jiang13-bbs/backend/service/fileallow.go

package service

import (
	"bytes"
	"image"
	_ "image/gif"
	_ "image/jpeg"
	_ "image/png"
	"io"
	"mime"
	"os"
	"path/filepath"
	"strings"
	"unicode/utf8"

	"golang.org/x/image/webp"
)

// 活跃内容扩展名：下载时强制 application/octet-stream，避免浏览器当页面执行
var activeContentExts = map[string]struct{}{
	".html": {}, ".htm": {}, ".xhtml": {},
	".svg": {}, ".xml": {},
	".js": {}, ".mjs": {}, ".css": {},
	".wasm": {},
}

// RasterImageExts 栅格图扩展名（可 DecodeConfig；伪装内容降为 octet-stream）
var RasterImageExts = map[string]struct{}{
	".jpg": {}, ".jpeg": {}, ".png": {}, ".webp": {}, ".gif": {},
}

const attachmentImageMaxDim = 4096

// IsActiveContentExt 是否为浏览器可能内联执行的扩展名
func IsActiveContentExt(ext string) bool {
	_, ok := activeContentExts[strings.ToLower(ext)]
	return ok
}

// MimeForAttachmentExt 按扩展名映射 MIME；未知或活跃内容走 octet-stream 由调用方再处理
func MimeForAttachmentExt(ext string) string {
	ext = strings.ToLower(ext)
	if IsActiveContentExt(ext) {
		return "application/octet-stream"
	}
	if t := mime.TypeByExtension(ext); t != "" {
		if i := strings.IndexByte(t, ';'); i >= 0 {
			t = strings.TrimSpace(t[:i])
		}
		return t
	}
	return "application/octet-stream"
}

// looksLikeMarkupOrScript 检测内容是否像 HTML/SVG/XML 伪装（前 512 字节）
func looksLikeMarkupOrScript(data []byte) bool {
	n := len(data)
	if n > 512 {
		n = 512
	}
	head := bytes.TrimLeft(data[:n], " \t\r\n")
	if len(head) == 0 {
		return false
	}
	lower := bytes.ToLower(head)
	prefixes := [][]byte{
		[]byte("<!doctype"),
		[]byte("<html"),
		[]byte("<svg"),
		[]byte("<?xml"),
		[]byte("<script"),
	}
	for _, p := range prefixes {
		if bytes.HasPrefix(lower, p) {
			return true
		}
	}
	return false
}

// ResolveAttachmentMIME 根据扩展名与内存内容决定 MIME（小样例/测试用）。
func ResolveAttachmentMIME(ext string, data []byte) string {
	ext = strings.ToLower(ext)
	if IsActiveContentExt(ext) {
		return "application/octet-stream"
	}
	if _, ok := RasterImageExts[ext]; ok {
		if looksLikeMarkupOrScript(data) {
			return "application/octet-stream"
		}
		if !validRasterImage(bytes.NewReader(data), ext) {
			return "application/octet-stream"
		}
		return MimeForAttachmentExt(ext)
	}
	return MimeForAttachmentExt(ext)
}

// ResolveAttachmentMIMEFile 只读文件头判定 MIME，不把整文件读进内存。
func ResolveAttachmentMIMEFile(ext, path string) string {
	ext = strings.ToLower(ext)
	if IsActiveContentExt(ext) {
		return "application/octet-stream"
	}
	f, err := os.Open(path)
	if err != nil {
		return "application/octet-stream"
	}
	defer f.Close()

	if _, ok := RasterImageExts[ext]; ok {
		head := make([]byte, 512)
		n, _ := io.ReadFull(f, head)
		if n > 0 && looksLikeMarkupOrScript(head[:n]) {
			return "application/octet-stream"
		}
		if _, err := f.Seek(0, io.SeekStart); err != nil {
			return "application/octet-stream"
		}
		if !validRasterImage(f, ext) {
			return "application/octet-stream"
		}
		return MimeForAttachmentExt(ext)
	}
	return MimeForAttachmentExt(ext)
}

func validRasterImage(r io.Reader, ext string) bool {
	var cfg image.Config
	var err error
	switch ext {
	case ".webp":
		cfg, err = webp.DecodeConfig(r)
	default:
		cfg, _, err = image.DecodeConfig(r)
	}
	if err != nil {
		return false
	}
	if cfg.Width < 1 || cfg.Height < 1 {
		return false
	}
	if cfg.Width > attachmentImageMaxDim || cfg.Height > attachmentImageMaxDim {
		return false
	}
	return true
}

// sanitizeFilename 清理上传文件名：去路径、危险字符、RTL 伪装、截断
func sanitizeFilename(name string) string {
	name = filepath.Base(strings.ReplaceAll(name, "\\", "/"))
	name = strings.TrimSpace(name)
	name = strings.Map(func(r rune) rune {
		switch r {
		case '/', '\\', '\x00', ':', '*', '?', '"', '<', '>', '|',
			'\u202e', '\u202d', '\u200f', '\u200e',
			'\ufeff':
			return '_'
		default:
			if r < 0x20 {
				return '_'
			}
			return r
		}
	}, name)
	if utf8.RuneCountInString(name) > 200 {
		runes := []rune(name)
		name = string(runes[:200])
	}
	return name
}

// ExtOfFilename 取小写扩展名（含点），无则空串
func ExtOfFilename(name string) string {
	return strings.ToLower(filepath.Ext(name))
}