From 1b8cdb7ef66d5ec53459c830e4a7941884564dad Mon Sep 17 00:00:00 2001
From: pleasedontbreak123 <nabeiyq@126.com>
Date: Thu, 7 Aug 2025 11:00:27 +0800
Subject: [PATCH 01/17] before change

---
 monobean/resources/gtk/code_page.ui           |  3 +-
 monobean/resources/gtk/window.ui              |  2 +-
 .../scalable/apps/status-normal-icon.svg      | 10 ++++++
 ....Web3Infrastructure.Monobean.gresource.xml |  2 ++
 monobean/src/components/code_page.rs          |  5 +--
 monobean/src/components/hello_page.rs         |  1 +
 monobean/src/window.rs                        | 31 ++++++++++++++++++-
 7 files changed, 49 insertions(+), 5 deletions(-)
 create mode 100644 monobean/resources/icons/scalable/apps/status-normal-icon.svg

diff --git a/monobean/resources/gtk/code_page.ui b/monobean/resources/gtk/code_page.ui
index 97c9a109f..9d108818f 100644
--- a/monobean/resources/gtk/code_page.ui
+++ b/monobean/resources/gtk/code_page.ui
@@ -104,6 +104,7 @@
           </object>
         </child>
 
+<!--        sidebar-->
         <child>
           <object class="GtkBox" id="sidebar">
             <property name="orientation">vertical</property>
@@ -118,7 +119,7 @@
                 <child>
                   <object class="GtkImage">
                     <property name="icon-name">monobean-url-clone-symbolic</property>
-                    <property name="icon-size">large</property> <!-- 或者使用数字：1=small, 2=normal, 3=large, 4=extra-large -->
+                    <property name="icon-size">large</property> 
                   </object>
                 </child>
                 <style>
diff --git a/monobean/resources/gtk/window.ui b/monobean/resources/gtk/window.ui
index a06419f05..c8d0f1a2d 100644
--- a/monobean/resources/gtk/window.ui
+++ b/monobean/resources/gtk/window.ui
@@ -217,7 +217,7 @@
                                     <property name="spacing">4</property>
 
                                     <child>
-                                      <object class="GtkImage">
+                                      <object class="GtkImage" id="status_icon">
                                         <property name="icon-name">dialog-warning</property>
                                       </object>
                                     </child>
diff --git a/monobean/resources/icons/scalable/apps/status-normal-icon.svg b/monobean/resources/icons/scalable/apps/status-normal-icon.svg
new file mode 100644
index 000000000..66e2b672a
--- /dev/null
+++ b/monobean/resources/icons/scalable/apps/status-normal-icon.svg
@@ -0,0 +1,10 @@
+<?xml version="1.0" encoding="utf-8"?>
+<!-- Uploaded to: SVG Repo, www.svgrepo.com, Generator: SVG Repo Mixer Tools -->
+<!DOCTYPE svg PUBLIC "-//W3C//DTD SVG 1.1//EN" "http://www.w3.org/Graphics/SVG/1.1/DTD/svg11.dtd">
+<svg height="800px" width="800px" version="1.1" id="Layer_1" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" 
+	 viewBox="0 0 345 345" enable-background="new 0 0 345 345" xml:space="preserve">
+<g>
+	<circle fill="#1D912F" cx="172.5" cy="172.5" r="172.5"/>
+	<circle fill="#29CB41" cx="172.5" cy="172.5" r="157.5"/>
+</g>
+</svg>
\ No newline at end of file
diff --git a/monobean/resources/org.Web3Infrastructure.Monobean.gresource.xml b/monobean/resources/org.Web3Infrastructure.Monobean.gresource.xml
index 99b2173ac..2608bb14d 100644
--- a/monobean/resources/org.Web3Infrastructure.Monobean.gresource.xml
+++ b/monobean/resources/org.Web3Infrastructure.Monobean.gresource.xml
@@ -36,6 +36,7 @@
         <file compressed="true">icons/scalable/apps/walrus-4.png</file>
         <file compressed="true">icons/scalable/apps/walrus-5.png</file>
         <file compressed="true">icons/scalable/apps/walrus-6.png</file>
+        <file compressed="true">icons/scalable/apps/status-normal-icon.svg</file>
 
         <file preprocess="xml-stripblanks" compressed="true">icons/symbolic/apps/monobean-archive-restore-symbolic.svg</file>
         <file preprocess="xml-stripblanks" compressed="true">icons/symbolic/apps/monobean-arrow-down-symbolic.svg</file>
@@ -71,5 +72,6 @@
         <file preprocess="xml-stripblanks" compressed="true">icons/symbolic/apps/monobean-rss-symbolic.svg</file>
         <file preprocess="xml-stripblanks" compressed="true">icons/symbolic/apps/monobean-css-symbolic.svg</file>
         <file preprocess="xml-stripblanks" compressed="true">icons/symbolic/apps/monobean-json-symbolic.svg</file>
+        
     </gresource>
 </gresources>
diff --git a/monobean/src/components/code_page.rs b/monobean/src/components/code_page.rs
index 2e90b25fa..0c00c3d5a 100644
--- a/monobean/src/components/code_page.rs
+++ b/monobean/src/components/code_page.rs
@@ -153,6 +153,7 @@ impl CodePage {
         list_view.set_factory(Some(&factory));
     }
 
+    // sidebar button
     fn setup_button(&self) {
         let imp = self.imp();
 
@@ -182,13 +183,13 @@ impl CodePage {
         let history_btn = imp.history_btn.get();
         let history_popover = imp.history_popover.get();
 
-        // 设置 Popover 在按钮左侧弹出
+        // set Popover 
         history_popover.set_position(gtk::PositionType::Left);
         history_popover.set_has_arrow(false);
         history_popover.set_parent(&history_btn);
         history_popover.set_autohide(false);
 
-        // 按钮点击处理 - 确保位置正确
+        // 
         history_btn.connect_clicked(clone!(
             #[weak]
             history_popover,
diff --git a/monobean/src/components/hello_page.rs b/monobean/src/components/hello_page.rs
index 87cb5593e..1aae8b1e1 100644
--- a/monobean/src/components/hello_page.rs
+++ b/monobean/src/components/hello_page.rs
@@ -289,6 +289,7 @@ impl HelloPage {
             self.imp().email_entry.text()
         );
 
+        // megacore need some time to init , so sleep 1.5s
         glib::MainContext::default().spawn_local(clone!(
             #[weak(rename_to=this)]
             self,
diff --git a/monobean/src/window.rs b/monobean/src/window.rs
index fc2eeca6d..48655648b 100644
--- a/monobean/src/window.rs
+++ b/monobean/src/window.rs
@@ -20,6 +20,7 @@ use gtk::glib;
 use gtk::prelude::{ButtonExt, GtkWindowExt, WidgetExt};
 use gtk::CompositeTemplate;
 use std::cell::OnceCell;
+use crate::core::mega_core::MegaCommands::CoreStatus;
 
 glib::wrapper! {
     pub struct MonobeanWindow(ObjectSubclass<imp::MonobeanWindow>)
@@ -69,9 +70,13 @@ mod imp {
         #[template_child]
         pub not_implemented: TemplateChild<NotImplemented>,
 
+        //bottom bar component
         #[template_child]
         pub mega_status_button: TemplateChild<gtk::Button>,
-
+        #[template_child]
+        pub status_icon: TemplateChild<gtk::Image>,
+        #[template_child]
+        pub status_label: TemplateChild<gtk::Label>,
         pub mega_popup: RefCell<Option<MegaTabWindow>>,
 
         pub sender: OnceCell<Sender<Action>>,
@@ -260,6 +265,8 @@ impl MonobeanWindow {
     fn setup_bottombar_widget(&self) {
         let imp = self.imp();
         let mega_status_button = imp.mega_status_button.get();
+        let status_icon = imp.status_icon.get();
+        let status_label = imp.status_label.get();
         mega_status_button.connect_clicked(clone!(
             #[weak(rename_to=window)]
             self,
@@ -284,6 +291,28 @@ impl MonobeanWindow {
                 imp.mega_popup.replace(Some(popup));
             }
         ));
+        
+        // mega status bar show 
+        let monobean_application:MonobeanApplication = self.application().unwrap().downcast().unwrap();
+        CONTEXT.spawn_local_with_priority(Priority::DEFAULT_IDLE, async move {
+            loop {
+                let rx = monobean_application.core_status();
+                let (core_started, _) = rx.await.unwrap();
+                if core_started {
+                    status_label.set_label("Mega started");
+                    status_icon.set_icon_name(Some("status-normal-icon"));
+                    //tracing::debug!("watching mage status----------")
+                } else {
+                    status_label.set_label("Mega stoped");
+                    status_icon.set_icon_name(Some("dialog-warning"));
+                    tracing::debug!("watching mage status faild----------")
+                }
+                glib::timeout_future(std::time::Duration::from_secs(5)).await;
+            }
+            
+        });
+        
+        //let action = Action::MegaCore(CoreStatus());
     }
 
     pub fn add_toast(&self, message: String) {

From 29200e8f863ba9c04d41332d009ac16744bc7461 Mon Sep 17 00:00:00 2001
From: pleasedontbreak123 <nabeiyq@126.com>
Date: Thu, 14 Aug 2025 09:42:46 +0800
Subject: [PATCH 02/17] adjust ui

---
 monobean/resources/gtk/preferences.ui         |  4 +-
 monobean/resources/gtk/window.ui              | 67 +++++++++++++------
 .../icons/scalable/apps/link-status-icon.svg  |  4 ++
 ....Web3Infrastructure.Monobean.gresource.xml |  1 +
 monobean/src/core/mega_core.rs                |  7 ++
 monobean/src/window.rs                        |  2 +
 6 files changed, 61 insertions(+), 24 deletions(-)
 create mode 100644 monobean/resources/icons/scalable/apps/link-status-icon.svg

diff --git a/monobean/resources/gtk/preferences.ui b/monobean/resources/gtk/preferences.ui
index cbf8f8cbd..a891f36ba 100644
--- a/monobean/resources/gtk/preferences.ui
+++ b/monobean/resources/gtk/preferences.ui
@@ -8,7 +8,7 @@
     <child>
       <object class="AdwPreferencesPage">
         <property name="title" translatable="yes">Mega Configuration</property>
-        <property name="icon_name">preferences-desktop-apps-symbolic</property>
+        <property name="icon_name">monobean-repo-symbolic</property>
         <child>
           <object class="AdwPreferencesGroup">
             <property name="title" translatable="yes">Base Settings</property>
@@ -483,7 +483,7 @@
     <child>
       <object class="AdwPreferencesPage">
         <property name="title" translatable="yes">P2P</property>
-        <property name="icon_name">user-available-symbolic</property>
+        <property name="icon_name">monobean-network-symbolic</property>
         <child>
           <object class="AdwPreferencesGroup">
             <property name="title" translatable="yes">Relay Settings</property>
diff --git a/monobean/resources/gtk/window.ui b/monobean/resources/gtk/window.ui
index c8d0f1a2d..b7249cbb2 100644
--- a/monobean/resources/gtk/window.ui
+++ b/monobean/resources/gtk/window.ui
@@ -73,35 +73,58 @@
 
 
                     <!-- control visable-->
-                    <child>
-                      <object class="GtkBox" id="search_container">
-                        <property name="orientation">horizontal</property>
-                        <property name="spacing">0</property>
+<!--                    <child>-->
+<!--                      <object class="GtkBox" id="search_container">-->
+<!--                        <property name="orientation">horizontal</property>-->
+<!--                        <property name="spacing">0</property>-->
+
+<!--                        &lt;!&ndash; 搜索框，拉伸 &ndash;&gt;-->
+<!--                        <child>-->
+<!--                          <object class="GtkSearchEntry" id="search_entry">-->
+<!--                            <property name="hexpand">true</property>-->
+<!--                            <property name="placeholder-text">请输入文本</property>-->
+<!--                          </object>-->
+<!--                        </child>-->
+
+<!--                        &lt;!&ndash; 搜索按钮 &ndash;&gt;-->
+<!--                        <child>-->
+<!--                          <object class="GtkButton" id="search_button">-->
+<!--                            <property name="label">搜索</property>-->
+<!--                          </object>-->
+<!--                        </child>-->
+
+<!--                      </object>-->
+<!--                    </child>-->
 
-                        <!-- 搜索框，拉伸 -->
-                        <child>
-                          <object class="GtkSearchEntry" id="search_entry">
-                            <property name="hexpand">true</property>
-                            <property name="placeholder-text">请输入文本</property>
-                          </object>
-                        </child>
+                  </object>
+                </child>
 
-                        <!-- 搜索按钮 -->
-                        <child>
-                          <object class="GtkButton" id="search_button">
-                            <property name="label">搜索</property>
-                          </object>
-                        </child>
 
+                <child type="title">
+                  <object class="GtkBox" id="search_container">
+                    <property name="orientation">horizontal</property>
+                    <property name="spacing">5</property>
+
+                    <child>
+                      <object class="GtkBox" >
+                        <property name="hexpand">true</property>
                       </object>
                     </child>
 
-                  </object>
-                </child>
+                    <child>
+                      <object class="GtkSearchEntry" id="search_entry">
+                        <property name="hexpand">true</property>
+                        <property name="placeholder-text">请输入文本</property>
+                      </object>
+                    </child>
 
+                    <child>
+                      <object class="GtkBox" >
+                        <property name="hexpand">true</property>
+                      </object>
+                    </child>
 
-                <child type="title">
-                  <object class="GtkBox"/>
+                  </object>
                 </child>
 
 
@@ -253,7 +276,7 @@
 
                                 <child>
                                   <object class="GtkImage">
-                                    <property name="icon-name">network-wired-symbolic</property>
+                                    <property name="icon-name">link-status-icon</property>
                                   </object>
                                 </child>
 
diff --git a/monobean/resources/icons/scalable/apps/link-status-icon.svg b/monobean/resources/icons/scalable/apps/link-status-icon.svg
new file mode 100644
index 000000000..cc8a4eae4
--- /dev/null
+++ b/monobean/resources/icons/scalable/apps/link-status-icon.svg
@@ -0,0 +1,4 @@
+<?xml version="1.0" encoding="utf-8"?><!-- Uploaded to: SVG Repo, www.svgrepo.com, Generator: SVG Repo Mixer Tools -->
+<svg width="800px" height="800px" viewBox="0 0 24 24" fill="none" xmlns="http://www.w3.org/2000/svg">
+<path d="M10 15H12C15.3137 15 18 12.3137 18 9C18 5.68629 15.3137 3 12 3H8C4.68629 3 2 5.68629 2 9C2 10.5367 2.57771 11.9385 3.52779 13M14 9H12C8.68629 9 6 11.6863 6 15C6 18.3137 8.68629 21 12 21H16C19.3137 21 22 18.3137 22 15C22 13.4633 21.4223 12.0615 20.4722 11" stroke="#1C274C" stroke-width="1.5" stroke-linecap="round"/>
+</svg>
\ No newline at end of file
diff --git a/monobean/resources/org.Web3Infrastructure.Monobean.gresource.xml b/monobean/resources/org.Web3Infrastructure.Monobean.gresource.xml
index 2608bb14d..1ad1f3af9 100644
--- a/monobean/resources/org.Web3Infrastructure.Monobean.gresource.xml
+++ b/monobean/resources/org.Web3Infrastructure.Monobean.gresource.xml
@@ -37,6 +37,7 @@
         <file compressed="true">icons/scalable/apps/walrus-5.png</file>
         <file compressed="true">icons/scalable/apps/walrus-6.png</file>
         <file compressed="true">icons/scalable/apps/status-normal-icon.svg</file>
+        <file compressed="true">icons/scalable/apps/link-status-icon.svg</file>
 
         <file preprocess="xml-stripblanks" compressed="true">icons/symbolic/apps/monobean-archive-restore-symbolic.svg</file>
         <file preprocess="xml-stripblanks" compressed="true">icons/symbolic/apps/monobean-arrow-down-symbolic.svg</file>
diff --git a/monobean/src/core/mega_core.rs b/monobean/src/core/mega_core.rs
index fc40e5b31..111bb895a 100644
--- a/monobean/src/core/mega_core.rs
+++ b/monobean/src/core/mega_core.rs
@@ -64,6 +64,9 @@ pub enum MegaCommands {
         chan: oneshot::Sender<MonoBeanResult<String>>,
         id: String,
     },
+    // GetRepoUrl {
+    //     chan: oneshot::Sender<MonoBeanResult<String>>,
+    // }
 }
 
 impl Debug for MegaCore {
@@ -184,6 +187,9 @@ impl MegaCore {
                 let content = self.load_blob(path).await;
                 chan.send(content).unwrap();
             }
+            // MegaCommands::GetRepoUrl => {
+            //
+            // }
         }
     }
 
@@ -287,6 +293,7 @@ impl MegaCore {
         *self.running_context.write().await = None;
     }
 
+    // choose ImportApiService or MonoApiService
     async fn api_handler(&self, path: impl AsRef<Path>) -> MonoBeanResult<Box<dyn ApiHandler>> {
         let ctx = self.running_context.read().await.clone();
         if ctx.is_none() {
diff --git a/monobean/src/window.rs b/monobean/src/window.rs
index 48655648b..6bd9c5b04 100644
--- a/monobean/src/window.rs
+++ b/monobean/src/window.rs
@@ -41,6 +41,7 @@ mod imp {
     #[derive(Default, CompositeTemplate)]
     #[template(resource = "/org/Web3Infrastructure/Monobean/gtk/window.ui")]
     pub struct MonobeanWindow {
+        // headbar components
         #[template_child]
         pub header_bar: TemplateChild<adw::HeaderBar>,
         #[template_child]
@@ -58,6 +59,7 @@ mod imp {
         #[template_child]
         pub search_container: TemplateChild<gtk::Box>,
 
+        // content page
         #[template_child]
         pub hello_page: TemplateChild<HelloPage>,
         // #[template_child]

From d4e5d4e0389c9c4c340fa89005e99f00d9a5ed30 Mon Sep 17 00:00:00 2001
From: pleasedontbreak123 <nabeiyq@126.com>
Date: Thu, 14 Aug 2025 22:21:17 +0800
Subject: [PATCH 03/17] before break change

---
 monobean/Cargo.toml                     |   1 +
 monobean/src/components/code_page.rs    | 104 ++++++++++++++++++++++--
 monobean/src/components/history_list.rs |  10 ++-
 monobean/src/core/mega_core.rs          |  38 ++++++++-
 monobean/src/window.rs                  |  37 +++++----
 5 files changed, 163 insertions(+), 27 deletions(-)

diff --git a/monobean/Cargo.toml b/monobean/Cargo.toml
index 19189cad7..25ceab47e 100644
--- a/monobean/Cargo.toml
+++ b/monobean/Cargo.toml
@@ -26,6 +26,7 @@ gateway = { path = "../gateway" }
 ceres = { path = "../ceres" }
 mercury = { path = "../mercury" }
 context = {path = "../context"}
+libra = {path = "../libra"}
 
 thiserror = { version = "2.0.11", features = ["default"] }
 tracing-subscriber = { version = "0.3.19", features = ["default", "env-filter"] }
diff --git a/monobean/src/components/code_page.rs b/monobean/src/components/code_page.rs
index 0c00c3d5a..02c78122d 100644
--- a/monobean/src/components/code_page.rs
+++ b/monobean/src/components/code_page.rs
@@ -15,7 +15,7 @@ use tokio::sync::oneshot;
 
 mod imp {
     use adw::subclass::prelude::BinImpl;
-    use tokio::sync::OnceCell;
+    use tokio::sync::{OnceCell};
 
     use crate::components::file_tree::FileTreeView;
 
@@ -52,6 +52,7 @@ mod imp {
         #[template_child(id = "history_listview")]
         pub history_listview: TemplateChild<gtk::ListView>,
 
+        pub cur_file_path: std::sync::RwLock<PathBuf>,
         pub sender: OnceCell<Sender<Action>>,
     }
 
@@ -105,21 +106,21 @@ impl CodePage {
 
         self.setup_paned();
         self.setup_button();
-        self.setup_history_list();
         self.setup_source_view(opened_file);
         self.setup_file_tree();
     }
 
     fn setup_history_list(&self) {
+        let imp = self.imp();
+        let file_path = imp.cur_file_path.read().unwrap().clone();
+        let sender = imp.sender.get().unwrap().clone();
+
         let list_view = self.imp().history_listview.clone();
 
         // 创建数据模型
         let store = ListStore::new::<HistoryItem>();
 
-        for i in 1..=20 {
-            store.append(&HistoryItem::new(&format!("commit信息 \n 提交人:{i}, sha")));
-        }
-        let selection_model = SingleSelection::new(Some(store));
+        let selection_model = SingleSelection::new(Some(store.clone()));
 
         let factory = SignalListItemFactory::new();
 
@@ -151,6 +152,83 @@ impl CodePage {
 
         list_view.set_model(Some(&selection_model));
         list_view.set_factory(Some(&factory));
+
+        CONTEXT.spawn_local(clone!(
+            #[weak] store,
+            async move{
+                let (tx, rx) = oneshot::channel();
+                sender
+                    .send(Action::MegaCore(MegaCommands::GetPathHistory {
+                        chan: tx,
+                        path: file_path.to_str().unwrap().to_string(),
+                    }))
+                    .await
+                    .unwrap();
+
+                match rx.await.unwrap() {
+                    Ok(commits)=>{
+                        tracing::debug!("Received commit history{:?}",commits);
+                        store.remove_all();
+                        for commit in commits {
+                            store.append(&HistoryItem::new(&commit.id._to_string(),&format!(
+                                "Commit:{}\nAuthor: {}\n message: {} \n Date: {}",
+                                commit.id, commit.author,commit.message,commit.committer.timestamp
+                            )));
+                        }
+                    }
+                    Err(e)=>{
+                        tracing::error!("get history commits error: {:?}", e);
+                        return;
+                    }
+
+                }
+
+
+
+            }
+        ));
+
+
+        // let list_view = self.imp().history_listview.clone();
+        //
+        // // 创建数据模型
+        // let store = ListStore::new::<HistoryItem>();
+        //
+        // for i in 1..=20 {
+        //     store.append(&HistoryItem::new(&format!("commit信息 \n 提交人:{i}, sha")));
+        // }
+        // let selection_model = SingleSelection::new(Some(store));
+        //
+        // let factory = SignalListItemFactory::new();
+        //
+        // factory.connect_setup(move |_, list_item| {
+        //     let label = Label::new(None);
+        //     list_item
+        //         .downcast_ref::<ListItem>()
+        //         .expect("Needs to be ListItem")
+        //         .set_child(Some(&label));
+        // });
+        //
+        // factory.connect_bind(move |_, list_item| {
+        //     let obj = list_item
+        //         .downcast_ref::<ListItem>()
+        //         .expect("Needs to be ListItem")
+        //         .item()
+        //         .and_downcast::<HistoryItem>()
+        //         .expect("Model should be HistoryItem");
+        //
+        //     let label = list_item
+        //         .downcast_ref::<ListItem>()
+        //         .expect("Needs to be ListItem")
+        //         .child()
+        //         .and_downcast::<Label>()
+        //         .expect("Child widget must be a Label");
+        //
+        //     label.set_label(&obj.text()); // 获取属性
+        // });
+        //
+        // list_view.set_model(Some(&selection_model));
+        // list_view.set_factory(Some(&factory));
     }
 
     // sidebar button
@@ -245,6 +323,20 @@ impl CodePage {
 
     pub fn show_editor_on(&self, hash: String, name: String, path: PathBuf) {
         let imp = self.imp();
+
+        tracing::debug!("show_editor_on: path: {:?}", path);
+
+        {
+            let mut cur_path = imp.cur_file_path.write().unwrap();
+            let mut git_path = path.to_string_lossy().replace('\\', "/");
+            if !git_path.starts_with('/') {
+                git_path = format!("/{}", git_path);
+            }
+            // *cur_path = git_path.parse().unwrap();
+            *cur_path = "/third-party/demo3".parse().unwrap();
+        }
+        self.setup_history_list();
+
         imp.file_path_label
             .set_text(&path.to_str().unwrap().replace(['/', '\\'], " > "));
 
diff --git a/monobean/src/components/history_list.rs b/monobean/src/components/history_list.rs
index b63a4c795..7597c878a 100644
--- a/monobean/src/components/history_list.rs
+++ b/monobean/src/components/history_list.rs
@@ -7,10 +7,13 @@ mod imp {
     use gtk::glib;
     use gtk::prelude::*;
     use gtk::subclass::prelude::*;
+ 
 
     #[derive(Properties, Default)]
     #[properties(wrapper_type = super::HistoryItem)]
     pub struct HistoryItem {
+        #[property(get, set)]
+        pub id: RefCell<String>,
         #[property(get, set)]
         pub text: RefCell<String>,
     }
@@ -30,7 +33,10 @@ glib::wrapper! {
 }
 
 impl HistoryItem {
-    pub fn new(text: &str) -> Self {
-        glib::Object::builder().property("text", text).build()
+    pub fn new(id: &str,text: &str) -> Self {
+        glib::Object::builder()
+            .property("text", text)
+            .property("id", id)
+            .build()
     }
 }
diff --git a/monobean/src/core/mega_core.rs b/monobean/src/core/mega_core.rs
index 111bb895a..278be84ff 100644
--- a/monobean/src/core/mega_core.rs
+++ b/monobean/src/core/mega_core.rs
@@ -20,6 +20,8 @@ use std::sync::Arc;
 use tokio::sync::{oneshot, OnceCell, RwLock};
 use vault::integration::vault_core::VaultCore;
 use vault::pgp::{SignedPublicKey, SignedSecretKey};
+use libra::command::log::{get_reachable_commits};
+use mercury::internal::object::commit::Commit;
 
 pub struct MegaCore {
     config: Arc<RwLock<Config>>,
@@ -67,6 +69,10 @@ pub enum MegaCommands {
     // GetRepoUrl {
     //     chan: oneshot::Sender<MonoBeanResult<String>>,
     // }
+    GetPathHistory{
+        chan: oneshot::Sender<MonoBeanResult<Vec<Commit>>>,
+        path: String,
+    },
 }
 
 impl Debug for MegaCore {
@@ -187,6 +193,10 @@ impl MegaCore {
                 let content = self.load_blob(path).await;
                 chan.send(content).unwrap();
             }
+            MegaCommands::GetPathHistory{ chan, path} => {
+                let commits = self.get_path_history(&path).await;
+                chan.send(commits).unwrap();
+            }
             // MegaCommands::GetRepoUrl => {
             //
             // }
@@ -293,7 +303,7 @@ impl MegaCore {
         *self.running_context.write().await = None;
     }
 
-    // choose ImportApiService or MonoApiService
+    /// choose ImportApiService or MonoApiService judging by path prefix
     async fn api_handler(&self, path: impl AsRef<Path>) -> MonoBeanResult<Box<dyn ApiHandler>> {
         let ctx = self.running_context.read().await.clone();
         if ctx.is_none() {
@@ -314,6 +324,7 @@ impl MegaCore {
                 .unwrap()
             {
                 let repo: Repo = model.into();
+                //tracing::debug!("@@@@@@@@@@@@@@@@@@@@@@@@Found repo {:?}", repo.repo_id);
                 return Ok(Box::new(ImportApiService {
                     storage: ctx.storage.clone(),
                     repo,
@@ -339,6 +350,7 @@ impl MegaCore {
                 acc + &e.as_os_str().to_string_lossy() + "/"
             });
         let path = PathBuf::from(path);
+        tracing::debug!("!!!!!!!!!!!!!!!!Loading tree from path: {}", path.display());
 
         let handler = self.api_handler(&path).await?;
         let tree = handler.search_tree_by_path(&path).await;
@@ -467,6 +479,24 @@ impl MegaCore {
     pub async fn is_core_running(&self) -> bool {
         self.running_context.read().await.is_some()
     }
+
+    async fn get_path_history(&self,path: impl AsRef<Path>) ->  MonoBeanResult<Vec<Commit>>{
+
+        let api_handler = self.api_handler(&path).await?;
+        let cur_commit = api_handler.get_latest_commit(path.as_ref().to_path_buf())
+            .await
+            .map_err(|e| MonoBeanError::MegaCoreError(format!("Failed to get latest commit: {:?}", e)))?;
+
+
+         let mut reachable_commits = get_reachable_commits(cur_commit.oid.clone()).await;
+        //let mut reachable_commits = get_reachable_commits("").await;
+        reachable_commits.sort_by_key(|c| c.committer.timestamp);
+        tracing::debug!("Reachable commits: {:?}", reachable_commits);
+
+
+
+        Ok(reachable_commits)
+    }
 }
 
 #[cfg(test)]
@@ -583,4 +613,10 @@ mod tests {
 
     #[tokio::test]
     async fn test_run_with_config() {}
+
+    #[tokio::test]
+    async fn test_get_history() {
+
+    }
+
 }
diff --git a/monobean/src/window.rs b/monobean/src/window.rs
index 6bd9c5b04..996e9990e 100644
--- a/monobean/src/window.rs
+++ b/monobean/src/window.rs
@@ -295,24 +295,25 @@ impl MonobeanWindow {
         ));
         
