ -3,7 +3,6 @@ package magic

 import (
 	"bytes"
-	"fmt"

 	"github.com/gabriel-vasile/mimetype/internal/scan"
 )
@@ -22,37 +21,20 @@ type (
 	}
 )

-// prefix creates a Detector which returns true if any of the provided signatures
-// is the prefix of the raw input.
-func prefix(sigs ...[]byte) Detector {
-	return func(raw []byte, limit uint32) bool {
-		for _, s := range sigs {
-			if bytes.HasPrefix(raw, s) {
-				return true
-			}
-		}
-		return false
-	}
-}
-
-// offset creates a Detector which returns true if the provided signature can be
+// offset returns true if the provided signature can be
 // found at offset in the raw input.
-func offset(sig []byte, offset int) Detector {
-	return func(raw []byte, limit uint32) bool {
-		return len(raw) > offset && bytes.HasPrefix(raw[offset:], sig)
-	}
+func offset(raw []byte, sig []byte, offset int) bool {
+	return len(raw) > offset && bytes.HasPrefix(raw[offset:], sig)
 }

 // ciPrefix is like prefix but the check is case insensitive.
-func ciPrefix(sigs ...[]byte) Detector {
-	return func(raw []byte, limit uint32) bool {
-		for _, s := range sigs {
-			if ciCheck(s, raw) {
-				return true
-			}
+func ciPrefix(raw []byte, sigs ...[]byte) bool {
+	for _, s := range sigs {
+		if ciCheck(s, raw) {
+			return true
 		}
-		return false
 	}
+	return false
 }
 func ciCheck(sig, raw []byte) bool {
 	if len(raw) < len(sig)+1 {
@@ -72,22 +54,18 @@ func ciCheck(sig, raw []byte) bool {
 	return true
 }

-// xml creates a Detector which returns true if any of the provided XML signatures
-// matches the raw input.
-func xml(sigs ...xmlSig) Detector {
-	return func(raw []byte, limit uint32) bool {
-		b := scan.Bytes(raw)
-		b.TrimLWS()
-		if len(b) == 0 {
-			return false
-		}
-		for _, s := range sigs {
-			if xmlCheck(s, b) {
-				return true
-			}
-		}
+// xml returns true if any of the provided XML signatures matches the raw input.
+func xml(b scan.Bytes, sigs ...xmlSig) bool {
+	b.TrimLWS()
+	if len(b) == 0 {
 		return false
 	}
+	for _, s := range sigs {
+		if xmlCheck(s, b) {
+			return true
+		}
+	}
+	return false
 }
 func xmlCheck(sig xmlSig, raw []byte) bool {
 	raw = raw[:min(len(raw), 512)]
@@ -103,28 +81,24 @@ func xmlCheck(sig xmlSig, raw []byte) bool {
 	return localNameIndex != -1 && localNameIndex < bytes.Index(raw, sig.xmlns)
 }

-// markup creates a Detector which returns true is any of the HTML signatures
-// matches the raw input.
-func markup(sigs ...[]byte) Detector {
-	return func(raw []byte, limit uint32) bool {
-		b := scan.Bytes(raw)
-		if bytes.HasPrefix(b, []byte{0xEF, 0xBB, 0xBF}) {
-			// We skip the UTF-8 BOM if present to ensure we correctly
-			// process any leading whitespace. The presence of the BOM
-			// is taken into account during charset detection in charset.go.
-			b.Advance(3)
-		}
-		b.TrimLWS()
-		if len(b) == 0 {
-			return false
-		}
-		for _, s := range sigs {
-			if markupCheck(s, b) {
-				return true
-			}
-		}
+// markup returns true is any of the HTML signatures matches the raw input.
+func markup(b scan.Bytes, sigs ...[]byte) bool {
+	if bytes.HasPrefix(b, []byte{0xEF, 0xBB, 0xBF}) {
+		// We skip the UTF-8 BOM if present to ensure we correctly
+		// process any leading whitespace. The presence of the BOM
+		// is taken into account during charset detection in charset.go.
+		b.Advance(3)
+	}
+	b.TrimLWS()
+	if len(b) == 0 {
 		return false
 	}
+	for _, s := range sigs {
+		if markupCheck(s, b) {
+			return true
+		}
+	}
+	return false
 }
 func markupCheck(sig, raw []byte) bool {
 	if len(raw) < len(sig)+1 {
@@ -149,29 +123,17 @@ func markupCheck(sig, raw []byte) bool {
 	return true
 }

-// ftyp creates a Detector which returns true if any of the FTYP signatures
-// matches the raw input.
-func ftyp(sigs ...[]byte) Detector {
-	return func(raw []byte, limit uint32) bool {
-		if len(raw) < 12 {
-			return false
-		}
-		for _, s := range sigs {
-			if bytes.Equal(raw[8:12], s) {
-				return true
-			}
-		}
+// ftyp returns true if any of the FTYP signatures matches the raw input.
+func ftyp(raw []byte, sigs ...[]byte) bool {
+	if len(raw) < 12 {
 		return false
 	}
-}
-
-func newXMLSig(localName, xmlns string) xmlSig {
-	ret := xmlSig{xmlns: []byte(xmlns)}
-	if localName != "" {
-		ret.localName = []byte(fmt.Sprintf("<%s", localName))
+	for _, s := range sigs {
+		if bytes.Equal(raw[8:12], s) {
+			return true
+		}
 	}
-
-	return ret
+	return false
 }

 // A valid shebang starts with the "#!" characters,
@@ -184,29 +146,17 @@ func newXMLSig(localName, xmlns string) xmlSig {
 //	#! /usr/bin/env php
 //
 // /usr/bin/env is the interpreter, php is the first and only argument.
-func shebang(sigs ...[]byte) Detector {
-	return func(raw []byte, limit uint32) bool {
-		b := scan.Bytes(raw)
-		line := b.Line()
-		for _, s := range sigs {
-			if shebangCheck(s, line) {
-				return true
-			}
+func shebang(b scan.Bytes, matchFlags scan.Flags, sigs ...[]byte) bool {
+	line := b.Line()
+	if len(line) < 2 || line[0] != '#' || line[1] != '!' {
+		return false
+	}
+	line = line[2:]
+	line.TrimLWS()
+	for _, s := range sigs {
+		if line.Match(s, matchFlags) != -1 {
+			return true
 		}
-		return false
 	}
-}
-
-func shebangCheck(sig []byte, raw scan.Bytes) bool {
-	if len(raw) < len(sig)+2 {
-		return false
-	}
-	if raw[0] != '#' || raw[1] != '!' {
-		return false
-	}
-
-	raw.Advance(2) // skip #! we checked above
-	raw.TrimLWS()
-	raw.TrimRWS()
-	return bytes.Equal(raw, sig)
+	return false
 }
@@ -44,17 +44,6 @@ func Ole(raw []byte, limit uint32) bool {
 	return bytes.HasPrefix(raw, []byte{0xD0, 0xCF, 0x11, 0xE0, 0xA1, 0xB1, 0x1A, 0xE1})
 }

-// Aaf matches an Advanced Authoring Format file.
-// See: https://pyaaf.readthedocs.io/en/latest/about.html
-// See: https://en.wikipedia.org/wiki/Advanced_Authoring_Format
-func Aaf(raw []byte, limit uint32) bool {
-	if len(raw) < 31 {
-		return false
-	}
-	return bytes.HasPrefix(raw[8:], []byte{0x41, 0x41, 0x46, 0x42, 0x0D, 0x00, 0x4F, 0x4D}) &&
-		(raw[30] == 0x09 || raw[30] == 0x0C)
-}
-
 // Doc matches a Microsoft Word 97-2003 file.
 // See: https://github.com/decalage2/oletools/blob/412ee36ae45e70f42123e835871bac956d958461/oletools/common/clsid.py
 func Doc(raw []byte, _ uint32) bool {
@@ -209,3 +198,14 @@ func matchOleClsid(in []byte, clsid []byte) bool {

 	return bytes.HasPrefix(in[clsidOffset:], clsid)
 }
+
+// WPD matches a WordPerfect document.
+func WPD(raw []byte, _ uint32) bool {
+	if len(raw) < 10 {
+		return false
+	}
+	if !bytes.HasPrefix(raw, []byte("\xffWPC")) {
+		return false
+	}
+	return raw[8] == 1 && raw[9] == 10
+}
@@ -10,9 +10,9 @@ import (
 	"github.com/gabriel-vasile/mimetype/internal/scan"
 )

