删除不需要的

Author: anyshu

tools/server/server-diffusion.cpp

@@ -722,7 +722,6 @@ struct callback_data {
     int32_t             n_input;
     void *              slot;           // For streaming updates (server_slot*)
     void *              ctx_server;     // For sending partial responses (server_context*)
-    std::string         last_sent_text; // Track last sent text for delta calculation
     llama_token *       last_tokens;   // Track last tokens for partial text decoding
 };
 
@@ -4115,17 +4114,6 @@ struct server_context {
                             // send the complete text as a single chunk before the final response
                             // In streaming mode with callbacks, the text was already sent incrementally
                             if (slot.params.stream) {
-                                // // Check if we need to send any remaining text that wasn't sent by callback
-                                // if (cb_data.last_sent_text != output_text && !output_text.empty()) {
-                                //     std::string remaining_text = output_text.substr(cb_data.last_sent_text.length());
-                                //     if (!remaining_text.empty()) {
-                                //         completion_token_output result;
-                                //         result.tok = -1;
-                                //         result.text_to_send = remaining_text;
-                                //         result.prob = 1.0f;
-                                //         send_partial_response(slot, result, false);
-                                //     }
-                                // }
                                 slot.generated_text = ""; // clear to avoid resending
                                 send_final_response(slot);
                             } else if (!output_text.empty()) {
@@ -4137,7 +4125,6 @@ struct server_context {
                                 send_partial_response(slot, result, false);
                                 send_final_response(slot);
                             }
-                            
 
                         } else {
                             send_error(slot, "Diffusion generation failed");
@@ -4816,11 +4803,6 @@ static bool diffusion_step_callback(int32_t             step,
         // Always send on first step, last step, or at regular intervals
         bool should_send = (step == 0) || 
                           (step == total_steps - 1);
-        
-        // // Also send if text has changed significantly (more tokens decoded)
-        // if (!should_send && current_text.length() > data->last_sent_text.length() + 10) {
-        //     should_send = true;
-        // }
 
         //for chat/completions
         if (true) {
@@ -4851,7 +4833,7 @@ static bool diffusion_step_callback(int32_t             step,
                     has_changes = true;
 
                     SRV_INF("Token changes at step %d: %zu positions changed\n", step, content_array.size());
-                    SRV_INF("Delta JSON: %s\n", delta_json.dump().c_str());
+                    SRV_DBG("Delta JSON: %s\n", delta_json.dump().c_str());
                 }
             }
 
@@ -4866,9 +4848,6 @@ static bool diffusion_step_callback(int32_t             step,
 
                 // Use is_progress=false to send actual content instead of progress info
                 ctx_server->send_partial_response(*slot, progress_token, false);
-                
-                // Update last sent text
-                data->last_sent_text = current_text;
             }
         }
     }