jj/lib/src/op_store.rs
Yuya Nishihara 4834d12c37 simple_op_store: serialize RefTarget in new format (breaks downgrades)
This is breaking change. Old jj binary will panic if it sees a view saved by
new jj. Alternatively, we can store both new and legacy data for backward
compatibility.
2023-07-27 15:32:48 +09:00

260 lines
8.2 KiB
Rust

// Copyright 2020 The Jujutsu Authors
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// https://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
#![allow(missing_docs)]
use std::collections::{BTreeMap, HashMap, HashSet};
use std::fmt::{Debug, Error, Formatter};
use once_cell::sync::Lazy;
use thiserror::Error;
use crate::backend::{id_type, CommitId, ObjectId, Timestamp};
use crate::conflicts::Conflict;
content_hash! {
#[derive(PartialEq, Eq, PartialOrd, Ord, Clone, Hash)]
pub struct WorkspaceId(String);
}
impl Debug for WorkspaceId {
fn fmt(&self, f: &mut Formatter<'_>) -> Result<(), Error> {
f.debug_tuple("WorkspaceId").field(&self.0).finish()
}
}
impl Default for WorkspaceId {
fn default() -> Self {
Self("default".to_string())
}
}
impl WorkspaceId {
pub fn new(value: String) -> Self {
Self(value)
}
pub fn as_str(&self) -> &str {
&self.0
}
}
id_type!(pub ViewId);
id_type!(pub OperationId);
content_hash! {
#[derive(PartialEq, Eq, Hash, Clone, Debug)]
pub struct RefTarget {
conflict: Conflict<Option<CommitId>>,
}
}
impl Default for RefTarget {
fn default() -> Self {
Self::absent()
}
}
impl RefTarget {
/// Creates non-conflicting target pointing to no commit.
pub fn absent() -> Self {
Self::from_conflict(Conflict::resolved(None))
}
/// Returns non-conflicting target pointing to no commit.
///
/// This will typically be used in place of `None` returned by map lookup.
pub fn absent_ref() -> &'static Self {
static TARGET: Lazy<RefTarget> = Lazy::new(RefTarget::absent);
&TARGET
}
/// Creates non-conflicting target pointing to a commit.
pub fn normal(id: CommitId) -> Self {
Self::from_conflict(Conflict::resolved(Some(id)))
}
/// Creates target from removed/added ids.
pub fn from_legacy_form(
removed_ids: impl IntoIterator<Item = CommitId>,
added_ids: impl IntoIterator<Item = CommitId>,
) -> Self {
Self::from_conflict(Conflict::from_legacy_form(removed_ids, added_ids))
}
pub fn from_conflict(conflict: Conflict<Option<CommitId>>) -> Self {
RefTarget { conflict }
}
/// Returns id if this target is non-conflicting and points to a commit.
pub fn as_normal(&self) -> Option<&CommitId> {
let maybe_id = self.conflict.as_resolved()?;
maybe_id.as_ref()
}
/// Returns true if this target points to no commit.
pub fn is_absent(&self) -> bool {
matches!(self.conflict.as_resolved(), Some(None))
}
/// Returns true if this target points to any commit. Conflicting target is
/// always "present" as it should have at least one commit id.
pub fn is_present(&self) -> bool {
!self.is_absent()
}
/// Whether this target has conflicts.
pub fn has_conflict(&self) -> bool {
!self.conflict.is_resolved()
}
pub fn removed_ids(&self) -> impl Iterator<Item = &CommitId> {
self.conflict.removes().iter().flatten()
}
pub fn added_ids(&self) -> impl Iterator<Item = &CommitId> {
self.conflict.adds().iter().flatten()
}
pub fn as_conflict(&self) -> &Conflict<Option<CommitId>> {
&self.conflict
}
}
/// Helper to strip redundant `Option<T>` from `RefTarget` lookup result.