-var (
-	// HTML matches a Hypertext Markup Language file.
-	HTML = markup(
+// HTML matches a Hypertext Markup Language file.
+func HTML(raw []byte, _ uint32) bool {
+	return markup(raw,
 		[]byte("<!DOCTYPE HTML"),
 		[]byte("<HTML"),
 		[]byte("<HEAD"),
@@ -31,133 +31,254 @@ var (
 		[]byte("<P"),
 		[]byte("<!--"),
 	)
-	// XML matches an Extensible Markup Language file.
-	XML = markup([]byte("<?XML"))
-	// Owl2 matches an Owl ontology file.
-	Owl2 = xml(newXMLSig("Ontology", `xmlns="http://www.w3.org/2002/07/owl#"`))
-	// Rss matches a Rich Site Summary file.
-	Rss = xml(newXMLSig("rss", ""))
-	// Atom matches an Atom Syndication Format file.
-	Atom = xml(newXMLSig("feed", `xmlns="http://www.w3.org/2005/Atom"`))
-	// Kml matches a Keyhole Markup Language file.
-	Kml = xml(
-		newXMLSig("kml", `xmlns="http://www.opengis.net/kml/2.2"`),
-		newXMLSig("kml", `xmlns="http://earth.google.com/kml/2.0"`),
-		newXMLSig("kml", `xmlns="http://earth.google.com/kml/2.1"`),
-		newXMLSig("kml", `xmlns="http://earth.google.com/kml/2.2"`),
+}
+
+// XML matches an Extensible Markup Language file.
+func XML(raw []byte, _ uint32) bool {
+	return markup(raw, []byte("<?XML"))
+}
+
+// Owl2 matches an Owl ontology file.
+func Owl2(raw []byte, _ uint32) bool {
+	return xml(raw,
+		xmlSig{[]byte("<Ontology"), []byte(`xmlns="http://www.w3.org/2002/07/owl#"`)},
 	)
-	// Xliff matches a XML Localization Interchange File Format file.
-	Xliff = xml(newXMLSig("xliff", `xmlns="urn:oasis:names:tc:xliff:document:1.2"`))
-	// Collada matches a COLLAborative Design Activity file.
-	Collada = xml(newXMLSig("COLLADA", `xmlns="http://www.collada.org/2005/11/COLLADASchema"`))
-	// Gml matches a Geography Markup Language file.
-	Gml = xml(
-		newXMLSig("", `xmlns:gml="http://www.opengis.net/gml"`),
-		newXMLSig("", `xmlns:gml="http://www.opengis.net/gml/3.2"`),
-		newXMLSig("", `xmlns:gml="http://www.opengis.net/gml/3.3/exr"`),
+}
+
+// Rss matches a Rich Site Summary file.
+func Rss(raw []byte, _ uint32) bool {
+	return xml(raw,
+		xmlSig{[]byte("<rss"), []byte{}},
 	)
-	// Gpx matches a GPS Exchange Format file.
-	Gpx = xml(newXMLSig("gpx", `xmlns="http://www.topografix.com/GPX/1/1"`))
-	// Tcx matches a Training Center XML file.
-	Tcx = xml(newXMLSig("TrainingCenterDatabase", `xmlns="http://www.garmin.com/xmlschemas/TrainingCenterDatabase/v2"`))
-	// X3d matches an Extensible 3D Graphics file.
-	X3d = xml(newXMLSig("X3D", `xmlns:xsd="http://www.w3.org/2001/XMLSchema-instance"`))
-	// Amf matches an Additive Manufacturing XML file.
-	Amf = xml(newXMLSig("amf", ""))
-	// Threemf matches a 3D Manufacturing Format file.
-	Threemf = xml(newXMLSig("model", `xmlns="http://schemas.microsoft.com/3dmanufacturing/core/2015/02"`))
-	// Xfdf matches a XML Forms Data Format file.
-	Xfdf = xml(newXMLSig("xfdf", `xmlns="http://ns.adobe.com/xfdf/"`))
-	// VCard matches a Virtual Contact File.
-	VCard = ciPrefix([]byte("BEGIN:VCARD\n"), []byte("BEGIN:VCARD\r\n"))
-	// ICalendar matches a iCalendar file.
-	ICalendar = ciPrefix([]byte("BEGIN:VCALENDAR\n"), []byte("BEGIN:VCALENDAR\r\n"))
-	phpPageF  = ciPrefix(
+}
+
+// Atom matches an Atom Syndication Format file.
+func Atom(raw []byte, _ uint32) bool {
+	return xml(raw,
+		xmlSig{[]byte("<feed"), []byte(`xmlns="http://www.w3.org/2005/Atom"`)},
+	)
+}
+
+// Kml matches a Keyhole Markup Language file.
+func Kml(raw []byte, _ uint32) bool {
+	return xml(raw,
+		xmlSig{[]byte("<kml"), []byte(`xmlns="http://www.opengis.net/kml/2.2"`)},
+		xmlSig{[]byte("<kml"), []byte(`xmlns="http://earth.google.com/kml/2.0"`)},
+		xmlSig{[]byte("<kml"), []byte(`xmlns="http://earth.google.com/kml/2.1"`)},
+		xmlSig{[]byte("<kml"), []byte(`xmlns="http://earth.google.com/kml/2.2"`)},
+	)
+}
+
+// Xliff matches a XML Localization Interchange File Format file.
+func Xliff(raw []byte, _ uint32) bool {
+	return xml(raw,
+		xmlSig{[]byte("<xliff"), []byte(`xmlns="urn:oasis:names:tc:xliff:document:1.2"`)},
+	)
+}
+
+// Collada matches a COLLAborative Design Activity file.
+func Collada(raw []byte, _ uint32) bool {
+	return xml(raw,
+		xmlSig{[]byte("<COLLADA"), []byte(`xmlns="http://www.collada.org/2005/11/COLLADASchema"`)},
+	)
+}
+
+// Gml matches a Geography Markup Language file.
+func Gml(raw []byte, _ uint32) bool {
+	return xml(raw,
+		xmlSig{[]byte{}, []byte(`xmlns:gml="http://www.opengis.net/gml"`)},
+		xmlSig{[]byte{}, []byte(`xmlns:gml="http://www.opengis.net/gml/3.2"`)},
+		xmlSig{[]byte{}, []byte(`xmlns:gml="http://www.opengis.net/gml/3.3/exr"`)},
+	)
+}
+
+// Gpx matches a GPS Exchange Format file.
+func Gpx(raw []byte, _ uint32) bool {
+	return xml(raw,
+		xmlSig{[]byte("<gpx"), []byte(`xmlns="http://www.topografix.com/GPX/1/1"`)},
+	)
+}
+
+// Tcx matches a Training Center XML file.
+func Tcx(raw []byte, _ uint32) bool {
+	return xml(raw,
+		xmlSig{[]byte("<TrainingCenterDatabase"), []byte(`xmlns="http://www.garmin.com/xmlschemas/TrainingCenterDatabase/v2"`)},
+	)
+}
+
+// X3d matches an Extensible 3D Graphics file.
+func X3d(raw []byte, _ uint32) bool {
+	return xml(raw,
+		xmlSig{[]byte("<X3D"), []byte(`xmlns:xsd="http://www.w3.org/2001/XMLSchema-instance"`)},
+	)
+}
+
+// Amf matches an Additive Manufacturing XML file.
+func Amf(raw []byte, _ uint32) bool {
+	return xml(raw, xmlSig{[]byte("<amf"), []byte{}})
+}
+
+// Threemf matches a 3D Manufacturing Format file.
+func Threemf(raw []byte, _ uint32) bool {
+	return xml(raw,
+		xmlSig{[]byte("<model"), []byte(`xmlns="http://schemas.microsoft.com/3dmanufacturing/core/2015/02"`)},
+	)
+}
+
+// Xfdf matches a XML Forms Data Format file.
+func Xfdf(raw []byte, _ uint32) bool {
+	return xml(raw, xmlSig{[]byte("<xfdf"), []byte(`xmlns="http://ns.adobe.com/xfdf/"`)})
+}
+
+// VCard matches a Virtual Contact File.
+func VCard(raw []byte, _ uint32) bool {
+	return ciPrefix(raw, []byte("BEGIN:VCARD\n"), []byte("BEGIN:VCARD\r\n"))
+}
+
+// ICalendar matches a iCalendar file.
+func ICalendar(raw []byte, _ uint32) bool {
+	return ciPrefix(raw, []byte("BEGIN:VCALENDAR\n"), []byte("BEGIN:VCALENDAR\r\n"))
+}
+func phpPageF(raw []byte, _ uint32) bool {
+	return ciPrefix(raw,
 		[]byte("<?PHP"),
 		[]byte("<?\n"),
 		[]byte("<?\r"),
 		[]byte("<? "),
 	)
-	phpScriptF = shebang(
+}
+func phpScriptF(raw []byte, _ uint32) bool {
+	return shebang(raw,
+		scan.CompactWS,
 		[]byte("/usr/local/bin/php"),
 		[]byte("/usr/bin/php"),
 		[]byte("/usr/bin/env php"),
+		[]byte("/usr/bin/env -S php"),
 	)
-	// Js matches a Javascript file.
-	Js = shebang(
+}
+
+// Js matches a Javascript file.
+func Js(raw []byte, _ uint32) bool {
+	return shebang(raw,
+		scan.CompactWS,
 		[]byte("/bin/node"),
 		[]byte("/usr/bin/node"),
 		[]byte("/bin/nodejs"),
 		[]byte("/usr/bin/nodejs"),
 		[]byte("/usr/bin/env node"),
+		[]byte("/usr/bin/env -S node"),
 		[]byte("/usr/bin/env nodejs"),
+		[]byte("/usr/bin/env -S nodejs"),
 	)
-	// Lua matches a Lua programming language file.
-	Lua = shebang(
+}
+
+// Lua matches a Lua programming language file.
+func Lua(raw []byte, _ uint32) bool {
+	return shebang(raw,
+		scan.CompactWS|scan.FullWord,
 		[]byte("/usr/bin/lua"),
 		[]byte("/usr/local/bin/lua"),
 		[]byte("/usr/bin/env lua"),
+		[]byte("/usr/bin/env -S lua"),
 	)
-	// Perl matches a Perl programming language file.
-	Perl = shebang(
+}
+
+// Perl matches a Perl programming language file.
+func Perl(raw []byte, _ uint32) bool {
+	return shebang(raw,
+		scan.CompactWS|scan.FullWord,
 		[]byte("/usr/bin/perl"),
 		[]byte("/usr/bin/env perl"),
+		[]byte("/usr/bin/env -S perl"),
 	)
-	// Python matches a Python programming language file.
-	Python = shebang(
+}
+
+// Python matches a Python programming language file.
+func Python(raw []byte, _ uint32) bool {
+	return shebang(raw,
+		scan.CompactWS,
 		[]byte("/usr/bin/python"),
 		[]byte("/usr/local/bin/python"),
 		[]byte("/usr/bin/env python"),
+		[]byte("/usr/bin/env -S python"),
 		[]byte("/usr/bin/python2"),
 		[]byte("/usr/local/bin/python2"),
 		[]byte("/usr/bin/env python2"),
+		[]byte("/usr/bin/env -S python2"),
 		[]byte("/usr/bin/python3"),
 		[]byte("/usr/local/bin/python3"),
 		[]byte("/usr/bin/env python3"),
+		[]byte("/usr/bin/env -S python3"),
 	)
-	// Ruby matches a Ruby programming language file.
-	Ruby = shebang(
+
+}
+
+// Ruby matches a Ruby programming language file.
+func Ruby(raw []byte, _ uint32) bool {
+	return shebang(raw,
+		scan.CompactWS,
 		[]byte("/usr/bin/ruby"),
 		[]byte("/usr/local/bin/ruby"),
 		[]byte("/usr/bin/env ruby"),
+		[]byte("/usr/bin/env -S ruby"),
 	)
-	// Tcl matches a Tcl programming language file.
-	Tcl = shebang(
+}
+
+// Tcl matches a Tcl programming language file.
+func Tcl(raw []byte, _ uint32) bool {
+	return shebang(raw,
+		scan.CompactWS,
 		[]byte("/usr/bin/tcl"),
 		[]byte("/usr/local/bin/tcl"),
 		[]byte("/usr/bin/env tcl"),
+		[]byte("/usr/bin/env -S tcl"),
 		[]byte("/usr/bin/tclsh"),
 		[]byte("/usr/local/bin/tclsh"),
 		[]byte("/usr/bin/env tclsh"),
+		[]byte("/usr/bin/env -S tclsh"),
 		[]byte("/usr/bin/wish"),
 		[]byte("/usr/local/bin/wish"),
 		[]byte("/usr/bin/env wish"),
+		[]byte("/usr/bin/env -S wish"),
 	)
-	// Rtf matches a Rich Text Format file.
-	Rtf = prefix([]byte("{\\rtf"))
-	// Shell matches a shell script file.
-	Shell = shebang(
+}
+
+// Rtf matches a Rich Text Format file.
+func Rtf(raw []byte, _ uint32) bool {
+	return bytes.HasPrefix(raw, []byte("{\\rtf"))
+}
+
+// Shell matches a shell script file.
+func Shell(raw []byte, _ uint32) bool {
+	return shebang(raw,
+		scan.CompactWS|scan.FullWord,
 		[]byte("/bin/sh"),
 		[]byte("/bin/bash"),
 		[]byte("/usr/local/bin/bash"),
 		[]byte("/usr/bin/env bash"),
+		[]byte("/usr/bin/env -S bash"),
 		[]byte("/bin/csh"),
 		[]byte("/usr/local/bin/csh"),
 		[]byte("/usr/bin/env csh"),
+		[]byte("/usr/bin/env -S csh"),
 		[]byte("/bin/dash"),
 		[]byte("/usr/local/bin/dash"),
 		[]byte("/usr/bin/env dash"),
+		[]byte("/usr/bin/env -S dash"),
 		[]byte("/bin/ksh"),
 		[]byte("/usr/local/bin/ksh"),
 		[]byte("/usr/bin/env ksh"),
+		[]byte("/usr/bin/env -S ksh"),
 		[]byte("/bin/tcsh"),
 		[]byte("/usr/local/bin/tcsh"),
 		[]byte("/usr/bin/env tcsh"),
+		[]byte("/usr/bin/env -S tcsh"),
 		[]byte("/bin/zsh"),
 		[]byte("/usr/local/bin/zsh"),
 		[]byte("/usr/bin/env zsh"),
+		[]byte("/usr/bin/env -S zsh"),
 	)
-)
+}

 // Text matches a plain text file.
 //
@@ -183,10 +304,14 @@ func Text(raw []byte, _ uint32) bool {

 // XHTML matches an XHTML file. This check depends on the XML check to have passed.
 func XHTML(raw []byte, limit uint32) bool {
-	raw = raw[:min(len(raw), 4096)]
+	raw = raw[:min(len(raw), 1024)]
 	b := scan.Bytes(raw)
-	return b.Search([]byte("<!DOCTYPE HTML"), scan.CompactWS|scan.IgnoreCase) != -1 ||
-		b.Search([]byte("<HTML XMLNS="), scan.CompactWS|scan.IgnoreCase) != -1
+	i, _ := b.Search([]byte("<!DOCTYPE HTML"), scan.CompactWS|scan.IgnoreCase)
+	if i != -1 {
+		return true
+	}
+	i, _ = b.Search([]byte("<HTML XMLNS="), scan.CompactWS|scan.IgnoreCase)
+	return i != -1
 }

 // Php matches a PHP: Hypertext Preprocessor file.
@@ -294,11 +419,12 @@ func svgWithoutXMLDeclaration(s scan.Bytes) bool {
 		return false
 	}

-	targetName, targetVal := "xmlns", "http://www.w3.org/2000/svg"
-	aName, aVal, hasMore := "", "", true
+	targetName, targetVal := []byte("xmlns"), []byte("http://www.w3.org/2000/svg")
+	var aName, aVal []byte
+	hasMore := true
 	for hasMore {
 		aName, aVal, hasMore = mkup.GetAnAttribute(&s)
-		if aName == targetName && aVal == targetVal {
+		if bytes.Equal(aName, targetName) && bytes.Equal(aVal, targetVal) {
 			return true
 		}
 		if !hasMore {
@@ -325,10 +451,11 @@ func svgWithXMLDeclaration(s scan.Bytes) bool {

 	// version is a required attribute for XML.
 	hasVersion := false
-	aName, hasMore := "", true
+	var aName []byte
+	hasMore := true
 	for hasMore {
 		aName, _, hasMore = mkup.GetAnAttribute(&s)
-		if aName == "version" {
+		if bytes.Equal(aName, []byte("version")) {
 			hasVersion = true
 			break
 		}
@@ -4,17 +4,23 @@ import (
 	"bytes"
 )

-var (
-	// Flv matches a Flash video file.
-	Flv = prefix([]byte("\x46\x4C\x56\x01"))
-	// Asf matches an Advanced Systems Format file.
-	Asf = prefix([]byte{
+// Flv matches a Flash video file.
+func Flv(raw []byte, _ uint32) bool {
+	return bytes.HasPrefix(raw, []byte("\x46\x4C\x56\x01"))
+}
+
+// Asf matches an Advanced Systems Format file.
+func Asf(raw []byte, _ uint32) bool {
+	return bytes.HasPrefix(raw, []byte{
 		0x30, 0x26, 0xB2, 0x75, 0x8E, 0x66, 0xCF, 0x11,
 		0xA6, 0xD9, 0x00, 0xAA, 0x00, 0x62, 0xCE, 0x6C,
 	})
-	// Rmvb matches a RealMedia Variable Bitrate file.
-	Rmvb = prefix([]byte{0x2E, 0x52, 0x4D, 0x46})
-)
+}
+
+// Rmvb matches a RealMedia Variable Bitrate file.
+func Rmvb(raw []byte, _ uint32) bool {
+	return bytes.HasPrefix(raw, []byte{0x2E, 0x52, 0x4D, 0x46})
+}

