Files
navidrome/adapters/gotaglib/gotaglib.go
T
Deluan Quintão 31d94acfe7 fix(scanner): widen WASM panic recovery to cover tag/property reading (#5223)
* fix(scanner): widen WASM panic recovery to cover tag/property reading

The panic recovery in gotaglib's extractMetadata was only inside
openFile(), which covers taglib.OpenStream(). Panics from f.AllTags()
and f.Properties() (e.g. readString crashes on malformed files) were
uncaught, crashing the scanner subprocess with exit status 2.

Move the recover() up to extractMetadata() so it covers the entire
tag reading lifecycle, matching the CGO taglib wrapper's approach.

Fixes #5220

* fix(scanner): use consistent log key "filePath" in panic recovery

* fix(scanner): include stack trace in WASM panic recovery log

Signed-off-by: Deluan <deluan@navidrome.org>

---------

Signed-off-by: Deluan <deluan@navidrome.org>
2026-03-18 08:03:46 -04:00

302 lines
9.2 KiB
Go

// Package gotaglib provides an alternative metadata extractor using go-taglib,
// a pure Go (WASM-based) implementation of TagLib.
//
// This extractor aims for parity with the CGO-based taglib extractor. It uses
// TagLib's PropertyMap interface for standard tags. The File handle API provides
// efficient access to format-specific tags (ID3v2 frames, MP4 atoms, ASF attributes)
// through a single file open operation.
//
// This extractor is registered under the name "taglib". It only works with a filesystem
// (fs.FS) and does not support direct local file paths. Files returned by the filesystem
// must implement io.ReadSeeker for go-taglib to read them.
package gotaglib
import (
"errors"
"fmt"
"io"
"io/fs"
"runtime/debug"
"strings"
"time"
"github.com/navidrome/navidrome/conf"
"github.com/navidrome/navidrome/core/storage/local"
"github.com/navidrome/navidrome/log"
"github.com/navidrome/navidrome/model/metadata"
"go.senan.xyz/taglib"
)
type extractor struct {
fs fs.FS
}
func (e extractor) Parse(files ...string) (map[string]metadata.Info, error) {
results := make(map[string]metadata.Info)
for _, path := range files {
props, err := e.extractMetadata(path)
if err != nil {
continue
}
results[path] = *props
}
return results, nil
}
func (e extractor) Version() string {
bi, ok := debug.ReadBuildInfo()
if ok {
for _, dep := range bi.Deps {
if dep.Path == "go.senan.xyz/taglib" {
if dep.Replace != nil {
return dep.Replace.Version
}
return dep.Version
}
}
}
return "unknown"
}
func (e extractor) extractMetadata(filePath string) (info *metadata.Info, err error) {
// Recover from panics in the WASM runtime that can occur during any taglib
// operation (opening, reading tags, or reading properties). This catches crashes
// from malformed files or WASM runtime issues (e.g., wazero mmap failures on
// hardened systems with MemoryDenyWriteExecute=true).
debug.SetPanicOnFault(true)
defer func() {
if r := recover(); r != nil {
log.Error("gotaglib: WASM runtime panic reading file. Skipping", "filePath", filePath, "panic", r)
debug.PrintStack()
err = fmt.Errorf("WASM runtime panic: %v", r)
}
}()
f, close, err := e.openFile(filePath)
if err != nil {
log.Warn("gotaglib: Error reading metadata from file. Skipping", "filePath", filePath, err)
return nil, err
}
defer close()
// Get all tags and properties in one go
allTags := f.AllTags()
props := f.Properties()
// Map properties to AudioProperties
ap := metadata.AudioProperties{
Duration: props.Length.Round(time.Millisecond * 10),
BitRate: int(props.Bitrate),
Channels: int(props.Channels),
SampleRate: int(props.SampleRate),
BitDepth: int(props.BitsPerSample),
Codec: props.Codec,
}
// Convert normalized tags to lowercase keys (go-taglib returns UPPERCASE keys)
normalizedTags := make(map[string][]string, len(allTags.Tags))
for key, values := range allTags.Tags {
lowerKey := strings.ToLower(key)
normalizedTags[lowerKey] = values
}
// Process format-specific raw tags
processRawTags(allTags, normalizedTags)
// Parse track/disc totals from "N/Total" format
parseTuple(normalizedTags, "track")
parseTuple(normalizedTags, "disc")
// Adjust some ID3 tags
parseLyrics(normalizedTags)
parseTIPL(normalizedTags)
delete(normalizedTags, "tmcl") // TMCL is already parsed by TagLib
// Determine if file has embedded picture
hasPicture := len(props.Images) > 0
return &metadata.Info{
Tags: normalizedTags,
AudioProperties: ap,
HasPicture: hasPicture,
}, nil
}
// openFile opens the file at filePath using the extractor's filesystem.
// It returns a TagLib File handle and a cleanup function to close resources.
func (e extractor) openFile(filePath string) (f *taglib.File, closeFunc func(), err error) {
// Open the file from the filesystem
file, err := e.fs.Open(filePath)
if err != nil {
return nil, nil, err
}
rs, isSeekable := file.(io.ReadSeeker)
if !isSeekable {
file.Close()
return nil, nil, errors.New("file is not seekable")
}
// WithFilename provides a format detection hint via the file extension,
// since OpenStream alone relies on content-sniffing which fails for some files.
