dirstate: Remove the flat Rust DirstateMap implementation
Before this changeset we had two Rust implementations of `DirstateMap`.
This removes the "flat" DirstateMap so that the "tree" DirstateMap is always
used when Rust enabled. This simplifies the code a lot, and will enable
(in the next changeset) further removal of a trait abstraction.
This is a performance regression when:
* Rust is enabled, and
* The repository uses the legacy dirstate-v1 file format, and
* For `hg status`, unknown files are not listed (such as with `-mard`)
The regression is about 100 milliseconds for `hg status -mard` on a
semi-large repository (mozilla-central), from ~320ms to ~420ms.
We deem this to be small enough to be worth it.
The new dirstate-v2 is still experimental at this point, but we aim to
stabilize it (though not yet enable it by default for new repositories)
in Mercurial 6.0. Eventually, upgrating repositories to dirsate-v2 will
eliminate this regression (and enable other performance improvements).
# Background
The flat DirstateMap was introduced with the first Rust implementation of the
status algorithm. It works similarly to the previous Python + C one, with a
single `HashMap` that associates file paths to a `DirstateEntry` (where Python
has a dict).
We later added the tree DirstateMap where the root of the tree contains nodes
for files and directories that are directly at the root of the repository,
and nodes for directories can contain child nodes representing the files and
directly that *they* contain directly. The shape of this tree mirrors that of
the working directory in the filesystem. This enables the status algorithm to
traverse this tree in tandem with traversing the filesystem tree, which in
turns enables a more efficient algorithm.
Furthermore, the new dirstate-v2 file format is also based on a tree of the
same shape. The tree DirstateMap can access a dirstate-v2 file without parsing
it: binary data in a single large (possibly memory-mapped) bytes buffer is
traversed on demand. This allows `DirstateMap` creation to take `O(1)` time.
(Mutation works by creating new in-memory nodes with copy-on-write semantics,
and serialization is append-mostly.)
The tradeoff is that for "legacy" repositories that use the dirstate-v1 file
format, parsing that file into a tree DirstateMap takes more time. Profiling
shows that this time is dominated by `HashMap`. For a dirstate containing `F`
files with an average `D` directory depth, the flat DirstateMap does parsing
in `O(F)` number of HashMap operations but the tree DirstateMap in `O(F × D)`
operations, since each node has its own HashMap containing its child nodes.
This slower costs ~140ms on an old snapshot of mozilla-central, and ~80ms
on an old snapshot of the Netbeans repository.
The status algorithm is faster, but with `-mard` (when not listing unknown
files) it is typically not faster *enough* to compensate the slower parsing.
Both Rust implementations are always faster than the Python + C implementation
# Benchmark results
All benchmarks are run on changeset 98c0408324e6, with repositories that use
the dirstate-v1 file format, on a server with 4 CPU cores and 4 CPU threads
(no HyperThreading).
`hg status` benchmarks show wall clock times of the entire command as the
average and standard deviation of serveral runs, collected by
https://github.com/sharkdp/hyperfine and reformated.
Parsing benchmarks are wall clock time of the Rust function that converts a
