Use wchar_t for TagLib filenames on Windows (#1310)

* Use wchar_t for tagLib filenames on Windows

* Make TagLib default extractor for all platforms.

* Organize imports

Co-authored-by: Deluan <deluan@navidrome.org>

conf/configuration.go

@@ -214,7 +214,7 @@ func init() {
 	viper.SetDefault("reverseproxyuserheader", "Remote-User")
 	viper.SetDefault("reverseproxywhitelist", "")
 
-	viper.SetDefault("scanner.extractor", DefaultScannerExtractor)
+	viper.SetDefault("scanner.extractor", consts.DefaultScannerExtractor)
 	viper.SetDefault("scanner.genreseparators", ";/,")
 
 	viper.SetDefault("agents", "lastfm,spotify")

conf/defaults.go

@@ -1,5 +0,0 @@
-// +build !windows
-
-package conf
-
-const DefaultScannerExtractor = "taglib"

conf/defaults_win.go

@@ -1,5 +0,0 @@
-// +build windows
-
-package conf
-
-const DefaultScannerExtractor = "ffmpeg"

consts/consts.go

@@ -47,6 +47,8 @@ const (
 	PlaceholderAvatar   = "logo-192x192.png"
 
 	DefaultHttpClientTimeOut = 10 * time.Second
+
+	DefaultScannerExtractor = "taglib"
 )
 
 // Cache options

scanner/metadata/metadata.go

@@ -10,13 +10,12 @@ import (
 	"strings"
 	"time"
 
-	"github.com/navidrome/navidrome/scanner/metadata/ffmpeg"
-
-	"github.com/navidrome/navidrome/scanner/metadata/taglib"
-
 	"github.com/google/uuid"
 	"github.com/navidrome/navidrome/conf"
+	"github.com/navidrome/navidrome/consts"
 	"github.com/navidrome/navidrome/log"
+	"github.com/navidrome/navidrome/scanner/metadata/ffmpeg"
+	"github.com/navidrome/navidrome/scanner/metadata/taglib"
 )
 
 type Parser interface {
@@ -32,8 +31,8 @@ func Extract(files ...string) (map[string]Tags, error) {
 	p, ok := parsers[conf.Server.Scanner.Extractor]
 	if !ok {
 		log.Warn("Invalid 'Scanner.Extractor' option. Using default", "requested", conf.Server.Scanner.Extractor,
-			"validOptions", "ffmpeg,taglib", "default", conf.DefaultScannerExtractor)
+			"validOptions", "ffmpeg,taglib", "default", consts.DefaultScannerExtractor)
-		p = parsers[conf.DefaultScannerExtractor]
+		p = parsers[consts.DefaultScannerExtractor]
 	}
 
 	extractedTags, err := p.Parse(files...)

scanner/metadata/taglib/get_filename.go

@@ -0,0 +1,9 @@
+// +build !windows
+
+package taglib
+
+import "C"
+
+func getFilename(s string) *C.char {
+	return C.CString(s)
+}

scanner/metadata/taglib/get_filename_win.go

@@ -0,0 +1,96 @@
+// +build windows
+
+package taglib
+
+// From https://github.com/orofarne/gowchar
+
+/*
+#include <wchar.h>
+
+const size_t SIZEOF_WCHAR_T = sizeof(wchar_t);
+
+void gowchar_set (wchar_t *arr, int pos, wchar_t val)
+{
+	arr[pos] = val;
+}
+
+wchar_t gowchar_get (wchar_t *arr, int pos)
+{
+	return arr[pos];
+}
+*/
+import "C"
+
+import (
+	"fmt"
+	"unicode/utf16"
+	"unicode/utf8"
+)
+
+var SIZEOF_WCHAR_T C.size_t = C.size_t(C.SIZEOF_WCHAR_T)
+
+func getFilename(s string) *C.wchar_t {
+	wstr, _ := StringToWcharT(s)
+	return wstr
+}
+
+func StringToWcharT(s string) (*C.wchar_t, C.size_t) {
+	switch SIZEOF_WCHAR_T {
+	case 2:
+		return stringToWchar2(s) // Windows
+	case 4:
+		return stringToWchar4(s) // Unix
+	default:
+		panic(fmt.Sprintf("Invalid sizeof(wchar_t) = %v", SIZEOF_WCHAR_T))
+	}
+	panic("?!!")
+}
+
+// Windows
+func stringToWchar2(s string) (*C.wchar_t, C.size_t) {
+	var slen int
+	s1 := s
+	for len(s1) > 0 {
+		r, size := utf8.DecodeRuneInString(s1)
+		if er, _ := utf16.EncodeRune(r); er == '\uFFFD' {
+			slen += 1
+		} else {
+			slen += 2
+		}
+		s1 = s1[size:]
+	}
+	slen++ // \0
+	res := C.malloc(C.size_t(slen) * SIZEOF_WCHAR_T)
+	var i int
+	for len(s) > 0 {
+		r, size := utf8.DecodeRuneInString(s)
+		if r1, r2 := utf16.EncodeRune(r); r1 != '\uFFFD' {
+			C.gowchar_set((*C.wchar_t)(res), C.int(i), C.wchar_t(r1))
+			i++
+			C.gowchar_set((*C.wchar_t)(res), C.int(i), C.wchar_t(r2))
+			i++
+		} else {
+			C.gowchar_set((*C.wchar_t)(res), C.int(i), C.wchar_t(r))
+			i++
+		}
+		s = s[size:]
+	}
+	C.gowchar_set((*C.wchar_t)(res), C.int(slen-1), C.wchar_t(0)) // \0
+	return (*C.wchar_t)(res), C.size_t(slen)
+}
+
+// Unix
+func stringToWchar4(s string) (*C.wchar_t, C.size_t) {
+	slen := utf8.RuneCountInString(s)
+	slen++ // \0
+	res := C.malloc(C.size_t(slen) * SIZEOF_WCHAR_T)
+	var i int
+	for len(s) > 0 {
+		r, size := utf8.DecodeRuneInString(s)
+		C.gowchar_set((*C.wchar_t)(res), C.int(i), C.wchar_t(r))
+		s = s[size:]
+		i++
+	}
+	C.gowchar_set((*C.wchar_t)(res), C.int(slen-1), C.wchar_t(0)) // \0
+	return (*C.wchar_t)(res), C.size_t(slen)
+}

scanner/metadata/taglib/taglib_wrapper.cpp

@@ -17,7 +17,11 @@
 
 char has_cover(const TagLib::FileRef f);
 
+#ifdef WIN32
+int taglib_read(const wchar_t *filename, unsigned long id) {
+#else
 int taglib_read(const char *filename, unsigned long id) {
+#endif
   TagLib::FileRef f(filename, true, TagLib::AudioProperties::Fast);
 
   if (f.isNull()) {

scanner/metadata/taglib/taglib_wrapper.go

@@ -21,7 +21,7 @@ import (
 )
 
 func Read(filename string) (map[string][]string, error) {
-	fp := C.CString(filename)
+	fp := getFilename(filename)
 	defer C.free(unsafe.Pointer(fp))
 	id, m := newMap()
 	defer deleteMap(id)

scanner/metadata/taglib/taglib_wrapper.h

@@ -8,7 +8,11 @@ extern "C" {
 extern void go_map_put_str(unsigned long id, char *key, char *val);
 extern void go_map_put_int(unsigned long id, char *key, int val);
 
+#ifdef WIN32
+int taglib_read(const wchar_t *filename, unsigned long id);
+#else
 int taglib_read(const char *filename, unsigned long id);
+#endif
 
 #ifdef __cplusplus
 }