 // WebM matches a WebM file.
 func WebM(raw []byte, limit uint32) bool {
@@ -6,32 +6,65 @@ import (
 	"github.com/gabriel-vasile/mimetype/internal/scan"
 )

-var (
-	// Odt matches an OpenDocument Text file.
-	Odt = offset([]byte("mimetypeapplication/vnd.oasis.opendocument.text"), 30)
-	// Ott matches an OpenDocument Text Template file.
-	Ott = offset([]byte("mimetypeapplication/vnd.oasis.opendocument.text-template"), 30)
-	// Ods matches an OpenDocument Spreadsheet file.
-	Ods = offset([]byte("mimetypeapplication/vnd.oasis.opendocument.spreadsheet"), 30)
-	// Ots matches an OpenDocument Spreadsheet Template file.
-	Ots = offset([]byte("mimetypeapplication/vnd.oasis.opendocument.spreadsheet-template"), 30)
-	// Odp matches an OpenDocument Presentation file.
-	Odp = offset([]byte("mimetypeapplication/vnd.oasis.opendocument.presentation"), 30)
-	// Otp matches an OpenDocument Presentation Template file.
-	Otp = offset([]byte("mimetypeapplication/vnd.oasis.opendocument.presentation-template"), 30)
-	// Odg matches an OpenDocument Drawing file.
-	Odg = offset([]byte("mimetypeapplication/vnd.oasis.opendocument.graphics"), 30)
-	// Otg matches an OpenDocument Drawing Template file.
-	Otg = offset([]byte("mimetypeapplication/vnd.oasis.opendocument.graphics-template"), 30)
-	// Odf matches an OpenDocument Formula file.
-	Odf = offset([]byte("mimetypeapplication/vnd.oasis.opendocument.formula"), 30)
-	// Odc matches an OpenDocument Chart file.
-	Odc = offset([]byte("mimetypeapplication/vnd.oasis.opendocument.chart"), 30)
-	// Epub matches an EPUB file.
-	Epub = offset([]byte("mimetypeapplication/epub+zip"), 30)
-	// Sxc matches an OpenOffice Spreadsheet file.
-	Sxc = offset([]byte("mimetypeapplication/vnd.sun.xml.calc"), 30)
-)
+// Odt matches an OpenDocument Text file.
+func Odt(raw []byte, _ uint32) bool {
+	return offset(raw, []byte("mimetypeapplication/vnd.oasis.opendocument.text"), 30)
+}
+
+// Ott matches an OpenDocument Text Template file.
+func Ott(raw []byte, _ uint32) bool {
+	return offset(raw, []byte("mimetypeapplication/vnd.oasis.opendocument.text-template"), 30)
+}
+
+// Ods matches an OpenDocument Spreadsheet file.
+func Ods(raw []byte, _ uint32) bool {
+	return offset(raw, []byte("mimetypeapplication/vnd.oasis.opendocument.spreadsheet"), 30)
+}
+
+// Ots matches an OpenDocument Spreadsheet Template file.
+func Ots(raw []byte, _ uint32) bool {
+	return offset(raw, []byte("mimetypeapplication/vnd.oasis.opendocument.spreadsheet-template"), 30)
+}
+
+// Odp matches an OpenDocument Presentation file.
+func Odp(raw []byte, _ uint32) bool {
+	return offset(raw, []byte("mimetypeapplication/vnd.oasis.opendocument.presentation"), 30)
+}
+
+// Otp matches an OpenDocument Presentation Template file.
+func Otp(raw []byte, _ uint32) bool {
+	return offset(raw, []byte("mimetypeapplication/vnd.oasis.opendocument.presentation-template"), 30)
+}
+
+// Odg matches an OpenDocument Drawing file.
+func Odg(raw []byte, _ uint32) bool {
+	return offset(raw, []byte("mimetypeapplication/vnd.oasis.opendocument.graphics"), 30)
+}
+
+// Otg matches an OpenDocument Drawing Template file.
+func Otg(raw []byte, _ uint32) bool {
+	return offset(raw, []byte("mimetypeapplication/vnd.oasis.opendocument.graphics-template"), 30)
+}
+
+// Odf matches an OpenDocument Formula file.
+func Odf(raw []byte, _ uint32) bool {
+	return offset(raw, []byte("mimetypeapplication/vnd.oasis.opendocument.formula"), 30)
+}
+
+// Odc matches an OpenDocument Chart file.
+func Odc(raw []byte, _ uint32) bool {
+	return offset(raw, []byte("mimetypeapplication/vnd.oasis.opendocument.chart"), 30)
+}
+
+// Epub matches an EPUB file.
+func Epub(raw []byte, _ uint32) bool {
+	return offset(raw, []byte("mimetypeapplication/epub+zip"), 30)
+}
+
+// Sxc matches an OpenOffice Spreadsheet file.
+func Sxc(raw []byte, _ uint32) bool {
+	return offset(raw, []byte("mimetypeapplication/vnd.sun.xml.calc"), 30)
+}

 // Zip matches a zip archive.
 func Zip(raw []byte, limit uint32) bool {
@@ -134,11 +167,11 @@ func msoxml(raw scan.Bytes, searchFor zipEntries, stopAfter int) bool {
 		// If the first is not one of the next usually expected entries,
 		// then abort this check.
 		if i == 0 {
-			if !bytes.Equal(f, []byte("[Content_Types].xml")) &&
-				!bytes.Equal(f, []byte("_rels/.rels")) &&
-				!bytes.Equal(f, []byte("docProps")) &&
-				!bytes.Equal(f, []byte("customXml")) &&
-				!bytes.Equal(f, []byte("[trash]")) {
+			if !bytes.Equal(f, []byte("[Content_Types].xml")) && // this is a file
+				!bytes.HasPrefix(f, []byte("_rels/")) && // these are directories
+				!bytes.HasPrefix(f, []byte("docProps/")) &&
+				!bytes.HasPrefix(f, []byte("customXml/")) &&
+				!bytes.HasPrefix(f, []byte("[trash]/")) {
 				return false
 			}
 		}
@@ -8,46 +8,48 @@ import (
 	"github.com/gabriel-vasile/mimetype/internal/scan"
 )

-func GetAnAttribute(s *scan.Bytes) (name, val string, hasMore bool) {
+// GetAnAttribute assumes we passed over an SGML tag and extracts first
+// attribute and its value.
+//
+// Initially, this code existed inside charset/charset.go, because it was part of
+// implementing the https://html.spec.whatwg.org/multipage/parsing.html#prescan-a-byte-stream-to-determine-its-encoding
+// algorithm. But because extracting an attribute from a tag is the same for
+// both HTML and XML, then the code was moved here.
+func GetAnAttribute(s *scan.Bytes) (name, val []byte, hasMore bool) {
 	for scan.ByteIsWS(s.Peek()) || s.Peek() == '/' {
 		s.Advance(1)
 	}
 	if s.Peek() == '>' {
-		return "", "", false
+		return nil, nil, false
 	}
-	// Allocate 10 to avoid resizes.
-	// Attribute names and values are continuous slices of bytes in input,
-	// so we could do without allocating and returning slices of input.
-	nameB := make([]byte, 0, 10)
+	origS, end := *s, 0
 	// step 4 and 5
 	for {
 		// bap means byte at position in the specification.
 		bap := s.Pop()
 		if bap == 0 {
-			return "", "", false
+			return nil, nil, false
 		}
-		if bap == '=' && len(nameB) > 0 {
+		if bap == '=' && end > 0 {
 			val, hasMore := getAValue(s)
-			return string(nameB), string(val), hasMore
+			return origS[:end], val, hasMore
 		} else if scan.ByteIsWS(bap) {
 			for scan.ByteIsWS(s.Peek()) {
 				s.Advance(1)
 			}
 			if s.Peek() != '=' {
-				return string(nameB), "", true
+				return origS[:end], nil, true
 			}
 			s.Advance(1)
 			for scan.ByteIsWS(s.Peek()) {
 				s.Advance(1)
 			}
 			val, hasMore := getAValue(s)
-			return string(nameB), string(val), hasMore
+			return origS[:end], val, hasMore
 		} else if bap == '/' || bap == '>' {
-			return string(nameB), "", false
-		} else if bap >= 'A' && bap <= 'Z' {
-			nameB = append(nameB, bap+0x20)
-		} else {
-			nameB = append(nameB, bap)
+			return origS[:end], nil, false
+		} else { // for any ASCII, non-ASCII, just advance
+			end++
 		}
 	}
 }
@@ -138,46 +138,85 @@ func (b *Bytes) Uint16() (uint16, bool) {
 	return v, true
 }