bytes buffer of the dirstate file into the `DirstateMap` data structure as
used by the status algorithm. A single run each, collected by running
`hg status` this environment variable:
RUST_LOG=hg::dirstate::dirstate_map=trace,hg::dirstate_tree::dirstate_map=trace
Benchmark 1: Rust flat DirstateMap → Rust tree DirstateMap
hg status
mozilla-clean 562.3 ms ± 2.0 ms → 462.5 ms ± 0.6 ms 1.22 ± 0.00 times faster
mozilla-dirty 859.6 ms ± 2.2 ms → 719.5 ms ± 3.2 ms 1.19 ± 0.01 times faster
mozilla-ignored 558.2 ms ± 3.0 ms → 457.9 ms ± 2.9 ms 1.22 ± 0.01 times faster
mozilla-unknowns 859.4 ms ± 5.7 ms → 716.0 ms ± 4.7 ms 1.20 ± 0.01 times faster
netbeans-clean 336.5 ms ± 0.9 ms → 339.5 ms ± 0.4 ms 0.99 ± 0.00 times faster
netbeans-dirty 491.4 ms ± 1.6 ms → 475.1 ms ± 1.2 ms 1.03 ± 0.00 times faster
netbeans-ignored 343.7 ms ± 1.0 ms → 347.8 ms ± 0.4 ms 0.99 ± 0.00 times faster
netbeans-unknowns 484.3 ms ± 1.0 ms → 466.0 ms ± 1.2 ms 1.04 ± 0.00 times faster
hg status -mard
mozilla-clean 317.3 ms ± 0.6 ms → 422.5 ms ± 1.2 ms 0.75 ± 0.00 times faster
mozilla-dirty 315.4 ms ± 0.6 ms → 417.7 ms ± 1.1 ms 0.76 ± 0.00 times faster
mozilla-ignored 314.6 ms ± 0.6 ms → 417.4 ms ± 1.0 ms 0.75 ± 0.00 times faster
mozilla-unknowns 312.9 ms ± 0.9 ms → 417.3 ms ± 1.6 ms 0.75 ± 0.00 times faster
netbeans-clean 212.0 ms ± 0.6 ms → 283.6 ms ± 0.8 ms 0.75 ± 0.00 times faster
netbeans-dirty 211.4 ms ± 1.0 ms → 283.4 ms ± 1.6 ms 0.75 ± 0.01 times faster
netbeans-ignored 211.4 ms ± 0.9 ms → 283.9 ms ± 0.8 ms 0.74 ± 0.01 times faster
netbeans-unknowns 211.1 ms ± 0.6 ms → 283.4 ms ± 1.0 ms 0.74 ± 0.00 times faster
Parsing
mozilla-clean 38.4ms → 177.6ms
mozilla-dirty 38.8ms → 177.0ms
mozilla-ignored 38.8ms → 178.0ms
mozilla-unknowns 38.7ms → 176.9ms
netbeans-clean 16.5ms → 97.3ms
netbeans-dirty 16.5ms → 98.4ms
netbeans-ignored 16.9ms → 97.4ms
netbeans-unknowns 16.9ms → 96.3ms
Benchmark 2: Python + C dirstatemap → Rust tree DirstateMap
hg status
mozilla-clean 1261.0 ms ± 3.6 ms → 461.1 ms ± 0.5 ms 2.73 ± 0.00 times faster
mozilla-dirty 2293.4 ms ± 9.1 ms → 719.6 ms ± 3.6 ms 3.19 ± 0.01 times faster
mozilla-ignored 1240.4 ms ± 2.3 ms → 457.7 ms ± 1.9 ms 2.71 ± 0.00 times faster
mozilla-unknowns 2283.3 ms ± 9.0 ms → 719.7 ms ± 3.8 ms 3.17 ± 0.01 times faster
netbeans-clean 879.7 ms ± 3.5 ms → 339.9 ms ± 0.5 ms 2.59 ± 0.00 times faster
netbeans-dirty 1257.3 ms ± 4.7 ms → 474.6 ms ± 1.6 ms 2.65 ± 0.01 times faster
netbeans-ignored 943.9 ms ± 1.9 ms → 347.3 ms ± 1.1 ms 2.72 ± 0.00 times faster
netbeans-unknowns 1188.1 ms ± 5.0 ms → 465.2 ms ± 2.3 ms 2.55 ± 0.01 times faster
hg status -mard
mozilla-clean 903.2 ms ± 3.6 ms → 423.4 ms ± 2.2 ms 2.13 ± 0.01 times faster
mozilla-dirty 884.6 ms ± 4.5 ms → 417.3 ms ± 1.4 ms 2.12 ± 0.01 times faster
mozilla-ignored 881.9 ms ± 1.3 ms → 417.3 ms ± 0.8 ms 2.11 ± 0.00 times faster
mozilla-unknowns 878.5 ms ± 1.9 ms → 416.4 ms ± 0.9 ms 2.11 ± 0.00 times faster
netbeans-clean 434.9 ms ± 1.8 ms → 284.0 ms ± 0.8 ms 1.53 ± 0.01 times faster
netbeans-dirty 434.1 ms ± 0.8 ms → 283.1 ms ± 0.8 ms 1.53 ± 0.00 times faster
netbeans-ignored 431.7 ms ± 1.1 ms → 283.6 ms ± 1.8 ms 1.52 ± 0.01 times faster
netbeans-unknowns 433.0 ms ± 1.3 ms → 283.5 ms ± 0.7 ms 1.53 ± 0.00 times faster
Differential Revision: https://phab.mercurial-scm.org/D11516
// dirstate_map.rs
//
// Copyright 2019 Raphaël Gomès <rgomes@octobus.net>
//
// This software may be used and distributed according to the terms of the
// GNU General Public License version 2 or any later version.
