Move to new persistence model for wt

Cargo.toml

@@ -11,16 +11,11 @@ repository = "https://github.com/pathscale/DataBucket"
 description = "DataBucket is container for WorkTable's data"
 
 [dependencies]
-data_bucket_derive = {path = "codegen" , version = "0.1.0"}
+data_bucket_derive = { path = "codegen", version = "0.1.0" }
 
 eyre = "0.6.12"
 derive_more = { version = "1.0.0", features = ["from", "error", "display", "into"] }
-rkyv = { version = "0.8.9", features = ["uuid-1"]}
-scc = "2.1.16"
+rkyv = { version = "0.8.9", features = ["uuid-1"] }
 lockfree = "0.5.1"
-serde = { version = "1.0.215", features = ["derive"] }
 uuid = { version = "1.11.0", features = ["v4"] }
-bitcode = "0.6.3"
-
-[dev-dependencies]
-mktemp = "0.4.0"
+indexset = { version = "0.11.1", features = ["concurrent", "cdc", "multimap"] }

src/lib.rs

@@ -1,3 +1,5 @@
+extern crate core;
+
 pub mod link;
 pub mod page;
 pub mod persistence;
@@ -8,11 +10,12 @@ pub use link::Link;
 
 pub use data_bucket_codegen::SizeMeasure;
 pub use page::{
-    map_data_pages_to_general, map_index_pages_to_general, map_tree_index,
-    parse_data_page, parse_index_page, parse_page, persist_page, read_data_pages, read_index_pages,
-    read_rows_schema, seek_by_link, seek_to_page_start, update_at, Data as DataPage,
-    General as GeneralPage, GeneralHeader, IndexPage as IndexData, IndexValue, Interval, PageType,
-    SpaceInfo as SpaceInfoData, DATA_VERSION, GENERAL_HEADER_SIZE, INNER_PAGE_SIZE, PAGE_SIZE,
+    get_index_page_size_from_data_length, map_data_pages_to_general, map_index_pages_to_general,
+    parse_data_page, parse_index_page, parse_page, persist_page, seek_by_link, seek_to_page_start,
+    update_at, DataPage, GeneralHeader, GeneralPage, IndexPage, IndexValue, Interval, PageType,
+    SpaceInfoPage, TableOfContentsPage, DATA_VERSION, GENERAL_HEADER_SIZE, INNER_PAGE_SIZE,
+    PAGE_SIZE,
 };
 pub use persistence::{PersistableIndex, PersistableTable};
-pub use util::{align, Persistable, SizeMeasurable};
+pub use space::Id as SpaceId;
+pub use util::{align, align8, align_vec, Persistable, SizeMeasurable};

src/page/data.rs

@@ -1,15 +1,14 @@
-use eyre::{eyre, Result};
-
 use crate::Link;
 use crate::Persistable;
+use eyre::{eyre, Result};
 
 #[derive(Debug)]
-pub struct Data<const DATA_LENGTH: usize> {
+pub struct DataPage<const DATA_LENGTH: usize> {
     pub length: u32,
     pub data: [u8; DATA_LENGTH],
 }
 
-impl<const DATA_LENGTH: usize> Data<DATA_LENGTH> {
+impl<const DATA_LENGTH: usize> DataPage<DATA_LENGTH> {
     pub fn update_at(&mut self, link: Link, new_data: &[u8]) -> Result<()> {
         if new_data.len() as u32 != link.length {
             return Err(eyre!(
@@ -52,10 +51,19 @@ impl<const DATA_LENGTH: usize> Data<DATA_LENGTH> {
     }
 }
 
-impl<const DATA_LENGTH: usize> Persistable for Data<DATA_LENGTH> {
+impl<const DATA_LENGTH: usize> Persistable for DataPage<DATA_LENGTH> {
     fn as_bytes(&self) -> impl AsRef<[u8]> {
         &self.data[..self.length as usize]
     }
+
+    fn from_bytes(bytes: &[u8]) -> Self {
+        let mut data = [0; DATA_LENGTH];
+        data.copy_from_slice(bytes);
+        Self {
+            length: bytes.len() as u32,
+            data,
+        }
+    }
 }
 