+type Flags int
+
 const (
-	CompactWS = 1 << iota
+	// CompactWS will make one whitespace from pattern to match one or more spaces from input.
+	CompactWS Flags = 1 << iota
+	// IgnoreCase will match lower case from pattern with lower case from input.
+	// IgnoreCase will match upper case from pattern with both lower and upper case from input.
+	// This flag is not really well named,
 	IgnoreCase
+	// FullWord ensures the input ends with a full word (it's followed by spaces.)
+	FullWord
 )

 // Search for occurences of pattern p inside b at any index.
-func (b Bytes) Search(p []byte, flags int) int {
+// It returns the index where p was found in b and how many bytes were needed
+// for matching the pattern.
+func (b Bytes) Search(p []byte, flags Flags) (i int, l int) {
+	lb, lp := len(b), len(p)
+	if lp == 0 {
+		return 0, 0
+	}
+	if lb == 0 {
+		return -1, 0
+	}
 	if flags == 0 {
-		return bytes.Index(b, p)
+		if i = bytes.Index(b, p); i == -1 {
+			return -1, 0
+		} else {
+			return i, lp
+		}
 	}

-	lb, lp := len(b), len(p)
 	for i := range b {
 		if lb-i < lp {
-			return -1
+			return -1, 0
 		}
-		if b[i:].Match(p, flags) {
-			return i
+		if l = b[i:].Match(p, flags); l != -1 {
+			return i, l
 		}
 	}

-	return 0
+	return -1, 0
 }

-// Match pattern p at index 0 of b.
-func (b Bytes) Match(p []byte, flags int) bool {
+// Match returns how many bytes were needed to match pattern p.
+// It returns -1 if p does not match b.
+func (b Bytes) Match(p []byte, flags Flags) int {
+	l := len(b)
+	if len(p) == 0 {
+		return 0
+	}
+	if l == 0 {
+		return -1
+	}
+	// If no flags, or scanning for full word at the end of pattern then
+	// do a fast HasPrefix check.
+	// For other flags it's not possible to use HasPrefix.
+	if flags == 0 || flags&FullWord > 0 {
+		if bytes.HasPrefix(b, p) {
+			b = b[len(p):]
+			p = p[len(p):]
+			goto out
+		}
+		return -1
+	}
 	for len(b) > 0 {
-		// If we finished all we we're looking for from p.
+		// If we finished all we were looking for from p.
 		if len(p) == 0 {
-			return true
+			goto out
 		}
 		if flags&IgnoreCase > 0 && isUpper(p[0]) {
 			if upper(b[0]) != p[0] {
-				return false
+				return -1
 			}
 			b, p = b[1:], p[1:]
 		} else if flags&CompactWS > 0 && ByteIsWS(p[0]) {
 			p = p[1:]
 			if !ByteIsWS(b[0]) {
-				return false
+				return -1
 			}
 			b = b[1:]
 			if !ByteIsWS(p[0]) {
@@ -185,12 +224,22 @@ func (b Bytes) Match(p []byte, flags int) bool {
 			}
 		} else {
 			if b[0] != p[0] {
-				return false
+				return -1
 			}
 			b, p = b[1:], p[1:]
 		}
 	}
-	return true
+out:
+	// If p still has leftover characters, it means it didn't fully match b.
+	if len(p) > 0 {
+		return -1
+	}
+	if flags&FullWord > 0 {
+		if len(b) > 0 && !ByteIsWS(b[0]) {
+			return -1
+		}
+	}
+	return l - len(b)
 }

 func isUpper(c byte) bool {
@@ -2,6 +2,7 @@ package mimetype

 import (
 	"mime"
+	"strings"

 	"github.com/gabriel-vasile/mimetype/internal/charset"
 	"github.com/gabriel-vasile/mimetype/internal/magic"
@@ -109,7 +110,7 @@ func (m *MIME) match(in []byte, readLimit uint32) *MIME {
 		// Limit the number of bytes searched for to 1024.
 		charset = f(in[:min(len(in), 1024)])
 	}
-	if m == root {
+	if m == root || charset == "" {
 		return m
 	}

@@ -126,6 +127,27 @@ func (m *MIME) flatten() []*MIME {
 	return out
 }

+// hierarchy returns an easy to read list of ancestors for m.
+// For example, application/json would return json>txt>root.
+func (m *MIME) hierarchy() string {
+	h := ""
+	for m := m; m != nil; m = m.Parent() {
+		e := strings.TrimPrefix(m.Extension(), ".")
+		if e == "" {
+			// There are some MIME without extensions. When generating the hierarchy,
+			// it would be confusing to use empty string as extension.
+			// Use the subtype instead; ex: application/x-executable -> x-executable.
+			e = strings.Split(m.String(), "/")[1]
+			if m.Is("application/octet-stream") {
+				// for octet-stream use root, because it's short and used in many places
+				e = "root"
+			}
+		}
+		h += ">" + e
+	}
+	return strings.TrimPrefix(h, ">")
+}
+
 // clone creates a new MIME with the provided optional MIME parameters.
 func (m *MIME) clone(charset string) *MIME {
 	clonedMIME := m.mime
@@ -155,7 +177,10 @@ func (m *MIME) cloneHierarchy(charset string) *MIME {
 }

 func (m *MIME) lookup(mime string) *MIME {
-	for _, n := range append(m.aliases, m.mime) {
+	if mime == m.mime {
+		return m
+	}
+	for _, n := range m.aliases {
 		if n == mime {
 			return m
 		}
@@ -12,7 +12,7 @@ import (
 	"sync/atomic"
 )

-var defaultLimit uint32 = 3072
+const defaultLimit uint32 = 3072

 // readLimit is the maximum number of bytes from the input used when detecting.
 var readLimit uint32 = defaultLimit
@@ -112,15 +112,18 @@ func SetLimit(limit uint32) {
 }

 // Extend adds detection for other file formats.
-// It is equivalent to calling Extend() on the root mime type "application/octet-stream".
+// It is equivalent to calling Extend() on the root MIME type "application/octet-stream".
 func Extend(detector func(raw []byte, limit uint32) bool, mime, extension string, aliases ...string) {
 	root.Extend(detector, mime, extension, aliases...)
 }

 // Lookup finds a MIME object by its string representation.
-// The representation can be the main mime type, or any of its aliases.
-func Lookup(mime string) *MIME {
+// The representation can be the main MIME type, or any of its aliases.
+func Lookup(m string) *MIME {
+	// We store the MIME types without optional params, so
+	// perform parsing to extract the target MIME type without optional params.
+	m, _, _ = mime.ParseMediaType(m)
 	mu.RLock()
 	defer mu.RUnlock()
-	return root.lookup(mime)
+	return root.lookup(m)
 }
@@ -1,196 +1,197 @@
-## 191 Supported MIME types
+## 192 Supported MIME types
 This file is automatically generated when running tests. Do not edit manually.