         // mega status bar show 
-        let monobean_application:MonobeanApplication = self.application().unwrap().downcast().unwrap();
-        CONTEXT.spawn_local_with_priority(Priority::DEFAULT_IDLE, async move {
-            loop {
-                let rx = monobean_application.core_status();
-                let (core_started, _) = rx.await.unwrap();
-                if core_started {
-                    status_label.set_label("Mega started");
-                    status_icon.set_icon_name(Some("status-normal-icon"));
-                    //tracing::debug!("watching mage status----------")
-                } else {
-                    status_label.set_label("Mega stoped");
-                    status_icon.set_icon_name(Some("dialog-warning"));
-                    tracing::debug!("watching mage status faild----------")
-                }
-                glib::timeout_future(std::time::Duration::from_secs(5)).await;
-            }
-            
-        });
+        // todo here produce tons of log ,need to adjust logic of generate log 
+        // let monobean_application:MonobeanApplication = self.application().unwrap().downcast().unwrap();
+        // CONTEXT.spawn_local_with_priority(Priority::DEFAULT_IDLE, async move {
+        //     loop {
+        //         let rx = monobean_application.core_status();
+        //         let (core_started, _) = rx.await.unwrap();
+        //         if core_started {
+        //             status_label.set_label("Mega started");
+        //             status_icon.set_icon_name(Some("status-normal-icon"));
+        //             //tracing::debug!("watching mage status----------")
+        //         } else {
+        //             status_label.set_label("Mega stoped");
+        //             status_icon.set_icon_name(Some("dialog-warning"));
+        //             tracing::debug!("watching mage status faild----------")
+        //         }
+        //         glib::timeout_future(std::time::Duration::from_secs(5)).await;
+        //     }
+        //     
+        // });
         
         //let action = Action::MegaCore(CoreStatus());
     }

From cf17417942683b30e2bfd22d48b5edaa5445b53e Mon Sep 17 00:00:00 2001
From: pleasedontbreak123 <nabeiyq@126.com>
Date: Sat, 16 Aug 2025 16:47:04 +0800
Subject: [PATCH 04/17] history commit browse & history file content display

---
 monobean/Cargo.toml                     |  11 +-
 monobean/resources/gtk/code_page.ui     |   4 +-
 monobean/src/components/code_page.rs    | 149 ++++++++++++++----------
 monobean/src/components/history_list.rs |   5 +-
 monobean/src/core/mega_core.rs          |  94 ++++++++++++---
 monobean/src/window.rs                  |  44 +++----
 6 files changed, 195 insertions(+), 112 deletions(-)

diff --git a/monobean/Cargo.toml b/monobean/Cargo.toml
index 25ceab47e..f3c179ea5 100644
--- a/monobean/Cargo.toml
+++ b/monobean/Cargo.toml
@@ -20,19 +20,18 @@ scv = { version = "0.9.0", package = "sourceview5", features = ["gtk_v4_12", "v5
 # To run mega core, we have to add these:
 mono = { path = "../mono" }
 common = { path = "../common" }
-jupiter = { path = "../jupiter" }
+
 vault = { path = "../vault" }
 gateway = { path = "../gateway" }
 ceres = { path = "../ceres" }
 mercury = { path = "../mercury" }
 context = {path = "../context"}
-libra = {path = "../libra"}
+
 
 thiserror = { version = "2.0.11", features = ["default"] }
 tracing-subscriber = { version = "0.3.19", features = ["default", "env-filter"] }
 tokio = "1.45.0"
 russh = "0.52"
-axum = "0.8.1"
 axum-server = "0.7.1"
 bytes = "1.10.0"
 async-channel = "2.3.1"
@@ -41,13 +40,11 @@ rustls = "0.23.23"
 gix-config = "0.45.1"
 home = "0.5.11"
 smallvec = "1.14.0"
-directories = "6.0.0"
 tracing-appender = "0.2.3"
 toml = "0.8.22"
-#once_cell = "1.21.3"
+chrono = "0.4"
+
 
-resvg = "0.27"
-tiny-skia = "0.11"
 
 [build-dependencies]
 glib-build-tools = "0.20.0"
diff --git a/monobean/resources/gtk/code_page.ui b/monobean/resources/gtk/code_page.ui
index 9d108818f..5bb1dfc0c 100644
--- a/monobean/resources/gtk/code_page.ui
+++ b/monobean/resources/gtk/code_page.ui
@@ -260,9 +260,9 @@
               <property name="vexpand">True</property>
               <property name="hexpand">True</property>
               <property name="propagate-natural-height">False</property>
-              <property name="min-content-height">400</property>
+              <property name="min-content-height">500</property>
               <property name="max-content-height">600</property>
-              <property name="min-content-width">150</property>
+              <property name="min-content-width">250</property>
               <property name="hscrollbar-policy">automatic</property>
               <property name="vscrollbar-policy">automatic</property>
               
diff --git a/monobean/src/components/code_page.rs b/monobean/src/components/code_page.rs
index 02c78122d..732de944b 100644
--- a/monobean/src/components/code_page.rs
+++ b/monobean/src/components/code_page.rs
@@ -12,6 +12,7 @@ use gtk::{glib, CompositeTemplate, Label, ListItem, SignalListItemFactory, Singl
 use scv::{prelude::*, Buffer};
 use std::path::{Path, PathBuf};
 use tokio::sync::oneshot;
+use chrono::{DateTime, Utc};
 
 mod imp {
     use adw::subclass::prelude::BinImpl;
@@ -114,16 +115,12 @@ impl CodePage {
         let imp = self.imp();
         let file_path = imp.cur_file_path.read().unwrap().clone();
         let sender = imp.sender.get().unwrap().clone();
-
+        
+        //init list view
         let list_view = self.imp().history_listview.clone();
-
-        // 创建数据模型
         let store = ListStore::new::<HistoryItem>();
-
         let selection_model = SingleSelection::new(Some(store.clone()));
-
         let factory = SignalListItemFactory::new();
-
         factory.connect_setup(move |_, list_item| {
             let label = Label::new(None);
             list_item
@@ -147,32 +144,44 @@ impl CodePage {
                 .and_downcast::<Label>()
                 .expect("Child widget must be a Label");
 
-            label.set_label(&obj.text()); // 获取属性
+            label.set_label(&obj.text());
         });
-
         list_view.set_model(Some(&selection_model));
         list_view.set_factory(Some(&factory));
 
+        // get list view content
+        let sender_clone = sender.clone();
+        let file_path_clone = file_path.to_str().unwrap().to_string();
         CONTEXT.spawn_local(clone!(
             #[weak] store,
             async move{
                 let (tx, rx) = oneshot::channel();
-                sender
-                    .send(Action::MegaCore(MegaCommands::GetPathHistory {
-                        chan: tx,
-                        path: file_path.to_str().unwrap().to_string(),
-                    }))
+                sender_clone
+                    .send(Action::MegaCore(MegaCommands::
+                        GetPathHistory {
+                            chan: tx,
+                            path: file_path_clone
+                        }))
                     .await
                     .unwrap();
 
                 match rx.await.unwrap() {
                     Ok(commits)=>{
-                        tracing::debug!("Received commit history{:?}",commits);
+                        //tracing::debug!("Received commit history{:?}",commits);
                         store.remove_all();
                         for commit in commits {
-                            store.append(&HistoryItem::new(&commit.id._to_string(),&format!(
-                                "Commit:{}\nAuthor: {}\n message: {} \n Date: {}",
-                                commit.id, commit.author,commit.message,commit.committer.timestamp
+                            let  full_message = commit.message;
+                            let parts: Vec<&str> = full_message.split("\n\n").collect();
+                            let main_message = parts.last().unwrap_or(&"");
+                            store.append(&HistoryItem::new(&commit.id._to_string(),&commit.tree_id.to_string(),&format!(
+                                "Commit:{}\nAuthor: {}\n message: {}  Date: {}",
+                                commit.id.to_string(),
+                                commit.author.name,
+                                main_message,
+                                //commit.committer.timestamp.to_string()
+                                 DateTime::<Utc>::from_timestamp(commit.committer.timestamp as i64, 0)
+                                  .map(|dt| dt.format("%Y-%m-%d %H:%M:%S UTC").to_string())
+                                  .unwrap_or_else(|| "Invalid Date".to_string())
                             )));
                         }
                     }
@@ -182,53 +191,63 @@ impl CodePage {
                     }
 
                 }
+            }
+        ));
 
+        let selection_model_clone = selection_model.clone();
+        let sender_clone = sender.clone();
+        let page_weak = self.downgrade();
+        list_view.connect_activate(move |_list_view, pos| {
+            tracing::debug!("##############点击了第 {} 项", pos);
+            selection_model_clone.set_selected(pos);
+            if let Some(item) = selection_model_clone.item(pos) {
+                if let Ok(history_item) = item.downcast::<HistoryItem>() {
+
+                    let tree_id = history_item.tree_id();
+                    let file_path_clone = file_path.to_str().unwrap().to_string();
+                    let sender = sender_clone.clone();
+                    let page_weak = page_weak.clone();
+
+                    CONTEXT.spawn_local(async move {
+                        let (tx, rx) = oneshot::channel();
+                        sender
+                            .send(Action::MegaCore(MegaCommands::GetHistoryBlobId {
+                                chan: tx,
+                                tree_id,
+                                path: file_path_clone.clone(),
+                            })).await.unwrap();
+
+                        match rx.await {
+                            Ok(Ok(blob_id)) => {
+                                let path = PathBuf::from(file_path_clone);
+                                if let Some(file_name) = path.file_name().and_then(|n| n.to_str()) {
+                                    // 升级弱引用
+                                    if let Some(page) = page_weak.upgrade() {
+                                        page.show_editor_on(blob_id, file_name.to_string(), path);
+                                    } else {
+                                        tracing::error!("Failed to upgrade weak reference to page");
+                                    }
+                                } else {
+                                    tracing::error!("Failed to extract file name from path");
+                                }
+                            }
+                            Ok(Err(e)) => {
+                                tracing::error!("Failed to get history blob: {:?}", e);
+                            }
+                            Err(e) => {
+                                tracing::error!("Channel error: {:?}", e);
+                            }
+                        }
+                    });
 
-
+                }
             }
-        ));
+        });
+
+
+
 
 
-        // let list_view = self.imp().history_listview.clone();
-        //
-        // // 创建数据模型
-        // let store = ListStore::new::<HistoryItem>();
-        //
-        // for i in 1..=20 {
-        //     store.append(&HistoryItem::new(&format!("commit信息 \n 提交人:{i}, sha")));
-        // }
-        // let selection_model = SingleSelection::new(Some(store));
-        //
-        // let factory = SignalListItemFactory::new();
-        //
-        // factory.connect_setup(move |_, list_item| {
-        //     let label = Label::new(None);
-        //     list_item
-        //         .downcast_ref::<ListItem>()
-        //         .expect("Needs to be ListItem")
-        //         .set_child(Some(&label));
-        // });
-        //
-        // factory.connect_bind(move |_, list_item| {
-        //     let obj = list_item
-        //         .downcast_ref::<ListItem>()
-        //         .expect("Needs to be ListItem")
-        //         .item()
-        //         .and_downcast::<HistoryItem>()
-        //         .expect("Model should be HistoryItem");
-        //
-        //     let label = list_item
-        //         .downcast_ref::<ListItem>()
-        //         .expect("Needs to be ListItem")
-        //         .child()
-        //         .and_downcast::<Label>()
-        //         .expect("Child widget must be a Label");
-        //
-        //     label.set_label(&obj.text()); // 获取属性
-        // });
-        //
-        // list_view.set_model(Some(&selection_model));
-        // list_view.set_factory(Some(&factory));
     }
 
     // sidebar button
@@ -265,16 +284,18 @@ impl CodePage {
         history_popover.set_position(gtk::PositionType::Left);
         history_popover.set_has_arrow(false);
         history_popover.set_parent(&history_btn);
-        history_popover.set_autohide(false);
+        history_popover.set_autohide(true);
 
-        // 
+        //let page_weak = self.downgrade();
         history_btn.connect_clicked(clone!(
+            #[weak(rename_to=page)] self,
             #[weak]
             history_popover,
             move |_| {
                 if history_popover.is_visible() {
                     history_popover.popdown();
                 } else {
+                    page.setup_history_list();
                     let rect = gdk::Rectangle::new(-5, 200, 100, 30);
 
                     history_popover.set_pointing_to(Some(&rect));
@@ -332,10 +353,10 @@ impl CodePage {
             if !git_path.starts_with('/') {
                 git_path = format!("/{}", git_path);
             }
-            // *cur_path = git_path.parse().unwrap();
-            *cur_path = "/third-party/demo3".parse().unwrap();
+             *cur_path = git_path.parse().unwrap();
+            
         }
-        self.setup_history_list();
+
 
         imp.file_path_label
             .set_text(&path.to_str().unwrap().replace(['/', '\\'], " > "));
diff --git a/monobean/src/components/history_list.rs b/monobean/src/components/history_list.rs
index 7597c878a..3f84cf352 100644
--- a/monobean/src/components/history_list.rs
+++ b/monobean/src/components/history_list.rs
@@ -15,6 +15,8 @@ mod imp {
         #[property(get, set)]
         pub id: RefCell<String>,
         #[property(get, set)]
+        pub tree_id: RefCell<String>,
+        #[property(get, set)]
         pub text: RefCell<String>,
     }
 
@@ -33,10 +35,11 @@ glib::wrapper! {
 }
 
 impl HistoryItem {
-    pub fn new(id: &str,text: &str) -> Self {
+    pub fn new(id: &str,tree_id: &str,text: &str) -> Self {
         glib::Object::builder()
             .property("text", text)
             .property("id", id)
+            .property("tree_id", tree_id)
             .build()
     }
 }
diff --git a/monobean/src/core/mega_core.rs b/monobean/src/core/mega_core.rs
index 278be84ff..aa637f2cf 100644
--- a/monobean/src/core/mega_core.rs
+++ b/monobean/src/core/mega_core.rs
@@ -1,3 +1,4 @@
+use std::collections::{HashSet, VecDeque};
 use crate::application::Action;
 use crate::core::servers::{HttpOptions, SshOptions};
 use crate::core::CoreConfigChanged;
@@ -10,7 +11,7 @@ use ceres::protocol::repo::Repo;
 use common::config::Config;
 use common::model::P2pOptions;
 use context::AppContext as MegaContext;
-use mercury::internal::object::tree::Tree;
+use mercury::internal::object::tree::{Tree};
 use std::fmt;
 use std::fmt::{Debug, Formatter};
 use std::net::SocketAddr;
@@ -20,7 +21,6 @@ use std::sync::Arc;
 use tokio::sync::{oneshot, OnceCell, RwLock};
 use vault::integration::vault_core::VaultCore;
 use vault::pgp::{SignedPublicKey, SignedSecretKey};
-use libra::command::log::{get_reachable_commits};
 use mercury::internal::object::commit::Commit;
 
 pub struct MegaCore {
@@ -64,7 +64,7 @@ pub enum MegaCommands {
     },
     LoadFileContent {
         chan: oneshot::Sender<MonoBeanResult<String>>,
-        id: String,
+        id: String, //blob_id
     },
     // GetRepoUrl {
     //     chan: oneshot::Sender<MonoBeanResult<String>>,
@@ -73,6 +73,11 @@ pub enum MegaCommands {
         chan: oneshot::Sender<MonoBeanResult<Vec<Commit>>>,
         path: String,
     },
+    GetHistoryBlobId {
+        chan: oneshot::Sender<MonoBeanResult<String>>,
+        tree_id: String,
+        path: String,
+    },
 }
 
 impl Debug for MegaCore {
@@ -197,6 +202,10 @@ impl MegaCore {
                 let commits = self.get_path_history(&path).await;
                 chan.send(commits).unwrap();
             }
+            MegaCommands::GetHistoryBlobId { chan, tree_id, path } => {
+                let blob_id = self.get_history_blob(tree_id, path).await;
+                chan.send(blob_id).unwrap();
+            }
             // MegaCommands::GetRepoUrl => {
             //
             // }
@@ -324,7 +333,6 @@ impl MegaCore {
                 .unwrap()
             {
                 let repo: Repo = model.into();
-                //tracing::debug!("@@@@@@@@@@@@@@@@@@@@@@@@Found repo {:?}", repo.repo_id);
                 return Ok(Box::new(ImportApiService {
                     storage: ctx.storage.clone(),
                     repo,
@@ -350,7 +358,6 @@ impl MegaCore {
                 acc + &e.as_os_str().to_string_lossy() + "/"
             });
         let path = PathBuf::from(path);
-        tracing::debug!("!!!!!!!!!!!!!!!!Loading tree from path: {}", path.display());
 
         let handler = self.api_handler(&path).await?;
         let tree = handler.search_tree_by_path(&path).await;
@@ -480,22 +487,77 @@ impl MegaCore {
         self.running_context.read().await.is_some()
     }
 
-    async fn get_path_history(&self,path: impl AsRef<Path>) ->  MonoBeanResult<Vec<Commit>>{
+    async fn get_path_history(&self,path: impl AsRef<Path>) ->  MonoBeanResult<Vec<Commit>> {
+        
+        // fistly get the file's latest commit
+        let mut path = PathBuf::from(path.as_ref());
+        let file_name = path.file_name()
+            .expect("path must have file name")
+            .to_string_lossy()
+            .to_string(); // OsString -> String
+        path.pop();
+        tracing::info!("get path:{:?} filename:{:?} history", path,file_name);
 
         let api_handler = self.api_handler(&path).await?;
-        let cur_commit = api_handler.get_latest_commit(path.as_ref().to_path_buf())
+        let map = api_handler.item_to_commit_map(path)
             .await
-            .map_err(|e| MonoBeanError::MegaCoreError(format!("Failed to get latest commit: {:?}", e)))?;
-
-
-         let mut reachable_commits = get_reachable_commits(cur_commit.oid.clone()).await;
-        //let mut reachable_commits = get_reachable_commits("").await;
-        reachable_commits.sort_by_key(|c| c.committer.timestamp);
-        tracing::debug!("Reachable commits: {:?}", reachable_commits);
+            .map_err(|e| MonoBeanError::MegaCoreError(e.to_string()))?;
+        //tracing::info!("get map {:?}",map);
+        let latest_commit_id = map.into_iter()
+            .find(|(tree_item, _)| tree_item.name == file_name)
+            .and_then(|(_, commit_opt)| commit_opt)
+            .map(|commit| commit.id)
+            .ok_or_else(|| MonoBeanError::MegaCoreError("no commit found in history".to_string()))?;
+        tracing::info!("get last commit id:{:?}",latest_commit_id.to_string());
+        
+        self.get_reachable_commits(&latest_commit_id.to_string(),api_handler).await
+        
+    }
 
+    async fn get_reachable_commits(&self, commit_id:&str, api_handler: Box<dyn ApiHandler>) -> MonoBeanResult<Vec<Commit>> {
+        let mut queue = VecDeque::new();
+        let mut commit_set: HashSet<String> = HashSet::new(); 
+        let mut reachable_commits: Vec<Commit> = Vec::new();
+        queue.push_back(commit_id.to_string());
+        while !queue.is_empty() {
+            let commit_id = queue.pop_front().unwrap();
+            if commit_set.contains(&commit_id) { continue;  }
+            commit_set.insert(commit_id.clone());
+
+            let commit = api_handler.get_commit_by_hash(&commit_id).await.unwrap();
+            let parent_ids = commit.clone().parent_commit_ids;
+            for parent_id in parent_ids.iter()  {
+                queue.push_back(parent_id.to_string())
+            }
+            reachable_commits.push(commit);
+        }
+        //tracing::info!("get reachable commits:{:?}",reachable_commits);
+        reachable_commits.sort_by(|a, b| b.committer.timestamp.cmp(&a.committer.timestamp));
+        Ok(reachable_commits)
+    }
 
+    async fn get_history_blob(&self, tree_id: String, path: String) -> MonoBeanResult<String> {
+        let path = PathBuf::from(path.as_str());
+        let api_handler = self.api_handler(&path).await?;
+        let tree = api_handler.get_tree_by_hash(&tree_id).await;
+        
+        // let tree_item = tree.tree_items.iter()
+        //     .find(|tree_item| tree_item.name == path.file_name().unwrap().to_str().unwrap()).unwrap();
+        let file_name = match path.file_name() {
+            Some(name) => name.to_str().unwrap_or(""),
+            None => {
+                tracing::error!("Invalid file path: {:?}", path);
+                return Err(MonoBeanError::MegaCoreError("Invalid file path".to_string()));
+            }
+        };
 
-        Ok(reachable_commits)
+        let tree_item = tree.tree_items.iter()
+            .find(|tree_item| tree_item.name == file_name).unwrap();
+        
+        let blob_id = tree_item.id.to_string().clone();
+        tracing::info!("id of file:{:?} is {:?}",path,blob_id);
+        
+        Ok(blob_id)
     }
 }
 
@@ -559,7 +621,7 @@ mod tests {
         // Later we will use a bit more complex mechanism to load config,
         // and this test will be able to detect if the loading mechanism is broken.
         // TODO: use `Config::load_sources` to load glib shcema
-        if let Some(cargo_dir) = std::option_env!("CARGO_MANIFEST_DIR") {
+        if let Some(cargo_dir) = option_env!("CARGO_MANIFEST_DIR") {
             std::env::set_current_dir(cargo_dir).expect("Failed to set workspace dir");
         }
         let resources =
diff --git a/monobean/src/window.rs b/monobean/src/window.rs
index 996e9990e..aefea7875 100644
--- a/monobean/src/window.rs
+++ b/monobean/src/window.rs
@@ -20,7 +20,7 @@ use gtk::glib;
 use gtk::prelude::{ButtonExt, GtkWindowExt, WidgetExt};
 use gtk::CompositeTemplate;
 use std::cell::OnceCell;
-use crate::core::mega_core::MegaCommands::CoreStatus;
+
 
 glib::wrapper! {
     pub struct MonobeanWindow(ObjectSubclass<imp::MonobeanWindow>)
@@ -210,7 +210,7 @@ impl MonobeanWindow {
         // 强制刷新文件树
         let code_page = self.imp().code_page.get();
         let file_tree = code_page.imp().file_tree_view.get();
-        CONTEXT.spawn_local_with_priority(adw::glib::Priority::LOW, async move {
+        CONTEXT.spawn_local_with_priority(Priority::LOW, async move {
             file_tree.refresh_root().await;
         });
     }
@@ -296,26 +296,26 @@ impl MonobeanWindow {
         
         // mega status bar show 
         // todo here produce tons of log ,need to adjust logic of generate log 
-        // let monobean_application:MonobeanApplication = self.application().unwrap().downcast().unwrap();
-        // CONTEXT.spawn_local_with_priority(Priority::DEFAULT_IDLE, async move {
-        //     loop {
-        //         let rx = monobean_application.core_status();
-        //         let (core_started, _) = rx.await.unwrap();
-        //         if core_started {
-        //             status_label.set_label("Mega started");
-        //             status_icon.set_icon_name(Some("status-normal-icon"));
-        //             //tracing::debug!("watching mage status----------")
-        //         } else {
-        //             status_label.set_label("Mega stoped");
-        //             status_icon.set_icon_name(Some("dialog-warning"));
-        //             tracing::debug!("watching mage status faild----------")
-        //         }
-        //         glib::timeout_future(std::time::Duration::from_secs(5)).await;
-        //     }
-        //     
-        // });
+        let monobean_application:MonobeanApplication = self.application().unwrap().downcast().unwrap();
+        CONTEXT.spawn_local_with_priority(Priority::DEFAULT_IDLE, async move {
+            loop {
+                let rx = monobean_application.core_status();
+                let (core_started, _) = rx.await.unwrap();
+                if core_started {
+                    status_label.set_label("Mega started");
+                    status_icon.set_icon_name(Some("status-normal-icon"));
+                    //tracing::debug!("watching mage status----------")
+                } else {
+                    status_label.set_label("Mega stoped");
+                    status_icon.set_icon_name(Some("dialog-warning"));
+                    tracing::debug!("watching mage status faild----------")
+                }
+                glib::timeout_future(std::time::Duration::from_secs(5)).await;
+            }
+            
+        });
+        
         
-        //let action = Action::MegaCore(CoreStatus());
     }
 
     pub fn add_toast(&self, message: String) {
@@ -332,7 +332,7 @@ impl MonobeanWindow {
         // seems that dismiss will clear something used by animation
         // cause adw_animation_skip emit 'done' segfault on closure(https://github.com/gmg137/netease-cloud-music-gtk/issues/202)
         // delay to wait for animation skipped/done
-        crate::CONTEXT.spawn_local_with_priority(Priority::DEFAULT_IDLE, async move {
+        CONTEXT.spawn_local_with_priority(Priority::DEFAULT_IDLE, async move {
             glib::timeout_future(std::time::Duration::from_millis(500)).await;
             // removed from overlay toast queue by signal
             pre.dismiss();

From cbd7a0ea7980654774fabdd36bdfdcf49bba0d6b Mon Sep 17 00:00:00 2001
From: pleasedontbreak123 <nabeiyq@126.com>
Date: Sat, 16 Aug 2025 20:00:01 +0800
Subject: [PATCH 05/17] fix clippy

---
 monobean/src/components/code_page.rs    | 83 +++++++++++----------
 monobean/src/components/history_list.rs |  3 +-
 monobean/src/core/mega_core.rs          | 98 +++++++++++++++----------
 monobean/src/window.rs                  | 13 ++--
 4 files changed, 109 insertions(+), 88 deletions(-)

diff --git a/monobean/src/components/code_page.rs b/monobean/src/components/code_page.rs
index 732de944b..8595ef565 100644
--- a/monobean/src/components/code_page.rs
+++ b/monobean/src/components/code_page.rs
@@ -5,6 +5,7 @@ use crate::CONTEXT;
 use adw::gdk;
 use adw::gio::ListStore;
 use async_channel::Sender;
+use chrono::{DateTime, Utc};
 use glib::clone;
 use gtk::prelude::*;
 use gtk::subclass::prelude::*;
@@ -12,11 +13,10 @@ use gtk::{glib, CompositeTemplate, Label, ListItem, SignalListItemFactory, Singl
 use scv::{prelude::*, Buffer};
 use std::path::{Path, PathBuf};
 use tokio::sync::oneshot;
-use chrono::{DateTime, Utc};
 
 mod imp {
     use adw::subclass::prelude::BinImpl;
-    use tokio::sync::{OnceCell};
+    use tokio::sync::OnceCell;
 
     use crate::components::file_tree::FileTreeView;
 
@@ -114,8 +114,9 @@ impl CodePage {
     fn setup_history_list(&self) {
         let imp = self.imp();
         let file_path = imp.cur_file_path.read().unwrap().clone();
+        tracing::info!("file_path: {:?}", file_path);
         let sender = imp.sender.get().unwrap().clone();
-        
+
         //init list view
         let list_view = self.imp().history_listview.clone();
         let store = ListStore::new::<HistoryItem>();
@@ -149,47 +150,58 @@ impl CodePage {
         list_view.set_model(Some(&selection_model));
         list_view.set_factory(Some(&factory));
 
+        if file_path.as_os_str().is_empty() {
+            store.append(&HistoryItem::new("", "", "尚未打开文件"));
+            return;
+        }
+
         // get list view content
         let sender_clone = sender.clone();
         let file_path_clone = file_path.to_str().unwrap().to_string();
         CONTEXT.spawn_local(clone!(
-            #[weak] store,
-            async move{
+            #[weak]
+            store,
+            async move {
                 let (tx, rx) = oneshot::channel();
                 sender_clone
-                    .send(Action::MegaCore(MegaCommands::
-                        GetPathHistory {
-                            chan: tx,
-                            path: file_path_clone
-                        }))
+                    .send(Action::MegaCore(MegaCommands::GetPathHistory {
+                        chan: tx,
+                        path: file_path_clone,
+                    }))
                     .await
                     .unwrap();
 
                 match rx.await.unwrap() {
-                    Ok(commits)=>{
+                    Ok(commits) => {
                         //tracing::debug!("Received commit history{:?}",commits);
                         store.remove_all();
                         for commit in commits {
-                            let  full_message = commit.message;
+                            let full_message = commit.message;
                             let parts: Vec<&str> = full_message.split("\n\n").collect();
                             let main_message = parts.last().unwrap_or(&"");
-                            store.append(&HistoryItem::new(&commit.id._to_string(),&commit.tree_id.to_string(),&format!(
-                                "Commit:{}\nAuthor: {}\n message: {}  Date: {}",
-                                commit.id.to_string(),
-                                commit.author.name,
-                                main_message,
-                                //commit.committer.timestamp.to_string()
-                                 DateTime::<Utc>::from_timestamp(commit.committer.timestamp as i64, 0)
-                                  .map(|dt| dt.format("%Y-%m-%d %H:%M:%S UTC").to_string())
-                                  .unwrap_or_else(|| "Invalid Date".to_string())
-                            )));
+                            store.append(&HistoryItem::new(
+                                &commit.id._to_string(),
+                                &commit.tree_id.to_string(),
+                                &format!(
+                                    "Commit:{}\nAuthor: {}\n message: {}  Date: {}",
+                                    commit.id,
+                                    commit.author.name,
+                                    main_message,
+                                    //commit.committer.timestamp.to_string()
+                                    DateTime::<Utc>::from_timestamp(
+                                        commit.committer.timestamp as i64,
+                                        0
+                                    )
+                                    .map(|dt| dt.format("%Y-%m-%d %H:%M:%S UTC").to_string())
+                                    .unwrap_or_else(|| "Invalid Date".to_string())
+                                ),
+                            ));
                         }
                     }
-                    Err(e)=>{
+                    Err(e) => {
                         tracing::error!("get history commits error: {:?}", e);
                         return;
                     }
-
                 }
             }
         ));
@@ -202,7 +214,6 @@ impl CodePage {
             selection_model_clone.set_selected(pos);
             if let Some(item) = selection_model_clone.item(pos) {
                 if let Ok(history_item) = item.downcast::<HistoryItem>() {
-
                     let tree_id = history_item.tree_id();
                     let file_path_clone = file_path.to_str().unwrap().to_string();
                     let sender = sender_clone.clone();
@@ -215,13 +226,14 @@ impl CodePage {
                                 chan: tx,
                                 tree_id,
                                 path: file_path_clone.clone(),
-                            })).await.unwrap();
+                            }))
+                            .await
+                            .unwrap();
 
