feat(backend): zip page index with natural sort, zip-slip rejection, content hash

backend/internal/bookfile/zip.go

@@ -0,0 +1,99 @@
+package bookfile
+
+import (
+	"archive/zip"
+	"errors"
+	"fmt"
+	"io"
+	"path"
+	"sort"
+	"strconv"
+	"strings"
+)
+
+var ErrNotZip = errors.New("not a readable zip")
+var ErrUnsafeZip = errors.New("unsafe zip entry")
+
+func unsafeEntry(n string) bool {
+	return strings.HasPrefix(n, "/") || strings.Contains(n, "..") || strings.ContainsRune(n, '\\')
+}
+
+func isImage(name string) bool {
+	switch strings.ToLower(path.Ext(name)) {
+	case ".jpg", ".jpeg", ".png", ".webp", ".gif", ".avif":
+		return true
+	}
+	return false
+}
+
+func PageIndex(f io.ReaderAt, size int64) ([]string, error) {
+	zr, err := zip.NewReader(f, size)
+	if err != nil {
+		return nil, ErrNotZip
+	}
+	var names []string
+	for _, zf := range zr.File {
+		if unsafeEntry(zf.Name) {
+			return nil, fmt.Errorf("%w: %s", ErrUnsafeZip, zf.Name)
+		}
+		if isImage(zf.Name) {
+			names = append(names, zf.Name)
+		}
+	}
+	sort.Slice(names, func(i, j int) bool { return NaturalLess(names[i], names[j]) })
+	return names, nil
+}
+
+func ReadEntry(f io.ReaderAt, size int64, name string) ([]byte, error) {
+	zr, err := zip.NewReader(f, size)
+	if err != nil {
+		return nil, ErrNotZip
+	}
+	for _, zf := range zr.File {
+		if zf.Name == name {
+			rc, err := zf.Open()
+			if err != nil {
+				return nil, err
+			}
+			defer rc.Close()
+			return io.ReadAll(io.LimitReader(rc, 64<<20))
+		}
+	}
+	return nil, errors.New("no such entry")
+}
+
+func NaturalLess(a, b string) bool {
+	i, j := 0, 0
+	for i < len(a) && j < len(b) {
+		da, db := isDigit(a[i]), isDigit(b[j])
+		switch {
+		case da && db:
+			si, sj := i, j
+			for i < len(a) && isDigit(a[i]) {
+				i++
+			}
+			for j < len(b) && isDigit(b[j]) {
+				j++
+			}
+			na, _ := strconv.Atoi(a[si:i])
+			nb, _ := strconv.Atoi(b[sj:j])
+			if na != nb {
+				return na < nb
+			}
+			if a[si:i] != b[sj:j] {
+				return a[si:i] < b[sj:j]
+			}
+		case !da && !db:
+			if a[i] != b[j] {
+				return a[i] < b[j]
+			}
+			i++
+			j++
+		default:
+			return da // 数字段排在字母前
+		}
+	}
+	return j < len(b)
+}
+
+func isDigit(c byte) bool { return c >= '0' && c <= '9' }