-Extension | MIME type | Aliases
--------- | --------- | -------
-**n/a** | application/octet-stream | -
-**.xpm** | image/x-xpixmap | -
-**.7z** | application/x-7z-compressed | -
-**.zip** | application/zip | application/x-zip, application/x-zip-compressed
-**.docx** | application/vnd.openxmlformats-officedocument.wordprocessingml.document | -
-**.pptx** | application/vnd.openxmlformats-officedocument.presentationml.presentation | -
-**.xlsx** | application/vnd.openxmlformats-officedocument.spreadsheetml.sheet | -
-**.epub** | application/epub+zip | -
-**.apk** | application/vnd.android.package-archive | -
-**.jar** | application/java-archive | application/jar, application/jar-archive, application/x-java-archive
-**.odt** | application/vnd.oasis.opendocument.text | application/x-vnd.oasis.opendocument.text
-**.ott** | application/vnd.oasis.opendocument.text-template | application/x-vnd.oasis.opendocument.text-template
-**.ods** | application/vnd.oasis.opendocument.spreadsheet | application/x-vnd.oasis.opendocument.spreadsheet
-**.ots** | application/vnd.oasis.opendocument.spreadsheet-template | application/x-vnd.oasis.opendocument.spreadsheet-template
-**.odp** | application/vnd.oasis.opendocument.presentation | application/x-vnd.oasis.opendocument.presentation
-**.otp** | application/vnd.oasis.opendocument.presentation-template | application/x-vnd.oasis.opendocument.presentation-template
-**.odg** | application/vnd.oasis.opendocument.graphics | application/x-vnd.oasis.opendocument.graphics
-**.otg** | application/vnd.oasis.opendocument.graphics-template | application/x-vnd.oasis.opendocument.graphics-template
-**.odf** | application/vnd.oasis.opendocument.formula | application/x-vnd.oasis.opendocument.formula
-**.odc** | application/vnd.oasis.opendocument.chart | application/x-vnd.oasis.opendocument.chart
-**.sxc** | application/vnd.sun.xml.calc | -
-**.kmz** | application/vnd.google-earth.kmz | -
-**.vsdx** | application/vnd.ms-visio.drawing.main+xml | -
-**.pdf** | application/pdf | application/x-pdf
-**.fdf** | application/vnd.fdf | -
-**n/a** | application/x-ole-storage | -
-**.msi** | application/x-ms-installer | application/x-windows-installer, application/x-msi
-**.aaf** | application/octet-stream | -
-**.msg** | application/vnd.ms-outlook | -
-**.xls** | application/vnd.ms-excel | application/msexcel
-**.pub** | application/vnd.ms-publisher | -
-**.ppt** | application/vnd.ms-powerpoint | application/mspowerpoint
-**.doc** | application/msword | application/vnd.ms-word
-**.ps** | application/postscript | -
-**.psd** | image/vnd.adobe.photoshop | image/x-psd, application/photoshop
-**.p7s** | application/pkcs7-signature | -
-**.ogg** | application/ogg | application/x-ogg
-**.oga** | audio/ogg | -
-**.ogv** | video/ogg | -
-**.png** | image/png | -
-**.png** | image/vnd.mozilla.apng | -
-**.jpg** | image/jpeg | -
-**.jxl** | image/jxl | -
-**.jp2** | image/jp2 | -
-**.jpf** | image/jpx | -
-**.jpm** | image/jpm | video/jpm
-**.jxs** | image/jxs | -
-**.gif** | image/gif | -
-**.webp** | image/webp | -
-**.exe** | application/vnd.microsoft.portable-executable | -
-**n/a** | application/x-elf | -
-**n/a** | application/x-object | -
-**n/a** | application/x-executable | -
-**.so** | application/x-sharedlib | -
-**n/a** | application/x-coredump | -
-**.a** | application/x-archive | application/x-unix-archive
-**.deb** | application/vnd.debian.binary-package | -
-**.tar** | application/x-tar | -
-**.xar** | application/x-xar | -
-**.bz2** | application/x-bzip2 | -
-**.fits** | application/fits | image/fits
-**.tiff** | image/tiff | -
-**.bmp** | image/bmp | image/x-bmp, image/x-ms-bmp
-**.123** | application/vnd.lotus-1-2-3 | -
-**.ico** | image/x-icon | -
-**.mp3** | audio/mpeg | audio/x-mpeg, audio/mp3
-**.flac** | audio/flac | -
-**.midi** | audio/midi | audio/mid, audio/sp-midi, audio/x-mid, audio/x-midi
-**.ape** | audio/ape | -
-**.mpc** | audio/musepack | -
-**.amr** | audio/amr | audio/amr-nb
-**.wav** | audio/wav | audio/x-wav, audio/vnd.wave, audio/wave
-**.aiff** | audio/aiff | audio/x-aiff
-**.au** | audio/basic | -
-**.mpeg** | video/mpeg | -
-**.mov** | video/quicktime | -
-**.mp4** | video/mp4 | -
-**.avif** | image/avif | -
-**.3gp** | video/3gpp | video/3gp, audio/3gpp
-**.3g2** | video/3gpp2 | video/3g2, audio/3gpp2
-**.mp4** | audio/mp4 | audio/x-mp4a
-**.mqv** | video/quicktime | -
-**.m4a** | audio/x-m4a | -
-**.m4v** | video/x-m4v | -
-**.heic** | image/heic | -
-**.heic** | image/heic-sequence | -
-**.heif** | image/heif | -
-**.heif** | image/heif-sequence | -
-**.mj2** | video/mj2 | -
-**.dvb** | video/vnd.dvb.file | -
-**.webm** | video/webm | audio/webm
-**.avi** | video/x-msvideo | video/avi, video/msvideo
-**.flv** | video/x-flv | -
-**.mkv** | video/x-matroska | -
-**.asf** | video/x-ms-asf | video/asf, video/x-ms-wmv
-**.aac** | audio/aac | -
-**.voc** | audio/x-unknown | -
-**.m3u** | application/vnd.apple.mpegurl | audio/mpegurl
-**.rmvb** | application/vnd.rn-realmedia-vbr | -
-**.gz** | application/gzip | application/x-gzip, application/x-gunzip, application/gzipped, application/gzip-compressed, application/x-gzip-compressed, gzip/document
-**.class** | application/x-java-applet | -
-**.swf** | application/x-shockwave-flash | -
-**.crx** | application/x-chrome-extension | -
-**.ttf** | font/ttf | font/sfnt, application/x-font-ttf, application/font-sfnt
-**.woff** | font/woff | -
-**.woff2** | font/woff2 | -
-**.otf** | font/otf | -
-**.ttc** | font/collection | -
-**.eot** | application/vnd.ms-fontobject | -
-**.wasm** | application/wasm | -
-**.shx** | application/vnd.shx | -
-**.shp** | application/vnd.shp | -
-**.dbf** | application/x-dbf | -
-**.dcm** | application/dicom | -
-**.rar** | application/x-rar-compressed | application/x-rar
-**.djvu** | image/vnd.djvu | -
-**.mobi** | application/x-mobipocket-ebook | -
-**.lit** | application/x-ms-reader | -
-**.bpg** | image/bpg | -
-**.cbor** | application/cbor | -
-**.sqlite** | application/vnd.sqlite3 | application/x-sqlite3
-**.dwg** | image/vnd.dwg | image/x-dwg, application/acad, application/x-acad, application/autocad_dwg, application/dwg, application/x-dwg, application/x-autocad, drawing/dwg
-**.nes** | application/vnd.nintendo.snes.rom | -
-**.lnk** | application/x-ms-shortcut | -
-**.macho** | application/x-mach-binary | -
-**.qcp** | audio/qcelp | -
-**.icns** | image/x-icns | -
-**.hdr** | image/vnd.radiance | -
-**.mrc** | application/marc | -
-**.mdb** | application/x-msaccess | -
-**.accdb** | application/x-msaccess | -
-**.zst** | application/zstd | -
-**.cab** | application/vnd.ms-cab-compressed | -
-**.rpm** | application/x-rpm | -
-**.xz** | application/x-xz | -
-**.lz** | application/lzip | application/x-lzip
-**.torrent** | application/x-bittorrent | -
-**.cpio** | application/x-cpio | -
-**n/a** | application/tzif | -
-**.xcf** | image/x-xcf | -
-**.pat** | image/x-gimp-pat | -
-**.gbr** | image/x-gimp-gbr | -
-**.glb** | model/gltf-binary | -
-**.cab** | application/x-installshield | -
-**.jxr** | image/jxr | image/vnd.ms-photo
-**.parquet** | application/vnd.apache.parquet | application/x-parquet
-**.one** | application/onenote | -
-**.chm** | application/vnd.ms-htmlhelp | -
-**.txt** | text/plain | -
-**.svg** | image/svg+xml | -
-**.html** | text/html | -
-**.xml** | text/xml | application/xml
-**.rss** | application/rss+xml | text/rss
-**.atom** | application/atom+xml | -
-**.x3d** | model/x3d+xml | -
-**.kml** | application/vnd.google-earth.kml+xml | -
-**.xlf** | application/x-xliff+xml | -
-**.dae** | model/vnd.collada+xml | -
-**.gml** | application/gml+xml | -
-**.gpx** | application/gpx+xml | -