                         match rx.await {
                             Ok(Ok(blob_id)) => {
                                 let path = PathBuf::from(file_path_clone);
                                 if let Some(file_name) = path.file_name().and_then(|n| n.to_str()) {
-                                    // 升级弱引用
                                     if let Some(page) = page_weak.upgrade() {
                                         page.show_editor_on(blob_id, file_name.to_string(), path);
                                     } else {
@@ -239,15 +251,9 @@ impl CodePage {
                             }
                         }
                     });
-
                 }
             }
         });
-
-
-
-
-
     }
 
     // sidebar button
@@ -280,7 +286,7 @@ impl CodePage {
         let history_btn = imp.history_btn.get();
         let history_popover = imp.history_popover.get();
 
-        // set Popover 
+        // set Popover
         history_popover.set_position(gtk::PositionType::Left);
         history_popover.set_has_arrow(false);
         history_popover.set_parent(&history_btn);
@@ -288,7 +294,8 @@ impl CodePage {
 
         //let page_weak = self.downgrade();
         history_btn.connect_clicked(clone!(
-            #[weak(rename_to=page)] self,
+            #[weak(rename_to=page)]
+            self,
             #[weak]
             history_popover,
             move |_| {
@@ -351,13 +358,11 @@ impl CodePage {
             let mut cur_path = imp.cur_file_path.write().unwrap();
             let mut git_path = path.to_string_lossy().replace('\\', "/");
             if !git_path.starts_with('/') {
-                git_path = format!("/{}", git_path);
+                git_path = format!("/{git_path}");
             }
-             *cur_path = git_path.parse().unwrap();
-            
+            *cur_path = git_path.parse().unwrap();
         }
 
-
         imp.file_path_label
             .set_text(&path.to_str().unwrap().replace(['/', '\\'], " > "));
 
diff --git a/monobean/src/components/history_list.rs b/monobean/src/components/history_list.rs
index 3f84cf352..cc39c43e7 100644
--- a/monobean/src/components/history_list.rs
+++ b/monobean/src/components/history_list.rs
@@ -7,7 +7,6 @@ mod imp {
     use gtk::glib;
     use gtk::prelude::*;
     use gtk::subclass::prelude::*;
- 
 
     #[derive(Properties, Default)]
     #[properties(wrapper_type = super::HistoryItem)]
@@ -35,7 +34,7 @@ glib::wrapper! {
 }
 
 impl HistoryItem {
-    pub fn new(id: &str,tree_id: &str,text: &str) -> Self {
+    pub fn new(id: &str, tree_id: &str, text: &str) -> Self {
         glib::Object::builder()
             .property("text", text)
             .property("id", id)
diff --git a/monobean/src/core/mega_core.rs b/monobean/src/core/mega_core.rs
index aa637f2cf..e6c7ba3f7 100644
--- a/monobean/src/core/mega_core.rs
+++ b/monobean/src/core/mega_core.rs
@@ -1,4 +1,3 @@
-use std::collections::{HashSet, VecDeque};
 use crate::application::Action;
 use crate::core::servers::{HttpOptions, SshOptions};
 use crate::core::CoreConfigChanged;
@@ -11,7 +10,9 @@ use ceres::protocol::repo::Repo;
 use common::config::Config;
 use common::model::P2pOptions;
 use context::AppContext as MegaContext;
-use mercury::internal::object::tree::{Tree};
+use mercury::internal::object::commit::Commit;
+use mercury::internal::object::tree::Tree;
+use std::collections::{HashSet, VecDeque};
 use std::fmt;
 use std::fmt::{Debug, Formatter};
 use std::net::SocketAddr;
@@ -21,7 +22,6 @@ use std::sync::Arc;
 use tokio::sync::{oneshot, OnceCell, RwLock};
 use vault::integration::vault_core::VaultCore;
 use vault::pgp::{SignedPublicKey, SignedSecretKey};
-use mercury::internal::object::commit::Commit;
 
 pub struct MegaCore {
     config: Arc<RwLock<Config>>,
@@ -69,7 +69,7 @@ pub enum MegaCommands {
     // GetRepoUrl {
     //     chan: oneshot::Sender<MonoBeanResult<String>>,
     // }
-    GetPathHistory{
+    GetPathHistory {
         chan: oneshot::Sender<MonoBeanResult<Vec<Commit>>>,
         path: String,
     },
@@ -198,17 +198,20 @@ impl MegaCore {
                 let content = self.load_blob(path).await;
                 chan.send(content).unwrap();
             }
-            MegaCommands::GetPathHistory{ chan, path} => {
+            MegaCommands::GetPathHistory { chan, path } => {
                 let commits = self.get_path_history(&path).await;
                 chan.send(commits).unwrap();
             }
-            MegaCommands::GetHistoryBlobId { chan, tree_id, path } => {
+            MegaCommands::GetHistoryBlobId {
+                chan,
+                tree_id,
+                path,
+            } => {
                 let blob_id = self.get_history_blob(tree_id, path).await;
                 chan.send(blob_id).unwrap();
-            }
-            // MegaCommands::GetRepoUrl => {
-            //
-            // }
+            } // MegaCommands::GetRepoUrl => {
+              //
+              // }
         }
     }
 
@@ -487,46 +490,56 @@ impl MegaCore {
         self.running_context.read().await.is_some()
     }
 
-    async fn get_path_history(&self,path: impl AsRef<Path>) ->  MonoBeanResult<Vec<Commit>> {
-        
+    async fn get_path_history(&self, path: impl AsRef<Path>) -> MonoBeanResult<Vec<Commit>> {
         // fistly get the file's latest commit
         let mut path = PathBuf::from(path.as_ref());
-        let file_name = path.file_name()
+        let file_name = path
+            .file_name()
             .expect("path must have file name")
             .to_string_lossy()
             .to_string(); // OsString -> String
         path.pop();
-        tracing::info!("get path:{:?} filename:{:?} history", path,file_name);
+        tracing::info!("get path:{:?} filename:{:?} history", path, file_name);
 
         let api_handler = self.api_handler(&path).await?;
-        let map = api_handler.item_to_commit_map(path)
+        let map = api_handler
+            .item_to_commit_map(path)
             .await
             .map_err(|e| MonoBeanError::MegaCoreError(e.to_string()))?;
         //tracing::info!("get map {:?}",map);
-        let latest_commit_id = map.into_iter()
+        let latest_commit_id = map
+            .into_iter()
             .find(|(tree_item, _)| tree_item.name == file_name)
             .and_then(|(_, commit_opt)| commit_opt)
             .map(|commit| commit.id)
-            .ok_or_else(|| MonoBeanError::MegaCoreError("no commit found in history".to_string()))?;
-        tracing::info!("get last commit id:{:?}",latest_commit_id.to_string());
-        
-        self.get_reachable_commits(&latest_commit_id.to_string(),api_handler).await
-        
+            .ok_or_else(|| {
+                MonoBeanError::MegaCoreError("no commit found in history".to_string())
+            })?;
+        tracing::info!("get last commit id:{:?}", latest_commit_id.to_string());
+
+        self.get_reachable_commits(&latest_commit_id.to_string(), api_handler)
+            .await
     }
 
-    async fn get_reachable_commits(&self, commit_id:&str, api_handler: Box<dyn ApiHandler>) -> MonoBeanResult<Vec<Commit>> {
+    async fn get_reachable_commits(
+        &self,
+        commit_id: &str,
+        api_handler: Box<dyn ApiHandler>,
+    ) -> MonoBeanResult<Vec<Commit>> {
         let mut queue = VecDeque::new();
-        let mut commit_set: HashSet<String> = HashSet::new(); 
+        let mut commit_set: HashSet<String> = HashSet::new();
         let mut reachable_commits: Vec<Commit> = Vec::new();
         queue.push_back(commit_id.to_string());
         while !queue.is_empty() {
             let commit_id = queue.pop_front().unwrap();
-            if commit_set.contains(&commit_id) { continue;  }
+            if commit_set.contains(&commit_id) {
+                continue;
+            }
             commit_set.insert(commit_id.clone());
 
             let commit = api_handler.get_commit_by_hash(&commit_id).await.unwrap();
             let parent_ids = commit.clone().parent_commit_ids;
-            for parent_id in parent_ids.iter()  {
+            for parent_id in parent_ids.iter() {
                 queue.push_back(parent_id.to_string())
             }
             reachable_commits.push(commit);
@@ -540,23 +553,36 @@ impl MegaCore {
         let path = PathBuf::from(path.as_str());
         let api_handler = self.api_handler(&path).await?;
         let tree = api_handler.get_tree_by_hash(&tree_id).await;
-        
+
         // let tree_item = tree.tree_items.iter()
         //     .find(|tree_item| tree_item.name == path.file_name().unwrap().to_str().unwrap()).unwrap();
         let file_name = match path.file_name() {
             Some(name) => name.to_str().unwrap_or(""),
             None => {
                 tracing::error!("Invalid file path: {:?}", path);
-                return Err(MonoBeanError::MegaCoreError("Invalid file path".to_string()));
+                return Err(MonoBeanError::MegaCoreError(
+                    "Invalid file path".to_string(),
+                ));
+            }
+        };
+
+        let tree_item = match tree
+            .tree_items
+            .iter()
+            .find(|tree_item| tree_item.name == file_name)
+        {
+            Some(item) => item,
+            None => {
+                tracing::error!("File not found in tree: {:?}", file_name);
+                return Err(MonoBeanError::MegaCoreError(
+                    "File not found in tree ".to_string(),
+                ));
             }
         };
 
-        let tree_item = tree.tree_items.iter()
-            .find(|tree_item| tree_item.name == file_name).unwrap();
-        
         let blob_id = tree_item.id.to_string().clone();
-        tracing::info!("id of file:{:?} is {:?}",path,blob_id);
-        
+        tracing::info!("id of file:{:?} is {:?}", path, blob_id);
+
         Ok(blob_id)
     }
 }
@@ -569,7 +595,7 @@ mod tests {
     use crate::config::MEGA_CONFIG_PATH;
     use async_channel::bounded;
     use common::config::LogConfig;
-    use common::config::{AuthConfig, BuildConfig,DbConfig, LFSConfig, MonoConfig, PackConfig};
+    use common::config::{AuthConfig, BuildConfig, DbConfig, LFSConfig, MonoConfig, PackConfig};
     use gtk::gio;
     use gtk::glib;
 
@@ -675,10 +701,4 @@ mod tests {
 
     #[tokio::test]
     async fn test_run_with_config() {}
-
-    #[tokio::test]
-    async fn test_get_history() {
-
-    }
-
 }
diff --git a/monobean/src/window.rs b/monobean/src/window.rs
index aefea7875..c9630fb99 100644
--- a/monobean/src/window.rs
+++ b/monobean/src/window.rs
@@ -21,7 +21,6 @@ use gtk::prelude::{ButtonExt, GtkWindowExt, WidgetExt};
 use gtk::CompositeTemplate;
 use std::cell::OnceCell;
 
-
 glib::wrapper! {
     pub struct MonobeanWindow(ObjectSubclass<imp::MonobeanWindow>)
         @extends gtk::Widget, gtk::Window, adw::ApplicationWindow,
@@ -293,10 +292,11 @@ impl MonobeanWindow {
                 imp.mega_popup.replace(Some(popup));
             }
         ));
-        
-        // mega status bar show 
-        // todo here produce tons of log ,need to adjust logic of generate log 
-        let monobean_application:MonobeanApplication = self.application().unwrap().downcast().unwrap();
+
+        // mega status bar show
+        // todo here produce tons of log ,need to adjust logic of generate log
+        let monobean_application: MonobeanApplication =
+            self.application().unwrap().downcast().unwrap();
         CONTEXT.spawn_local_with_priority(Priority::DEFAULT_IDLE, async move {
             loop {
                 let rx = monobean_application.core_status();
@@ -312,10 +312,7 @@ impl MonobeanWindow {
                 }
                 glib::timeout_future(std::time::Duration::from_secs(5)).await;
             }
-            
         });
-        
-        
     }
 
     pub fn add_toast(&self, message: String) {

From 8f1223a90f800aa03168a239f3d522c943554224 Mon Sep 17 00:00:00 2001
From: pleasedontbreak123 <nabeiyq@126.com>
Date: Sun, 17 Aug 2025 10:10:41 +0800
Subject: [PATCH 06/17] fix async execute bug

---
 monobean/src/components/code_page.rs    | 149 +++++++++++++++---------
 monobean/src/components/history_list.rs |   5 +-
 2 files changed, 95 insertions(+), 59 deletions(-)

diff --git a/monobean/src/components/code_page.rs b/monobean/src/components/code_page.rs
index 8595ef565..030bc17f3 100644
--- a/monobean/src/components/code_page.rs
+++ b/monobean/src/components/code_page.rs
@@ -54,6 +54,8 @@ mod imp {
         pub history_listview: TemplateChild<gtk::ListView>,
 
         pub cur_file_path: std::sync::RwLock<PathBuf>,
+        pub history_selection_signal: std::sync::RwLock<Option<glib::SignalHandlerId>>,
+        pub history_store: std::sync::RwLock<Option<glib::WeakRef<ListStore>>>,
         pub sender: OnceCell<Sender<Action>>,
     }
 
@@ -107,21 +109,20 @@ impl CodePage {
 
         self.setup_paned();
         self.setup_button();
+        self.setup_history_list();
         self.setup_source_view(opened_file);
         self.setup_file_tree();
     }
 
     fn setup_history_list(&self) {
         let imp = self.imp();
-        let file_path = imp.cur_file_path.read().unwrap().clone();
-        tracing::info!("file_path: {:?}", file_path);
-        let sender = imp.sender.get().unwrap().clone();
-
-        //init list view
         let list_view = self.imp().history_listview.clone();
+
+        // 初始化列表视图组件
         let store = ListStore::new::<HistoryItem>();
         let selection_model = SingleSelection::new(Some(store.clone()));
         let factory = SignalListItemFactory::new();
+
         factory.connect_setup(move |_, list_item| {
             let label = Label::new(None);
             list_item
@@ -147,15 +148,95 @@ impl CodePage {
 
             label.set_label(&obj.text());
         });
+
         list_view.set_model(Some(&selection_model));
         list_view.set_factory(Some(&factory));
 
+        {
+            let mut store_guard = imp.history_store.write().unwrap();
+            *store_guard = Some(store.downgrade());
+        }
+
+        {
+            let mut signal_guard = imp.history_selection_signal.write().unwrap();
+            if let Some(handler_id) = signal_guard.take() {
+                list_view.disconnect(handler_id);
+            }
+        }
+
+        let selection_model_clone = selection_model.clone();
+        let sender = imp.sender.get().unwrap().clone();
+        let page_weak = self.downgrade();
+
+        let handler_id = list_view.connect_activate(move |_list_view, pos| {
+            tracing::debug!("##############点击了第 {} 项", pos);
+            selection_model_clone.set_selected(pos);
+            if let Some(item) = selection_model_clone.item(pos) {
+                if let Ok(history_item) = item.downcast::<HistoryItem>() {
+                    let tree_id = history_item.tree_id();
+                    let file_path = history_item.file_path();
+                    let sender = sender.clone();
+                    let page_weak = page_weak.clone();
+
+                    CONTEXT.spawn_local(async move {
+                        let (tx, rx) = oneshot::channel();
+                        sender
+                            .send(Action::MegaCore(MegaCommands::GetHistoryBlobId {
+                                chan: tx,
+                                tree_id,
+                                path: file_path.clone(),
+                            }))
+                            .await
+                            .unwrap();
+
+                        match rx.await {
+                            Ok(Ok(blob_id)) => {
+                                let path = PathBuf::from(file_path);
+                                if let Some(file_name) = path.file_name().and_then(|n| n.to_str()) {
+                                    if let Some(page) = page_weak.upgrade() {
+                                        page.show_editor_on(blob_id, file_name.to_string(), path);
+                                    } else {
+                                        tracing::error!("Failed to upgrade weak reference to page");
+                                    }
+                                } else {
+                                    tracing::error!("Failed to extract file name from path");
+                                }
+                            }
+                            Ok(Err(e)) => {
+                                tracing::error!("Failed to get history blob: {:?}", e);
+                            }
+                            Err(e) => {
+                                tracing::error!("Channel error: {:?}", e);
+                            }
+                        }
+                    });
+                }
+            }
+        });
+
+        // 保存信号连接ID
+        {
+            let mut signal_guard = imp.history_selection_signal.write().unwrap();
+            *signal_guard = Some(handler_id);
+        }
+    }
+
+    fn update_history_list(&self) {
+        let imp = self.imp();
+        let file_path = imp.cur_file_path.read().unwrap().clone();
+        tracing::info!("file_path: {:?}", file_path);
+
+        // 获取store
+        let store_ref = imp.history_store.read().unwrap();
+        let store = store_ref.as_ref().and_then(|weak| weak.upgrade()).unwrap();
+
         if file_path.as_os_str().is_empty() {
-            store.append(&HistoryItem::new("", "", "尚未打开文件"));
+            store.append(&HistoryItem::new("", "", "", "尚未打开文件"));
             return;
         }
 
-        // get list view content
+        // 获取列表视图内容
+        let sender = imp.sender.get().unwrap().clone();
         let sender_clone = sender.clone();
         let file_path_clone = file_path.to_str().unwrap().to_string();
         CONTEXT.spawn_local(clone!(
@@ -166,7 +247,7 @@ impl CodePage {
                 sender_clone
                     .send(Action::MegaCore(MegaCommands::GetPathHistory {
                         chan: tx,
-                        path: file_path_clone,
+                        path: file_path_clone.clone(),
                     }))
                     .await
                     .unwrap();
@@ -182,6 +263,7 @@ impl CodePage {
                             store.append(&HistoryItem::new(
                                 &commit.id._to_string(),
                                 &commit.tree_id.to_string(),
+                                &file_path_clone,
                                 &format!(
                                     "Commit:{}\nAuthor: {}\n message: {}  Date: {}",
                                     commit.id,
@@ -205,55 +287,6 @@ impl CodePage {
                 }
             }
         ));
-
-        let selection_model_clone = selection_model.clone();
-        let sender_clone = sender.clone();
-        let page_weak = self.downgrade();
-        list_view.connect_activate(move |_list_view, pos| {
-            tracing::debug!("##############点击了第 {} 项", pos);
-            selection_model_clone.set_selected(pos);
-            if let Some(item) = selection_model_clone.item(pos) {
-                if let Ok(history_item) = item.downcast::<HistoryItem>() {
-                    let tree_id = history_item.tree_id();
-                    let file_path_clone = file_path.to_str().unwrap().to_string();
-                    let sender = sender_clone.clone();
-                    let page_weak = page_weak.clone();
-
-                    CONTEXT.spawn_local(async move {
-                        let (tx, rx) = oneshot::channel();
-                        sender
-                            .send(Action::MegaCore(MegaCommands::GetHistoryBlobId {
-                                chan: tx,
-                                tree_id,
-                                path: file_path_clone.clone(),
-                            }))
-                            .await
-                            .unwrap();
-
-                        match rx.await {
-                            Ok(Ok(blob_id)) => {
-                                let path = PathBuf::from(file_path_clone);
-                                if let Some(file_name) = path.file_name().and_then(|n| n.to_str()) {
-                                    if let Some(page) = page_weak.upgrade() {
-                                        page.show_editor_on(blob_id, file_name.to_string(), path);
-                                    } else {
-                                        tracing::error!("Failed to upgrade weak reference to page");
-                                    }
-                                } else {
-                                    tracing::error!("Failed to extract file name from path");
-                                }
-                            }
-                            Ok(Err(e)) => {
-                                tracing::error!("Failed to get history blob: {:?}", e);
-                            }
-                            Err(e) => {
-                                tracing::error!("Channel error: {:?}", e);
-                            }
-                        }
-                    });
-                }
-            }
-        });
     }
 
     // sidebar button
@@ -302,7 +335,7 @@ impl CodePage {
                 if history_popover.is_visible() {
                     history_popover.popdown();
                 } else {
-                    page.setup_history_list();
+                    page.update_history_list();
                     let rect = gdk::Rectangle::new(-5, 200, 100, 30);
 
                     history_popover.set_pointing_to(Some(&rect));
diff --git a/monobean/src/components/history_list.rs b/monobean/src/components/history_list.rs
index cc39c43e7..024a27fa9 100644
--- a/monobean/src/components/history_list.rs
+++ b/monobean/src/components/history_list.rs
@@ -16,6 +16,8 @@ mod imp {
         #[property(get, set)]
         pub tree_id: RefCell<String>,
         #[property(get, set)]
+        pub file_path: RefCell<String>,
+        #[property(get, set)]
         pub text: RefCell<String>,
     }
 
@@ -34,10 +36,11 @@ glib::wrapper! {
 }
 
 impl HistoryItem {
-    pub fn new(id: &str, tree_id: &str, text: &str) -> Self {
+    pub fn new(id: &str, tree_id: &str, file_path: &str, text: &str) -> Self {
         glib::Object::builder()
             .property("text", text)
             .property("id", id)
+            .property("file_path", file_path)
             .property("tree_id", tree_id)
             .build()
     }

From 34b8834e13461b60e4c40bd9fcffea500c68e211 Mon Sep 17 00:00:00 2001
From: pleasedontbreak123 <nabeiyq@126.com>
Date: Wed, 27 Aug 2025 21:43:38 +0800
Subject: [PATCH 07/17] magic sort & delta chain

---
 mercury/delta/Cargo.toml                  |   2 +-
 mercury/delta/src/lib.rs                  | 147 ++++++++-
 mercury/src/internal/pack/cache_object.rs |   1 +
 mercury/src/internal/pack/encode.rs       | 378 +++++++++++++++++++---
 mercury/src/internal/pack/entry.rs        |   5 +
 5 files changed, 487 insertions(+), 46 deletions(-)

diff --git a/mercury/delta/Cargo.toml b/mercury/delta/Cargo.toml
index 822d017ec..7e98d47c0 100644
--- a/mercury/delta/Cargo.toml
+++ b/mercury/delta/Cargo.toml
@@ -9,7 +9,7 @@ edition = "2021"
 diffs = { workspace = true }
 thiserror = { workspace = true }
 flate2 = { workspace = true, features = ["zlib"] }
-
+rayon = { workspace = true }
 [features]
 default = ["diff_mydrs"]
 diff_mydrs = []
diff --git a/mercury/delta/src/lib.rs b/mercury/delta/src/lib.rs
index 978a74bd2..1f7cdc615 100644
--- a/mercury/delta/src/lib.rs
+++ b/mercury/delta/src/lib.rs
@@ -1,4 +1,6 @@
+use std::hash::{DefaultHasher, Hash, Hasher};
 use encode::DeltaDiff;
+use rayon::prelude::*;
 
 mod decode;
 mod encode;
@@ -7,6 +9,89 @@ mod utils;
 
 pub use decode::delta_decode as decode;
 
+const SAMPLE_STEP: usize = 32;      // 每隔 64 字节取样
+const MIN_DELTA_RATE: f64 = 0.5;    // 最小 delta rate
+
+/// 计算两个对象的 delta rate（近似）
+pub fn heuristic_encode_rate(old_data: &[u8], new_data: &[u8]) -> f64 {
+    let old_len = old_data.len();
+    let new_len = new_data.len();
+
+    if old_len == 0 && new_len == 0 {
+        return 1.0; // 两个空 slice 完全匹配
+    }
+    if old_len == 0 || new_len == 0 {
+        return 0.0; // 一个空一个非空，匹配率为 0
+    }
+
+    let step = SAMPLE_STEP;
+    let mut match_count = 0;
+    let mut sample_count = 0;
+
+    // 对两个对象取样，计算 hash 匹配数
+    let min_len = old_len.min(new_len);
+
+    let total_samples = (min_len + step - 1) / step;
+    let mut i = 0;
+    while i < min_len {
+        let old_chunk = &old_data[i..(i + step).min(old_len)];
+        let new_chunk = &new_data[i..(i + step).min(new_len)];
+
+        if hash_chunk(old_chunk) == hash_chunk(new_chunk) {
+            match_count += 1;
+        }
+        sample_count += 1;
+
+        // 早停判断：
+        // 剩余样本即使全匹配也达不到 MIN_DELTA_RATE，则提前返回 0
+        let remaining_samples = total_samples - sample_count;
+        let max_possible_rate = (match_count + remaining_samples) as f64 / total_samples as f64;
+        if max_possible_rate < MIN_DELTA_RATE {
+            return 0.0;
+        }
+
+        i += step;
+    }
+
+    // 返回匹配率
+    match_count as f64 / sample_count as f64
+}
+
+fn hash_chunk(chunk: &[u8]) -> u64 {
+    let mut hasher = DefaultHasher::new();
+    chunk.hash(&mut hasher);
+    hasher.finish()
+}
+
+pub fn heuristic_encode_rate_parallel(old_data: &[u8], new_data: &[u8]) -> f64 {
+    let old_len = old_data.len();
+    let new_len = new_data.len();
+
+    if old_len == 0 && new_len == 0 { return 1.0; }
+    if old_len == 0 || new_len == 0 { return 0.0; }
+
+    let min_len = old_len.min(new_len);
+
+    let step = if min_len > 1_000_000 {
+        512
+    } else if min_len > 100_000 {
+        128
+    } else {
+        16
+    };
+
+    let chunks: Vec<_> = old_data[..min_len].chunks(step)
+        .zip(new_data[..min_len].chunks(step))
+        .collect();
+
+    let match_count: usize = chunks.par_iter()
+        .filter(|(a, b)| a == b)
+        .count();
+
+    let rate = match_count as f64 / chunks.len() as f64;
+    if rate < MIN_DELTA_RATE { 0.0 } else { rate }
+}
+
 pub fn encode_rate(old_data: &[u8], new_data: &[u8]) -> f64 {
     let differ = DeltaDiff::new(old_data, new_data);
     differ.get_ssam_rate()
@@ -18,4 +103,64 @@ pub fn encode(old_data: &[u8], new_data: &[u8]) -> Vec<u8> {
 }
 
 #[cfg(test)]
-mod tests {}
+mod tests {
+    use crate::{ encode_rate, heuristic_encode_rate, heuristic_encode_rate_parallel};
+
+    #[test]
+    fn test_heuristic_encode_rate() {
+        // 完全相同的数据
+        let data1 = b"hello world, this is a test for delta rate";
+        let data2 = b"hello world, this is a test for delta rate";
+        let rate = heuristic_encode_rate(data1, data2);
+        println!("rate = {}", rate);
+        assert!((rate - 1.0).abs() < 1e-6, "Expected 1.0 for identical data");
+
+        // 部分相似的数据
+        let data3 = b"hello world, this is a test for delta rate";
+        let data4 = b"hello worll, this is a test for delta rate";
+        let rate = encode_rate(data3, data4);
+        println!("rate = {}", rate);
+        assert!(rate > 0.5 && rate < 1.0, "Expected partial match for similar data");
+
+        // 完全不同的数据
+        let data5 = b"abcdefghijklmno";
+        let data6 = b"1234567890!@#";
+        let rate = heuristic_encode_rate(data5, data6);
+        println!("rate = {}", rate);
+        assert!(rate < 0.2, "Expected low match rate for different data");
+
+        // 空数据
+        let data7 = b"";
+        let data8 = b"";
+        let rate = heuristic_encode_rate(data7, data8);
+        println!("rate = {}", rate);
+        assert_eq!(rate, 1.0, "Empty slices should be fully matching");
+
+        // 一个空一个非空
+        let rate = heuristic_encode_rate(data7, data1);
+        assert_eq!(rate, 0.0, "Empty vs non-empty should give 0 rate");
+    }
+
+    #[test]
+    fn test_heuristic_encode_rate_large_files() {
+        // 大文件完全不匹配 → 触发早停
+        let data1 = vec![0u8; 1_000_00];
+        let data2 = vec![1u8; 1_000_00];
+        let rate = heuristic_encode_rate(&data1, &data2);
+        println!("Large non-matching data rate = {}", rate);
+        assert_eq!(rate, 0.0, "Large completely different data should early stop with 0 rate");
+
+        // 大文件部分匹配 → 保留部分匹配
+        let mut data3 = vec![0u8; 100];
+        let mut data4 = vec![0u8; 100];
+        // 在一部分修改一些数据
+        for i in 0..2 {
+            data4[i] = 1;
+        }
+        let rate1 = heuristic_encode_rate_parallel(&data3, &data4);
+        let rate2 = encode_rate(&data3, &data4);
+        println!("Large partially matching data rate = {}, accurate rate = {}", rate1,rate2);
+        
+        assert!((rate2-rate1).abs() < 0.2, "Large partially matching data should preserve partial rate");
+    }
+}
diff --git a/mercury/src/internal/pack/cache_object.rs b/mercury/src/internal/pack/cache_object.rs
index 0f3d9e1a2..a1a3ad0d6 100644
--- a/mercury/src/internal/pack/cache_object.rs
+++ b/mercury/src/internal/pack/cache_object.rs
@@ -230,6 +230,7 @@ impl CacheObject {
                 obj_type,
                 data: self.data_decompressed.clone(),
                 hash,
+                chain_len: 0,
             },
             _ => {
                 unreachable!("delta object should not persist!")
diff --git a/mercury/src/internal/pack/encode.rs b/mercury/src/internal/pack/encode.rs
index fc3f463e7..b9aa1566f 100644
--- a/mercury/src/internal/pack/encode.rs
+++ b/mercury/src/internal/pack/encode.rs
@@ -1,3 +1,4 @@
+use std::cmp::Ordering;
 use std::collections::VecDeque;
 use std::io::Write;
 
