[mercury]: Add comments (#1232)

* [mercury]: Add comments

Signed-off-by: Xiaoyang Han <lux1an@qq.com>

* [mercury]: Add comments

Signed-off-by: Xiaoyang Han <lux1an@qq.com>

* [refact]: for scorpio

Signed-off-by: Han Xiaoyang <lux1an@qq.com>

---------

Signed-off-by: Xiaoyang Han <lux1an@qq.com>
Signed-off-by: Han Xiaoyang <lux1an@qq.com>

Author: Ivanbeethoven

mercury/src/hash.rs

@@ -5,6 +5,7 @@
 
 use std::{fmt::Display, io};
 
+use bincode::{Encode, Decode};
 use colored::Colorize;
 use serde::{Deserialize, Serialize};
 use sha1::Digest;
@@ -28,6 +29,7 @@ use crate::internal::object::types::ObjectType;
 ///
 #[derive(
     Clone, Copy, Debug, PartialEq, Eq, Hash, PartialOrd, Ord, Default, Deserialize, Serialize,
+    Encode, Decode
 )]
 pub struct SHA1(pub [u8; 20]);
 

mercury/src/internal/object/blob.rs

@@ -36,6 +36,7 @@ use crate::internal::object::ObjectTrait;
 
 /// **The Blob Object**
 #[derive(Eq, Debug, Clone)]
+#[non_exhaustive]
 pub struct Blob {
     pub id: SHA1,
     pub data: Vec<u8>,
@@ -113,4 +114,5 @@ mod tests {
             "5dd01c177f5d7d1be5346a5bc18a569a7410c2ef"
         );
     }
+
 }

mercury/src/internal/object/commit.rs

@@ -19,6 +19,7 @@ use crate::hash::SHA1;
 use crate::internal::object::signature::Signature;
 use crate::internal::object::ObjectTrait;
 use crate::internal::object::ObjectType;
+use bincode::{Decode, Encode};
 use bstr::ByteSlice;
 use callisto::git_commit;
 use callisto::mega_commit;
@@ -35,7 +36,7 @@ use serde::Serialize;
 ///   history of a repository with a single commit object at its root.
 /// - The author and committer fields contain the name, email address, timestamp and timezone.
 /// - The message field contains the commit message, which maybe include signed or DCO.
-#[derive(Eq, Debug, Clone, Serialize, Deserialize)]
+#[derive(Eq, Debug, Clone, Serialize, Deserialize, Decode, Encode)]
 #[non_exhaustive]
 pub struct Commit {
     pub id: SHA1,
@@ -119,6 +120,8 @@ impl Commit {
         Commit::new(author, committer, tree_id, parent_commit_ids, message)
     }
 
+    /// Formats the commit message by extracting the first line of the message.
+    /// If the message contains a PGP signature, it will return the first line after the signature.
     pub fn format_message(&self) -> String {
         let mut has_signature = false;
         for line in self.message.lines() {
@@ -148,21 +151,26 @@ impl ObjectTrait for Commit {
         // Find the tree id and remove it from the data
         let tree_end = commit.find_byte(0x0a).unwrap();
         let tree_id: SHA1 = SHA1::from_str(
-            String::from_utf8(commit[5..tree_end].to_owned())
+            String::from_utf8(commit[5..tree_end].to_owned()) // 5 is the length of "tree "
                 .unwrap()
                 .as_str(),
         )
         .unwrap();
-        let binding = commit[tree_end + 1..].to_vec();
+        let binding = commit[tree_end + 1..].to_vec(); // Move past the tree id 
         commit = &binding;
 
         // Find the parent commit ids and remove them from the data
         let author_begin = commit.find("author").unwrap();
+        // Find all parent commit ids
+        // The parent commit ids are all the lines that start with "parent "
+        // We can use find_iter to find all occurrences of "parent "
+        // and then extract the SHA1 hashes from them.
         let parent_commit_ids: Vec<SHA1> = commit[..author_begin]
             .find_iter("parent")
             .map(|parent| {
                 let parent_end = commit[parent..].find_byte(0x0a).unwrap();
                 SHA1::from_str(
+                    // 7 is the length of "parent "
                     String::from_utf8(commit[parent + 7..parent + parent_end].to_owned())
                         .unwrap()
                         .as_str(),
@@ -174,8 +182,10 @@ impl ObjectTrait for Commit {
         commit = &binding;
 
         // Find the author and committer and remove them from the data
+        // 0x0a is the newline character
         let author =
-            Signature::from_data(commit[..commit.find_byte(0x0a).unwrap()].to_vec()).unwrap();
+            Signature::from_data(commit[..commit.find_byte(0x0a).unwrap()].to_vec()).unwrap(); 
+        
         let binding = commit[commit.find_byte(0x0a).unwrap() + 1..].to_vec();
         commit = &binding;
         let committer =

mercury/src/internal/object/signature.rs

@@ -11,6 +11,7 @@
 //!
 use std::{fmt::Display, str::FromStr};
 