pub trait RefTargetOptionExt {
type Value;
fn flatten(self) -> Self::Value;
}
impl RefTargetOptionExt for Option<RefTarget> {
type Value = RefTarget;
fn flatten(self) -> Self::Value {
self.unwrap_or_else(RefTarget::absent)
}
}
impl<'a> RefTargetOptionExt for Option<&'a RefTarget> {
type Value = &'a RefTarget;
fn flatten(self) -> Self::Value {
self.unwrap_or_else(|| RefTarget::absent_ref())
}
}
content_hash! {
#[derive(Default, PartialEq, Eq, Clone, Debug)]
pub struct BranchTarget {
/// The commit the branch points to locally. `None` if the branch has been
/// deleted locally.
pub local_target: RefTarget,
// TODO: Do we need to support tombstones for remote branches? For example, if the branch
// has been deleted locally and you pull from a remote, maybe it should make a difference
// whether the branch is known to have existed on the remote. We may not want to resurrect
// the branch if the branch's state on the remote was just not known.
pub remote_targets: BTreeMap<String, RefTarget>,
}
}
content_hash! {
/// Represents the way the repo looks at a given time, just like how a Tree
/// object represents how the file system looks at a given time.
#[derive(PartialEq, Eq, Clone, Debug, Default)]
pub struct View {
/// All head commits
pub head_ids: HashSet<CommitId>,
/// Heads of the set of public commits.
pub public_head_ids: HashSet<CommitId>,
pub branches: BTreeMap<String, BranchTarget>,
pub tags: BTreeMap<String, RefTarget>,
pub git_refs: BTreeMap<String, RefTarget>,
/// The commit the Git HEAD points to.
// TODO: Support multiple Git worktrees?
// TODO: Do we want to store the current branch name too?
pub git_head: RefTarget,
// The commit that *should be* checked out in the workspace. Note that the working copy
// (.jj/working_copy/) has the source of truth about which commit *is* checked out (to be
// precise: the commit to which we most recently completed an update to).
pub wc_commit_ids: HashMap<WorkspaceId, CommitId>,
}
}
content_hash! {
/// Represents an operation (transaction) on the repo view, just like how a
/// Commit object represents an operation on the tree.
///
/// Operations and views are not meant to be exchanged between repos or users;
/// they represent local state and history.
///
/// The operation history will almost always be linear. It will only have
/// forks when parallel operations occurred. The parent is determined when
/// the transaction starts. When the transaction commits, a lock will be
/// taken and it will be checked that the current head of the operation
/// graph is unchanged. If the current head has changed, there has been
/// concurrent operation.
#[derive(PartialEq, Eq, Clone, Debug)]
pub struct Operation {
pub view_id: ViewId,
pub parents: Vec<OperationId>,
pub metadata: OperationMetadata,
}
}
content_hash! {
#[derive(PartialEq, Eq, Clone, Debug)]
pub struct OperationMetadata {
pub start_time: Timestamp,
pub end_time: Timestamp,
// Whatever is useful to the user, such as exact command line call
pub description: String,
pub hostname: String,
pub username: String,
pub tags: HashMap<String, String>,
}
}
#[derive(Debug, Error)]
pub enum OpStoreError {
#[error("Operation not found")]
NotFound,
#[error("Error when reading object {hash} of type {object_type}: {source}")]
ReadObject {
object_type: String,
hash: String,
source: Box<dyn std::error::Error + Send + Sync>,
},
#[error("Could not write object of type {object_type}: {source}")]
WriteObject {
object_type: &'static str,
source: Box<dyn std::error::Error + Send + Sync>,
},
#[error(transparent)]
Other(Box<dyn std::error::Error + Send + Sync>),
}
pub type OpStoreResult<T> = Result<T, OpStoreError>;
pub trait OpStore: Send + Sync + Debug {
fn name(&self) -> &str;
fn read_view(&self, id: &ViewId) -> OpStoreResult<View>;
fn write_view(&self, contents: &View) -> OpStoreResult<ViewId>;
fn read_operation(&self, id: &OperationId) -> OpStoreResult<Operation>;
fn write_operation(&self, contents: &Operation) -> OpStoreResult<OperationId>;
}