@@ -11,7 +12,9 @@ use crate::internal::object::types::ObjectType;
 use crate::time_it;
 use crate::{errors::GitError, hash::SHA1, internal::pack::entry::Entry};
 
-const MIN_DELTA_RATE: f64 = 0.5; // minimum delta rate can accept
+const MAX_CHAIN_LEN: usize = 50;
+const MIN_DELTA_RATE: f64 = 0.5; // minimum delta rate can acceptbase 必须和 entry 类型一致
+
 
 /// A encoder for generating pack files with delta objects.
 pub struct PackEncoder {
@@ -105,6 +108,23 @@ fn encode_one_object(entry: &Entry, offset: Option<usize>) -> Result<Vec<u8>, Gi
     Ok(encoded_data)
 }
 
+
+fn magic_sort(a:&Entry,b:&Entry)  -> Ordering {
+    let ord = b.obj_type.to_u8().cmp(&a.obj_type.to_u8());
+    if ord != Ordering::Equal { return ord; }
+
+    // the hash should be file hash not content hash
+    // todo the feature need larger refactor
+    // let ord = b.hash.cmp(&a.hash);
+    // if ord != Ordering::Equal { return ord; }
+
+    let ord = b.data.len().cmp(&a.data.len());
+    if ord != Ordering::Equal { return ord; }
+
+    // fallback pointer order (newest first)
+    (a as *const Entry).cmp(&(b as *const Entry))
+}
+
 impl PackEncoder {
     pub fn new(object_number: usize, window_size: usize, sender: mpsc::Sender<Vec<u8>>) -> Self {
         PackEncoder {
@@ -170,37 +190,77 @@ impl PackEncoder {
                 "encoding operation is already in progress".to_string(),
             ));
         }
-        loop {
-            match entry_rx.recv().await {
-                Some(entry) => {
-                    self.process_index += 1;
-                    // push window after encode to void diff by self
-                    let offset = self.inner_offset;
-                    let mut try_delta_entry = entry.clone();
-                    let try_delfa_offset = match enable_zstdelta {
-                        true => self.try_as_offset_zstdelta(&mut try_delta_entry),
-                        false => self.try_as_offset_delta(&mut try_delta_entry),
-                    };
-                    let obj_data = encode_one_object(&try_delta_entry, try_delfa_offset)?;
 
-                    self.write_all_and_update(&obj_data).await;
-                    self.window.push_back((entry, offset));
-                    if self.window.len() > self.window_size {
-                        self.window.pop_front();
-                    }
-                }
-                None => {
-                    if self.process_index != self.object_number {
-                        panic!(
-                            "not all objects are encoded, process:{}, total:{}",
-                            self.process_index, self.object_number
-                        );
-                    }
-                    break;
-                }
+        // collect entries before magic sort
+        let mut all_entries = Vec::with_capacity(self.object_number);
+        while let Some(entry) = entry_rx.recv().await {
+            all_entries.push(entry);
+        }
+        
+        if all_entries.len() != self.object_number {
+            panic!(
+                "not all objects are encoded, process:{}, total:{}",
+                self.process_index, self.object_number
+            );
+        }
+        
+        
+        all_entries.par_sort_by(magic_sort);
+        
+        
+        for mut entry in all_entries {
+            self.process_index += 1;
+            let offset = self.inner_offset;
+            let mut entry_for_window = entry.clone();
+            let try_offset = if enable_zstdelta {
+                self.try_as_offset_zstdelta(&mut entry)
+            } else {
+                self.try_as_offset_delta(&mut entry)
+            };
+            let obj_data = encode_one_object(&entry, try_offset)?;
+            self.write_all_and_update(&obj_data).await;
+            entry_for_window.chain_len = entry.chain_len;
+        
+            self.window.push_back((entry_for_window, offset));
+            if self.window.len() > self.window_size {
+                self.window.pop_front();
             }
         }
 
+
+        // loop {
+        //     match entry_rx.recv().await {
+        //         Some(entry) => {
+        //             self.process_index += 1;
+        //             // push window after encode to void diff by self
+        //             let offset = self.inner_offset;
+        //             let mut try_delta_entry = entry.clone();
+        //             let try_delfa_offset = match enable_zstdelta {
+        //                 true => self.try_as_offset_zstdelta(&mut try_delta_entry),
+        //                 false => self.try_as_offset_delta(&mut try_delta_entry),
+        //             };
+        //             let obj_data = encode_one_object(&try_delta_entry, try_delfa_offset)?;
+        // 
+        //             self.write_all_and_update(&obj_data).await;
+        //             self.window.push_back((entry, offset));
+        //             if self.window.len() > self.window_size {
+        //                 self.window.pop_front();
+        //             }
+        //         }
+        //         None => {
+        //             if self.process_index != self.object_number {
+        //                 panic!(
+        //                     "not all objects are encoded, process:{}, total:{}",
+        //                     self.process_index, self.object_number
+        //                 );
+        //             }
+        //             break;
+        //         }
+        //     }
+        // }
+
+
+
         // hash signature
         let hash_result = self.inner_hash.clone().finalize();
         self.final_hash = Some(SHA1::from_bytes(&hash_result));
@@ -288,21 +348,66 @@ impl PackEncoder {
     fn try_as_offset_delta(&self, entry: &mut Entry) -> Option<usize> {
         let mut best_base: Option<&(Entry, usize)> = None;
         let mut best_rate: f64 = 0.0;
+        let mut best_chain_len: usize = 0;
+        let mut best_base_offset: usize = 0;
+        let tie_epsilon: f64 = 0.02;
         for try_base in self.window.iter() {
+            // tracing::debug!(
+            //     "entry={} best_base={:?} rate={:?}",
+            //     entry.hash,
+            //     best_base.map(|b| b.0.hash),
+            //     best_rate
+            // );
             if try_base.0.obj_type != entry.obj_type {
                 continue;
             }
-            let rate = delta::encode_rate(&try_base.0.data, &entry.data);
-            if rate > MIN_DELTA_RATE && rate > best_rate {
-                best_rate = rate;
-                best_base = Some(try_base);
+
+            if try_base.0.chain_len >= MAX_CHAIN_LEN {
+                continue;
+            }
+
+            if try_base.0.hash == entry.hash {
+                continue;
+            }
+
+            // 这是导致运行时间爆炸的原因
+            // when data is small, heuristic_encode_rate is not accurate
+            let rate = if (try_base.0.data.len() +  entry.data.len()) / 2 > 20 {
+               delta::heuristic_encode_rate_parallel(&try_base.0.data, &entry.data)
+            }else { 
+                delta::encode_rate(&try_base.0.data, &entry.data)
+            };
+            
+
+            if rate > MIN_DELTA_RATE {
+                let is_better = if rate > best_rate + tie_epsilon {
+                    true
+                } else if (rate - best_rate).abs() <= tie_epsilon {
+                    let base_chain = try_base.0.chain_len;
+                    let base_offset = self.inner_offset - try_base.1;
+                    base_chain > best_chain_len || (base_chain == best_chain_len && base_offset > best_base_offset)
+                } else {
+                    false
+                };
+
+                if is_better {
+                    best_rate = rate;
+                    best_chain_len = try_base.0.chain_len;
+                    best_base_offset = self.inner_offset - try_base.1;
+                    best_base = Some(try_base);
+                }
             }
         }
+
         best_base.map(|best_base| {
+            // 修正2：delta::encode 返回 Vec<u8>，不是 Result
             let delta = delta::encode(&best_base.0.data, &entry.data);
+
             let offset = self.inner_offset - best_base.1;
             entry.obj_type = ObjectType::OffsetDelta;
             entry.data = delta;
+            entry.chain_len = best_base.0.chain_len + 1;
+            tracing::debug!("chain_len: {:?}", entry.chain_len);
             offset
         })
     }
@@ -315,27 +420,57 @@ impl PackEncoder {
     fn try_as_offset_zstdelta(&self, entry: &mut Entry) -> Option<usize> {
         let mut best_delta_obj = None; // obj, offset
         let mut best_rate: f64 = 0.0;
+        let mut best_chain_len: usize = 0;
+        let mut best_base_offset: usize = 0;
+        let tie_epsilon: f64 = 0.02; // when rates are close, prefer deeper/newer base
+        
+        if entry.data.is_empty() {
+            return None;
+        }
+
         for try_base in self.window.iter() {
             if try_base.0.obj_type != entry.obj_type {
                 continue;
             }
-            let try_delta_obj = zstdelta::diff(&try_base.0.data, &entry.data);
-            if try_delta_obj.is_err() {
-                return None; // IO error, can't handle
+            if try_base.0.chain_len >= MAX_CHAIN_LEN {
+                continue;
             }
-            let try_delta_obj = try_delta_obj.unwrap();
+            let try_delta_obj = match zstdelta::diff(&try_base.0.data, &entry.data) {
+                Ok(v) => v,
+                Err(_) => continue, // IO error for this base, try next
+            };
             // The DeltaDiff algorithm uses the proportion of the delta part as the delta rate,
             // while zstdelta does not have this data and directly uses the overall compression rate.
             let rate = 1.0 - try_delta_obj.len() as f64 / entry.data.len() as f64;
-            if rate > MIN_DELTA_RATE && rate > best_rate {
-                best_rate = rate;
-                best_delta_obj = Some((try_delta_obj, try_base.1));
+            // if rate > MIN_DELTA_RATE && rate > best_rate {
+            //     best_rate = rate;
+            //     best_delta_obj = Some((try_delta_obj, try_base.1));
+            // }
+
+            if rate > MIN_DELTA_RATE {
+                let is_better = if rate > best_rate + tie_epsilon {
+                    true
+                } else if (rate - best_rate).abs() <= tie_epsilon {
+                    let base_chain = try_base.0.chain_len;
+                    let base_offset = self.inner_offset - try_base.1;
+                    base_chain > best_chain_len || (base_chain == best_chain_len && base_offset > best_base_offset)
+                } else {
+                    false
+                };
+
+                if is_better {
+                    best_rate = rate;
+                    best_chain_len = try_base.0.chain_len;
+                    best_base_offset = self.inner_offset - try_base.1;
+                    best_delta_obj = Some((try_delta_obj, try_base.1));
+                }
             }
         }
         best_delta_obj.map(|(best_delta_data, best_base_offset)| {
             let offset = self.inner_offset - best_base_offset;
             entry.obj_type = ObjectType::OffsetZstdelta;
             entry.data = best_delta_data;
+            entry.chain_len = best_chain_len + 1;
             // tracing::debug!("use base {:?} to delta, rate {:?}", offset, best_rate);
             offset
         })
@@ -372,11 +507,20 @@ impl PackEncoder {
             self.encode_with_zstdelta(rx).await.unwrap()
         }))
     }
+
+    #[cfg(test)]
+    pub fn get_window_info(&self) -> Vec<(ObjectType, usize, usize)> {
+        self.window
+            .iter()
+            .map(|(entry, offset)| (entry.obj_type, entry.chain_len, *offset))
+            .collect()
+    }
 }
 
 #[cfg(test)]
 mod tests {
     use std::env;
+    use std::collections::HashSet;
     use std::sync::Arc;
     use std::{io::Cursor, path::PathBuf};
 
@@ -438,6 +582,56 @@ mod tests {
         check_format(&pack_with_delta);
     }
 
+    #[tokio::test]
+    async fn test_encode_then_decode_delta_chain() {
+        init_logger();
+
+        // 1) 构造一组相似的内容，鼓励形成 delta 链
+        let mut contents: Vec<Vec<u8>> = Vec::new();
+        let mut base = String::from("base line 0\n");
+        for i in 1..=8usize {
+            base.push_str(&format!("line {i}: some repeated text with slight changes...\n"));
+            contents.push(base.as_bytes().to_vec());
+        }
+
+        // 2) 用 PackEncoder（window_size>0）编码，尽量产生 offset-delta
+        let (tx, mut rx) = mpsc::channel(10_000);
+        let (entry_tx, entry_rx) = mpsc::channel::<Entry>(10_000);
+
+        let encoder = PackEncoder::new(contents.len(), 10, tx);
+        encoder.encode_async(entry_rx).await.unwrap();
+
+        for data in contents.clone() {
+            let blob = crate::internal::object::blob::Blob::from_content_bytes(data);
+            let entry: Entry = blob.into();
+            entry_tx.send(entry).await.unwrap();
+        }
+        drop(entry_tx);
+
+        let mut pack_bytes = Vec::new();
+        while let Some(chunk) = rx.recv().await {
+            pack_bytes.extend(chunk);
+        }
+
+        // 3) 用 Pack::decode 解析，重建为完整对象
+        let mut p = Pack::new(None, Some(1024 * 1024 * 20), Some(PathBuf::from("/tmp/.cache_temp")), true);
+        let mut reader = Cursor::new(pack_bytes);
+        let decoded = Arc::new(tokio::sync::Mutex::new(Vec::<Entry>::new()));
+        let decoded_c = decoded.clone();
+        p.decode(&mut reader, move |entry, _| {
+            let mut guard = decoded_c.blocking_lock();
+            guard.push(entry);
+        }).expect("pack decode should succeed");
+
+        let decoded_entries = decoded.lock().await;
+        assert_eq!(decoded_entries.len(), contents.len());
+
+        // 4) 校验内容集合相等（忽略顺序）
+        let expected: HashSet<Vec<u8>> = contents.into_iter().collect();
+        let got: HashSet<Vec<u8>> = decoded_entries.iter().map(|e| e.data.clone()).collect();
+        assert_eq!(expected, got, "decoded contents should equal original contents");
+    }
+
     async fn get_entries_for_test() -> Arc<Mutex<Vec<Entry>>> {
         let mut source = PathBuf::from(env::current_dir().unwrap().parent().unwrap());
         source.push("tests/data/packs/pack-f8bbb573cef7d851957caceb491c073ee8e8de41.pack");
@@ -508,6 +702,11 @@ mod tests {
         let entries = get_entries_for_test().await;
         let entries_number = entries.lock().await.len();
 
+        // 计算原始总大小
+        let total_original_size: usize = entries.lock().await.iter()
+            .map(|entry| entry.data.len())
+            .sum();
+
         // encode entries
         let (tx, mut rx) = mpsc::channel(100_000);
         let (entry_tx, entry_rx) = mpsc::channel::<Entry>(100_000);
@@ -529,10 +728,29 @@ mod tests {
             tracing::info!("all entries sent");
         });
 
-        // only receive data
-        while (rx.recv().await).is_some() {
-            // do nothing
+        // // only receive data
+        // while (rx.recv().await).is_some() {
+        //     // do nothing
+        // }
+
+        // 收集数据并计算压缩率
+        let mut result = Vec::new();
+        while let Some(chunk) = rx.recv().await {
+            result.extend(chunk);
         }
+
+        // 计算压缩率
+        let pack_size = result.len();
+        let compression_rate = if total_original_size > 0 {
+            1.0 - (pack_size as f64 / total_original_size as f64)
+        } else {
+            0.0
+        };
+
+        tracing::info!("new pack file size: {}", pack_size);
+        tracing::info!("original total size: {}", total_original_size);
+        tracing::info!("compression rate: {:.2}%", compression_rate * 100.0);
+        tracing::info!("space saved: {} bytes", total_original_size.saturating_sub(pack_size));
     }
 
     #[tokio::test]
@@ -540,6 +758,12 @@ mod tests {
         init_logger();
         let entries = get_entries_for_test().await;
         let entries_number = entries.lock().await.len();
+        
+        // 计算原始总大小
+        let total_original_size: usize = entries.lock().await.iter()
+            .map(|entry| entry.data.len())
+            .sum();
+        
         let (tx, mut rx) = mpsc::channel(100_000);
         let (entry_tx, entry_rx) = mpsc::channel::<Entry>(100_000);
 
@@ -556,12 +780,25 @@ mod tests {
             tracing::info!("all entries sent");
         });
 
-        // only receive data
+        // 收集数据并计算压缩率
         let mut result = Vec::new();
         while let Some(chunk) = rx.recv().await {
             result.extend(chunk);
         }
-        tracing::info!("new pack file size: {}", result.len());
+        
+        // 计算压缩率
+        let pack_size = result.len();
+        let compression_rate = if total_original_size > 0 {
+            1.0 - (pack_size as f64 / total_original_size as f64)
+        } else {
+            0.0
+        };
+        
+        tracing::info!("new pack file size: {}", pack_size);
+        tracing::info!("original total size: {}", total_original_size);
+        tracing::info!("compression rate: {:.2}%", compression_rate * 100.0);
+        tracing::info!("space saved: {} bytes", total_original_size.saturating_sub(pack_size));
+        
         // check format
         check_format(&result);
     }
@@ -578,4 +815,57 @@ mod tests {
         println!("result: {result}");
         assert_eq!(result, value as u64);
     }
+
+    
+
+    #[tokio::test]
+    async fn test_pack_encoder_large_file_with_delta() {
+        init_logger();
+        let entries = get_entries_for_test().await;
+        let entries_number = entries.lock().await.len();
+
+        // 计算原始总大小
+        let total_original_size: usize = entries.lock().await.iter()
+            .map(|entry| entry.data.len())
+            .sum();
+
+        let (tx, mut rx) = mpsc::channel(100_000);
+        let (entry_tx, entry_rx) = mpsc::channel::<Entry>(100_000);
+
+        let mut encoder = PackEncoder::new(entries_number, 10, tx);
+        encoder.encode_async(entry_rx).await.unwrap();
+
+        // spawn a task to send entries
+        tokio::spawn(async move {
+            let entries = entries.lock().await;
+            for entry in entries.iter() {
+                entry_tx.send(entry.clone()).await.unwrap();
+            }
+            drop(entry_tx);
+            tracing::info!("all entries sent");
+        });
+
+        // 收集数据并计算压缩率
+        let mut result = Vec::new();
+        while let Some(chunk) = rx.recv().await {
+            result.extend(chunk);
+        }
+
+        // 计算压缩率
+        let pack_size = result.len();
+        let compression_rate = if total_original_size > 0 {
+            1.0 - (pack_size as f64 / total_original_size as f64)
+        } else {
+            0.0
+        };
+
+        tracing::info!("new pack file size: {}", pack_size);
+        tracing::info!("original total size: {}", total_original_size);
+        tracing::info!("compression rate: {:.2}%", compression_rate * 100.0);
+        tracing::info!("space saved: {} bytes", total_original_size.saturating_sub(pack_size));
+
+        // check format
+        check_format(&result);
+    }
+
 }
diff --git a/mercury/src/internal/pack/entry.rs b/mercury/src/internal/pack/entry.rs
index 2dd4afb9b..9c66e407d 100644
--- a/mercury/src/internal/pack/entry.rs
+++ b/mercury/src/internal/pack/entry.rs
@@ -18,6 +18,7 @@ pub struct Entry {
     pub obj_type: ObjectType,
     pub data: Vec<u8>,
     pub hash: SHA1,
+    pub chain_len: usize
 }
 
 impl PartialEq for Entry {
@@ -54,6 +55,7 @@ impl From<Blob> for Entry {
             obj_type: ObjectType::Blob,
             data: value.data,
             hash: value.id,
+            chain_len: 0,
         }
     }
 }
@@ -64,6 +66,7 @@ impl From<Commit> for Entry {
             obj_type: ObjectType::Commit,
             data: value.to_data().unwrap(),
             hash: value.id,
+            chain_len: 0,
         }
     }
 }
@@ -74,6 +77,7 @@ impl From<Tree> for Entry {
             obj_type: ObjectType::Tree,
             data: value.to_data().unwrap(),
             hash: value.id,
+            chain_len: 0,
         }
     }
 }
@@ -84,6 +88,7 @@ impl From<Tag> for Entry {
             obj_type: ObjectType::Tag,
             data: value.to_data().unwrap(),
             hash: value.id,
+            chain_len: 0,
         }
     }
 }

From 4143f47c93971b1691908e00ce60e0af3721a912 Mon Sep 17 00:00:00 2001
From: pleasedontbreak123 <nabeiyq@126.com>
Date: Fri, 29 Aug 2025 15:29:44 +0800
Subject: [PATCH 08/17] before breaking change

---
 mercury/Cargo.toml                  |   2 +-
 mercury/delta/src/lib.rs            |   4 +-
 mercury/src/internal/pack/encode.rs | 234 ++++++++++++++++++++--------
 3 files changed, 174 insertions(+), 66 deletions(-)

diff --git a/mercury/Cargo.toml b/mercury/Cargo.toml
index 460078837..1621f0a8a 100644
--- a/mercury/Cargo.toml
+++ b/mercury/Cargo.toml
@@ -10,7 +10,6 @@ delta = { path = "delta" }
 sapling-zstdelta = { path = "zstdelta" }
 common = { workspace = true }
 callisto = { workspace = true }
-
 flate2 = { workspace = true, features = ["zlib"] } # enable linking against the libz(C lib); better performance
 serde = { workspace = true, features = ["derive"] }
 bstr = { workspace = true }
@@ -38,6 +37,7 @@ rayon = { workspace = true }
 reqwest = { workspace = true }
 ring = { workspace = true }
 serde_json.workspace = true
+ahash = "0.8.12"
 
 [dev-dependencies]
 tokio = { workspace = true, features = ["full"] }
diff --git a/mercury/delta/src/lib.rs b/mercury/delta/src/lib.rs
index 1f7cdc615..6cd72030a 100644
--- a/mercury/delta/src/lib.rs
+++ b/mercury/delta/src/lib.rs
@@ -9,7 +9,7 @@ mod utils;
 
 pub use decode::delta_decode as decode;
 
-const SAMPLE_STEP: usize = 32;      // 每隔 64 字节取样
+const SAMPLE_STEP: usize = 64;      // 每隔 64 字节取样
 const MIN_DELTA_RATE: f64 = 0.5;    // 最小 delta rate
 
 /// 计算两个对象的 delta rate（近似）
@@ -151,7 +151,7 @@ mod tests {
         assert_eq!(rate, 0.0, "Large completely different data should early stop with 0 rate");
 
         // 大文件部分匹配 → 保留部分匹配
-        let mut data3 = vec![0u8; 100];
+        let  data3 = vec![0u8; 100];
         let mut data4 = vec![0u8; 100];
         // 在一部分修改一些数据
         for i in 0..2 {
diff --git a/mercury/src/internal/pack/encode.rs b/mercury/src/internal/pack/encode.rs
index b9aa1566f..54767e8d6 100644
--- a/mercury/src/internal/pack/encode.rs
+++ b/mercury/src/internal/pack/encode.rs
@@ -7,13 +7,16 @@ use rayon::prelude::*;
 use sha1::{Digest, Sha1};
 use tokio::sync::mpsc;
 use tokio::task::JoinHandle;
+use std::hash::{Hash, Hasher};
+use ahash::AHasher;
 
 use crate::internal::object::types::ObjectType;
 use crate::time_it;
 use crate::{errors::GitError, hash::SHA1, internal::pack::entry::Entry};
 
 const MAX_CHAIN_LEN: usize = 50;
-const MIN_DELTA_RATE: f64 = 0.5; // minimum delta rate can acceptbase 必须和 entry 类型一致
+const MIN_DELTA_RATE: f64 = 0.5; // minimum delta rate 
+const MAX_ZSTDELTA_CHAIN_LEN: usize = 50;
 
 
 /// A encoder for generating pack files with delta objects.
@@ -125,6 +128,95 @@ fn magic_sort(a:&Entry,b:&Entry)  -> Ordering {
     (a as *const Entry).cmp(&(b as *const Entry))
 }
 
+// pub fn cheap_similar(a: &[u8], b: &[u8]) -> bool {
+//     // 取前 64 字节做预判
+//     let n = a.len().min(b.len()).min(64);
+//
+//     // 数据太小就直接认为可能相似
+//     if n < 8 {
+//         return true;
+//     }
+//
+//     // 内部函数：对切片计算 AHasher hash
+//     fn calc_hash(data: &[u8]) -> u64 {
+//         let mut hasher = AHasher::default();
+//         data.hash(&mut hasher);
+//         hasher.finish()
+//     }
+//
+//     // 计算前 n 字节的 hash
+//     let ha = calc_hash(&a[..n]);
+//     let hb = calc_hash(&b[..n]);
+//
+//     // 简单阈值判断：
+//     // - 如果两者 trailing_zeros 很少，说明不相似
+//     // - 你可以根据经验调节 2
+//     ha.trailing_zeros().min(hb.trailing_zeros()) > 2
+// }
+
+/// 计算单片段 hash
+fn calc_hash(data: &[u8]) -> u64 {
+    let mut hasher = AHasher::default();
+    data.hash(&mut hasher);
+    hasher.finish()
+}
+
+/// 分片 hash 函数，将数据分成 n_segments 片段并计算每片 hash
+fn calc_hashes(data: &[u8], n_segments: usize) -> Vec<u64> {
+    let mut res = Vec::with_capacity(n_segments);
+    let step = (data.len() + n_segments - 1) / n_segments; // 向上取整
+    for i in 0..n_segments {
+        let start = i * step;
+        let end = ((i + 1) * step).min(data.len());
+        if start >= end { break; }
+        res.push(calc_hash(&data[start..end]));
+    }
+    res
+}
+
+/// cheap_similar 最终优化版
+pub fn cheap_similar(a: &[u8], b: &[u8]) -> bool {
+    let min_len = a.len().min(b.len());
+
+    // 对小对象直接全量比较
+    if min_len < 8 {
+        return a == b;
+    }
+
+    // 前缀长度，可调
+    let prefix_len = min_len.min(128);
+    let a_prefix = &a[..prefix_len];
+    let b_prefix = &b[..prefix_len];
+
+    // 动态分片数，每段约 32 字节
+    let n_segments = (prefix_len / 32).max(1).min(4);
+
+    let a_hashes = calc_hashes(a_prefix, n_segments);
+    let b_hashes = calc_hashes(b_prefix, n_segments);
+
+    // Hamming 距离阈值
+    let hamming_threshold = 8;
+
+    // 对小前缀顺序比较
+    if prefix_len <= 64 {
+        let mut diff_bits = 0;
+        for (x, y) in a_hashes.iter().zip(&b_hashes) {
+            diff_bits += (x ^ y).count_ones();
+            if diff_bits > hamming_threshold {
+                return false;
+            }
+        }
+        true
+    } else {
+        // 大对象使用并行比较
+        a_hashes
+            .par_iter()
+            .zip(&b_hashes)
+            .map(|(x, y)| (x ^ y).count_ones())
+            .sum::<u32>() <= hamming_threshold
+    }
+}
+
 impl PackEncoder {
     pub fn new(object_number: usize, window_size: usize, sender: mpsc::Sender<Vec<u8>>) -> Self {
         PackEncoder {
@@ -226,41 +318,7 @@ impl PackEncoder {
                 self.window.pop_front();
             }
         }
-
-
-        // loop {
-        //     match entry_rx.recv().await {
-        //         Some(entry) => {
-        //             self.process_index += 1;
-        //             // push window after encode to void diff by self
-        //             let offset = self.inner_offset;
-        //             let mut try_delta_entry = entry.clone();
-        //             let try_delfa_offset = match enable_zstdelta {
-        //                 true => self.try_as_offset_zstdelta(&mut try_delta_entry),
-        //                 false => self.try_as_offset_delta(&mut try_delta_entry),
-        //             };
-        //             let obj_data = encode_one_object(&try_delta_entry, try_delfa_offset)?;
-        // 
-        //             self.write_all_and_update(&obj_data).await;
-        //             self.window.push_back((entry, offset));
-        //             if self.window.len() > self.window_size {
-        //                 self.window.pop_front();
-        //             }
-        //         }
-        //         None => {
-        //             if self.process_index != self.object_number {
-        //                 panic!(
-        //                     "not all objects are encoded, process:{}, total:{}",
-        //                     self.process_index, self.object_number
-        //                 );
-        //             }
-        //             break;
-        //         }
-        //     }
-        // }
-
-
-
+        
         // hash signature
         let hash_result = self.inner_hash.clone().finalize();
         self.final_hash = Some(SHA1::from_bytes(&hash_result));
@@ -350,14 +408,9 @@ impl PackEncoder {
         let mut best_rate: f64 = 0.0;
         let mut best_chain_len: usize = 0;
         let mut best_base_offset: usize = 0;
-        let tie_epsilon: f64 = 0.02;
+        let tie_epsilon: f64 = 0.15;
+
         for try_base in self.window.iter() {
-            // tracing::debug!(
-            //     "entry={} best_base={:?} rate={:?}",
-            //     entry.hash,
-            //     best_base.map(|b| b.0.hash),
-            //     best_rate
-            // );
             if try_base.0.obj_type != entry.obj_type {
                 continue;
             }
@@ -370,44 +423,71 @@ impl PackEncoder {
                 continue;
             }
 
