AI automatically translates all comments in the code into English (#3917)

src/Common/PacketCache.h

@@ -15,20 +15,26 @@
 #include "Util/List.h"
 
 namespace mediakit {
-/// 缓存刷新策略类
+// / 缓存刷新策略类  [AUTO-TRANSLATED:bd941d15]
+// / Cache refresh strategy class
 class FlushPolicy {
 public:
     bool isFlushAble(bool is_video, bool is_key, uint64_t new_stamp, size_t cache_size);
 
 private:
-    // 音视频的最后时间戳
+    // 音视频的最后时间戳  [AUTO-TRANSLATED:957d18ed]
+    // Last timestamp of audio and video
     uint64_t _last_stamp[2] = { 0, 0 };
 };
 
-/// 合并写缓存模板
-/// \tparam packet 包类型
-/// \tparam policy 刷新缓存策略
-/// \tparam packet_list 包缓存类型
+// / 合并写缓存模板  [AUTO-TRANSLATED:25cde944]
+// / Merge write cache template
+// / \tparam packet 包类型  [AUTO-TRANSLATED:43085d9b]
+// / \tparam packet Packet type
+// / \tparam policy 刷新缓存策略  [AUTO-TRANSLATED:c5ac29a7]
+// / \tparam policy Refresh cache strategy
+// / \tparam packet_list 包缓存类型  [AUTO-TRANSLATED:a434e7fe]
+// / \tparam packet_list Packet cache type
 template<typename packet, typename policy = FlushPolicy, typename packet_list = toolkit::List<std::shared_ptr<packet> > >
 class PacketCache {
 public:
@@ -42,7 +48,8 @@ public:
             flush();
         }
 
-        //追加数据到最后
+        // 追加数据到最后  [AUTO-TRANSLATED:e24ccfb6]
+        // Append data to the end
         _cache->emplace_back(std::move(pkt));
         if (key_pos) {
             _key_pos = key_pos;
@@ -70,10 +77,14 @@ public:
 
 private:
     bool flushImmediatelyWhenCloseMerge() {
-        // 一般的协议关闭合并写时，立即刷新缓存，这样可以减少一帧的延时，但是rtp例外
-        // 因为rtp的包很小，一个RtpPacket包中也不是完整的一帧图像，所以在关闭合并写时，
-        // 还是有必要缓冲一帧的rtp(也就是时间戳相同的rtp)再输出，这样虽然会增加一帧的延时
-        // 但是却对性能提升很大，这样做还是比较划算的
+        // 一般的协议关闭合并写时，立即刷新缓存，这样可以减少一帧的延时，但是rtp例外  [AUTO-TRANSLATED:54eba701]
+        // Generally, when the protocol closes the merge write, the cache is refreshed immediately, which can reduce the delay of one frame, but RTP is an exception.
+        // 因为rtp的包很小，一个RtpPacket包中也不是完整的一帧图像，所以在关闭合并写时，  [AUTO-TRANSLATED:b219082d]
+        // Because the RTP packet is very small, and a RtpPacket does not contain a complete frame of image, so when closing the merge write,
+        // 还是有必要缓冲一帧的rtp(也就是时间戳相同的rtp)再输出，这样虽然会增加一帧的延时  [AUTO-TRANSLATED:27c7ee8b]
+        // It is still necessary to buffer one frame of RTP (that is, RTP with the same timestamp) before outputting. Although this will increase the delay of one frame,
+        // 但是却对性能提升很大，这样做还是比较划算的  [AUTO-TRANSLATED:80eab719]
+        // But it greatly improves performance, so it is still worthwhile to do so.
 
         GET_CONFIG(int, mergeWriteMS, General::kMergeWriteMS);
         GET_CONFIG(int, rtspLowLatency, Rtsp::kLowLatency);