f, err = taglib.OpenStream(rs,
taglib.WithReadStyle(taglib.ReadStyleFast),
taglib.WithFilename(filePath),
)
if err != nil {
file.Close()
return nil, nil, err
}
closeFunc = func() {
f.Close()
file.Close()
}
return f, closeFunc, nil
}
// parseTuple parses track/disc numbers in "N/Total" format and separates them.
// For example, tracknumber="2/10" becomes tracknumber="2" and tracktotal="10".
func parseTuple(tags map[string][]string, prop string) {
tagName := prop + "number"
tagTotal := prop + "total"
if value, ok := tags[tagName]; ok && len(value) > 0 {
parts := strings.Split(value[0], "/")
tags[tagName] = []string{parts[0]}
if len(parts) == 2 {
tags[tagTotal] = []string{parts[1]}
}
}
}
// parseLyrics ensures lyrics tags have a language code.
// If lyrics exist without a language code, they are moved to "lyrics:xxx".
func parseLyrics(tags map[string][]string) {
lyrics := tags["lyrics"]
if len(lyrics) > 0 {
tags["lyrics:xxx"] = lyrics
delete(tags, "lyrics")
}
}
// processRawTags processes format-specific raw tags based on the detected file format.
// This handles ID3v2 frames (MP3/WAV/AIFF), MP4 atoms, and ASF attributes.
func processRawTags(allTags taglib.AllTags, normalizedTags map[string][]string) {
switch allTags.Format {
case taglib.FormatMPEG, taglib.FormatWAV, taglib.FormatAIFF:
parseID3v2Frames(allTags.Raw, normalizedTags)
case taglib.FormatMP4:
parseMP4Atoms(allTags.Raw, normalizedTags)
case taglib.FormatASF:
parseASFAttributes(allTags.Raw, normalizedTags)
}
}
// parseID3v2Frames processes ID3v2 raw frames to extract USLT/SYLT with language codes.
// This extracts language-specific lyrics that the standard Tags() doesn't provide.
func parseID3v2Frames(rawFrames map[string][]string, tags map[string][]string) {
// Process frames that have language-specific data
for key, values := range rawFrames {
lowerKey := strings.ToLower(key)
// Handle USLT:xxx and SYLT:xxx (lyrics with language codes)
if strings.HasPrefix(lowerKey, "uslt:") || strings.HasPrefix(lowerKey, "sylt:") {
parts := strings.SplitN(lowerKey, ":", 2)
if len(parts) == 2 && parts[1] != "" {
lang := parts[1]
lyricsKey := "lyrics:" + lang
tags[lyricsKey] = append(tags[lyricsKey], values...)
}
}
}
// If we found any language-specific lyrics from ID3v2 frames, remove the generic lyrics
for key := range tags {
if strings.HasPrefix(key, "lyrics:") && key != "lyrics" {
delete(tags, "lyrics")
break
}
}
}
const iTunesKeyPrefix = "----:com.apple.iTunes:"
// parseMP4Atoms processes MP4 raw atoms to get iTunes-specific tags.
func parseMP4Atoms(rawAtoms map[string][]string, tags map[string][]string) {
// Process all atoms and add them to tags
for key, values := range rawAtoms {
// Strip iTunes prefix and convert to lowercase
normalizedKey := strings.TrimPrefix(key, iTunesKeyPrefix)
normalizedKey = strings.ToLower(normalizedKey)
// Only add if the tag doesn't already exist (avoid duplication with PropertyMap)
if _, exists := tags[normalizedKey]; !exists {
tags[normalizedKey] = values
}
}
}
// parseASFAttributes processes ASF raw attributes to get WMA-specific tags.
func parseASFAttributes(rawAttrs map[string][]string, tags map[string][]string) {
// Process all attributes and add them to tags
for key, values := range rawAttrs {
normalizedKey := strings.ToLower(key)
// Only add if the tag doesn't already exist (avoid duplication with PropertyMap)
if _, exists := tags[normalizedKey]; !exists {
tags[normalizedKey] = values
}
}
}
// These are the only roles we support, based on Picard's tag map:
// https://picard-docs.musicbrainz.org/downloads/MusicBrainz_Picard_Tag_Map.html
var tiplMapping = map[string]string{
"arranger": "arranger",
"engineer": "engineer",
"producer": "producer",
"mix": "mixer",
"DJ-mix": "djmixer",
}
// parseTIPL parses the ID3v2.4 TIPL frame string, which is received from TagLib in the format:
//
// "arranger Andrew Powell engineer Chris Blair engineer Pat Stapley producer Eric Woolfson".
//
// and breaks it down into a map of roles and names, e.g.:
//
// {"arranger": ["Andrew Powell"], "engineer": ["Chris Blair", "Pat Stapley"], "producer": ["Eric Woolfson"]}.
func parseTIPL(tags map[string][]string) {
tipl := tags["tipl"]
if len(tipl) == 0 {
return
}
addRole := func(currentRole string, currentValue []string) {
if currentRole != "" && len(currentValue) > 0 {
role := tiplMapping[currentRole]
tags[role] = append(tags[role], strings.Join(currentValue, " "))
}
}
var currentRole string
var currentValue []string
for part := range strings.SplitSeq(tipl[0], " ") {
if _, ok := tiplMapping[part]; ok {
addRole(currentRole, currentValue)
currentRole = part
currentValue = nil
continue
}
currentValue = append(currentValue, part)
}
addRole(currentRole, currentValue)
delete(tags, "tipl")
}
var _ local.Extractor = (*extractor)(nil)
func init() {
local.RegisterExtractor("taglib", func(fsys fs.FS, baseDir string) local.Extractor {
return &extractor{fsys}
})
conf.AddHook(func() {
log.Debug("go-taglib version", "version", extractor{}.Version())
})
}