-            // 这是导致运行时间爆炸的原因
+            // 尺寸差异过大时直接跳过
+            let sym_ratio = (try_base.0.data.len().min(entry.data.len()) as f64) / (try_base.0.data.len().max(entry.data.len()) as f64);
+            if sym_ratio < 0.5 {
+                continue;
+            }
+
+            // 轻量 hash 预判（例如比较前 64 字节的 hash）
+            if !cheap_similar(&try_base.0.data, &entry.data) {
+                continue;
+            }
+
+
             // when data is small, heuristic_encode_rate is not accurate
             let rate = if (try_base.0.data.len() +  entry.data.len()) / 2 > 20 {
                delta::heuristic_encode_rate_parallel(&try_base.0.data, &entry.data)
             }else { 
                 delta::encode_rate(&try_base.0.data, &entry.data)
             };
-            
 
+            
             if rate > MIN_DELTA_RATE {
                 let is_better = if rate > best_rate + tie_epsilon {
-                    true
-                } else if (rate - best_rate).abs() <= tie_epsilon {
-                    let base_chain = try_base.0.chain_len;
-                    let base_offset = self.inner_offset - try_base.1;
-                    base_chain > best_chain_len || (base_chain == best_chain_len && base_offset > best_base_offset)
-                } else {
-                    false
-                };
+                        // rate 明显更大，直接优先
+                        true
+                    } else if (rate - best_rate).abs() <= tie_epsilon {
+                        // 构造候选 Key
+                        let cand_key = (
+                            try_base.0.chain_len,
+                            self.inner_offset - try_base.1,
+                        );
+
+                        // 当前最佳 Key
+                        let best_key = (
+                            best_chain_len,
+                            best_base_offset,
+                        );
+                        cand_key > best_key
+                    } else {
+                        false
+                    };
 
                 if is_better {
                     best_rate = rate;
                     best_chain_len = try_base.0.chain_len;
                     best_base_offset = self.inner_offset - try_base.1;
                     best_base = Some(try_base);
+
+                    if rate > 0.90 {
+                        break;
+                    }
                 }
+
             }
+
         }
 
         best_base.map(|best_base| {
-            // 修正2：delta::encode 返回 Vec<u8>，不是 Result
+
             let delta = delta::encode(&best_base.0.data, &entry.data);
 
             let offset = self.inner_offset - best_base.1;
             entry.obj_type = ObjectType::OffsetDelta;
             entry.data = delta;
             entry.chain_len = best_base.0.chain_len + 1;
-            tracing::debug!("chain_len: {:?}", entry.chain_len);
+            //tracing::debug!("chain_len: {:?}", entry.chain_len);
             offset
         })
     }
@@ -432,9 +512,23 @@ impl PackEncoder {
             if try_base.0.obj_type != entry.obj_type {
                 continue;
             }
-            if try_base.0.chain_len >= MAX_CHAIN_LEN {
+            if try_base.0.chain_len >= MAX_ZSTDELTA_CHAIN_LEN {
                 continue;
             }
+
+            // 尺寸差异过大时直接跳过
+            let size_a = try_base.0.data.len();
+            let size_b = entry.data.len();
+            if size_a * 10 < size_b || size_b * 10 < size_a {
+                continue;
+            }
+
+            // 轻量 hash 预判（例如比较前 64 字节的 hash）
+            if !cheap_similar(&try_base.0.data, &entry.data) {
+                continue;
+            }
+            
+            
             let try_delta_obj = match zstdelta::diff(&try_base.0.data, &entry.data) {
                 Ok(v) => v,
                 Err(_) => continue, // IO error for this base, try next
@@ -485,6 +579,9 @@ impl PackEncoder {
 
     /// async version of encode, result data will be returned by JoinHandle.
     /// It will consume PackEncoder, so you can't use it after calling this function.
+    /// when window_size = 0, it executes parallel_encode which retains stream transmission
+    /// when window_size = 0,it executes encode which uses magic sort and delta. 
+    /// It seems that all other modules rely on this api
     pub async fn encode_async(
         mut self,
         rx: mpsc::Receiver<Entry>,
@@ -523,7 +620,7 @@ mod tests {
     use std::collections::HashSet;
     use std::sync::Arc;
     use std::{io::Cursor, path::PathBuf};
-
+    use std::time::Instant;
     use tokio::sync::Mutex;
 
     use crate::internal::object::blob::Blob;
@@ -599,10 +696,10 @@ mod tests {
         let (entry_tx, entry_rx) = mpsc::channel::<Entry>(10_000);
 
         let encoder = PackEncoder::new(contents.len(), 10, tx);
-        encoder.encode_async(entry_rx).await.unwrap();
+        encoder.encode_async_with_zstdelta(entry_rx).await.unwrap();
 
         for data in contents.clone() {
-            let blob = crate::internal::object::blob::Blob::from_content_bytes(data);
+            let blob = Blob::from_content_bytes(data);
             let entry: Entry = blob.into();
             entry_tx.send(entry).await.unwrap();
         }
@@ -636,7 +733,7 @@ mod tests {
         let mut source = PathBuf::from(env::current_dir().unwrap().parent().unwrap());
         source.push("tests/data/packs/pack-f8bbb573cef7d851957caceb491c073ee8e8de41.pack");
         // let file_map = crate::test_utils::setup_lfs_file().await;
-        // let source = file_map
+        // let source = file_mapa
         //     .get("git-2d187177923cd618a75da6c6db45bb89d92bd504.pack")
         //     .unwrap();
         // decode pack file to get entries
@@ -707,6 +804,7 @@ mod tests {
             .map(|entry| entry.data.len())
             .sum();
 
+        let start = Instant::now(); // 开始时间
         // encode entries
         let (tx, mut rx) = mpsc::channel(100_000);
         let (entry_tx, entry_rx) = mpsc::channel::<Entry>(100_000);
@@ -747,6 +845,8 @@ mod tests {
             0.0
         };
 
+        let duration = start.elapsed();
+        tracing::info!("test executed in: {:.2?}", duration);
         tracing::info!("new pack file size: {}", pack_size);
         tracing::info!("original total size: {}", total_original_size);
         tracing::info!("compression rate: {:.2}%", compression_rate * 100.0);
@@ -763,7 +863,8 @@ mod tests {
         let total_original_size: usize = entries.lock().await.iter()
             .map(|entry| entry.data.len())
             .sum();
-        
+
+        let start = Instant::now();
         let (tx, mut rx) = mpsc::channel(100_000);
         let (entry_tx, entry_rx) = mpsc::channel::<Entry>(100_000);
 
@@ -793,7 +894,9 @@ mod tests {
         } else {
             0.0
         };
-        
+
+        let duration = start.elapsed();
+        tracing::info!("test executed in: {:.2?}", duration);
         tracing::info!("new pack file size: {}", pack_size);
         tracing::info!("original total size: {}", total_original_size);
         tracing::info!("compression rate: {:.2}%", compression_rate * 100.0);
@@ -832,7 +935,9 @@ mod tests {
         let (tx, mut rx) = mpsc::channel(100_000);
         let (entry_tx, entry_rx) = mpsc::channel::<Entry>(100_000);
 
-        let mut encoder = PackEncoder::new(entries_number, 10, tx);
+        let encoder = PackEncoder::new(entries_number, 10, tx);
+
+        let start = Instant::now(); // 开始时间
         encoder.encode_async(entry_rx).await.unwrap();
 
         // spawn a task to send entries
@@ -859,6 +964,9 @@ mod tests {
             0.0
         };
 
+
+        let duration = start.elapsed();
+        tracing::info!("test executed in: {:.2?}", duration);
         tracing::info!("new pack file size: {}", pack_size);
         tracing::info!("original total size: {}", total_original_size);
         tracing::info!("compression rate: {:.2}%", compression_rate * 100.0);

From 8756bbce7af019e63a91a997bbc0fa6e12587218 Mon Sep 17 00:00:00 2001
From: pleasedontbreak123 <nabeiyq@126.com>
Date: Sat, 30 Aug 2025 14:52:03 +0800
Subject: [PATCH 09/17] before breaking change

---
 mercury/delta/src/lib.rs            |   4 +-
 mercury/src/internal/pack/encode.rs | 343 +++++++++++++++++-----------
 2 files changed, 215 insertions(+), 132 deletions(-)

diff --git a/mercury/delta/src/lib.rs b/mercury/delta/src/lib.rs
index 6cd72030a..f7161d499 100644
--- a/mercury/delta/src/lib.rs
+++ b/mercury/delta/src/lib.rs
@@ -72,7 +72,9 @@ pub fn heuristic_encode_rate_parallel(old_data: &[u8], new_data: &[u8]) -> f64 {
 
     let min_len = old_len.min(new_len);
 
-    let step = if min_len > 1_000_000 {
+    let step = if min_len > 10_000_000{
+       1024  
+    } else if min_len > 1_000_000 {
         512
     } else if min_len > 100_000 {
         128
diff --git a/mercury/src/internal/pack/encode.rs b/mercury/src/internal/pack/encode.rs
index 54767e8d6..7e523c368 100644
--- a/mercury/src/internal/pack/encode.rs
+++ b/mercury/src/internal/pack/encode.rs
@@ -154,69 +154,120 @@ fn magic_sort(a:&Entry,b:&Entry)  -> Ordering {
 //     ha.trailing_zeros().min(hb.trailing_zeros()) > 2
 // }
 
-/// 计算单片段 hash
+// /// 计算单片段 hash
 fn calc_hash(data: &[u8]) -> u64 {
     let mut hasher = AHasher::default();
     data.hash(&mut hasher);
     hasher.finish()
 }
+//
+// /// 分片 hash 函数，将数据分成 n_segments 片段并计算每片 hash
+// fn calc_hashes(data: &[u8], n_segments: usize) -> Vec<u64> {
+//     let mut res = Vec::with_capacity(n_segments);
+//     let step = (data.len() + n_segments - 1) / n_segments; // 向上取整
+//     for i in 0..n_segments {
+//         let start = i * step;
+//         let end = ((i + 1) * step).min(data.len());
+//         if start >= end { break; }
+//         res.push(calc_hash(&data[start..end]));
+//     }
+//     res
+// }
+//
+// /// cheap_similar 最终优化版
+// pub fn cheap_similar(a: &[u8], b: &[u8]) -> bool {
+//     let min_len = a.len().min(b.len());
+//
+//     // 对小对象直接全量比较
+//     if min_len < 8 {
+//         return a == b;
+//     }
+//
+//     // 前缀长度，可调
+//     let prefix_len = min_len.min(128);
+//     let a_prefix = &a[..prefix_len];
+//     let b_prefix = &b[..prefix_len];
+//
+//     // 动态分片数，每段约 32 字节
+//     let n_segments = (prefix_len / 32).max(1).min(4);
+//
+//     let a_hashes = calc_hashes(a_prefix, n_segments);
+//     let b_hashes = calc_hashes(b_prefix, n_segments);
+//
+//     // Hamming 距离阈值
+//     let hamming_threshold = 4;
+//
+//     // 对小前缀顺序比较
+//     if prefix_len <= 64 {
+//         let mut diff_bits = 0;
+//         for (x, y) in a_hashes.iter().zip(&b_hashes) {
+//             diff_bits += (x ^ y).count_ones();
+//             if diff_bits > hamming_threshold {
+//                 return false;
+//             }
+//         }
+//         true
+//     } else {
+//         // 大对象使用并行比较
+//         a_hashes
+//             .par_iter()
+//             .zip(&b_hashes)
+//             .map(|(x, y)| (x ^ y).count_ones())
+//             .sum::<u32>() <= hamming_threshold
+//     }
+// }
 
-/// 分片 hash 函数，将数据分成 n_segments 片段并计算每片 hash
-fn calc_hashes(data: &[u8], n_segments: usize) -> Vec<u64> {
-    let mut res = Vec::with_capacity(n_segments);
-    let step = (data.len() + n_segments - 1) / n_segments; // 向上取整
-    for i in 0..n_segments {
-        let start = i * step;
-        let end = ((i + 1) * step).min(data.len());
-        if start >= end { break; }
-        res.push(calc_hash(&data[start..end]));
-    }
-    res
-}
-
-/// cheap_similar 最终优化版
-pub fn cheap_similar(a: &[u8], b: &[u8]) -> bool {
-    let min_len = a.len().min(b.len());
-
-    // 对小对象直接全量比较
-    if min_len < 8 {
-        return a == b;
-    }
-
-    // 前缀长度，可调
-    let prefix_len = min_len.min(128);
-    let a_prefix = &a[..prefix_len];
-    let b_prefix = &b[..prefix_len];
-
-    // 动态分片数，每段约 32 字节
-    let n_segments = (prefix_len / 32).max(1).min(4);
-
-    let a_hashes = calc_hashes(a_prefix, n_segments);
-    let b_hashes = calc_hashes(b_prefix, n_segments);
+// pub fn cheap_similar(a: &[u8], b: &[u8]) -> bool {
+//     let min_len = a.len().min(b.len());
+//     if min_len < 64 {
+//         // 小对象直接要求前后各16字节完全相等，严格但便宜
+//         let k = min_len.min(16);
+//         return a.get(..k) == b.get(..k) && a.get(min_len.saturating_sub(k)..) == b.get(min_len.saturating_sub(k)..);
+//     }
+// 
+//     // 允许的最大不匹配字节数（锚点总计），越大越宽松
+//     let mut mismatches = 0usize;
+//     let budget = 8; // 可调
+// 
+//     // 比对前/中/后三个区域，各取16字节
+//     let anchor_len = 16;
+//     let mids = [
+//         0,                                     // 前
+//         (min_len / 2).saturating_sub(anchor_len / 2), // 中
+//         min_len.saturating_sub(anchor_len),    // 后
+//     ];
+// 
+//     for &pos in &mids {
+//         let a_seg = &a[pos..pos + anchor_len];
+//         let b_seg = &b[pos..pos + anchor_len];
+//         // 逐字节计数（16字节很快，且避免未对齐/未定义读取）
+//         for i in 0..anchor_len {
+//             if a_seg[i] != b_seg[i] {
+//                 mismatches += 1;
+//                 if mismatches > budget {
+//                     return false; // 早停
+//                 }
+//             }
+//         }
+//     }
+// 
+//     // 再加一个极便宜的长度差门槛（避免尺寸相差太大的情况）
+//     let sym_ratio = (a.len().min(b.len()) as f64) / (a.len().max(b.len()) as f64);
+//     if sym_ratio < 0.5 {
+//         return false;
+//     }
+// 
+//     true
+// }
 
-    // Hamming 距离阈值
-    let hamming_threshold = 8;
 
-    // 对小前缀顺序比较
-    if prefix_len <= 64 {
-        let mut diff_bits = 0;
-        for (x, y) in a_hashes.iter().zip(&b_hashes) {
-            diff_bits += (x ^ y).count_ones();
-            if diff_bits > hamming_threshold {
-                return false;
-            }
-        }
-        true
-    } else {
-        // 大对象使用并行比较
-        a_hashes
-            .par_iter()
-            .zip(&b_hashes)
-            .map(|(x, y)| (x ^ y).count_ones())
-            .sum::<u32>() <= hamming_threshold
-    }
+fn cheap_similar(a: &[u8], b: &[u8]) -> bool {
+    let k = a.len().min(b.len()).min(128);
+    if k == 0 { return false; }
+    calc_hash(&a[..k]) == calc_hash(&b[..k])
 }
 
+
 impl PackEncoder {
     pub fn new(object_number: usize, window_size: usize, sender: mpsc::Sender<Vec<u8>>) -> Self {
         PackEncoder {
@@ -284,22 +335,29 @@ impl PackEncoder {
         }
 
         // collect entries before magic sort
+
         let mut all_entries = Vec::with_capacity(self.object_number);
-        while let Some(entry) = entry_rx.recv().await {
-            all_entries.push(entry);
-        }
-        
-        if all_entries.len() != self.object_number {
-            panic!(
-                "not all objects are encoded, process:{}, total:{}",
-                self.process_index, self.object_number
-            );
-        }
-        
-        
-        all_entries.par_sort_by(magic_sort);
-        
-        
+        time_it!( "%%%&&&&&  receive cost :",{
+            //let mut all_entries = Vec::with_capacity(self.object_number);
+            while let Some(entry) = entry_rx.recv().await {
+                all_entries.push(entry);
+            }
+
+
+            if all_entries.len() != self.object_number {
+                panic!(
+                    "not all objects are encoded, process:{}, total:{}",
+                    self.process_index, self.object_number
+                );
+            }
+
+                all_entries.sort_by(magic_sort);
+            });
+        //all_entries.par_sort_by(magic_sort);
+
+
+
+
         for mut entry in all_entries {
             self.process_index += 1;
             let offset = self.inner_offset;
@@ -406,90 +464,92 @@ impl PackEncoder {
     fn try_as_offset_delta(&self, entry: &mut Entry) -> Option<usize> {
         let mut best_base: Option<&(Entry, usize)> = None;
         let mut best_rate: f64 = 0.0;
-        let mut best_chain_len: usize = 0;
         let mut best_base_offset: usize = 0;
         let tie_epsilon: f64 = 0.15;
 
-        for try_base in self.window.iter() {
-            if try_base.0.obj_type != entry.obj_type {
-                continue;
-            }
-
-            if try_base.0.chain_len >= MAX_CHAIN_LEN {
-                continue;
-            }
+        let candidates: Vec<_> = self.window
+            .par_iter()
+            .with_min_len(3)
+            .filter_map(|try_base| {
+                if try_base.0.obj_type != entry.obj_type {
+                    return None;
+                }
 
-            if try_base.0.hash == entry.hash {
-                continue;
-            }
+                if try_base.0.chain_len >= MAX_CHAIN_LEN {
+                    return None;
+                }
 
-            // 尺寸差异过大时直接跳过
-            let sym_ratio = (try_base.0.data.len().min(entry.data.len()) as f64) / (try_base.0.data.len().max(entry.data.len()) as f64);
-            if sym_ratio < 0.5 {
-                continue;
-            }
+                if try_base.0.hash == entry.hash {
+                    return None;
+                }
+                let sym_ratio = (try_base.0.data.len().min(entry.data.len()) as f64) / (try_base.0.data.len().max(entry.data.len()) as f64);
+                if sym_ratio < 0.5 {
+                    return None;
+                }
+                // 轻量 hash 预判（例如比较前 64 字节的 hash）
+                if !cheap_similar(&try_base.0.data, &entry.data) {
+                    return None;
+                }
 
-            // 轻量 hash 预判（例如比较前 64 字节的 hash）
-            if !cheap_similar(&try_base.0.data, &entry.data) {
-                continue;
-            }
+                let rate = if (try_base.0.data.len() +  entry.data.len()) / 2 > 20 {
+                           delta::heuristic_encode_rate_parallel(&try_base.0.data, &entry.data)
+                        }else {
+                            delta::encode_rate(&try_base.0.data, &entry.data)
+                        };
 
+                if rate > MIN_DELTA_RATE {
+                    Some((rate, try_base))
+                } else {
+                    None
+                }
+            })
+            .collect();
 
-            // when data is small, heuristic_encode_rate is not accurate
-            let rate = if (try_base.0.data.len() +  entry.data.len()) / 2 > 20 {
-               delta::heuristic_encode_rate_parallel(&try_base.0.data, &entry.data)
-            }else { 
-                delta::encode_rate(&try_base.0.data, &entry.data)
-            };
 
-            
-            if rate > MIN_DELTA_RATE {
-                let is_better = if rate > best_rate + tie_epsilon {
-                        // rate 明显更大，直接优先
+        for (rate, try_base) in candidates {
+            match best_base {
+                None => {
+                    best_rate = rate;
+                    best_base_offset = self.inner_offset - try_base.1;
+                    best_base = Some(try_base);
+                }
+                Some(best_base_ref) => {
+                    let is_better = if rate > best_rate + tie_epsilon {
                         true
                     } else if (rate - best_rate).abs() <= tie_epsilon {
-                        // 构造候选 Key
-                        let cand_key = (
-                            try_base.0.chain_len,
-                            self.inner_offset - try_base.1,
-                        );
-
-                        // 当前最佳 Key
-                        let best_key = (
-                            best_chain_len,
-                            best_base_offset,
-                        );
-                        cand_key > best_key
+
+                        if try_base.0.chain_len != best_base_ref.0.chain_len {
+                            try_base.0.chain_len > best_base_ref.0.chain_len
+                        } else {
+                            let base_offset = self.inner_offset - try_base.1;
+                            let best_offset = self.inner_offset - best_base_ref.1;
+                            base_offset > best_offset
+                        }
                     } else {
                         false
                     };
 
-                if is_better {
-                    best_rate = rate;
-                    best_chain_len = try_base.0.chain_len;
-                    best_base_offset = self.inner_offset - try_base.1;
-                    best_base = Some(try_base);
-
-                    if rate > 0.90 {
-                        break;
+                    if is_better {
+                        best_rate = rate;
+                        best_base_offset = self.inner_offset - try_base.1;
+                        best_base = Some(try_base);
                     }
                 }
-
             }
-
         }
 
+
         best_base.map(|best_base| {
+                let delta = delta::encode(&best_base.0.data, &entry.data);
 
-            let delta = delta::encode(&best_base.0.data, &entry.data);
+                let offset = self.inner_offset - best_base.1;
+                entry.obj_type = ObjectType::OffsetDelta;
+                entry.data = delta;
+                entry.chain_len = best_base.0.chain_len + 1;
+                //tracing::debug!("chain_len: {:?}", entry.chain_len);
+                offset
+            })
 
-            let offset = self.inner_offset - best_base.1;
-            entry.obj_type = ObjectType::OffsetDelta;
-            entry.data = delta;
-            entry.chain_len = best_base.0.chain_len + 1;
-            //tracing::debug!("chain_len: {:?}", entry.chain_len);
-            offset
-        })
     }
 
     /// Try to encode as sapling-zstdelta using objects in window
@@ -759,9 +819,17 @@ mod tests {
     #[tokio::test]
     async fn test_pack_encoder_parallel_large_file() {
         init_logger();
+
+
+        let start = Instant::now(); // 开始时间
         let entries = get_entries_for_test().await;
         let entries_number = entries.lock().await.len();
 
+        // 计算原始总大小
+        let total_original_size: usize = entries.lock().await.iter()
+            .map(|entry| entry.data.len())
+            .sum();
+
         // encode entries with parallel
         let (tx, mut rx) = mpsc::channel(1_000_000);
         let (entry_tx, entry_rx) = mpsc::channel::<Entry>(1_000_000);
@@ -787,8 +855,21 @@ mod tests {
         while let Some(chunk) = rx.recv().await {
             result.extend(chunk);
         }
-        tracing::info!("new pack file size: {}", result.len());
 
+        // 计算压缩率
+        let pack_size = result.len();
+        let compression_rate = if total_original_size > 0 {
+            1.0 - (pack_size as f64 / total_original_size as f64)
+        } else {
+            0.0
+        };
+        
+        
+
+        let duration = start.elapsed();
+        tracing::info!("test executed in: {:.2?}", duration);
+        tracing::info!("new pack file size: {}", result.len());
+        tracing::info!("compression rate: {:.2}%", compression_rate * 100.0);
         // check format
         check_format(&result);
     }
@@ -809,7 +890,7 @@ mod tests {
         let (tx, mut rx) = mpsc::channel(100_000);
         let (entry_tx, entry_rx) = mpsc::channel::<Entry>(100_000);
 
-        let mut encoder = PackEncoder::new(entries_number, 0, tx);
+        let mut encoder = PackEncoder::new(entries_number, 10, tx);
         tokio::spawn(async move {
             time_it!("test encode no parallel", {
                 encoder.encode(entry_rx).await.unwrap();

From 13ae8ff292ed59d8b7bd207c57d2993db4ffd6a0 Mon Sep 17 00:00:00 2001
From: pleasedontbreak123 <nabeiyq@126.com>
Date: Sat, 30 Aug 2025 17:53:40 +0800
Subject: [PATCH 10/17] test

---
 mercury/src/internal/pack/encode.rs | 662 +++++++++++++++-------------
 mercury/src/internal/pack/entry.rs  |   2 +-
 2 files changed, 362 insertions(+), 302 deletions(-)

diff --git a/mercury/src/internal/pack/encode.rs b/mercury/src/internal/pack/encode.rs
index 7e523c368..46649b527 100644
--- a/mercury/src/internal/pack/encode.rs
+++ b/mercury/src/internal/pack/encode.rs
@@ -2,23 +2,23 @@ use std::cmp::Ordering;
 use std::collections::VecDeque;
 use std::io::Write;
 
+use crate::internal::object::types::ObjectType;
+use crate::time_it;
+use crate::{errors::GitError, hash::SHA1, internal::pack::entry::Entry};
+use ahash::AHasher;
 use flate2::write::ZlibEncoder;
+use futures_util::future::join_all;
 use rayon::prelude::*;
 use sha1::{Digest, Sha1};
+use std::hash::{Hash, Hasher};
 use tokio::sync::mpsc;
+use tokio::task;
 use tokio::task::JoinHandle;
-use std::hash::{Hash, Hasher};
-use ahash::AHasher;
-
-use crate::internal::object::types::ObjectType;
-use crate::time_it;
-use crate::{errors::GitError, hash::SHA1, internal::pack::entry::Entry};
 
 const MAX_CHAIN_LEN: usize = 50;
-const MIN_DELTA_RATE: f64 = 0.5; // minimum delta rate 
+const MIN_DELTA_RATE: f64 = 0.5; // minimum delta rate
 const MAX_ZSTDELTA_CHAIN_LEN: usize = 50;
 
-
 /// A encoder for generating pack files with delta objects.
 pub struct PackEncoder {
     object_number: usize,
@@ -111,10 +111,11 @@ fn encode_one_object(entry: &Entry, offset: Option<usize>) -> Result<Vec<u8>, Gi
     Ok(encoded_data)
 }
 
-
-fn magic_sort(a:&Entry,b:&Entry)  -> Ordering {
+fn magic_sort(a: &Entry, b: &Entry) -> Ordering {
     let ord = b.obj_type.to_u8().cmp(&a.obj_type.to_u8());
-    if ord != Ordering::Equal { return ord; }
+    if ord != Ordering::Equal {
+        return ord;
+    }
 
     // the hash should be file hash not content hash
     // todo the feature need larger refactor
@@ -122,152 +123,29 @@ fn magic_sort(a:&Entry,b:&Entry)  -> Ordering {
     // if ord != Ordering::Equal { return ord; }
 
     let ord = b.data.len().cmp(&a.data.len());
-    if ord != Ordering::Equal { return ord; }
+    if ord != Ordering::Equal {
+        return ord;
+    }
 
     // fallback pointer order (newest first)
     (a as *const Entry).cmp(&(b as *const Entry))
 }
 
-// pub fn cheap_similar(a: &[u8], b: &[u8]) -> bool {
-//     // 取前 64 字节做预判
-//     let n = a.len().min(b.len()).min(64);
-//
-//     // 数据太小就直接认为可能相似
-//     if n < 8 {
-//         return true;
-//     }
-//
-//     // 内部函数：对切片计算 AHasher hash
-//     fn calc_hash(data: &[u8]) -> u64 {
-//         let mut hasher = AHasher::default();
-//         data.hash(&mut hasher);
-//         hasher.finish()
-//     }
-//
-//     // 计算前 n 字节的 hash
-//     let ha = calc_hash(&a[..n]);
-//     let hb = calc_hash(&b[..n]);
-//
-//     // 简单阈值判断：
-//     // - 如果两者 trailing_zeros 很少，说明不相似
-//     // - 你可以根据经验调节 2
-//     ha.trailing_zeros().min(hb.trailing_zeros()) > 2
-// }
-
 // /// 计算单片段 hash
 fn calc_hash(data: &[u8]) -> u64 {
     let mut hasher = AHasher::default();
     data.hash(&mut hasher);
     hasher.finish()
 }
-//
-// /// 分片 hash 函数，将数据分成 n_segments 片段并计算每片 hash
-// fn calc_hashes(data: &[u8], n_segments: usize) -> Vec<u64> {
-//     let mut res = Vec::with_capacity(n_segments);
-//     let step = (data.len() + n_segments - 1) / n_segments; // 向上取整
-//     for i in 0..n_segments {
-//         let start = i * step;
-//         let end = ((i + 1) * step).min(data.len());
-//         if start >= end { break; }
-//         res.push(calc_hash(&data[start..end]));
-//     }
-//     res
-// }
-//
-// /// cheap_similar 最终优化版
-// pub fn cheap_similar(a: &[u8], b: &[u8]) -> bool {
-//     let min_len = a.len().min(b.len());
-//
-//     // 对小对象直接全量比较
-//     if min_len < 8 {
-//         return a == b;
-//     }
-//
-//     // 前缀长度，可调
-//     let prefix_len = min_len.min(128);
-//     let a_prefix = &a[..prefix_len];
-//     let b_prefix = &b[..prefix_len];
-//
-//     // 动态分片数，每段约 32 字节
-//     let n_segments = (prefix_len / 32).max(1).min(4);
-//
-//     let a_hashes = calc_hashes(a_prefix, n_segments);
-//     let b_hashes = calc_hashes(b_prefix, n_segments);
-//
-//     // Hamming 距离阈值
-//     let hamming_threshold = 4;
-//
-//     // 对小前缀顺序比较
-//     if prefix_len <= 64 {
-//         let mut diff_bits = 0;
-//         for (x, y) in a_hashes.iter().zip(&b_hashes) {
-//             diff_bits += (x ^ y).count_ones();
-//             if diff_bits > hamming_threshold {
-//                 return false;
-//             }
-//         }
-//         true
-//     } else {
-//         // 大对象使用并行比较
-//         a_hashes
-//             .par_iter()
-//             .zip(&b_hashes)
-//             .map(|(x, y)| (x ^ y).count_ones())
-//             .sum::<u32>() <= hamming_threshold
-//     }
-// }
-
-// pub fn cheap_similar(a: &[u8], b: &[u8]) -> bool {
-//     let min_len = a.len().min(b.len());
-//     if min_len < 64 {
-//         // 小对象直接要求前后各16字节完全相等，严格但便宜
-//         let k = min_len.min(16);
-//         return a.get(..k) == b.get(..k) && a.get(min_len.saturating_sub(k)..) == b.get(min_len.saturating_sub(k)..);
-//     }
-// 
-//     // 允许的最大不匹配字节数（锚点总计），越大越宽松
-//     let mut mismatches = 0usize;
-//     let budget = 8; // 可调
-// 
-//     // 比对前/中/后三个区域，各取16字节
-//     let anchor_len = 16;
-//     let mids = [
-//         0,                                     // 前
-//         (min_len / 2).saturating_sub(anchor_len / 2), // 中
-//         min_len.saturating_sub(anchor_len),    // 后
-//     ];
-// 
-//     for &pos in &mids {
-//         let a_seg = &a[pos..pos + anchor_len];
-//         let b_seg = &b[pos..pos + anchor_len];
-//         // 逐字节计数（16字节很快，且避免未对齐/未定义读取）
-//         for i in 0..anchor_len {
-//             if a_seg[i] != b_seg[i] {
-//                 mismatches += 1;
-//                 if mismatches > budget {
-//                     return false; // 早停
-//                 }
-//             }
-//         }
-//     }
-// 
-//     // 再加一个极便宜的长度差门槛（避免尺寸相差太大的情况）
-//     let sym_ratio = (a.len().min(b.len()) as f64) / (a.len().max(b.len()) as f64);
-//     if sym_ratio < 0.5 {
-//         return false;
-//     }
-// 
-//     true
-// }
-
 
 fn cheap_similar(a: &[u8], b: &[u8]) -> bool {
     let k = a.len().min(b.len()).min(128);
-    if k == 0 { return false; }
+    if k == 0 {
+        return false;
+    }
     calc_hash(&a[..k]) == calc_hash(&b[..k])
 }
 