+use bincode::{Decode, Encode};
 use bstr::ByteSlice;
 use chrono::Offset;
 use serde::{Deserialize, Serialize};
@@ -30,7 +31,7 @@ use crate::errors::GitError;
 /// ```
 ///
 /// So, we design a `SignatureType` enum to indicate the signature type.
-#[derive(PartialEq, Eq, Debug, Clone, Serialize, Deserialize)]
+#[derive(PartialEq, Eq, Debug, Clone, Serialize, Deserialize, Decode, Encode)]
 pub enum SignatureType {
     Author,
     Committer,
@@ -75,7 +76,7 @@ impl SignatureType {
     }
 }
 
-#[derive(PartialEq, Eq, Debug, Clone, Serialize, Deserialize)]
+#[derive(PartialEq, Eq, Debug, Clone, Serialize, Deserialize,Decode,Encode)]
 pub struct Signature {
     pub signature_type: SignatureType,
     pub name: String,

mercury/src/internal/object/tag.rs

@@ -49,6 +49,7 @@ use crate::internal::object::ObjectType;
 
 /// The tag object is used to Annotated tag
 #[derive(Eq, Debug, Clone)]
+#[non_exhaustive]
 pub struct Tag {
     pub id: SHA1,
     pub object_hash: SHA1,
@@ -123,7 +124,7 @@ impl ObjectTrait for Tag {
         let tagger = Signature::from_data(tagger_data).unwrap();
         data = &data[data.find_byte(0x0a).unwrap() + 1..];
 
-        let message = unsafe {
+        let message = unsafe { // There may be non-UTF-8 characters, so we use `to_str_unchecked` for conversion.
             data[data.find_byte(0x0a).unwrap()..]
                 .to_vec()
                 .to_str_unchecked()

mercury/src/internal/object/tree.rs

@@ -18,6 +18,7 @@ use crate::errors::GitError;
 use crate::hash::SHA1;
 use crate::internal::object::ObjectTrait;
 use crate::internal::object::ObjectType;
+use bincode::{Encode, Decode};
 use colored::Colorize;
 use encoding_rs::GBK;
 use serde::Deserialize;
@@ -28,7 +29,7 @@ use std::fmt::Display;
 /// that entry. The mode is a three-digit octal number that encodes both the permissions and the
 /// type of the object. The first digit specifies the object type, and the remaining two digits
 /// specify the file mode or permissions.
-#[derive(PartialEq, Eq, Debug, Clone, Copy, Serialize, Deserialize, Hash)]
+#[derive(PartialEq, Eq, Debug, Clone, Copy, Serialize, Deserialize, Hash, Encode, Decode)]
 pub enum TreeItemMode {
     Blob,
     BlobExecutable,
@@ -129,7 +130,7 @@ impl TreeItemMode {
 /// 100644 hello-world\0<blob object ID>
 /// 040000 data\0<tree object ID>
 /// ```
-#[derive(PartialEq, Eq, Debug, Clone, Serialize, Deserialize, Hash)]
+#[derive(PartialEq, Eq, Debug, Clone, Serialize, Deserialize, Hash, Encode, Decode)]
 pub struct TreeItem {
     pub mode: TreeItemMode,
     pub id: SHA1,
@@ -220,7 +221,8 @@ impl TreeItem {
 
 /// A tree object is a Git object that represents a directory. It contains a list of entries, one
 /// for each file or directory in the tree.
-#[derive(Eq, Debug, Clone, Serialize, Deserialize)]
+#[derive(Eq, Debug, Clone, Serialize, Deserialize, Encode, Decode)]
+#[non_exhaustive]
 pub struct Tree {
     pub id: SHA1,
     pub tree_items: Vec<TreeItem>,

mercury/src/internal/pack/cache.rs

@@ -95,7 +95,7 @@ impl Caches {
         let hash_str = hash._to_string();
         path.push(&hash_str[..2]); // use first 2 chars as the directory
         self.path_prefixes[hash.as_ref()[0] as usize].call_once(|| {
-            // 检查目录是否存在，只有在不存在时才创建
+            // Check if the directory exists, if not, create it
             if !path.exists() {
                 fs::create_dir_all(&path).unwrap();
             }

mercury/src/internal/pack/encode.rs

@@ -153,6 +153,8 @@ impl PackEncoder {
         return self.inner_encode(entry_rx, true).await;
     }
 