-**.tcx** | application/vnd.garmin.tcx+xml | -
-**.amf** | application/x-amf | -
-**.3mf** | application/vnd.ms-package.3dmanufacturing-3dmodel+xml | -
-**.xfdf** | application/vnd.adobe.xfdf | -
-**.owl** | application/owl+xml | -
-**.html** | application/xhtml+xml | -
-**.php** | text/x-php | -
-**.js** | text/javascript | application/x-javascript, application/javascript
-**.lua** | text/x-lua | -
-**.pl** | text/x-perl | -
-**.py** | text/x-python | text/x-script.python, application/x-python
-**.rb** | text/x-ruby | application/x-ruby
-**.json** | application/json | -
-**.geojson** | application/geo+json | -
-**.har** | application/json | -
-**.gltf** | model/gltf+json | -
-**.ndjson** | application/x-ndjson | -
-**.rtf** | text/rtf | application/rtf
-**.srt** | application/x-subrip | application/x-srt, text/x-srt
-**.tcl** | text/x-tcl | application/x-tcl
-**.csv** | text/csv | -
-**.tsv** | text/tab-separated-values | -
-**.vcf** | text/vcard | -
-**.ics** | text/calendar | -
-**.warc** | application/warc | -
-**.vtt** | text/vtt | -
-**.sh** | text/x-shellscript | text/x-sh, application/x-shellscript, application/x-sh
-**.pbm** | image/x-portable-bitmap | -
-**.pgm** | image/x-portable-graymap | -
-**.ppm** | image/x-portable-pixmap | -
-**.pam** | image/x-portable-arbitrarymap | -
+Extension | MIME type <br> Aliases | Hierarchy
+--------- | ---------------------- | ---------
+**n/a** | **application/octet-stream** | root
+**.xpm** | **image/x-xpixmap** | xpm>root
+**.7z** | **application/x-7z-compressed** | 7z>root
+**.zip** | **application/zip** <br> application/x-zip, application/x-zip-compressed | zip>root
+**.docx** | **application/vnd.openxmlformats-officedocument.wordprocessingml.document** | docx>zip>root
+**.pptx** | **application/vnd.openxmlformats-officedocument.presentationml.presentation** | pptx>zip>root
+**.xlsx** | **application/vnd.openxmlformats-officedocument.spreadsheetml.sheet** | xlsx>zip>root
+**.epub** | **application/epub+zip** | epub>zip>root
+**.apk** | **application/vnd.android.package-archive** | apk>zip>root
+**.jar** | **application/java-archive** <br> application/jar, application/jar-archive, application/x-java-archive | jar>zip>root
+**.odt** | **application/vnd.oasis.opendocument.text** <br> application/x-vnd.oasis.opendocument.text | odt>zip>root
+**.ott** | **application/vnd.oasis.opendocument.text-template** <br> application/x-vnd.oasis.opendocument.text-template | ott>odt>zip>root
+**.ods** | **application/vnd.oasis.opendocument.spreadsheet** <br> application/x-vnd.oasis.opendocument.spreadsheet | ods>zip>root
+**.ots** | **application/vnd.oasis.opendocument.spreadsheet-template** <br> application/x-vnd.oasis.opendocument.spreadsheet-template | ots>ods>zip>root
+**.odp** | **application/vnd.oasis.opendocument.presentation** <br> application/x-vnd.oasis.opendocument.presentation | odp>zip>root
+**.otp** | **application/vnd.oasis.opendocument.presentation-template** <br> application/x-vnd.oasis.opendocument.presentation-template | otp>odp>zip>root
+**.odg** | **application/vnd.oasis.opendocument.graphics** <br> application/x-vnd.oasis.opendocument.graphics | odg>zip>root
+**.otg** | **application/vnd.oasis.opendocument.graphics-template** <br> application/x-vnd.oasis.opendocument.graphics-template | otg>odg>zip>root
+**.odf** | **application/vnd.oasis.opendocument.formula** <br> application/x-vnd.oasis.opendocument.formula | odf>zip>root
+**.odc** | **application/vnd.oasis.opendocument.chart** <br> application/x-vnd.oasis.opendocument.chart | odc>zip>root
+**.sxc** | **application/vnd.sun.xml.calc** | sxc>zip>root
+**.kmz** | **application/vnd.google-earth.kmz** | kmz>zip>root
+**.vsdx** | **application/vnd.ms-visio.drawing.main+xml** | vsdx>zip>root
+**.pdf** | **application/pdf** <br> application/x-pdf | pdf>root
+**.fdf** | **application/vnd.fdf** | fdf>root
+**n/a** | **application/x-ole-storage** | x-ole-storage>root
+**.msi** | **application/x-ms-installer** <br> application/x-windows-installer, application/x-msi | msi>x-ole-storage>root
+**.msg** | **application/vnd.ms-outlook** | msg>x-ole-storage>root
+**.xls** | **application/vnd.ms-excel** <br> application/msexcel | xls>x-ole-storage>root
+**.pub** | **application/vnd.ms-publisher** | pub>x-ole-storage>root
+**.ppt** | **application/vnd.ms-powerpoint** <br> application/mspowerpoint | ppt>x-ole-storage>root
+**.doc** | **application/msword** <br> application/vnd.ms-word | doc>x-ole-storage>root
+**.ps** | **application/postscript** | ps>root
+**.psd** | **image/vnd.adobe.photoshop** <br> image/x-psd, application/photoshop | psd>root
+**.p7s** | **application/pkcs7-signature** | p7s>root
+**.ogg** | **application/ogg** <br> application/x-ogg | ogg>root
+**.oga** | **audio/ogg** | oga>ogg>root
+**.ogv** | **video/ogg** | ogv>ogg>root
+**.png** | **image/png** | png>root
+**.png** | **image/vnd.mozilla.apng** | png>png>root
+**.jpg** | **image/jpeg** | jpg>root
+**.jxl** | **image/jxl** | jxl>root
+**.jp2** | **image/jp2** | jp2>root
+**.jpf** | **image/jpx** | jpf>root
+**.jpm** | **image/jpm** <br> video/jpm | jpm>root
+**.jxs** | **image/jxs** | jxs>root
+**.gif** | **image/gif** | gif>root
+**.webp** | **image/webp** | webp>root
+**.exe** | **application/vnd.microsoft.portable-executable** | exe>root
+**n/a** | **application/x-elf** | x-elf>root
+**n/a** | **application/x-object** | x-object>x-elf>root
+**n/a** | **application/x-executable** | x-executable>x-elf>root
+**.so** | **application/x-sharedlib** | so>x-elf>root
+**n/a** | **application/x-coredump** | x-coredump>x-elf>root
+**.a** | **application/x-archive** <br> application/x-unix-archive | a>root
+**.deb** | **application/vnd.debian.binary-package** | deb>a>root
+**.tar** | **application/x-tar** | tar>root
+**.xar** | **application/x-xar** | xar>root
+**.bz2** | **application/x-bzip2** | bz2>root
+**.fits** | **application/fits** <br> image/fits | fits>root
+**.tiff** | **image/tiff** | tiff>root
+**.bmp** | **image/bmp** <br> image/x-bmp, image/x-ms-bmp | bmp>root
+**.123** | **application/vnd.lotus-1-2-3** | 123>root
+**.ico** | **image/x-icon** | ico>root
+**.mp3** | **audio/mpeg** <br> audio/x-mpeg, audio/mp3 | mp3>root
+**.flac** | **audio/flac** | flac>root
+**.midi** | **audio/midi** <br> audio/mid, audio/sp-midi, audio/x-mid, audio/x-midi | midi>root
+**.ape** | **audio/ape** | ape>root
+**.mpc** | **audio/musepack** | mpc>root
+**.amr** | **audio/amr** <br> audio/amr-nb | amr>root
+**.wav** | **audio/wav** <br> audio/x-wav, audio/vnd.wave, audio/wave | wav>root
+**.aiff** | **audio/aiff** <br> audio/x-aiff | aiff>root
+**.au** | **audio/basic** | au>root
+**.mpeg** | **video/mpeg** | mpeg>root
+**.mov** | **video/quicktime** | mov>root
+**.mp4** | **video/mp4** | mp4>root
+**.avif** | **image/avif** | avif>mp4>root
+**.3gp** | **video/3gpp** <br> video/3gp, audio/3gpp | 3gp>mp4>root
+**.3g2** | **video/3gpp2** <br> video/3g2, audio/3gpp2 | 3g2>mp4>root
+**.mp4** | **audio/mp4** <br> audio/x-mp4a | mp4>mp4>root
+**.mqv** | **video/quicktime** | mqv>mp4>root
+**.m4a** | **audio/x-m4a** | m4a>mp4>root
+**.m4v** | **video/x-m4v** | m4v>mp4>root
+**.heic** | **image/heic** | heic>mp4>root
+**.heic** | **image/heic-sequence** | heic>mp4>root
+**.heif** | **image/heif** | heif>mp4>root
+**.heif** | **image/heif-sequence** | heif>mp4>root
+**.mj2** | **video/mj2** | mj2>mp4>root
+**.dvb** | **video/vnd.dvb.file** | dvb>mp4>root
+**.webm** | **video/webm** <br> audio/webm | webm>root
+**.avi** | **video/x-msvideo** <br> video/avi, video/msvideo | avi>root
+**.flv** | **video/x-flv** | flv>root
+**.mkv** | **video/x-matroska** | mkv>root