-
 impl PackEncoder {
     pub fn new(object_number: usize, window_size: usize, sender: mpsc::Sender<Vec<u8>>) -> Self {
         PackEncoder {
@@ -334,55 +212,212 @@ impl PackEncoder {
             ));
         }
 
-        // collect entries before magic sort
-
-        let mut all_entries = Vec::with_capacity(self.object_number);
-        time_it!( "%%%&&&&&  receive cost :",{
-            //let mut all_entries = Vec::with_capacity(self.object_number);
-            while let Some(entry) = entry_rx.recv().await {
-                all_entries.push(entry);
+        let mut commits: Vec<Entry> = Vec::new();
+        let mut trees: Vec<Entry> = Vec::new();
+        let mut blobs: Vec<Entry> = Vec::new();
+        let mut tags: Vec<Entry> = Vec::new();
+        while let Some(entry) = entry_rx.recv().await {
+            match entry.obj_type {
+                ObjectType::Commit => {
+                    commits.push(entry);
+                }
+                ObjectType::Tree => {
+                    trees.push(entry);
+                }
+                ObjectType::Blob => {
+                    blobs.push(entry);
+                }
+                ObjectType::Tag => {
+                    tags.push(entry);
+                }
+                _ => {}
             }
+        }
 
+        commits.sort_by(magic_sort);
+        trees.sort_by(magic_sort);
+        blobs.sort_by(magic_sort);
+        tags.sort_by(magic_sort);
+        tracing::info!(
+            "numbers :  commits: {:?} trees: {:?} blobs:{:?} tag :{:?}",
+            commits.len(),
+            trees.len(),
+            blobs.len(),
+            tags.len()
+        );
 
-            if all_entries.len() != self.object_number {
-                panic!(
-                    "not all objects are encoded, process:{}, total:{}",
-                    self.process_index, self.object_number
-                );
-            }
+        // 并行处理不同类型的 vec，使用 spawn_blocking 因为 try_as_offset_delta 是同步函数
+        let (commit_results, tree_results, blob_results, tag_results) = tokio::try_join!(
+            tokio::task::spawn_blocking(move || { Self::try_as_offset_delta(commits, 10) }),
+            tokio::task::spawn_blocking(move || { Self::try_as_offset_delta(trees, 10) }),
+            tokio::task::spawn_blocking(move || { Self::try_as_offset_delta(blobs, 10) }),
+            tokio::task::spawn_blocking(move || { Self::try_as_offset_delta(tags, 10) }),
+        )
+        .map_err(|e| GitError::PackEncodeError(format!("Task join error: {}", e)))?;
+
+        // 收集并合并结果
+        let all_encoded_data = [
+            commit_results
+                .map_err(|e| GitError::PackEncodeError(format!("Commit encoding error: {}", e)))?,
+            tree_results
+                .map_err(|e| GitError::PackEncodeError(format!("Tree encoding error: {}", e)))?,
+            blob_results
+                .map_err(|e| GitError::PackEncodeError(format!("Blob encoding error: {}", e)))?,
+            tag_results
+                .map_err(|e| GitError::PackEncodeError(format!("Tag encoding error: {}", e)))?,
+        ]
+        .concat();
+
+        // 按顺序发送合并后的结果
+        for data in all_encoded_data {
+            self.write_all_and_update(&data).await;
+        }
 
-                all_entries.sort_by(magic_sort);
-            });
-        //all_entries.par_sort_by(magic_sort);
+        // Hash signature
+        let hash_result = self.inner_hash.clone().finalize();
+        self.final_hash = Some(SHA1::from_bytes(&hash_result));
+        self.send_data(hash_result.to_vec()).await;
 
+        self.drop_sender();
+        Ok(())
+    }
 
+    /// Try to encode as delta using objects in window
+    /// # Returns
+    /// - Return (offset) if success make delta
+    /// - Return (None) if didn't delta,
+    fn try_as_offset_delta(
+        mut bucket: Vec<Entry>,
+        window_size: usize,
+    ) -> Result<Vec<Vec<u8>>, GitError> {
+        let mut current_offset = 0usize; // 当前类型的局部偏移
+        let mut window: VecDeque<(Entry, usize)> = VecDeque::with_capacity(window_size);
+        let mut res: Vec<Vec<u8>> = Vec::new();
+
+        for entry in bucket.iter_mut() {
+            
+            let entry_for_window = entry.clone();
+            // 每次循环重置最佳基对象选择
+            let mut best_base: Option<&(Entry, usize)> = None;
+            let mut best_rate: f64 = 0.0;
+            //let mut best_base_offset: usize = 0;
+            let tie_epsilon: f64 = 0.15;
+
+            // 在窗口内寻找最佳基对象
+            let candidates: Vec<_> = window
+                .par_iter()
+                .with_min_len(3)
+                .filter_map(|try_base| {
+                    
+                    
+                    if try_base.0.obj_type != entry.obj_type {
+                        return None;
+                    }
 
+                    if try_base.0.chain_len >= MAX_CHAIN_LEN {
+                        return None;
+                    }
+
+                    if try_base.0.hash == entry.hash {
+                        return None;
+                    }
+
+                    let sym_ratio = (try_base.0.data.len().min(entry.data.len()) as f64)
+                        / (try_base.0.data.len().max(entry.data.len()) as f64);
+                    if sym_ratio < 0.5 {
+                        return None;
+                    }
+
+                    if !cheap_similar(&try_base.0.data, &entry.data) {
+                        return None;
+                    }
+
+                    let rate = if (try_base.0.data.len() + entry.data.len()) / 2 > 20 {
+                        delta::heuristic_encode_rate_parallel(&try_base.0.data, &entry.data)
+                    } else {
+                        delta::encode_rate(&try_base.0.data, &entry.data)
+                    };
+
+                    if rate > MIN_DELTA_RATE {
+                        Some((rate, try_base))
+                    } else {
+                        None
+                    }
+                })
+                .collect();
+
+            // 选择最佳基对象
+            for (rate, try_base) in candidates {
+                match best_base {
+                    None => {
+                        best_rate = rate;
+                        //best_base_offset = current_offset - try_base.1; // 使用局部偏移
+                        best_base = Some(try_base);
+                    }
+                    Some(best_base_ref) => {
+                        let is_better = if rate > best_rate + tie_epsilon {
+                            true
+                        } else if (rate - best_rate).abs() <= tie_epsilon {
+                            
+                            try_base.0.chain_len > best_base_ref.0.chain_len
+                            
+                        } else {
+                            false
+                        };
 
-        for mut entry in all_entries {
-            self.process_index += 1;
-            let offset = self.inner_offset;
-            let mut entry_for_window = entry.clone();
-            let try_offset = if enable_zstdelta {
-                self.try_as_offset_zstdelta(&mut entry)
+                        if is_better {
+                            best_rate = rate;
+                            //best_base_offset = current_offset - try_base.1;
+                            best_base = Some(try_base);
+                        }
+                    }
+                }
+            }
+
+            // 应用 delta 并更新窗口
+            if let Some(best_base) = best_base {
+                // 保存原始数据用于窗口
+                let original_data = entry.data.clone();
+                let original_obj_type = entry.obj_type;
+                
+                let delta = delta::encode(&best_base.0.data, &original_data);
+                let offset = best_base.1-entry.1; // 基对象在pack文件中的绝对位置
+
+                entry.obj_type = ObjectType::OffsetDelta;
+                entry.data = delta;
+                entry.chain_len = best_base.0.chain_len + 1;
+
+                // 将原始对象添加到窗口（不是 delta 对象）
+                let original_entry = Entry {
+                    obj_type: original_obj_type,
+                    data: original_data, // 保存原始数据
+                    hash: entry.hash,
+                    chain_len: entry.chain_len,
+                };
+                
+                // 在添加到窗口之前检查大小
+                if window.len() >= window_size {
+                    window.pop_front();
+                }
+                window.push_back((original_entry, current_offset));
             } else {
-                self.try_as_offset_delta(&mut entry)
-            };
-            let obj_data = encode_one_object(&entry, try_offset)?;
-            self.write_all_and_update(&obj_data).await;
-            entry_for_window.chain_len = entry.chain_len;
-        
-            self.window.push_back((entry_for_window, offset));
-            if self.window.len() > self.window_size {
-                self.window.pop_front();
+                // 没有找到合适的基对象，保持原样
+                // 在添加到窗口之前检查大小
+                if window.len() >= window_size {
+                    window.pop_front();
+                }
+                window.push_back((entry.clone(), current_offset));
             }
+
+            // 编码对象（使用局部偏移）
+            let obj_data = encode_one_object(entry, Some(current_offset)).unwrap();
+            res.push(obj_data.clone());
+
+            // 更新局部偏移
+            current_offset += obj_data.len();
         }
-        
-        // hash signature
-        let hash_result = self.inner_hash.clone().finalize();
-        self.final_hash = Some(SHA1::from_bytes(&hash_result));
-        self.send_data((hash_result).to_vec()).await;
-        self.drop_sender();
-        Ok(())
+
+        Ok(res)
     }
 
     /// Parallel encode with rayon, only works when window_size == 0 (no delta)
@@ -457,100 +492,100 @@ impl PackEncoder {
         Ok(())
     }
 
-    /// Try to encode as delta using objects in window
-    /// # Returns
-    /// - Return (offset) if success make delta
-    /// - Return (None) if didn't delta,
-    fn try_as_offset_delta(&self, entry: &mut Entry) -> Option<usize> {
-        let mut best_base: Option<&(Entry, usize)> = None;
-        let mut best_rate: f64 = 0.0;
-        let mut best_base_offset: usize = 0;
-        let tie_epsilon: f64 = 0.15;
-
-        let candidates: Vec<_> = self.window
-            .par_iter()
-            .with_min_len(3)
-            .filter_map(|try_base| {
-                if try_base.0.obj_type != entry.obj_type {
-                    return None;
-                }
-
-                if try_base.0.chain_len >= MAX_CHAIN_LEN {
-                    return None;
-                }
-
-                if try_base.0.hash == entry.hash {
-                    return None;
-                }
-                let sym_ratio = (try_base.0.data.len().min(entry.data.len()) as f64) / (try_base.0.data.len().max(entry.data.len()) as f64);
-                if sym_ratio < 0.5 {
-                    return None;
-                }
-                // 轻量 hash 预判（例如比较前 64 字节的 hash）
-                if !cheap_similar(&try_base.0.data, &entry.data) {
-                    return None;
-                }
-
-                let rate = if (try_base.0.data.len() +  entry.data.len()) / 2 > 20 {
-                           delta::heuristic_encode_rate_parallel(&try_base.0.data, &entry.data)
-                        }else {
-                            delta::encode_rate(&try_base.0.data, &entry.data)
-                        };
-
-                if rate > MIN_DELTA_RATE {
-                    Some((rate, try_base))
-                } else {
-                    None
-                }
-            })
-            .collect();
-
-
-        for (rate, try_base) in candidates {
-            match best_base {
-                None => {
-                    best_rate = rate;
-                    best_base_offset = self.inner_offset - try_base.1;
-                    best_base = Some(try_base);
-                }
-                Some(best_base_ref) => {
-                    let is_better = if rate > best_rate + tie_epsilon {
-                        true
-                    } else if (rate - best_rate).abs() <= tie_epsilon {
-
-                        if try_base.0.chain_len != best_base_ref.0.chain_len {
-                            try_base.0.chain_len > best_base_ref.0.chain_len
-                        } else {
-                            let base_offset = self.inner_offset - try_base.1;
-                            let best_offset = self.inner_offset - best_base_ref.1;
-                            base_offset > best_offset
-                        }
-                    } else {
-                        false
-                    };
-
-                    if is_better {
-                        best_rate = rate;
-                        best_base_offset = self.inner_offset - try_base.1;
-                        best_base = Some(try_base);
-                    }
-                }
-            }
-        }
-
-
-        best_base.map(|best_base| {
-                let delta = delta::encode(&best_base.0.data, &entry.data);
-
-                let offset = self.inner_offset - best_base.1;
-                entry.obj_type = ObjectType::OffsetDelta;
-                entry.data = delta;
-                entry.chain_len = best_base.0.chain_len + 1;
-                //tracing::debug!("chain_len: {:?}", entry.chain_len);
-                offset
-            })
-
-    }
+    // /// Try to encode as delta using objects in window
+    // /// # Returns
+    // /// - Return (offset) if success make delta
+    // /// - Return (None) if didn't delta,
+    // fn try_as_offset_delta(&self, entry: &mut Entry) -> Option<usize> {
+    //     let mut best_base: Option<&(Entry, usize)> = None;
+    //     let mut best_rate: f64 = 0.0;
+    //     let mut best_base_offset: usize = 0;
+    //     let tie_epsilon: f64 = 0.15;
+    //
+    //     let candidates: Vec<_> = self.window
+    //         .par_iter()
+    //         .with_min_len(3)
+    //         .filter_map(|try_base| {
+    //             if try_base.0.obj_type != entry.obj_type {
+    //                 return None;
+    //             }
+    //
+    //             if try_base.0.chain_len >= MAX_CHAIN_LEN {
+    //                 return None;
+    //             }
+    //
+    //             if try_base.0.hash == entry.hash {
+    //                 return None;
+    //             }
+    //             let sym_ratio = (try_base.0.data.len().min(entry.data.len()) as f64) / (try_base.0.data.len().max(entry.data.len()) as f64);
+    //             if sym_ratio < 0.5 {
+    //                 return None;
+    //             }
+    //             // 轻量 hash 预判（例如比较前 64 字节的 hash）
+    //             if !cheap_similar(&try_base.0.data, &entry.data) {
+    //                 return None;
+    //             }
+    //
+    //             let rate = if (try_base.0.data.len() +  entry.data.len()) / 2 > 20 {
+    //                        delta::heuristic_encode_rate_parallel(&try_base.0.data, &entry.data)
+    //                     }else {
+    //                         delta::encode_rate(&try_base.0.data, &entry.data)
+    //                     };
+    //
+    //             if rate > MIN_DELTA_RATE {
+    //                 Some((rate, try_base))
+    //             } else {
+    //                 None
+    //             }
+    //         })
+    //         .collect();
+    //
+    //
+    //     for (rate, try_base) in candidates {
+    //         match best_base {
+    //             None => {
+    //                 best_rate = rate;
+    //                 best_base_offset = self.inner_offset - try_base.1;
+    //                 best_base = Some(try_base);
+    //             }
+    //             Some(best_base_ref) => {
+    //                 let is_better = if rate > best_rate + tie_epsilon {
+    //                     true
+    //                 } else if (rate - best_rate).abs() <= tie_epsilon {
+    //
+    //                     if try_base.0.chain_len != best_base_ref.0.chain_len {
+    //                         try_base.0.chain_len > best_base_ref.0.chain_len
+    //                     } else {
+    //                         let base_offset = self.inner_offset - try_base.1;
+    //                         let best_offset = self.inner_offset - best_base_ref.1;
+    //                         base_offset > best_offset
+    //                     }
+    //                 } else {
+    //                     false
+    //                 };
+    //
+    //                 if is_better {
+    //                     best_rate = rate;
+    //                     best_base_offset = self.inner_offset - try_base.1;
+    //                     best_base = Some(try_base);
+    //                 }
+    //             }
+    //         }
+    //     }
+    //
+    //
+    //     best_base.map(|best_base| {
+    //             let delta = delta::encode(&best_base.0.data, &entry.data);
+    //
+    //             let offset = self.inner_offset - best_base.1;
+    //             entry.obj_type = ObjectType::OffsetDelta;
+    //             entry.data = delta;
+    //             entry.chain_len = best_base.0.chain_len + 1;
+    //             //tracing::debug!("chain_len: {:?}", entry.chain_len);
+    //             offset
+    //         })
+    //
+    // }
 
     /// Try to encode as sapling-zstdelta using objects in window
     /// Refs: https://sapling-scm.com/docs/dev/internals/zstdelta/
@@ -563,7 +598,7 @@ impl PackEncoder {
         let mut best_chain_len: usize = 0;
         let mut best_base_offset: usize = 0;
         let tie_epsilon: f64 = 0.02; // when rates are close, prefer deeper/newer base
-        
+
         if entry.data.is_empty() {
             return None;
         }
@@ -587,8 +622,7 @@ impl PackEncoder {
             if !cheap_similar(&try_base.0.data, &entry.data) {
                 continue;
             }
-            
-            
+
             let try_delta_obj = match zstdelta::diff(&try_base.0.data, &entry.data) {
                 Ok(v) => v,
                 Err(_) => continue, // IO error for this base, try next
@@ -607,7 +641,8 @@ impl PackEncoder {
                 } else if (rate - best_rate).abs() <= tie_epsilon {
                     let base_chain = try_base.0.chain_len;
                     let base_offset = self.inner_offset - try_base.1;
-                    base_chain > best_chain_len || (base_chain == best_chain_len && base_offset > best_base_offset)
+                    base_chain > best_chain_len
+                        || (base_chain == best_chain_len && base_offset > best_base_offset)
                 } else {
                     false
                 };
@@ -640,7 +675,7 @@ impl PackEncoder {
     /// async version of encode, result data will be returned by JoinHandle.
     /// It will consume PackEncoder, so you can't use it after calling this function.
     /// when window_size = 0, it executes parallel_encode which retains stream transmission
-    /// when window_size = 0,it executes encode which uses magic sort and delta. 
+    /// when window_size = 0,it executes encode which uses magic sort and delta.
     /// It seems that all other modules rely on this api
     pub async fn encode_async(
         mut self,
@@ -676,11 +711,11 @@ impl PackEncoder {
 
 #[cfg(test)]
 mod tests {
-    use std::env;
     use std::collections::HashSet;
+    use std::env;
     use std::sync::Arc;
-    use std::{io::Cursor, path::PathBuf};
     use std::time::Instant;
+    use std::{io::Cursor, path::PathBuf};
     use tokio::sync::Mutex;
 
     use crate::internal::object::blob::Blob;
@@ -747,7 +782,9 @@ mod tests {
         let mut contents: Vec<Vec<u8>> = Vec::new();
         let mut base = String::from("base line 0\n");
         for i in 1..=8usize {
-            base.push_str(&format!("line {i}: some repeated text with slight changes...\n"));
+            base.push_str(&format!(
+                "line {i}: some repeated text with slight changes...\n"
+            ));
             contents.push(base.as_bytes().to_vec());
         }
 
@@ -771,14 +808,20 @@ mod tests {
         }
 
         // 3) 用 Pack::decode 解析，重建为完整对象
-        let mut p = Pack::new(None, Some(1024 * 1024 * 20), Some(PathBuf::from("/tmp/.cache_temp")), true);
+        let mut p = Pack::new(
+            None,
+            Some(1024 * 1024 * 20),
+            Some(PathBuf::from("/tmp/.cache_temp")),
+            true,
+        );
         let mut reader = Cursor::new(pack_bytes);
         let decoded = Arc::new(tokio::sync::Mutex::new(Vec::<Entry>::new()));
         let decoded_c = decoded.clone();
         p.decode(&mut reader, move |entry, _| {
             let mut guard = decoded_c.blocking_lock();
             guard.push(entry);
-        }).expect("pack decode should succeed");
+        })
+        .expect("pack decode should succeed");
 
         let decoded_entries = decoded.lock().await;
         assert_eq!(decoded_entries.len(), contents.len());
@@ -786,7 +829,10 @@ mod tests {
         // 4) 校验内容集合相等（忽略顺序）
         let expected: HashSet<Vec<u8>> = contents.into_iter().collect();
         let got: HashSet<Vec<u8>> = decoded_entries.iter().map(|e| e.data.clone()).collect();
-        assert_eq!(expected, got, "decoded contents should equal original contents");
+        assert_eq!(
+            expected, got,
+            "decoded contents should equal original contents"
+        );
     }
 
     async fn get_entries_for_test() -> Arc<Mutex<Vec<Entry>>> {
@@ -820,13 +866,15 @@ mod tests {
     async fn test_pack_encoder_parallel_large_file() {
         init_logger();
 
-
         let start = Instant::now(); // 开始时间
         let entries = get_entries_for_test().await;
         let entries_number = entries.lock().await.len();
 
         // 计算原始总大小
-        let total_original_size: usize = entries.lock().await.iter()
+        let total_original_size: usize = entries
+            .lock()
+            .await
+            .iter()
             .map(|entry| entry.data.len())
             .sum();
 
@@ -863,8 +911,6 @@ mod tests {
         } else {
             0.0
         };
-        
-        
 
         let duration = start.elapsed();
         tracing::info!("test executed in: {:.2?}", duration);
@@ -881,12 +927,15 @@ mod tests {
         let entries_number = entries.lock().await.len();
 
         // 计算原始总大小
-        let total_original_size: usize = entries.lock().await.iter()
+        let total_original_size: usize = entries
+            .lock()
+            .await
+            .iter()
             .map(|entry| entry.data.len())
             .sum();
 
         let start = Instant::now(); // 开始时间
-        // encode entries
+                                    // encode entries
         let (tx, mut rx) = mpsc::channel(100_000);
         let (entry_tx, entry_rx) = mpsc::channel::<Entry>(100_000);
 
@@ -931,7 +980,10 @@ mod tests {
         tracing::info!("new pack file size: {}", pack_size);
         tracing::info!("original total size: {}", total_original_size);
         tracing::info!("compression rate: {:.2}%", compression_rate * 100.0);
-        tracing::info!("space saved: {} bytes", total_original_size.saturating_sub(pack_size));
+        tracing::info!(
+            "space saved: {} bytes",
+            total_original_size.saturating_sub(pack_size)
+        );
     }
 
     #[tokio::test]
@@ -939,9 +991,12 @@ mod tests {
         init_logger();
         let entries = get_entries_for_test().await;
         let entries_number = entries.lock().await.len();
-        
+
         // 计算原始总大小
-        let total_original_size: usize = entries.lock().await.iter()
+        let total_original_size: usize = entries
+            .lock()
+            .await
+            .iter()
             .map(|entry| entry.data.len())
             .sum();
 
@@ -967,7 +1022,7 @@ mod tests {
         while let Some(chunk) = rx.recv().await {
             result.extend(chunk);
         }
-        
+
         // 计算压缩率
         let pack_size = result.len();
         let compression_rate = if total_original_size > 0 {
@@ -981,8 +1036,11 @@ mod tests {
         tracing::info!("new pack file size: {}", pack_size);
         tracing::info!("original total size: {}", total_original_size);
         tracing::info!("compression rate: {:.2}%", compression_rate * 100.0);
-        tracing::info!("space saved: {} bytes", total_original_size.saturating_sub(pack_size));
-        
+        tracing::info!(
+            "space saved: {} bytes",
+            total_original_size.saturating_sub(pack_size)
+        );
+
         // check format
         check_format(&result);
     }
@@ -1000,8 +1058,6 @@ mod tests {
         assert_eq!(result, value as u64);
     }
 
-    
-
     #[tokio::test]
     async fn test_pack_encoder_large_file_with_delta() {
         init_logger();
@@ -1009,7 +1065,10 @@ mod tests {
         let entries_number = entries.lock().await.len();
 
         // 计算原始总大小
-        let total_original_size: usize = entries.lock().await.iter()
+        let total_original_size: usize = entries
+            .lock()
+            .await
+            .iter()
             .map(|entry| entry.data.len())
             .sum();
 
@@ -1045,16 +1104,17 @@ mod tests {
             0.0
         };
 
-
         let duration = start.elapsed();
         tracing::info!("test executed in: {:.2?}", duration);
         tracing::info!("new pack file size: {}", pack_size);
         tracing::info!("original total size: {}", total_original_size);
         tracing::info!("compression rate: {:.2}%", compression_rate * 100.0);
-        tracing::info!("space saved: {} bytes", total_original_size.saturating_sub(pack_size));
+        tracing::info!(
+            "space saved: {} bytes",
+            total_original_size.saturating_sub(pack_size)
+        );
 
         // check format
         check_format(&result);
     }
-
 }
diff --git a/mercury/src/internal/pack/entry.rs b/mercury/src/internal/pack/entry.rs
index 9c66e407d..4afebc195 100644
--- a/mercury/src/internal/pack/entry.rs
+++ b/mercury/src/internal/pack/entry.rs
@@ -18,7 +18,7 @@ pub struct Entry {
     pub obj_type: ObjectType,
     pub data: Vec<u8>,
     pub hash: SHA1,
-    pub chain_len: usize
+    pub chain_len: usize,
 }
 
 impl PartialEq for Entry {

From a4da0426e7b3e338fafc3818e295dcbe2d69d850 Mon Sep 17 00:00:00 2001
From: pleasedontbreak123 <nabeiyq@126.com>
Date: Sun, 31 Aug 2025 17:26:15 +0800
Subject: [PATCH 11/17] complete encode

---
 mercury/src/internal/pack/encode.rs | 375 ++++++++++------------------
 1 file changed, 133 insertions(+), 242 deletions(-)

diff --git a/mercury/src/internal/pack/encode.rs b/mercury/src/internal/pack/encode.rs
index 46649b527..a5653d752 100644
--- a/mercury/src/internal/pack/encode.rs
+++ b/mercury/src/internal/pack/encode.rs
@@ -7,24 +7,22 @@ use crate::time_it;
 use crate::{errors::GitError, hash::SHA1, internal::pack::entry::Entry};
 use ahash::AHasher;
 use flate2::write::ZlibEncoder;
-use futures_util::future::join_all;
 use rayon::prelude::*;
 use sha1::{Digest, Sha1};
 use std::hash::{Hash, Hasher};
 use tokio::sync::mpsc;
-use tokio::task;
 use tokio::task::JoinHandle;
 
 const MAX_CHAIN_LEN: usize = 50;
 const MIN_DELTA_RATE: f64 = 0.5; // minimum delta rate
-const MAX_ZSTDELTA_CHAIN_LEN: usize = 50;
+//const MAX_ZSTDELTA_CHAIN_LEN: usize = 50;
 
 /// A encoder for generating pack files with delta objects.
 pub struct PackEncoder {
     object_number: usize,
     process_index: usize,
     window_size: usize,
-    window: VecDeque<(Entry, usize)>, // entry and offset
+   // window: VecDeque<(Entry, usize)>, // entry and offset
     sender: Option<mpsc::Sender<Vec<u8>>>,
     inner_offset: usize, // offset of current entry
     inner_hash: Sha1,    // Not SHA1 because need update trait
@@ -112,10 +110,10 @@ fn encode_one_object(entry: &Entry, offset: Option<usize>) -> Result<Vec<u8>, Gi
 }
 
 fn magic_sort(a: &Entry, b: &Entry) -> Ordering {
-    let ord = b.obj_type.to_u8().cmp(&a.obj_type.to_u8());
-    if ord != Ordering::Equal {
-        return ord;
-    }
+    // let ord = b.obj_type.to_u8().cmp(&a.obj_type.to_u8());
+    // if ord != Ordering::Equal {
+    //     return ord;
+    // }
 
     // the hash should be file hash not content hash
     // todo the feature need larger refactor
@@ -152,7 +150,7 @@ impl PackEncoder {
             object_number,
             window_size,
             process_index: 0,
-            window: VecDeque::with_capacity(window_size),
+           // window: VecDeque::with_capacity(window_size),
             sender: Some(sender),
             inner_offset: 12, // 12 bytes header
             inner_hash: Sha1::new(),
@@ -184,7 +182,7 @@ impl PackEncoder {
     /// - Returns a `GitError` if there is a failure during the encoding process.
     /// - Returns `PackEncodeError` if an encoding operation is already in progress.
     pub async fn encode(&mut self, entry_rx: mpsc::Receiver<Entry>) -> Result<(), GitError> {
-        return self.inner_encode(entry_rx, false).await;
+         self.inner_encode(entry_rx, false).await
     }
 
     pub async fn encode_with_zstdelta(
@@ -246,12 +244,12 @@ impl PackEncoder {
             tags.len()
         );
 