+    /// Delta selection heuristics are based on:
+    ///   https://github.com/git/git/blob/master/Documentation/technical/pack-heuristics.adoc
     async fn inner_encode(
         &mut self,
         mut entry_rx: mpsc::Receiver<Entry>,

scorpio/Cargo.toml

@@ -12,10 +12,8 @@ reqwest =  { version = "0.12.7", features = ["json","blocking"] }
 serde = { version = "1.0.210", features = ["derive"] }
 fuse-backend-rs = { version = "0.12.0", features = ["fusedev","async-io"]}
 tokio = { version = "1.40.0", features = ["full"] }
-vm-memory = { version = "0.15.0", features = ["backend-mmap", "backend-bitmap"] }
-axum = { version =  "0.7.7",features=["macros"]}
+axum = { version =  "0.8.4",features=["macros"]}
 rfuse3 = { version = "0.0.2" ,features = ["tokio-runtime","unprivileged"]}
-futures-util = { version = "0.3.30", features = ["sink"] }
 syn = { version = "2.0.98", features = ["full", "extra-traits"] }
 clap = { version = "4.0", features = ["derive"] }
 
@@ -30,11 +28,10 @@ once_cell = "1.19.0"
 arc-swap = "1.7.1"
 env_logger = "0.11.5"
 sled = "0.34.7"
-bincode = "1.3.3"
+bincode = { workspace = true , features = ["serde"] }
 async-recursion = "1.1.1"
 bytes = "1.7.2"
 futures = "0.3.31"
-vmm-sys-util = "0.11"
 quote = "1.0.38"
 proc-macro2 = "1.0.93"
 uuid = "1.14.0"

scorpio/src/dicfuse/tree_store.rs

@@ -1,4 +1,5 @@
 use crate::util::{config, GPath};
+use bincode::{Decode, Encode};
 use rfuse3::raw::reply::ReplyEntry;
 use rfuse3::FileType;
 use serde::{Deserialize, Serialize};
@@ -14,7 +15,7 @@ pub struct TreeStorage {
     db: Db,
 }
 
-#[derive(Serialize, Deserialize, Clone)]
+#[derive(Serialize, Deserialize, Clone,Encode,Decode)]
 pub struct StorageItem {
     inode: u64,
     parent: u64,
@@ -80,12 +81,12 @@ impl TreeStorage {
             children: Vec::new(),
             hash: item.hash,
         };
-
+        let config = bincode::config::standard();
         // Insert an item into db and update the parent item's children list.
         self.db
             .insert(
                 inode.to_be_bytes(),
-                bincode::serialize(&storage_item).map_err(Error::other)?,
+                bincode::encode_to_vec(&storage_item,config).map_err(Error::other)?,
             )
             .map_err(Error::other)?;
 
@@ -97,7 +98,7 @@ impl TreeStorage {
             self.db
                 .insert(
                     parent.to_be_bytes(),
-                    bincode::serialize(&parent_item).map_err(Error::other)?,
+                    bincode::encode_to_vec(&parent_item,config).map_err(Error::other)?,
                 )
                 .map_err(Error::other)?;
         }
@@ -122,10 +123,12 @@ impl TreeStorage {
             if storage_item.parent != 0 {
                 let mut parent_item: StorageItem = self.get_storage_item(storage_item.parent)?;
                 parent_item.children.retain(|&x| x != inode);
+                let config = bincode::config::standard();
+
                 self.db
                     .insert(
                         storage_item.parent.to_be_bytes(),
-                        bincode::serialize(&parent_item).map_err(Error::other)?,
+                       bincode::encode_to_vec(&parent_item, config).map_err(Error::other)?,
                     )
                     .map_err(Error::other)?;
             }
@@ -140,10 +143,11 @@ impl TreeStorage {
     pub fn append_child(&self, parent: u64, inode: u64) -> io::Result<()> {
         let mut st = self.get_storage_item(parent)?;
         st.children.push(inode);
+        let config = bincode::config::standard();
         self.db
             .insert(
                 parent.to_be_bytes(),
-                bincode::serialize(&st).map_err(Error::other)?,
+                bincode::encode_to_vec(&st, config).map_err(Error::other)?,
             )
             .map_err(Error::other)?;
         Ok(())
@@ -163,7 +167,8 @@ impl TreeStorage {
     pub fn get_storage_item(&self, inode: u64) -> io::Result<StorageItem> {
         match self.db.get(inode.to_be_bytes())? {
             Some(value) => {
-                let item: StorageItem = bincode::deserialize(&value).map_err(Error::other)?;
+                let config = bincode::config::standard();
+                let (item ,_) = bincode::decode_from_slice(&value,config).map_err(Error::other)?;
                 Ok(item)
             }
             None => Err(Error::new(ErrorKind::NotFound, "Item not found")),
@@ -184,10 +189,11 @@ impl TreeStorage {
     pub fn update_item_hash(&self, inode: u64, hash: String) -> io::Result<()> {
         let mut item = self.get_storage_item(inode)?;
         item.hash = hash;
+        let config = bincode::config::standard();
         self.db
             .insert(
                 inode.to_be_bytes(),
-                bincode::serialize(&item).map_err(Error::other)?,
+                bincode::encode_to_vec(&item,config).map_err(Error::other)?,
             )
             .map_err(Error::other)?;
         Ok(())