 #[cfg(test)]
@@ -64,7 +72,7 @@ mod tests {
 
     #[test]
     fn test_update_at_success() {
-        let mut data = Data {
+        let mut data = DataPage {
             length: 0,
             data: [0; 100],
         };
@@ -82,7 +90,7 @@ mod tests {
 
     #[test]
     fn test_update_at_wrong_length() {
-        let mut data = Data {
+        let mut data = DataPage {
             length: 0,
             data: [0; 100],
         };
@@ -101,7 +109,7 @@ mod tests {
 
     #[test]
     fn test_update_at_out_of_bounds() {
-        let mut data = Data {
+        let mut data = DataPage {
             length: 0,
             data: [0; 100],
         };
@@ -120,7 +128,7 @@ mod tests {
 
     #[test]
     fn test_get_at_out_of_bounds() {
-        let data = Data {
+        let data = DataPage {
             length: 0,
             data: [0; 100],
         };

src/page/header.rs

@@ -3,9 +3,10 @@
 use rkyv::{Archive, Deserialize, Serialize};
 
 use crate::page::ty::PageType;
+use crate::page::PageId;
 use crate::space;
 use crate::util::Persistable;
-use crate::{page, PAGE_SIZE};
+use crate::PAGE_SIZE;
 
 pub const DATA_VERSION: u32 = 1u32;
 
@@ -16,15 +17,15 @@ pub const DATA_VERSION: u32 = 1u32;
 pub struct GeneralHeader {
     pub data_version: u32,
     pub space_id: space::Id,
-    pub page_id: page::PageId,
-    pub previous_id: page::PageId,
-    pub next_id: page::PageId,
+    pub page_id: PageId,
+    pub previous_id: PageId,
+    pub next_id: PageId,
     pub page_type: PageType,
     pub data_length: u32,
 }
 
 impl GeneralHeader {
-    pub fn new(page_id: page::PageId, type_: PageType, space_id: space::Id) -> Self {
+    pub fn new(page_id: PageId, type_: PageType, space_id: space::Id) -> Self {
         Self {
             data_version: DATA_VERSION,
             page_id,
@@ -38,7 +39,7 @@ impl GeneralHeader {
 
     /// Creates a new [`GeneralHeader`] for a page that follows page with given
     /// header. It means that [`PageType`] and [`space::Id`] are same and
-    /// old [`page::PageId`] will be `previous_id`.
+    /// old [`PageId`] will be `previous_id`.
     pub fn follow(&mut self) -> Self {
         self.next_id = self.page_id.next();
         Self {
@@ -54,7 +55,7 @@ impl GeneralHeader {
 
     /// Creates a new [`GeneralHeader`] for a page that follows page with given
     /// header but with different [`PageType`]. [`space::Id`] is same and old
-    /// [`page::PageId`] will be `previous_id`.
+    /// [`PageId`] will be `previous_id`.
     pub fn follow_with(&mut self, page_type: PageType) -> Self {
         self.next_id = self.page_id.next();
         Self {
@@ -67,12 +68,33 @@ impl GeneralHeader {
             data_length: PAGE_SIZE as u32,
         }
     }
+
+    /// Creates a new [`GeneralHeader`] for a page that follows page with given
+    /// header with provided [`PageId`], but with different [`PageType`].
+    /// [`space::Id`] is same and old [`PageId`] will be `previous_id`.
+    pub fn follow_with_page_id(&mut self, page_id: PageId) -> Self {
+        self.next_id = page_id;
+        Self {
+            data_version: DATA_VERSION,
+            page_id,
+            previous_id: self.page_id,
+            next_id: 0.into(),
+            page_type: self.page_type,
+            space_id: self.space_id,
+            data_length: PAGE_SIZE as u32,
+        }
+    }
 }
 
 impl Persistable for GeneralHeader {
     fn as_bytes(&self) -> impl AsRef<[u8]> {
         rkyv::to_bytes::<rkyv::rancor::Error>(self).unwrap()
     }
+
+    fn from_bytes(bytes: &[u8]) -> Self {
+        let archived = unsafe { rkyv::access_unchecked::<<Self as Archive>::Archived>(bytes) };
+        rkyv::deserialize::<_, rkyv::rancor::Error>(archived).expect("data should be valid")
+    }
 }
 
 #[cfg(test)]