-        // 并行处理不同类型的 vec，使用 spawn_blocking 因为 try_as_offset_delta 是同步函数
+        // parallel encoding vec with different object_type
         let (commit_results, tree_results, blob_results, tag_results) = tokio::try_join!(
-            tokio::task::spawn_blocking(move || { Self::try_as_offset_delta(commits, 10) }),
-            tokio::task::spawn_blocking(move || { Self::try_as_offset_delta(trees, 10) }),
-            tokio::task::spawn_blocking(move || { Self::try_as_offset_delta(blobs, 10) }),
-            tokio::task::spawn_blocking(move || { Self::try_as_offset_delta(tags, 10) }),
+            tokio::task::spawn_blocking(move || { Self::try_as_offset_delta(commits, 10,enable_zstdelta) }),
+            tokio::task::spawn_blocking(move || { Self::try_as_offset_delta(trees, 10,enable_zstdelta) }),
+            tokio::task::spawn_blocking(move || { Self::try_as_offset_delta(blobs, 10,enable_zstdelta) }),
+            tokio::task::spawn_blocking(move || { Self::try_as_offset_delta(tags, 10,enable_zstdelta) }),
         )
         .map_err(|e| GitError::PackEncodeError(format!("Task join error: {}", e)))?;
 
@@ -283,27 +281,30 @@ impl PackEncoder {
     }
 
     /// Try to encode as delta using objects in window
+    /// delta & zstdelta have been gathered here 
+    /// Refs: https://sapling-scm.com/docs/dev/internals/zstdelta/
+    /// the sliding window was moved here
     /// # Returns
-    /// - Return (offset) if success make delta
+    /// - Return (Vec<Vec<u8>) if success make delta
     /// - Return (None) if didn't delta,
     fn try_as_offset_delta(
         mut bucket: Vec<Entry>,
         window_size: usize,
+        enable_zstdelta: bool,
     ) -> Result<Vec<Vec<u8>>, GitError> {
-        let mut current_offset = 0usize; // 当前类型的局部偏移
+        let mut current_offset = 0usize; 
         let mut window: VecDeque<(Entry, usize)> = VecDeque::with_capacity(window_size);
         let mut res: Vec<Vec<u8>> = Vec::new();
 
         for entry in bucket.iter_mut() {
             
-            let entry_for_window = entry.clone();
+            //let entry_for_window = entry.clone();
             // 每次循环重置最佳基对象选择
             let mut best_base: Option<&(Entry, usize)> = None;
             let mut best_rate: f64 = 0.0;
-            //let mut best_base_offset: usize = 0;
             let tie_epsilon: f64 = 0.15;
 
-            // 在窗口内寻找最佳基对象
+            
             let candidates: Vec<_> = window
                 .par_iter()
                 .with_min_len(3)
@@ -332,11 +333,15 @@ impl PackEncoder {
                         return None;
                     }
 
-                    let rate = if (try_base.0.data.len() + entry.data.len()) / 2 > 20 {
-                        delta::heuristic_encode_rate_parallel(&try_base.0.data, &entry.data)
-                    } else {
-                        delta::encode_rate(&try_base.0.data, &entry.data)
-                    };
+
+                    let rate = 
+                        if (try_base.0.data.len() + entry.data.len()) / 2 > 20 {
+                            delta::heuristic_encode_rate_parallel(&try_base.0.data, &entry.data)
+                        } else {
+                            delta::encode_rate(&try_base.0.data, &entry.data)
+                        };
+                    
+
 
                     if rate > MIN_DELTA_RATE {
                         Some((rate, try_base))
@@ -346,80 +351,66 @@ impl PackEncoder {
                 })
                 .collect();
 
-            // 选择最佳基对象
+           
             for (rate, try_base) in candidates {
                 match best_base {
                     None => {
                         best_rate = rate;
-                        //best_base_offset = current_offset - try_base.1; // 使用局部偏移
+                        //best_base_offset = current_offset - try_base.1; 
                         best_base = Some(try_base);
                     }
                     Some(best_base_ref) => {
                         let is_better = if rate > best_rate + tie_epsilon {
                             true
                         } else if (rate - best_rate).abs() <= tie_epsilon {
-                            
                             try_base.0.chain_len > best_base_ref.0.chain_len
-                            
                         } else {
                             false
                         };
 
                         if is_better {
                             best_rate = rate;
-                            //best_base_offset = current_offset - try_base.1;
                             best_base = Some(try_base);
                         }
                     }
                 }
             }
-
-            // 应用 delta 并更新窗口
-            if let Some(best_base) = best_base {
-                // 保存原始数据用于窗口
-                let original_data = entry.data.clone();
-                let original_obj_type = entry.obj_type;
-                
-                let delta = delta::encode(&best_base.0.data, &original_data);
-                let offset = best_base.1-entry.1; // 基对象在pack文件中的绝对位置
-
-                entry.obj_type = ObjectType::OffsetDelta;
-                entry.data = delta;
-                entry.chain_len = best_base.0.chain_len + 1;
-
-                // 将原始对象添加到窗口（不是 delta 对象）
-                let original_entry = Entry {
-                    obj_type: original_obj_type,
-                    data: original_data, // 保存原始数据
-                    hash: entry.hash,
-                    chain_len: entry.chain_len,
+            
+            let mut entry_for_window = entry.clone();
+
+            let offset = best_base.map(|best_base| { 
+                let delta = if enable_zstdelta{
+                    entry.obj_type = ObjectType::OffsetZstdelta;
+                    zstdelta::diff(&best_base.0.data,&entry.data)
+                        .map_err(|e| GitError::DeltaObjectError(format!("zstdelta diff failed: {}", e))).unwrap()
+                }else {
+                    entry.obj_type = ObjectType::OffsetDelta;
+                    delta::encode(&best_base.0.data, &entry.data)
                 };
-                
-                // 在添加到窗口之前检查大小
-                if window.len() >= window_size {
-                    window.pop_front();
-                }
-                window.push_back((original_entry, current_offset));
-            } else {
-                // 没有找到合适的基对象，保持原样
-                // 在添加到窗口之前检查大小
-                if window.len() >= window_size {
-                    window.pop_front();
-                }
-                window.push_back((entry.clone(), current_offset));
+                //entry.obj_type = ObjectType::OffsetDelta; 
+                entry.data = delta; 
+                entry.chain_len = best_base.0.chain_len + 1;
+                let offset = current_offset - best_base.1;
+                offset
+            }); 
+            
+            entry_for_window.chain_len = entry.chain_len;
+            let obj_data = encode_one_object(&entry, offset)?;
+            window.push_back((entry_for_window, current_offset));
+            if window.len() > window_size {
+                window.pop_front();
             }
-
-            // 编码对象（使用局部偏移）
-            let obj_data = encode_one_object(entry, Some(current_offset)).unwrap();
-            res.push(obj_data.clone());
-
-            // 更新局部偏移
-            current_offset += obj_data.len();
+            current_offset = current_offset + obj_data.len();
+            res.push(obj_data);
+            
+            
         }
-
         Ok(res)
     }
 
+
+   
+
     /// Parallel encode with rayon, only works when window_size == 0 (no delta)
     pub async fn parallel_encode(
         &mut self,
@@ -492,179 +483,85 @@ impl PackEncoder {
         Ok(())
     }
 
-    // /// Try to encode as delta using objects in window
-    // /// # Returns
-    // /// - Return (offset) if success make delta
-    // /// - Return (None) if didn't delta,
-    // fn try_as_offset_delta(&self, entry: &mut Entry) -> Option<usize> {
-    //     let mut best_base: Option<&(Entry, usize)> = None;
+    
+    /// Try to encode as sapling-zstdelta using objects in window
+    /// Refs: https://sapling-scm.com/docs/dev/internals/zstdelta/
+    /// # Returns
+    /// - Return (offset) if success make delta
+    /// - Return (None) if didn't delta,
+    // fn try_as_offset_zstdelta(&self, entry: &mut Entry) -> Option<usize> {
+    //     let mut best_delta_obj = None; // obj, offset
     //     let mut best_rate: f64 = 0.0;
+    //     let mut best_chain_len: usize = 0;
     //     let mut best_base_offset: usize = 0;
-    //     let tie_epsilon: f64 = 0.15;
-    //
-    //     let candidates: Vec<_> = self.window
-    //         .par_iter()
-    //         .with_min_len(3)
-    //         .filter_map(|try_base| {
-    //             if try_base.0.obj_type != entry.obj_type {
-    //                 return None;
-    //             }
-    //
-    //             if try_base.0.chain_len >= MAX_CHAIN_LEN {
-    //                 return None;
-    //             }
-    //
-    //             if try_base.0.hash == entry.hash {
-    //                 return None;
-    //             }
-    //             let sym_ratio = (try_base.0.data.len().min(entry.data.len()) as f64) / (try_base.0.data.len().max(entry.data.len()) as f64);
-    //             if sym_ratio < 0.5 {
-    //                 return None;
-    //             }
-    //             // 轻量 hash 预判（例如比较前 64 字节的 hash）
-    //             if !cheap_similar(&try_base.0.data, &entry.data) {
-    //                 return None;
-    //             }
-    //
-    //             let rate = if (try_base.0.data.len() +  entry.data.len()) / 2 > 20 {
-    //                        delta::heuristic_encode_rate_parallel(&try_base.0.data, &entry.data)
-    //                     }else {
-    //                         delta::encode_rate(&try_base.0.data, &entry.data)
-    //                     };
-    //
-    //             if rate > MIN_DELTA_RATE {
-    //                 Some((rate, try_base))
+    //     let tie_epsilon: f64 = 0.02; // when rates are close, prefer deeper/newer base
+    // 
+    //     if entry.data.is_empty() {
+    //         return None;
+    //     }
+    // 
+    //     for try_base in self.window.iter() {
+    //         if try_base.0.obj_type != entry.obj_type {
+    //             continue;
+    //         }
+    //         if try_base.0.chain_len >= MAX_ZSTDELTA_CHAIN_LEN {
+    //             continue;
+    //         }
+    // 
+    //         // 尺寸差异过大时直接跳过
+    //         let size_a = try_base.0.data.len();
+    //         let size_b = entry.data.len();
+    //         if size_a * 10 < size_b || size_b * 10 < size_a {
+    //             continue;
+    //         }
+    // 
+    //         // 轻量 hash 预判（例如比较前 64 字节的 hash）
+    //         if !cheap_similar(&try_base.0.data, &entry.data) {
+    //             continue;
+    //         }
+    // 
+    //         let try_delta_obj = match zstdelta::diff(&try_base.0.data, &entry.data) {
+    //             Ok(v) => v,
+    //             Err(_) => continue, // IO error for this base, try next
+    //         };
+    //         // The DeltaDiff algorithm uses the proportion of the delta part as the delta rate,
+    //         // while zstdelta does not have this data and directly uses the overall compression rate.
+    //         let rate = 1.0 - try_delta_obj.len() as f64 / entry.data.len() as f64;
+    //         // if rate > MIN_DELTA_RATE && rate > best_rate {
+    //         //     best_rate = rate;
+    //         //     best_delta_obj = Some((try_delta_obj, try_base.1));
+    //         // }
+    // 
+    //         if rate > MIN_DELTA_RATE {
+    //             let is_better = if rate > best_rate + tie_epsilon {
+    //                 true
+    //             } else if (rate - best_rate).abs() <= tie_epsilon {
+    //                 let base_chain = try_base.0.chain_len;
+    //                 let base_offset = self.inner_offset - try_base.1;
+    //                 base_chain > best_chain_len
+    //                     || (base_chain == best_chain_len && base_offset > best_base_offset)
     //             } else {
-    //                 None
-    //             }
-    //         })
-    //         .collect();
-    //
-    //
-    //     for (rate, try_base) in candidates {
-    //         match best_base {
-    //             None => {
+    //                 false
+    //             };
+    // 
+    //             if is_better {
     //                 best_rate = rate;
+    //                 best_chain_len = try_base.0.chain_len;
     //                 best_base_offset = self.inner_offset - try_base.1;
-    //                 best_base = Some(try_base);
-    //             }
-    //             Some(best_base_ref) => {
-    //                 let is_better = if rate > best_rate + tie_epsilon {
-    //                     true
-    //                 } else if (rate - best_rate).abs() <= tie_epsilon {
-    //
-    //                     if try_base.0.chain_len != best_base_ref.0.chain_len {
-    //                         try_base.0.chain_len > best_base_ref.0.chain_len
-    //                     } else {
-    //                         let base_offset = self.inner_offset - try_base.1;
-    //                         let best_offset = self.inner_offset - best_base_ref.1;
-    //                         base_offset > best_offset
-    //                     }
-    //                 } else {
-    //                     false
-    //                 };
-    //
-    //                 if is_better {
-    //                     best_rate = rate;
-    //                     best_base_offset = self.inner_offset - try_base.1;
-    //                     best_base = Some(try_base);
-    //                 }
+    //                 best_delta_obj = Some((try_delta_obj, try_base.1));
     //             }
     //         }
     //     }
-    //
-    //
-    //     best_base.map(|best_base| {
-    //             let delta = delta::encode(&best_base.0.data, &entry.data);
-    //
-    //             let offset = self.inner_offset - best_base.1;
-    //             entry.obj_type = ObjectType::OffsetDelta;
-    //             entry.data = delta;
-    //             entry.chain_len = best_base.0.chain_len + 1;
-    //             //tracing::debug!("chain_len: {:?}", entry.chain_len);
-    //             offset
-    //         })
-    //
+    //     best_delta_obj.map(|(best_delta_data, best_base_offset)| {
+    //         let offset = self.inner_offset - best_base_offset;
+    //         entry.obj_type = ObjectType::OffsetZstdelta;
+    //         entry.data = best_delta_data;
+    //         entry.chain_len = best_chain_len + 1;
+    //         // tracing::debug!("use base {:?} to delta, rate {:?}", offset, best_rate);
+    //         offset
+    //     })
     // }
 
-    /// Try to encode as sapling-zstdelta using objects in window
-    /// Refs: https://sapling-scm.com/docs/dev/internals/zstdelta/
-    /// # Returns
-    /// - Return (offset) if success make delta
-    /// - Return (None) if didn't delta,
-    fn try_as_offset_zstdelta(&self, entry: &mut Entry) -> Option<usize> {
-        let mut best_delta_obj = None; // obj, offset
-        let mut best_rate: f64 = 0.0;
-        let mut best_chain_len: usize = 0;
-        let mut best_base_offset: usize = 0;
-        let tie_epsilon: f64 = 0.02; // when rates are close, prefer deeper/newer base
-
-        if entry.data.is_empty() {
-            return None;
-        }
-
-        for try_base in self.window.iter() {
-            if try_base.0.obj_type != entry.obj_type {
-                continue;
-            }
-            if try_base.0.chain_len >= MAX_ZSTDELTA_CHAIN_LEN {
-                continue;
-            }
-
-            // 尺寸差异过大时直接跳过
-            let size_a = try_base.0.data.len();
-            let size_b = entry.data.len();
-            if size_a * 10 < size_b || size_b * 10 < size_a {
-                continue;
-            }
-
-            // 轻量 hash 预判（例如比较前 64 字节的 hash）
-            if !cheap_similar(&try_base.0.data, &entry.data) {
-                continue;
-            }
-
-            let try_delta_obj = match zstdelta::diff(&try_base.0.data, &entry.data) {
-                Ok(v) => v,
-                Err(_) => continue, // IO error for this base, try next
-            };
-            // The DeltaDiff algorithm uses the proportion of the delta part as the delta rate,
-            // while zstdelta does not have this data and directly uses the overall compression rate.
-            let rate = 1.0 - try_delta_obj.len() as f64 / entry.data.len() as f64;
-            // if rate > MIN_DELTA_RATE && rate > best_rate {
-            //     best_rate = rate;
-            //     best_delta_obj = Some((try_delta_obj, try_base.1));
-            // }
-
-            if rate > MIN_DELTA_RATE {
-                let is_better = if rate > best_rate + tie_epsilon {
-                    true
-                } else if (rate - best_rate).abs() <= tie_epsilon {
-                    let base_chain = try_base.0.chain_len;
-                    let base_offset = self.inner_offset - try_base.1;
-                    base_chain > best_chain_len
-                        || (base_chain == best_chain_len && base_offset > best_base_offset)
-                } else {
-                    false
-                };
-
-                if is_better {
-                    best_rate = rate;
-                    best_chain_len = try_base.0.chain_len;
-                    best_base_offset = self.inner_offset - try_base.1;
-                    best_delta_obj = Some((try_delta_obj, try_base.1));
-                }
-            }
-        }
-        best_delta_obj.map(|(best_delta_data, best_base_offset)| {
-            let offset = self.inner_offset - best_base_offset;
-            entry.obj_type = ObjectType::OffsetZstdelta;
-            entry.data = best_delta_data;
-            entry.chain_len = best_chain_len + 1;
-            // tracing::debug!("use base {:?} to delta, rate {:?}", offset, best_rate);
-            offset
-        })
-    }
-
     /// Write data to writer and update hash & offset
     async fn write_all_and_update(&mut self, data: &[u8]) {
         self.inner_hash.update(data);
@@ -700,13 +597,7 @@ impl PackEncoder {
         }))
     }
 
-    #[cfg(test)]
-    pub fn get_window_info(&self) -> Vec<(ObjectType, usize, usize)> {
-        self.window
-            .iter()
-            .map(|(entry, offset)| (entry.obj_type, entry.chain_len, *offset))
-            .collect()
-    }
+   
 }
 
 #[cfg(test)]
@@ -837,7 +728,7 @@ mod tests {
 
     async fn get_entries_for_test() -> Arc<Mutex<Vec<Entry>>> {
         let mut source = PathBuf::from(env::current_dir().unwrap().parent().unwrap());
-        source.push("tests/data/packs/pack-f8bbb573cef7d851957caceb491c073ee8e8de41.pack");
+        source.push("tests/data/packs/pack-a3b9e4e7e3a73634813f2b7f0239a9c19c914cd2.pack");
         // let file_map = crate::test_utils::setup_lfs_file().await;
         // let source = file_mapa
         //     .get("git-2d187177923cd618a75da6c6db45bb89d92bd504.pack")

From 380081027a17f1b36cccbf8396849f0bc38b57d0 Mon Sep 17 00:00:00 2001
From: pleasedontbreak123 <nabeiyq@126.com>
Date: Wed, 3 Sep 2025 16:50:29 +0800
Subject: [PATCH 12/17] complete encode

---
 mercury/delta/src/lib.rs            |  34 ++--
 mercury/src/internal/pack/decode.rs |  14 +-
 mercury/src/internal/pack/encode.rs | 248 ++++++----------------------
 mercury/src/utils.rs                |  41 ++++-
 4 files changed, 116 insertions(+), 221 deletions(-)

diff --git a/mercury/delta/src/lib.rs b/mercury/delta/src/lib.rs
index f7161d499..7c33e5293 100644
--- a/mercury/delta/src/lib.rs
+++ b/mercury/delta/src/lib.rs
@@ -9,29 +9,31 @@ mod utils;
 
 pub use decode::delta_decode as decode;
 
-const SAMPLE_STEP: usize = 64;      // 每隔 64 字节取样
-const MIN_DELTA_RATE: f64 = 0.5;    // 最小 delta rate
+const SAMPLE_STEP: usize = 64;     
+const MIN_DELTA_RATE: f64 = 0.5;    
 
-/// 计算两个对象的 delta rate（近似）
+/// approximate calculation delta rate to enhance efficiency
+#[allow(clippy::manual_div_ceil)]
 pub fn heuristic_encode_rate(old_data: &[u8], new_data: &[u8]) -> f64 {
     let old_len = old_data.len();
     let new_len = new_data.len();
 
     if old_len == 0 && new_len == 0 {
-        return 1.0; // 两个空 slice 完全匹配
+        return 1.0; 
     }
     if old_len == 0 || new_len == 0 {
-        return 0.0; // 一个空一个非空，匹配率为 0
+        return 0.0; 
     }
 
     let step = SAMPLE_STEP;
     let mut match_count = 0;
     let mut sample_count = 0;
 
-    // 对两个对象取样，计算 hash 匹配数
+    
     let min_len = old_len.min(new_len);
 
-    let total_samples = (min_len + step - 1) / step;
+   
+    let total_samples =  (min_len + step - 1) / step;
     let mut i = 0;
     while i < min_len {
         let old_chunk = &old_data[i..(i + step).min(old_len)];
@@ -53,7 +55,7 @@ pub fn heuristic_encode_rate(old_data: &[u8], new_data: &[u8]) -> f64 {
         i += step;
     }
 
-    // 返回匹配率
+    
     match_count as f64 / sample_count as f64
 }
 
@@ -110,52 +112,52 @@ mod tests {
 
     #[test]
     fn test_heuristic_encode_rate() {
-        // 完全相同的数据
+        
         let data1 = b"hello world, this is a test for delta rate";
         let data2 = b"hello world, this is a test for delta rate";
         let rate = heuristic_encode_rate(data1, data2);
         println!("rate = {}", rate);
         assert!((rate - 1.0).abs() < 1e-6, "Expected 1.0 for identical data");
 
-        // 部分相似的数据
+        
         let data3 = b"hello world, this is a test for delta rate";
         let data4 = b"hello worll, this is a test for delta rate";
         let rate = encode_rate(data3, data4);
         println!("rate = {}", rate);
         assert!(rate > 0.5 && rate < 1.0, "Expected partial match for similar data");
 
-        // 完全不同的数据
+       
         let data5 = b"abcdefghijklmno";
         let data6 = b"1234567890!@#";
         let rate = heuristic_encode_rate(data5, data6);
         println!("rate = {}", rate);
         assert!(rate < 0.2, "Expected low match rate for different data");
 
-        // 空数据
+        
         let data7 = b"";
         let data8 = b"";
         let rate = heuristic_encode_rate(data7, data8);
         println!("rate = {}", rate);
         assert_eq!(rate, 1.0, "Empty slices should be fully matching");
 
-        // 一个空一个非空
+        
         let rate = heuristic_encode_rate(data7, data1);
         assert_eq!(rate, 0.0, "Empty vs non-empty should give 0 rate");
     }
 
     #[test]
     fn test_heuristic_encode_rate_large_files() {
-        // 大文件完全不匹配 → 触发早停
+       
         let data1 = vec![0u8; 1_000_00];
         let data2 = vec![1u8; 1_000_00];
         let rate = heuristic_encode_rate(&data1, &data2);
         println!("Large non-matching data rate = {}", rate);
         assert_eq!(rate, 0.0, "Large completely different data should early stop with 0 rate");
 
-        // 大文件部分匹配 → 保留部分匹配
+        
         let  data3 = vec![0u8; 100];
         let mut data4 = vec![0u8; 100];
-        // 在一部分修改一些数据
+       
         for i in 0..2 {
             data4[i] = 1;
         }
diff --git a/mercury/src/internal/pack/decode.rs b/mercury/src/internal/pack/decode.rs
index 4b25feb5d..464ec3fb5 100644
--- a/mercury/src/internal/pack/decode.rs
+++ b/mercury/src/internal/pack/decode.rs
@@ -17,6 +17,7 @@ use crate::errors::GitError;
 use crate::hash::SHA1;
 use crate::internal::object::types::ObjectType;
 
+use super::cache_object::CacheObjectInfo;
 use crate::internal::pack::cache::Caches;
 use crate::internal::pack::cache::_Cache;
 use crate::internal::pack::cache_object::{CacheObject, MemSizeRecorder};
@@ -25,8 +26,7 @@ use crate::internal::pack::entry::Entry;
 use crate::internal::pack::waitlist::Waitlist;
 use crate::internal::pack::wrapper::Wrapper;
 use crate::internal::pack::{utils, Pack, DEFAULT_TMP_DIR};
-
-use super::cache_object::CacheObjectInfo;
+use crate::utils::CountingReader;
 
 /// For the convenience of passing parameters
 struct SharedParams {
@@ -204,9 +204,11 @@ impl Pack {
     ) -> Result<(Vec<u8>, usize), GitError> {
         // Create a buffer with the expected size for the decompressed data
         let mut buf = Vec::with_capacity(expected_size);
-        // Create a new Zlib decoder with the original data
-        let mut deflate = ZlibDecoder::new(pack);
 
+        let mut counting_reader = CountingReader::new(pack);
+        // Create a new Zlib decoder with the original data
+        //let mut deflate = ZlibDecoder::new(pack);
+        let mut deflate = ZlibDecoder::new(&mut counting_reader);
         // Attempt to read data to the end of the buffer
         match deflate.read_to_end(&mut buf) {
             Ok(_) => {
@@ -219,8 +221,8 @@ impl Pack {
                     )))
                 } else {
                     // If everything is as expected, return the buffer, the original data, and the total number of input bytes processed
-                    Ok((buf, deflate.total_in() as usize))
-                    // TODO this will likely be smaller than what the decompressor actually read from the underlying stream due to buffering.
+                    let actual_input_bytes = counting_reader.bytes_read as usize;
+                    Ok((buf, actual_input_bytes))
                 }
             }
             Err(e) => {
diff --git a/mercury/src/internal/pack/encode.rs b/mercury/src/internal/pack/encode.rs
index a5653d752..6385ef314 100644
--- a/mercury/src/internal/pack/encode.rs
+++ b/mercury/src/internal/pack/encode.rs
@@ -15,14 +15,14 @@ use tokio::task::JoinHandle;
 
 const MAX_CHAIN_LEN: usize = 50;
 const MIN_DELTA_RATE: f64 = 0.5; // minimum delta rate
-//const MAX_ZSTDELTA_CHAIN_LEN: usize = 50;
+                                 //const MAX_ZSTDELTA_CHAIN_LEN: usize = 50;
 
 /// A encoder for generating pack files with delta objects.
 pub struct PackEncoder {
     object_number: usize,
     process_index: usize,
     window_size: usize,
-   // window: VecDeque<(Entry, usize)>, // entry and offset
+    // window: VecDeque<(Entry, usize)>, // entry and offset
     sender: Option<mpsc::Sender<Vec<u8>>>,
     inner_offset: usize, // offset of current entry
     inner_hash: Sha1,    // Not SHA1 because need update trait
@@ -150,7 +150,7 @@ impl PackEncoder {
             object_number,
             window_size,
             process_index: 0,
-           // window: VecDeque::with_capacity(window_size),
+            // window: VecDeque::with_capacity(window_size),
             sender: Some(sender),
             inner_offset: 12, // 12 bytes header
             inner_hash: Sha1::new(),
@@ -182,7 +182,7 @@ impl PackEncoder {
     /// - Returns a `GitError` if there is a failure during the encoding process.
     /// - Returns `PackEncodeError` if an encoding operation is already in progress.
     pub async fn encode(&mut self, entry_rx: mpsc::Receiver<Entry>) -> Result<(), GitError> {
-         self.inner_encode(entry_rx, false).await
+        self.inner_encode(entry_rx, false).await
     }
 
     pub async fn encode_with_zstdelta(
@@ -246,23 +246,31 @@ impl PackEncoder {
 
         // parallel encoding vec with different object_type
         let (commit_results, tree_results, blob_results, tag_results) = tokio::try_join!(
-            tokio::task::spawn_blocking(move || { Self::try_as_offset_delta(commits, 10,enable_zstdelta) }),
-            tokio::task::spawn_blocking(move || { Self::try_as_offset_delta(trees, 10,enable_zstdelta) }),
-            tokio::task::spawn_blocking(move || { Self::try_as_offset_delta(blobs, 10,enable_zstdelta) }),
-            tokio::task::spawn_blocking(move || { Self::try_as_offset_delta(tags, 10,enable_zstdelta) }),
+            tokio::task::spawn_blocking(move || {
+                Self::try_as_offset_delta(commits, 10, enable_zstdelta)
+            }),
+            tokio::task::spawn_blocking(move || {
+                Self::try_as_offset_delta(trees, 10, enable_zstdelta)
+            }),
+            tokio::task::spawn_blocking(move || {
+                Self::try_as_offset_delta(blobs, 10, enable_zstdelta)
+            }),
+            tokio::task::spawn_blocking(move || {
+                Self::try_as_offset_delta(tags, 10, enable_zstdelta)
+            }),
         )
-        .map_err(|e| GitError::PackEncodeError(format!("Task join error: {}", e)))?;
+        .map_err(|e| GitError::PackEncodeError(format!("Task join error: {e}")))?;
 
         // 收集并合并结果
         let all_encoded_data = [
             commit_results
-                .map_err(|e| GitError::PackEncodeError(format!("Commit encoding error: {}", e)))?,
+                .map_err(|e| GitError::PackEncodeError(format!("Commit encoding error: {e}")))?,
             tree_results
-                .map_err(|e| GitError::PackEncodeError(format!("Tree encoding error: {}", e)))?,
+                .map_err(|e| GitError::PackEncodeError(format!("Tree encoding error: {e}")))?,
             blob_results
-                .map_err(|e| GitError::PackEncodeError(format!("Blob encoding error: {}", e)))?,
+                .map_err(|e| GitError::PackEncodeError(format!("Blob encoding error: {e}")))?,
             tag_results
-                .map_err(|e| GitError::PackEncodeError(format!("Tag encoding error: {}", e)))?,
+                .map_err(|e| GitError::PackEncodeError(format!("Tag encoding error: {e}")))?,
         ]
         .concat();
 