//! Bindings for the `hg::dirstate::dirstate_map` file provided by the
//! `hg-core` package.
use std::cell::{RefCell, RefMut};
use std::convert::TryInto;
use cpython::{
exc, PyBool, PyBytes, PyClone, PyDict, PyErr, PyList, PyNone, PyObject,
PyResult, Python, PythonObject, ToPyObject, UnsafePyLeaked,
};
use crate::{
dirstate::copymap::{CopyMap, CopyMapItemsIterator, CopyMapKeysIterator},
dirstate::item::DirstateItem,
pybytes_deref::PyBytesDeref,
};
use hg::{
dirstate::parsers::Timestamp,
dirstate::StateMapIter,
dirstate_tree::dirstate_map::DirstateMap as TreeDirstateMap,
dirstate_tree::dispatch::DirstateMapMethods,
dirstate_tree::on_disk::DirstateV2ParseError,
dirstate_tree::owning::OwningDirstateMap,
revlog::Node,
utils::files::normalize_case,
utils::hg_path::{HgPath, HgPathBuf},
DirstateEntry, DirstateError, DirstateParents, EntryState,
};
// TODO
// This object needs to share references to multiple members of its Rust
// inner struct, namely `copy_map`, `dirs` and `all_dirs`.
// Right now `CopyMap` is done, but it needs to have an explicit reference
// to `RustDirstateMap` which itself needs to have an encapsulation for
// every method in `CopyMap` (copymapcopy, etc.).
// This is ugly and hard to maintain.
// The same logic applies to `dirs` and `all_dirs`, however the `Dirs`
// `py_class!` is already implemented and does not mention
// `RustDirstateMap`, rightfully so.
// All attributes also have to have a separate refcount data attribute for
// leaks, with all methods that go along for reference sharing.
py_class!(pub class DirstateMap |py| {
@shared data inner: Box<dyn DirstateMapMethods + Send>;
/// Returns a `(dirstate_map, parents)` tuple
@staticmethod
def new_v1(
on_disk: PyBytes,
) -> PyResult<PyObject> {
let on_disk = PyBytesDeref::new(py, on_disk);
let mut map = OwningDirstateMap::new_empty(on_disk);
let (on_disk, map_placeholder) = map.get_mut_pair();
let (actual_map, parents) = TreeDirstateMap::new_v1(on_disk)
.map_err(|e| dirstate_error(py, e))?;
*map_placeholder = actual_map;
let map = Self::create_instance(py, Box::new(map))?;
let parents = parents.map(|p| {
let p1 = PyBytes::new(py, p.p1.as_bytes());
let p2 = PyBytes::new(py, p.p2.as_bytes());
(p1, p2)
});
Ok((map, parents).to_py_object(py).into_object())
}
/// Returns a DirstateMap
@staticmethod
def new_v2(
on_disk: PyBytes,
data_size: usize,
tree_metadata: PyBytes,
) -> PyResult<PyObject> {
let dirstate_error = |e: DirstateError| {
PyErr::new::<exc::OSError, _>(py, format!("Dirstate error: {:?}", e))
};
let on_disk = PyBytesDeref::new(py, on_disk);
let mut map = OwningDirstateMap::new_empty(on_disk);
let (on_disk, map_placeholder) = map.get_mut_pair();
*map_placeholder = TreeDirstateMap::new_v2(
on_disk, data_size, tree_metadata.data(py),
).map_err(dirstate_error)?;
let map = Self::create_instance(py, Box::new(map))?;
Ok(map.into_object())
}
def clear(&self) -> PyResult<PyObject> {
self.inner(py).borrow_mut().clear();
Ok(py.None())
}
def get(
&self,
key: PyObject,
default: Option<PyObject> = None
) -> PyResult<Option<PyObject>> {
let key = key.extract::<PyBytes>(py)?;
match self
.inner(py)
.borrow()
.get(HgPath::new(key.data(py)))
.map_err(|e| v2_error(py, e))?
{
Some(entry) => {
Ok(Some(DirstateItem::new_as_pyobject(py, entry)?))