+**.asf** | **video/x-ms-asf** <br> video/asf, video/x-ms-wmv | asf>root
+**.aac** | **audio/aac** | aac>root
+**.voc** | **audio/x-unknown** | voc>root
+**.m3u** | **application/vnd.apple.mpegurl** <br> audio/mpegurl | m3u>root
+**.rmvb** | **application/vnd.rn-realmedia-vbr** | rmvb>root
+**.gz** | **application/gzip** <br> application/x-gzip, application/x-gunzip, application/gzipped, application/gzip-compressed, application/x-gzip-compressed, gzip/document | gz>root
+**.class** | **application/x-java-applet** | class>root
+**.swf** | **application/x-shockwave-flash** | swf>root
+**.crx** | **application/x-chrome-extension** | crx>root
+**.ttf** | **font/ttf** <br> font/sfnt, application/x-font-ttf, application/font-sfnt | ttf>root
+**.woff** | **font/woff** | woff>root
+**.woff2** | **font/woff2** | woff2>root
+**.otf** | **font/otf** | otf>root
+**.ttc** | **font/collection** | ttc>root
+**.eot** | **application/vnd.ms-fontobject** | eot>root
+**.wasm** | **application/wasm** | wasm>root
+**.shx** | **application/vnd.shx** | shx>root
+**.shp** | **application/vnd.shp** | shp>shx>root
+**.dbf** | **application/x-dbf** | dbf>root
+**.dcm** | **application/dicom** | dcm>root
+**.rar** | **application/x-rar-compressed** <br> application/x-rar | rar>root
+**.djvu** | **image/vnd.djvu** | djvu>root
+**.mobi** | **application/x-mobipocket-ebook** | mobi>root
+**.lit** | **application/x-ms-reader** | lit>root
+**.bpg** | **image/bpg** | bpg>root
+**.cbor** | **application/cbor** | cbor>root
+**.sqlite** | **application/vnd.sqlite3** <br> application/x-sqlite3 | sqlite>root
+**.dwg** | **image/vnd.dwg** <br> image/x-dwg, application/acad, application/x-acad, application/autocad_dwg, application/dwg, application/x-dwg, application/x-autocad, drawing/dwg | dwg>root
+**.nes** | **application/vnd.nintendo.snes.rom** | nes>root
+**.lnk** | **application/x-ms-shortcut** | lnk>root
+**.macho** | **application/x-mach-binary** | macho>root
+**.qcp** | **audio/qcelp** | qcp>root
+**.icns** | **image/x-icns** | icns>root
+**.hdr** | **image/vnd.radiance** | hdr>root
+**.mrc** | **application/marc** | mrc>root
+**.mdb** | **application/x-msaccess** | mdb>root
+**.accdb** | **application/x-msaccess** | accdb>root
+**.zst** | **application/zstd** | zst>root
+**.cab** | **application/vnd.ms-cab-compressed** | cab>root
+**.rpm** | **application/x-rpm** | rpm>root
+**.xz** | **application/x-xz** | xz>root
+**.lz** | **application/lzip** <br> application/x-lzip | lz>root
+**.torrent** | **application/x-bittorrent** | torrent>root
+**.cpio** | **application/x-cpio** | cpio>root
+**n/a** | **application/tzif** | tzif>root
+**.xcf** | **image/x-xcf** | xcf>root
+**.pat** | **image/x-gimp-pat** | pat>root
+**.gbr** | **image/x-gimp-gbr** | gbr>root
+**.glb** | **model/gltf-binary** | glb>root
+**.cab** | **application/x-installshield** | cab>root
+**.jxr** | **image/jxr** <br> image/vnd.ms-photo | jxr>root
+**.parquet** | **application/vnd.apache.parquet** <br> application/x-parquet | parquet>root
+**.one** | **application/onenote** | one>root
+**.chm** | **application/vnd.ms-htmlhelp** | chm>root
+**.wpd** | **application/vnd.wordperfect** | wpd>root
+**.dxf** | **image/vnd.dxf** | dxf>root
+**.txt** | **text/plain** | txt>root
+**.svg** | **image/svg+xml** | svg>txt>root
+**.html** | **text/html** | html>txt>root
+**.xml** | **text/xml** <br> application/xml | xml>txt>root
+**.rss** | **application/rss+xml** <br> text/rss | rss>xml>txt>root
+**.atom** | **application/atom+xml** | atom>xml>txt>root
+**.x3d** | **model/x3d+xml** | x3d>xml>txt>root
+**.kml** | **application/vnd.google-earth.kml+xml** | kml>xml>txt>root
+**.xlf** | **application/x-xliff+xml** | xlf>xml>txt>root
+**.dae** | **model/vnd.collada+xml** | dae>xml>txt>root
+**.gml** | **application/gml+xml** | gml>xml>txt>root
+**.gpx** | **application/gpx+xml** | gpx>xml>txt>root
+**.tcx** | **application/vnd.garmin.tcx+xml** | tcx>xml>txt>root
+**.amf** | **application/x-amf** | amf>xml>txt>root
+**.3mf** | **application/vnd.ms-package.3dmanufacturing-3dmodel+xml** | 3mf>xml>txt>root
+**.xfdf** | **application/vnd.adobe.xfdf** | xfdf>xml>txt>root
+**.owl** | **application/owl+xml** | owl>xml>txt>root
+**.html** | **application/xhtml+xml** | html>xml>txt>root
+**.php** | **text/x-php** | php>txt>root
+**.js** | **text/javascript** <br> application/x-javascript, application/javascript | js>txt>root
+**.lua** | **text/x-lua** | lua>txt>root
+**.pl** | **text/x-perl** | pl>txt>root
+**.py** | **text/x-python** <br> text/x-script.python, application/x-python | py>txt>root
+**.rb** | **text/x-ruby** <br> application/x-ruby | rb>txt>root
+**.json** | **application/json** | json>txt>root
+**.geojson** | **application/geo+json** | geojson>json>txt>root
+**.har** | **application/json** | har>json>txt>root
+**.gltf** | **model/gltf+json** | gltf>json>txt>root
+**.ndjson** | **application/x-ndjson** | ndjson>txt>root
+**.rtf** | **text/rtf** <br> application/rtf | rtf>txt>root
+**.srt** | **application/x-subrip** <br> application/x-srt, text/x-srt | srt>txt>root
+**.tcl** | **text/x-tcl** <br> application/x-tcl | tcl>txt>root
+**.csv** | **text/csv** | csv>txt>root
+**.tsv** | **text/tab-separated-values** | tsv>txt>root
+**.vcf** | **text/vcard** | vcf>txt>root
+**.ics** | **text/calendar** | ics>txt>root
+**.warc** | **application/warc** | warc>txt>root
+**.vtt** | **text/vtt** | vtt>txt>root
+**.sh** | **text/x-shellscript** <br> text/x-sh, application/x-shellscript, application/x-sh | sh>txt>root
+**.pbm** | **image/x-portable-bitmap** | pbm>txt>root
+**.pgm** | **image/x-portable-graymap** | pgm>txt>root
+**.ppm** | **image/x-portable-pixmap** | ppm>txt>root
+**.pam** | **image/x-portable-arbitrarymap** | pam>txt>root
@@ -24,7 +24,7 @@ var root = newMIME("application/octet-stream", "",
 	woff2, otf, ttc, eot, wasm, shx, dbf, dcm, rar, djvu, mobi, lit, bpg, cbor,
 	sqlite3, dwg, nes, lnk, macho, qcp, icns, hdr, mrc, mdb, accdb, zstd, cab,
 	rpm, xz, lzip, torrent, cpio, tzif, xcf, pat, gbr, glb, cabIS, jxr, parquet,
-	oneNote, chm,
+	oneNote, chm, wpd, dxf,
 	// Keep text last because it is the slowest check.
 	text,
 )
@@ -65,10 +65,9 @@ var (
 	jar   = newMIME("application/java-archive", ".jar", magic.Jar).
 		alias("application/jar", "application/jar-archive", "application/x-java-archive")
 	apk = newMIME("application/vnd.android.package-archive", ".apk", magic.APK)
-	ole = newMIME("application/x-ole-storage", "", magic.Ole, msi, aaf, msg, xls, pub, ppt, doc)
+	ole = newMIME("application/x-ole-storage", "", magic.Ole, msi, msg, xls, pub, ppt, doc)
 	msi = newMIME("application/x-ms-installer", ".msi", magic.Msi).
 		alias("application/x-windows-installer", "application/x-msi")
-	aaf = newMIME("application/octet-stream", ".aaf", magic.Aaf)
 	doc = newMIME("application/msword", ".doc", magic.Doc).
 		alias("application/vnd.ms-word")
 	ppt = newMIME("application/vnd.ms-powerpoint", ".ppt", magic.Ppt).
@@ -286,4 +285,6 @@ var (
 	cbor    = newMIME("application/cbor", ".cbor", magic.CBOR)
 	oneNote = newMIME("application/onenote", ".one", magic.One)
 	chm     = newMIME("application/vnd.ms-htmlhelp", ".chm", magic.CHM)
+	wpd     = newMIME("application/vnd.wordperfect", ".wpd", magic.WPD)
+	dxf     = newMIME("image/vnd.dxf", ".dxf", magic.DXF)
 )
@@ -444,7 +444,7 @@ github.com/felixge/httpsnoop
 ## explicit; go 1.17
 github.com/fsnotify/fsnotify
 github.com/fsnotify/fsnotify/internal
-# github.com/gabriel-vasile/mimetype v1.4.10
+# github.com/gabriel-vasile/mimetype v1.4.11
 ## explicit; go 1.21
 github.com/gabriel-vasile/mimetype
 github.com/gabriel-vasile/mimetype/internal/charset