@@ -281,7 +289,7 @@ impl PackEncoder {
     }
 
     /// Try to encode as delta using objects in window
-    /// delta & zstdelta have been gathered here 
+    /// delta & zstdelta have been gathered here
     /// Refs: https://sapling-scm.com/docs/dev/internals/zstdelta/
     /// the sliding window was moved here
     /// # Returns
@@ -292,25 +300,21 @@ impl PackEncoder {
         window_size: usize,
         enable_zstdelta: bool,
     ) -> Result<Vec<Vec<u8>>, GitError> {
-        let mut current_offset = 0usize; 
+        let mut current_offset = 0usize;
         let mut window: VecDeque<(Entry, usize)> = VecDeque::with_capacity(window_size);
         let mut res: Vec<Vec<u8>> = Vec::new();
 
         for entry in bucket.iter_mut() {
-            
             //let entry_for_window = entry.clone();
             // 每次循环重置最佳基对象选择
             let mut best_base: Option<&(Entry, usize)> = None;
             let mut best_rate: f64 = 0.0;
             let tie_epsilon: f64 = 0.15;
 
-            
             let candidates: Vec<_> = window
                 .par_iter()
                 .with_min_len(3)
                 .filter_map(|try_base| {
-                    
-                    
                     if try_base.0.obj_type != entry.obj_type {
                         return None;
                     }
@@ -333,15 +337,13 @@ impl PackEncoder {
                         return None;
                     }
 
-
-                    let rate = 
-                        if (try_base.0.data.len() + entry.data.len()) / 2 > 20 {
-                            delta::heuristic_encode_rate_parallel(&try_base.0.data, &entry.data)
-                        } else {
-                            delta::encode_rate(&try_base.0.data, &entry.data)
-                        };
-                    
-
+                    let rate = if (try_base.0.data.len() + entry.data.len()) / 2 > 64 {
+                        delta::heuristic_encode_rate_parallel(&try_base.0.data, &entry.data)
+                    } else {
+                        delta::encode_rate(&try_base.0.data, &entry.data)
+                        // let try_delta_obj = zstdelta::diff(&try_base.0.data, &entry.data).unwrap();
+                        // 1.0 - try_delta_obj.len() as f64 / entry.data.len() as f64
+                    };
 
                     if rate > MIN_DELTA_RATE {
                         Some((rate, try_base))
@@ -351,12 +353,11 @@ impl PackEncoder {
                 })
                 .collect();
 
-           
             for (rate, try_base) in candidates {
                 match best_base {
                     None => {
                         best_rate = rate;
-                        //best_base_offset = current_offset - try_base.1; 
+                        //best_base_offset = current_offset - try_base.1;
                         best_base = Some(try_base);
                     }
                     Some(best_base_ref) => {
@@ -375,42 +376,39 @@ impl PackEncoder {
                     }
                 }
             }
-            
+
             let mut entry_for_window = entry.clone();
 
-            let offset = best_base.map(|best_base| { 
-                let delta = if enable_zstdelta{
+            let offset = best_base.map(|best_base| {
+                let delta = if enable_zstdelta {
                     entry.obj_type = ObjectType::OffsetZstdelta;
-                    zstdelta::diff(&best_base.0.data,&entry.data)
-                        .map_err(|e| GitError::DeltaObjectError(format!("zstdelta diff failed: {}", e))).unwrap()
-                }else {
+                    zstdelta::diff(&best_base.0.data, &entry.data)
+                        .map_err(|e| {
+                            GitError::DeltaObjectError(format!("zstdelta diff failed: {e}"))
+                        })
+                        .unwrap()
+                } else {
                     entry.obj_type = ObjectType::OffsetDelta;
                     delta::encode(&best_base.0.data, &entry.data)
                 };
-                //entry.obj_type = ObjectType::OffsetDelta; 
-                entry.data = delta; 
+                //entry.obj_type = ObjectType::OffsetDelta;
+                entry.data = delta;
                 entry.chain_len = best_base.0.chain_len + 1;
-                let offset = current_offset - best_base.1;
-                offset
-            }); 
-            
+                current_offset - best_base.1
+            });
+
             entry_for_window.chain_len = entry.chain_len;
-            let obj_data = encode_one_object(&entry, offset)?;
+            let obj_data = encode_one_object(entry, offset)?;
             window.push_back((entry_for_window, current_offset));
             if window.len() > window_size {
                 window.pop_front();
             }
-            current_offset = current_offset + obj_data.len();
+            current_offset += obj_data.len();
             res.push(obj_data);
-            
-            
         }
         Ok(res)
     }
 
-
-   
-
     /// Parallel encode with rayon, only works when window_size == 0 (no delta)
     pub async fn parallel_encode(
         &mut self,
@@ -483,85 +481,6 @@ impl PackEncoder {
         Ok(())
     }
 
-    
-    /// Try to encode as sapling-zstdelta using objects in window
-    /// Refs: https://sapling-scm.com/docs/dev/internals/zstdelta/
-    /// # Returns
-    /// - Return (offset) if success make delta
-    /// - Return (None) if didn't delta,
-    // fn try_as_offset_zstdelta(&self, entry: &mut Entry) -> Option<usize> {
-    //     let mut best_delta_obj = None; // obj, offset
-    //     let mut best_rate: f64 = 0.0;
-    //     let mut best_chain_len: usize = 0;
-    //     let mut best_base_offset: usize = 0;
-    //     let tie_epsilon: f64 = 0.02; // when rates are close, prefer deeper/newer base
-    // 
-    //     if entry.data.is_empty() {
-    //         return None;
-    //     }
-    // 
-    //     for try_base in self.window.iter() {
-    //         if try_base.0.obj_type != entry.obj_type {
-    //             continue;
-    //         }
-    //         if try_base.0.chain_len >= MAX_ZSTDELTA_CHAIN_LEN {
-    //             continue;
-    //         }
-    // 
-    //         // 尺寸差异过大时直接跳过
-    //         let size_a = try_base.0.data.len();
-    //         let size_b = entry.data.len();
-    //         if size_a * 10 < size_b || size_b * 10 < size_a {
-    //             continue;
-    //         }
-    // 
-    //         // 轻量 hash 预判（例如比较前 64 字节的 hash）
-    //         if !cheap_similar(&try_base.0.data, &entry.data) {
-    //             continue;
-    //         }
-    // 
-    //         let try_delta_obj = match zstdelta::diff(&try_base.0.data, &entry.data) {
-    //             Ok(v) => v,
-    //             Err(_) => continue, // IO error for this base, try next
-    //         };
-    //         // The DeltaDiff algorithm uses the proportion of the delta part as the delta rate,
-    //         // while zstdelta does not have this data and directly uses the overall compression rate.
-    //         let rate = 1.0 - try_delta_obj.len() as f64 / entry.data.len() as f64;
-    //         // if rate > MIN_DELTA_RATE && rate > best_rate {
-    //         //     best_rate = rate;
-    //         //     best_delta_obj = Some((try_delta_obj, try_base.1));
-    //         // }
-    // 
-    //         if rate > MIN_DELTA_RATE {
-    //             let is_better = if rate > best_rate + tie_epsilon {
-    //                 true
-    //             } else if (rate - best_rate).abs() <= tie_epsilon {
-    //                 let base_chain = try_base.0.chain_len;
-    //                 let base_offset = self.inner_offset - try_base.1;
-    //                 base_chain > best_chain_len
-    //                     || (base_chain == best_chain_len && base_offset > best_base_offset)
-    //             } else {
-    //                 false
-    //             };
-    // 
-    //             if is_better {
-    //                 best_rate = rate;
-    //                 best_chain_len = try_base.0.chain_len;
-    //                 best_base_offset = self.inner_offset - try_base.1;
-    //                 best_delta_obj = Some((try_delta_obj, try_base.1));
-    //             }
-    //         }
-    //     }
-    //     best_delta_obj.map(|(best_delta_data, best_base_offset)| {
-    //         let offset = self.inner_offset - best_base_offset;
-    //         entry.obj_type = ObjectType::OffsetZstdelta;
-    //         entry.data = best_delta_data;
-    //         entry.chain_len = best_chain_len + 1;
-    //         // tracing::debug!("use base {:?} to delta, rate {:?}", offset, best_rate);
-    //         offset
-    //     })
-    // }
-
     /// Write data to writer and update hash & offset
     async fn write_all_and_update(&mut self, data: &[u8]) {
         self.inner_hash.update(data);
@@ -596,8 +515,6 @@ impl PackEncoder {
             self.encode_with_zstdelta(rx).await.unwrap()
         }))
     }
-
-   
 }
 
 #[cfg(test)]
@@ -636,7 +553,7 @@ mod tests {
             let (entry_tx, entry_rx) = mpsc::channel::<Entry>(1);
 
             // make some different objects, or decode will fail
-            let str_vec = vec!["hello, code,", "hello, world.", "!", "123141251251"];
+            let str_vec = vec!["hello, word", "hello, world.", "!", "123141251251"];
             let encoder = PackEncoder::new(str_vec.len(), window_size, tx);
             encoder.encode_async(entry_rx).await.unwrap();
 
@@ -660,75 +577,14 @@ mod tests {
         check_format(&pack_without_delta);
 
         // with delta
-        let pack_with_delta = encode_once(3).await;
-        assert_ne!(pack_with_delta.len(), pack_without_delta_size);
+        let pack_with_delta = encode_once(4).await;
+        assert!(pack_with_delta.len() <= pack_without_delta_size);
         check_format(&pack_with_delta);
     }
 
-    #[tokio::test]
-    async fn test_encode_then_decode_delta_chain() {
-        init_logger();
-
-        // 1) 构造一组相似的内容，鼓励形成 delta 链
-        let mut contents: Vec<Vec<u8>> = Vec::new();
-        let mut base = String::from("base line 0\n");
-        for i in 1..=8usize {
-            base.push_str(&format!(
-                "line {i}: some repeated text with slight changes...\n"
-            ));
-            contents.push(base.as_bytes().to_vec());
-        }
-
-        // 2) 用 PackEncoder（window_size>0）编码，尽量产生 offset-delta
-        let (tx, mut rx) = mpsc::channel(10_000);
-        let (entry_tx, entry_rx) = mpsc::channel::<Entry>(10_000);
-
-        let encoder = PackEncoder::new(contents.len(), 10, tx);
-        encoder.encode_async_with_zstdelta(entry_rx).await.unwrap();
-
-        for data in contents.clone() {
-            let blob = Blob::from_content_bytes(data);
-            let entry: Entry = blob.into();
-            entry_tx.send(entry).await.unwrap();
-        }
-        drop(entry_tx);
-
-        let mut pack_bytes = Vec::new();
-        while let Some(chunk) = rx.recv().await {
-            pack_bytes.extend(chunk);
-        }
-
-        // 3) 用 Pack::decode 解析，重建为完整对象
-        let mut p = Pack::new(
-            None,
-            Some(1024 * 1024 * 20),
-            Some(PathBuf::from("/tmp/.cache_temp")),
-            true,
-        );
-        let mut reader = Cursor::new(pack_bytes);
-        let decoded = Arc::new(tokio::sync::Mutex::new(Vec::<Entry>::new()));
-        let decoded_c = decoded.clone();
-        p.decode(&mut reader, move |entry, _| {
-            let mut guard = decoded_c.blocking_lock();
-            guard.push(entry);
-        })
-        .expect("pack decode should succeed");
-
-        let decoded_entries = decoded.lock().await;
-        assert_eq!(decoded_entries.len(), contents.len());
-
-        // 4) 校验内容集合相等（忽略顺序）
-        let expected: HashSet<Vec<u8>> = contents.into_iter().collect();
-        let got: HashSet<Vec<u8>> = decoded_entries.iter().map(|e| e.data.clone()).collect();
-        assert_eq!(
-            expected, got,
-            "decoded contents should equal original contents"
-        );
-    }
-
     async fn get_entries_for_test() -> Arc<Mutex<Vec<Entry>>> {
         let mut source = PathBuf::from(env::current_dir().unwrap().parent().unwrap());
-        source.push("tests/data/packs/pack-a3b9e4e7e3a73634813f2b7f0239a9c19c914cd2.pack");
+        source.push("tests/data/packs/pack-f8bbb573cef7d851957caceb491c073ee8e8de41.pack");
         // let file_map = crate::test_utils::setup_lfs_file().await;
         // let source = file_mapa
         //     .get("git-2d187177923cd618a75da6c6db45bb89d92bd504.pack")
@@ -830,7 +686,7 @@ mod tests {
         let (tx, mut rx) = mpsc::channel(100_000);
         let (entry_tx, entry_rx) = mpsc::channel::<Entry>(100_000);
 
-        let mut encoder = PackEncoder::new(entries_number, 10, tx);
+        let mut encoder = PackEncoder::new(entries_number, 0, tx);
         tokio::spawn(async move {
             time_it!("test encode no parallel", {
                 encoder.encode(entry_rx).await.unwrap();
diff --git a/mercury/src/utils.rs b/mercury/src/utils.rs
index 17ae9e0c4..3ac7613fc 100644
--- a/mercury/src/utils.rs
+++ b/mercury/src/utils.rs
@@ -1,7 +1,6 @@
-use std::io;
-use std::io::Read;
-
 use crate::hash::SHA1;
+use std::io;
+use std::io::{BufRead, Read};
 
 pub fn read_bytes(file: &mut impl Read, len: usize) -> io::Result<Vec<u8>> {
     let mut buf = vec![0; len];
@@ -12,3 +11,39 @@ pub fn read_bytes(file: &mut impl Read, len: usize) -> io::Result<Vec<u8>> {
 pub fn read_sha1(file: &mut impl Read) -> io::Result<SHA1> {
     SHA1::from_stream(file)
 }
+
+/// A lightweight wrapper that counts bytes read from the underlying reader.
+/// replace deflate.intotal() in decompress_data
+pub struct CountingReader<R> {
+    pub inner: R,
+    pub bytes_read: u64,
+}
+
+impl<R> CountingReader<R> {
+    /// Creates a new `CountingReader` wrapping the given reader.
+    pub fn new(inner: R) -> Self {
+        Self {
+            inner,
+            bytes_read: 0,
+        }
+    }
+}
+
+impl<R: Read> Read for CountingReader<R> {
+    fn read(&mut self, buf: &mut [u8]) -> std::io::Result<usize> {
+        let n = self.inner.read(buf)?;
+        self.bytes_read += n as u64;
+        Ok(n)
+    }
+}
+
+impl<R: BufRead> BufRead for CountingReader<R> {
+    fn fill_buf(&mut self) -> std::io::Result<&[u8]> {
+        self.inner.fill_buf()
+    }
+
+    fn consume(&mut self, amt: usize) {
+        self.bytes_read += amt as u64; // 关键：统计通过缓冲区消耗的字节
+        self.inner.consume(amt);
+    }
+}

From 762e46fde99c8cb0514cca01a8258f48284d9c16 Mon Sep 17 00:00:00 2001
From: pleasedontbreak123 <nabeiyq@126.com>
Date: Wed, 3 Sep 2025 16:53:58 +0800
Subject: [PATCH 13/17] complete encode

---
 mercury/src/internal/pack/encode.rs | 23 +++++++++++------------
 1 file changed, 11 insertions(+), 12 deletions(-)

diff --git a/mercury/src/internal/pack/encode.rs b/mercury/src/internal/pack/encode.rs
index 6385ef314..173a52203 100644
--- a/mercury/src/internal/pack/encode.rs
+++ b/mercury/src/internal/pack/encode.rs
@@ -613,11 +613,11 @@ mod tests {
     async fn test_pack_encoder_parallel_large_file() {
         init_logger();
 
-        let start = Instant::now(); // 开始时间
+        let start = Instant::now();
         let entries = get_entries_for_test().await;
         let entries_number = entries.lock().await.len();
 
-        // 计算原始总大小
+      
         let total_original_size: usize = entries
             .lock()
             .await
@@ -651,7 +651,7 @@ mod tests {
             result.extend(chunk);
         }
 
-        // 计算压缩率
+        
         let pack_size = result.len();
         let compression_rate = if total_original_size > 0 {
             1.0 - (pack_size as f64 / total_original_size as f64)
@@ -681,7 +681,7 @@ mod tests {
             .map(|entry| entry.data.len())
             .sum();
 
-        let start = Instant::now(); // 开始时间
+        let start = Instant::now(); 
                                     // encode entries
         let (tx, mut rx) = mpsc::channel(100_000);
         let (entry_tx, entry_rx) = mpsc::channel::<Entry>(100_000);
@@ -708,13 +708,13 @@ mod tests {
         //     // do nothing
         // }
 
-        // 收集数据并计算压缩率
+       
         let mut result = Vec::new();
         while let Some(chunk) = rx.recv().await {
             result.extend(chunk);
         }
 
-        // 计算压缩率
+       
         let pack_size = result.len();
         let compression_rate = if total_original_size > 0 {
             1.0 - (pack_size as f64 / total_original_size as f64)
@@ -739,7 +739,7 @@ mod tests {
         let entries = get_entries_for_test().await;
         let entries_number = entries.lock().await.len();
 
-        // 计算原始总大小
+        
         let total_original_size: usize = entries
             .lock()
             .await
@@ -764,13 +764,12 @@ mod tests {
             tracing::info!("all entries sent");
         });
 
-        // 收集数据并计算压缩率
+        
         let mut result = Vec::new();
         while let Some(chunk) = rx.recv().await {
             result.extend(chunk);
         }
 
-        // 计算压缩率
         let pack_size = result.len();
         let compression_rate = if total_original_size > 0 {
             1.0 - (pack_size as f64 / total_original_size as f64)
@@ -811,7 +810,7 @@ mod tests {
         let entries = get_entries_for_test().await;
         let entries_number = entries.lock().await.len();
 
-        // 计算原始总大小
+       
         let total_original_size: usize = entries
             .lock()
             .await
@@ -837,13 +836,13 @@ mod tests {
             tracing::info!("all entries sent");
         });
 
-        // 收集数据并计算压缩率
+        
         let mut result = Vec::new();
         while let Some(chunk) = rx.recv().await {
             result.extend(chunk);
         }
 
-        // 计算压缩率
+        
         let pack_size = result.len();
         let compression_rate = if total_original_size > 0 {
             1.0 - (pack_size as f64 / total_original_size as f64)

From 86255c2eac80632a457e330a1ca5dfb74381d033 Mon Sep 17 00:00:00 2001
From: pleasedontbreak123 <nabeiyq@126.com>
Date: Wed, 3 Sep 2025 16:55:06 +0800
Subject: [PATCH 14/17] fmt

---
 mercury/src/internal/pack/encode.rs | 13 ++-----------
 1 file changed, 2 insertions(+), 11 deletions(-)

diff --git a/mercury/src/internal/pack/encode.rs b/mercury/src/internal/pack/encode.rs
index 173a52203..71eac5f89 100644
--- a/mercury/src/internal/pack/encode.rs
+++ b/mercury/src/internal/pack/encode.rs
@@ -617,7 +617,6 @@ mod tests {
         let entries = get_entries_for_test().await;
         let entries_number = entries.lock().await.len();
 
-      
         let total_original_size: usize = entries
             .lock()
             .await
@@ -651,7 +650,6 @@ mod tests {
             result.extend(chunk);
         }
 
-        
         let pack_size = result.len();
         let compression_rate = if total_original_size > 0 {
             1.0 - (pack_size as f64 / total_original_size as f64)
@@ -681,8 +679,8 @@ mod tests {
             .map(|entry| entry.data.len())
             .sum();
 
-        let start = Instant::now(); 
-                                    // encode entries
+        let start = Instant::now();
+        // encode entries
         let (tx, mut rx) = mpsc::channel(100_000);
         let (entry_tx, entry_rx) = mpsc::channel::<Entry>(100_000);
 
@@ -708,13 +706,11 @@ mod tests {
         //     // do nothing
         // }
 
-       
         let mut result = Vec::new();
         while let Some(chunk) = rx.recv().await {
             result.extend(chunk);
         }
 
-       
         let pack_size = result.len();
         let compression_rate = if total_original_size > 0 {
             1.0 - (pack_size as f64 / total_original_size as f64)
@@ -739,7 +735,6 @@ mod tests {
         let entries = get_entries_for_test().await;
         let entries_number = entries.lock().await.len();
 
-        
         let total_original_size: usize = entries
             .lock()
             .await
@@ -764,7 +759,6 @@ mod tests {
             tracing::info!("all entries sent");
         });
 
-        
         let mut result = Vec::new();
         while let Some(chunk) = rx.recv().await {
             result.extend(chunk);
@@ -810,7 +804,6 @@ mod tests {
         let entries = get_entries_for_test().await;
         let entries_number = entries.lock().await.len();
 
-       
         let total_original_size: usize = entries
             .lock()
             .await
@@ -836,13 +829,11 @@ mod tests {
             tracing::info!("all entries sent");
         });
 
-        
         let mut result = Vec::new();
         while let Some(chunk) = rx.recv().await {
             result.extend(chunk);
         }
 
-        
         let pack_size = result.len();
         let compression_rate = if total_original_size > 0 {
             1.0 - (pack_size as f64 / total_original_size as f64)

From a8a130348cd5be2d5fa9b89729074ad69e6c74b3 Mon Sep 17 00:00:00 2001
From: pleasedontbreak123 <nabeiyq@126.com>
Date: Wed, 3 Sep 2025 17:30:20 +0800
Subject: [PATCH 15/17] fix clippy

---
 mercury/src/internal/pack/encode.rs | 5 ++---
 mercury/src/utils.rs                | 4 ++--
 2 files changed, 4 insertions(+), 5 deletions(-)

diff --git a/mercury/src/internal/pack/encode.rs b/mercury/src/internal/pack/encode.rs
index 71eac5f89..72190592a 100644
--- a/mercury/src/internal/pack/encode.rs
+++ b/mercury/src/internal/pack/encode.rs
@@ -189,7 +189,7 @@ impl PackEncoder {
         &mut self,
         entry_rx: mpsc::Receiver<Entry>,
     ) -> Result<(), GitError> {
-        return self.inner_encode(entry_rx, true).await;
+        self.inner_encode(entry_rx, true).await
     }
 
     /// Delta selection heuristics are based on:
@@ -476,7 +476,7 @@ impl PackEncoder {
         // hash signature
         let hash_result = self.inner_hash.clone().finalize();
         self.final_hash = Some(SHA1::from_bytes(&hash_result));
-        self.send_data((hash_result).to_vec()).await;
+        self.send_data(hash_result.to_vec()).await;
         self.drop_sender();
         Ok(())
     }
@@ -519,7 +519,6 @@ impl PackEncoder {
 
 #[cfg(test)]
 mod tests {
-    use std::collections::HashSet;
     use std::env;
     use std::sync::Arc;
     use std::time::Instant;
diff --git a/mercury/src/utils.rs b/mercury/src/utils.rs
index 3ac7613fc..bba34df8d 100644
--- a/mercury/src/utils.rs
+++ b/mercury/src/utils.rs
@@ -30,7 +30,7 @@ impl<R> CountingReader<R> {
 }
 
 impl<R: Read> Read for CountingReader<R> {
-    fn read(&mut self, buf: &mut [u8]) -> std::io::Result<usize> {
+    fn read(&mut self, buf: &mut [u8]) -> io::Result<usize> {
         let n = self.inner.read(buf)?;
         self.bytes_read += n as u64;
         Ok(n)
@@ -38,7 +38,7 @@ impl<R: Read> Read for CountingReader<R> {
 }
 
 impl<R: BufRead> BufRead for CountingReader<R> {
-    fn fill_buf(&mut self) -> std::io::Result<&[u8]> {
+    fn fill_buf(&mut self) -> io::Result<&[u8]> {
         self.inner.fill_buf()
     }
 

From 30ec8795e48ef6a204de1fcb9f004bf12dc6d510 Mon Sep 17 00:00:00 2001
From: pleasedontbreak123 <nabeiyq@126.com>
Date: Wed, 3 Sep 2025 22:01:08 +0800
Subject: [PATCH 16/17] fix code annotation

---
 mercury/delta/src/lib.rs            | 4 ++--
 mercury/src/internal/pack/encode.rs | 4 ++--
 mercury/src/utils.rs                | 2 +-
 3 files changed, 5 insertions(+), 5 deletions(-)

diff --git a/mercury/delta/src/lib.rs b/mercury/delta/src/lib.rs
index 7c33e5293..10db4acce 100644
--- a/mercury/delta/src/lib.rs
+++ b/mercury/delta/src/lib.rs
@@ -44,8 +44,8 @@ pub fn heuristic_encode_rate(old_data: &[u8], new_data: &[u8]) -> f64 {
         }
         sample_count += 1;
 
-        // 早停判断：
-        // 剩余样本即使全匹配也达不到 MIN_DELTA_RATE，则提前返回 0
+        // Early stopping condition:
+        // If all remaining samples matched, the rate would still not reach MIN_DELTA_RATE, so return 0 early
         let remaining_samples = total_samples - sample_count;
         let max_possible_rate = (match_count + remaining_samples) as f64 / total_samples as f64;
         if max_possible_rate < MIN_DELTA_RATE {
diff --git a/mercury/src/internal/pack/encode.rs b/mercury/src/internal/pack/encode.rs
index 72190592a..88e8dae56 100644
--- a/mercury/src/internal/pack/encode.rs
+++ b/mercury/src/internal/pack/encode.rs
@@ -261,7 +261,7 @@ impl PackEncoder {
         )
         .map_err(|e| GitError::PackEncodeError(format!("Task join error: {e}")))?;
 
-        // 收集并合并结果
+        
         let all_encoded_data = [
             commit_results
                 .map_err(|e| GitError::PackEncodeError(format!("Commit encoding error: {e}")))?,
@@ -670,7 +670,7 @@ mod tests {
         let entries = get_entries_for_test().await;
         let entries_number = entries.lock().await.len();
 
-        // 计算原始总大小
+       
         let total_original_size: usize = entries
             .lock()
             .await
diff --git a/mercury/src/utils.rs b/mercury/src/utils.rs
index bba34df8d..d80577cc7 100644
--- a/mercury/src/utils.rs
+++ b/mercury/src/utils.rs
@@ -43,7 +43,7 @@ impl<R: BufRead> BufRead for CountingReader<R> {
     }
 
     fn consume(&mut self, amt: usize) {
-        self.bytes_read += amt as u64; // 关键：统计通过缓冲区消耗的字节
+        self.bytes_read += amt as u64; 
         self.inner.consume(amt);
     }
 }

From 8cf64e01a4edb6d79118afdfd53913ce7457b18c Mon Sep 17 00:00:00 2001
From: pleasedontbreak123 <nabeiyq@126.com>
Date: Thu, 4 Sep 2025 12:03:48 +0800
Subject: [PATCH 17/17] fix code annotation

---
 mercury/delta/src/lib.rs            | 4 ++--
 mercury/src/internal/pack/encode.rs | 4 ++--
 mercury/src/utils.rs                | 3 +++
 3 files changed, 7 insertions(+), 4 deletions(-)

diff --git a/mercury/delta/src/lib.rs b/mercury/delta/src/lib.rs
index 10db4acce..9593b44e2 100644
--- a/mercury/delta/src/lib.rs
+++ b/mercury/delta/src/lib.rs
@@ -148,8 +148,8 @@ mod tests {
     #[test]
     fn test_heuristic_encode_rate_large_files() {
        
-        let data1 = vec![0u8; 1_000_00];
-        let data2 = vec![1u8; 1_000_00];
+        let data1 = vec![0u8; 100_000];
+        let data2 = vec![1u8; 100_000];
         let rate = heuristic_encode_rate(&data1, &data2);
         println!("Large non-matching data rate = {}", rate);
         assert_eq!(rate, 0.0, "Large completely different data should early stop with 0 rate");
diff --git a/mercury/src/internal/pack/encode.rs b/mercury/src/internal/pack/encode.rs
index 88e8dae56..18ea699ea 100644
--- a/mercury/src/internal/pack/encode.rs
+++ b/mercury/src/internal/pack/encode.rs
@@ -129,7 +129,7 @@ fn magic_sort(a: &Entry, b: &Entry) -> Ordering {
     (a as *const Entry).cmp(&(b as *const Entry))
 }
 
-// /// 计算单片段 hash
+
 fn calc_hash(data: &[u8]) -> u64 {
     let mut hasher = AHasher::default();
     data.hash(&mut hasher);
@@ -585,7 +585,7 @@ mod tests {
         let mut source = PathBuf::from(env::current_dir().unwrap().parent().unwrap());
         source.push("tests/data/packs/pack-f8bbb573cef7d851957caceb491c073ee8e8de41.pack");
         // let file_map = crate::test_utils::setup_lfs_file().await;
-        // let source = file_mapa
+        // let source = file_map
         //     .get("git-2d187177923cd618a75da6c6db45bb89d92bd504.pack")
         //     .unwrap();
         // decode pack file to get entries
diff --git a/mercury/src/utils.rs b/mercury/src/utils.rs
index d80577cc7..a36ee8297 100644
--- a/mercury/src/utils.rs
+++ b/mercury/src/utils.rs
@@ -12,6 +12,9 @@ pub fn read_sha1(file: &mut impl Read) -> io::Result<SHA1> {
     SHA1::from_stream(file)
 }
 
+
+
+
 /// A lightweight wrapper that counts bytes read from the underlying reader.
 /// replace deflate.intotal() in decompress_data
 pub struct CountingReader<R> {