fix: resolve multipart/form-data boundary issue in HTTP Request compo nent #22880 (#23008)

Co-authored-by: crazywoola <427733928@qq.com>

api/core/workflow/nodes/http_request/executor.py

@@ -277,6 +277,22 @@ class Executor:
             elif self.auth.config.type == "custom":
                 headers[authorization.config.header] = authorization.config.api_key or ""
 
+        # Handle Content-Type for multipart/form-data requests
+        # Fix for issue #22880: Missing boundary when using multipart/form-data
+        body = self.node_data.body
+        if body and body.type == "form-data":
+            # For multipart/form-data with files, let httpx handle the boundary automatically
+            # by not setting Content-Type header when files are present
+            if not self.files or all(f[0] == "__multipart_placeholder__" for f in self.files):
+                # Only set Content-Type when there are no actual files
+                # This ensures httpx generates the correct boundary
+                if "content-type" not in (k.lower() for k in headers):
+                    headers["Content-Type"] = "multipart/form-data"
+        elif body and body.type in BODY_TYPE_TO_CONTENT_TYPE:
+            # Set Content-Type for other body types
+            if "content-type" not in (k.lower() for k in headers):
+                headers["Content-Type"] = BODY_TYPE_TO_CONTENT_TYPE[body.type]
+
         return headers
 
     def _validate_and_parse_response(self, response: httpx.Response) -> Response:
@@ -384,15 +400,24 @@ class Executor:
         # '__multipart_placeholder__' is inserted to force multipart encoding but is not a real file.
         # This prevents logging meaningless placeholder entries.
         if self.files and not all(f[0] == "__multipart_placeholder__" for f in self.files):
-            for key, (filename, content, mime_type) in self.files:
+            for file_entry in self.files:
+                # file_entry should be (key, (filename, content, mime_type)), but handle edge cases
+                if len(file_entry) != 2 or not isinstance(file_entry[1], tuple) or len(file_entry[1]) < 2:
+                    continue  # skip malformed entries
+                key = file_entry[0]
+                content = file_entry[1][1]
                 body_string += f"--{boundary}\r\n"
                 body_string += f'Content-Disposition: form-data; name="{key}"\r\n\r\n'
-                # decode content
+                # decode content safely
+                if isinstance(content, bytes):
                     try:
                         body_string += content.decode("utf-8")
                     except UnicodeDecodeError:
-                    # fix: decode binary content
-                    pass
+                        body_string += content.decode("utf-8", errors="replace")
+                elif isinstance(content, str):
+                    body_string += content
+                else:
+                    body_string += f"[Unsupported content type: {type(content).__name__}]"
                 body_string += "\r\n"
             body_string += f"--{boundary}--\r\n"
         elif self.node_data.body: