feat: 添加附件路径解析功能，支持非 UTF-8 编码文件名的回退机制

src/tools/session_send.rs

@@ -1,16 +1,20 @@
 use std::io::Read;
-use std::path::Path;
+use std::path::{Path, PathBuf};
 use std::sync::Arc;
 
 use anyhow::anyhow;
 use async_trait::async_trait;
 use base64::Engine;
+use encoding_rs::GBK;
 use serde_json::json;
 
 use crate::bus::MediaItem;
 
 use super::traits::{Tool, ToolContext, ToolResult};
 
+#[cfg(unix)]
+use std::os::unix::ffi::OsStringExt;
+
 #[derive(Debug, Clone)]
 pub struct SessionSendRequest {
     pub text: Option<String>,
@@ -165,6 +169,42 @@ fn validate_context(context: &ToolContext) -> anyhow::Result<()> {
     Ok(())
 }
 
+/// 解析附件文件路径，支持非 UTF-8 编码的文件名回退。
+///
+/// 在 Linux 上，文件名可能是 GBK 等非 UTF-8 编码。当 LLM 用 UTF-8 字符串
+/// 引用这些文件时，直接访问会失败。此函数先按原样（UTF-8）尝试访问，
+/// 如果文件不存在，则将文件名部分用 GBK 重新编码后再尝试。
+#[cfg_attr(not(unix), allow(unused_variables))]
+fn resolve_attachment_path(raw_path: &str) -> PathBuf {
+    let path = Path::new(raw_path);
+
+    // 先按原样（UTF-8）尝试
+    if path.exists() {
+        return path.to_path_buf();
+    }
+
+    // 提取父目录和文件名，只对文件名做编码回退
+    if let (Some(parent), Some(filename_os)) = (path.parent(), path.file_name()) {
+        let filename_str = filename_os.to_string_lossy();
+
+        // 尝试 GBK 编码回退
+        let (gbk_bytes, _, had_errors) = GBK.encode(&filename_str);
+        if !had_errors {
+            #[cfg(unix)]
+            {
+                let os_filename = std::ffi::OsString::from_vec(gbk_bytes.into_owned());
+                let resolved = parent.join(os_filename);
+                if resolved.exists() {
+                    return resolved;
+                }
+            }
+        }
+    }
+
+    // 回退失败，返回原路径（让调用方报错）
+    path.to_path_buf()
+}
+
 fn parse_attachments(value: &serde_json::Value) -> anyhow::Result<Vec<MediaItem>> {
     // 支持两种格式：实际数组 或 字符串化的 JSON 数组
     let paths = if let Some(arr) = value.as_array() {
@@ -198,7 +238,11 @@ fn parse_attachments(value: &serde_json::Value) -> anyhow::Result<Vec<MediaItem>
 
     let mut attachments = Vec::with_capacity(paths.len());
     for raw_path in paths {
-        let metadata = std::fs::metadata(&raw_path)
+        // 解析路径（含编码回退），确保能正确访问非 UTF-8 文件名的文件
+        let resolved_path = resolve_attachment_path(&raw_path);
+        let resolved_path_str = resolved_path.to_string_lossy().to_string();
+
+        let metadata = std::fs::metadata(&resolved_path)
             .map_err(|err| anyhow!("failed to access attachment '{}': {}", raw_path, err))?;
         if !metadata.is_file() {
             return Err(anyhow!("attachment path is not a file: {}", raw_path));
@@ -209,21 +253,20 @@ fn parse_attachments(value: &serde_json::Value) -> anyhow::Result<Vec<MediaItem>
 
         let content_base64 = (metadata.len() <= 50 * 1024 * 1024)
             .then(|| {
-                let mut file = std::fs::File::open(&raw_path)?;
+                let mut file = std::fs::File::open(&resolved_path)?;
                 let mut buf = Vec::with_capacity(metadata.len() as usize);
                 file.read_to_end(&mut buf)?;
                 Ok::<_, anyhow::Error>(base64::engine::general_purpose::STANDARD.encode(&buf))
             })
             .transpose()?;
 
-        let file_name = Path::new(&raw_path)
+        let file_name = Path::new(&resolved_path)
             .file_name()
-            .and_then(|n| n.to_str())
-            .map(ToOwned::to_owned);
+            .map(|n| n.to_string_lossy().to_string());
 
-        let media_type = infer_media_type(&raw_path);
-        let mut item = MediaItem::new(raw_path.to_string(), media_type);
-        item.mime_type = mime_guess::from_path(&raw_path)
+        let media_type = infer_media_type(&resolved_path_str);
+        let mut item = MediaItem::new(resolved_path_str, media_type);
+        item.mime_type = mime_guess::from_path(&resolved_path)
             .first_raw()
             .map(ToOwned::to_owned);
         item.content_base64 = content_base64;
@@ -371,4 +414,36 @@ mod tests {
         assert_eq!(attachments.len(), 1);
         assert_eq!(attachments[0].media_type, "file");
     }
+
+    #[test]
+    #[cfg(unix)]
+    fn parse_attachments_resolves_gbk_encoded_filename() {
+        use encoding_rs::GBK;
+        use std::os::unix::ffi::OsStringExt;
+
+        // UTF-8 中文文件名
+        let utf8_filename = "测试文件.txt";
+
+        // 编码为 GBK 字节
+        let (gbk_bytes, _, had_errors) = GBK.encode(utf8_filename);
+        assert!(!had_errors);
+
+        // 在临时目录中创建 GBK 编码文件名的文件
+        let tmp_dir = tempfile::tempdir().unwrap();
+        let gbk_osstring = std::ffi::OsString::from_vec(gbk_bytes.into_owned());
+        let gbk_path = tmp_dir.path().join(gbk_osstring);
+
+        std::fs::write(&gbk_path, b"test content").unwrap();
+
+        // 用 UTF-8 路径调用 parse_attachments
+        let utf8_path = tmp_dir.path().join(utf8_filename);
+        let utf8_path_str = utf8_path.to_string_lossy().to_string();
+
+        let attachments = parse_attachments(&json!([utf8_path_str])).unwrap();
+
+        assert_eq!(attachments.len(), 1);
+        assert_eq!(attachments[0].media_type, "file");
+        // 验证文件名能正确提取（用 lossy 方式，因为是 GBK 编码）
+        assert!(attachments[0].file_name.is_some());
+    }
 }