},
None => Ok(default)
}
}
def set_dirstate_item(
&self,
path: PyObject,
item: DirstateItem
) -> PyResult<PyObject> {
let f = path.extract::<PyBytes>(py)?;
let filename = HgPath::new(f.data(py));
self.inner(py)
.borrow_mut()
.set_entry(filename, item.get_entry(py))
.map_err(|e| v2_error(py, e))?;
Ok(py.None())
}
def addfile(
&self,
f: PyBytes,
item: DirstateItem,
) -> PyResult<PyNone> {
let filename = HgPath::new(f.data(py));
let entry = item.get_entry(py);
self.inner(py)
.borrow_mut()
.add_file(filename, entry)
.map_err(|e |dirstate_error(py, e))?;
Ok(PyNone)
}
def removefile(
&self,
f: PyObject,
in_merge: PyObject
) -> PyResult<PyObject> {
self.inner(py).borrow_mut()
.remove_file(
HgPath::new(f.extract::<PyBytes>(py)?.data(py)),
in_merge.extract::<PyBool>(py)?.is_true(),
)
.or_else(|_| {
Err(PyErr::new::<exc::OSError, _>(
py,
"Dirstate error".to_string(),
))
})?;
Ok(py.None())
}
def drop_item_and_copy_source(
&self,
f: PyBytes,
) -> PyResult<PyNone> {
self.inner(py)
.borrow_mut()
.drop_entry_and_copy_source(HgPath::new(f.data(py)))
.map_err(|e |dirstate_error(py, e))?;
Ok(PyNone)
}
def hastrackeddir(&self, d: PyObject) -> PyResult<PyBool> {
let d = d.extract::<PyBytes>(py)?;
Ok(self.inner(py).borrow_mut()
.has_tracked_dir(HgPath::new(d.data(py)))
.map_err(|e| {
PyErr::new::<exc::ValueError, _>(py, e.to_string())
})?
.to_py_object(py))
}
def hasdir(&self, d: PyObject) -> PyResult<PyBool> {
let d = d.extract::<PyBytes>(py)?;
Ok(self.inner(py).borrow_mut()
.has_dir(HgPath::new(d.data(py)))
.map_err(|e| {
PyErr::new::<exc::ValueError, _>(py, e.to_string())
})?
.to_py_object(py))
}
def write_v1(
&self,
p1: PyObject,
p2: PyObject,
now: PyObject
) -> PyResult<PyBytes> {
let now = Timestamp(now.extract(py)?);
let mut inner = self.inner(py).borrow_mut();
let parents = DirstateParents {
p1: extract_node_id(py, &p1)?,
p2: extract_node_id(py, &p2)?,
};
let result = inner.pack_v1(parents, now);
match result {
Ok(packed) => Ok(PyBytes::new(py, &packed)),
Err(_) => Err(PyErr::new::<exc::OSError, _>(
py,
"Dirstate error".to_string(),
)),
}
}
/// Returns new data together with whether that data should be appended to
/// the existing data file whose content is at `self.on_disk` (True),
/// instead of written to a new data file (False).
def write_v2(
&self,
now: PyObject,
can_append: bool,
) -> PyResult<PyObject> {
let now = Timestamp(now.extract(py)?);
let mut inner = self.inner(py).borrow_mut();
let result = inner.pack_v2(now, can_append);
match result {
Ok((packed, tree_metadata, append)) => {
let packed = PyBytes::new(py, &packed);
let tree_metadata = PyBytes::new(py, &tree_metadata);
let tuple = (packed, tree_metadata, append);
Ok(tuple.to_py_object(py).into_object())
},
Err(_) => Err(PyErr::new::<exc::OSError, _>(
py,
"Dirstate error".to_string(),
)),
}
}
def filefoldmapasdict(&self) -> PyResult<PyDict> {
let dict = PyDict::new(py);
for item in self.inner(py).borrow_mut().iter() {
let (path, entry) = item.map_err(|e| v2_error(py, e))?;
if entry.state() != EntryState::Removed {
let key = normalize_case(path);
let value = path;
dict.set_item(
py,
PyBytes::new(py, key.as_bytes()).into_object(),
PyBytes::new(py, value.as_bytes()).into_object(),
)?;
}
}
Ok(dict)
}
def __len__(&self) -> PyResult<usize> {
Ok(self.inner(py).borrow().len())
}
def __contains__(&self, key: PyObject) -> PyResult<bool> {
let key = key.extract::<PyBytes>(py)?;
self.inner(py)
.borrow()
.contains_key(HgPath::new(key.data(py)))
.map_err(|e| v2_error(py, e))
}
def __getitem__(&self, key: PyObject) -> PyResult<PyObject> {
let key = key.extract::<PyBytes>(py)?;
let key = HgPath::new(key.data(py));
match self
.inner(py)
.borrow()
.get(key)
.map_err(|e| v2_error(py, e))?
{
Some(entry) => {
Ok(DirstateItem::new_as_pyobject(py, entry)?)
},
None => Err(PyErr::new::<exc::KeyError, _>(
py,
String::from_utf8_lossy(key.as_bytes()),
)),
}
}
def keys(&self) -> PyResult<DirstateMapKeysIterator> {
let leaked_ref = self.inner(py).leak_immutable();
DirstateMapKeysIterator::from_inner(
py,
unsafe { leaked_ref.map(py, |o| o.iter()) },
)
}
def items(&self) -> PyResult<DirstateMapItemsIterator> {
let leaked_ref = self.inner(py).leak_immutable();
DirstateMapItemsIterator::from_inner(
py,
unsafe { leaked_ref.map(py, |o| o.iter()) },
)
}
def __iter__(&self) -> PyResult<DirstateMapKeysIterator> {
let leaked_ref = self.inner(py).leak_immutable();
DirstateMapKeysIterator::from_inner(
py,
unsafe { leaked_ref.map(py, |o| o.iter()) },
)
}
// TODO all copymap* methods, see docstring above
def copymapcopy(&self) -> PyResult<PyDict> {
let dict = PyDict::new(py);
for item in self.inner(py).borrow().copy_map_iter() {
let (key, value) = item.map_err(|e| v2_error(py, e))?;
dict.set_item(
py,
PyBytes::new(py, key.as_bytes()),
PyBytes::new(py, value.as_bytes()),
)?;
}
Ok(dict)
}
def copymapgetitem(&self, key: PyObject) -> PyResult<PyBytes> {
let key = key.extract::<PyBytes>(py)?;
match self
.inner(py)
.borrow()
.copy_map_get(HgPath::new(key.data(py)))
.map_err(|e| v2_error(py, e))?
{
Some(copy) => Ok(PyBytes::new(py, copy.as_bytes())),
None => Err(PyErr::new::<exc::KeyError, _>(
py,
String::from_utf8_lossy(key.data(py)),
)),
}
}
def copymap(&self) -> PyResult<CopyMap> {
CopyMap::from_inner(py, self.clone_ref(py))
}
def copymaplen(&self) -> PyResult<usize> {
Ok(self.inner(py).borrow().copy_map_len())
}
def copymapcontains(&self, key: PyObject) -> PyResult<bool> {
let key = key.extract::<PyBytes>(py)?;
self.inner(py)
.borrow()
.copy_map_contains_key(HgPath::new(key.data(py)))
.map_err(|e| v2_error(py, e))
}
def copymapget(
&self,
key: PyObject,
default: Option<PyObject>
) -> PyResult<Option<PyObject>> {
let key = key.extract::<PyBytes>(py)?;
match self
.inner(py)
.borrow()
.copy_map_get(HgPath::new(key.data(py)))
.map_err(|e| v2_error(py, e))?
{
Some(copy) => Ok(Some(
PyBytes::new(py, copy.as_bytes()).into_object(),
)),
None => Ok(default),
}
}
def copymapsetitem(
&self,
key: PyObject,
value: PyObject
) -> PyResult<PyObject> {
let key = key.extract::<PyBytes>(py)?;
let value = value.extract::<PyBytes>(py)?;
self.inner(py)
.borrow_mut()
.copy_map_insert(
HgPathBuf::from_bytes(key.data(py)),
HgPathBuf::from_bytes(value.data(py)),
)
.map_err(|e| v2_error(py, e))?;
Ok(py.None())
}
def copymappop(
&self,
key: PyObject,
default: Option<PyObject>
) -> PyResult<Option<PyObject>> {
let key = key.extract::<PyBytes>(py)?;
match self
.inner(py)
.borrow_mut()
.copy_map_remove(HgPath::new(key.data(py)))
.map_err(|e| v2_error(py, e))?
{
Some(_) => Ok(None),
None => Ok(default),
}
}
def copymapiter(&self) -> PyResult<CopyMapKeysIterator> {
let leaked_ref = self.inner(py).leak_immutable();
CopyMapKeysIterator::from_inner(
py,
unsafe { leaked_ref.map(py, |o| o.copy_map_iter()) },
)
}
def copymapitemsiter(&self) -> PyResult<CopyMapItemsIterator> {
let leaked_ref = self.inner(py).leak_immutable();
CopyMapItemsIterator::from_inner(
py,
unsafe { leaked_ref.map(py, |o| o.copy_map_iter()) },
)
}
def tracked_dirs(&self) -> PyResult<PyList> {
let dirs = PyList::new(py, &[]);
for path in self.inner(py).borrow_mut().iter_tracked_dirs()
.map_err(|e |dirstate_error(py, e))?
{
let path = path.map_err(|e| v2_error(py, e))?;
let path = PyBytes::new(py, path.as_bytes());
dirs.append(py, path.into_object())
}
Ok(dirs)
}
def debug_iter(&self, all: bool) -> PyResult<PyList> {
let dirs = PyList::new(py, &[]);
for item in self.inner(py).borrow().debug_iter(all) {
let (path, (state, mode, size, mtime)) =
item.map_err(|e| v2_error(py, e))?;
let path = PyBytes::new(py, path.as_bytes());
let item = (path, state, mode, size, mtime);
dirs.append(py, item.to_py_object(py).into_object())
}
Ok(dirs)
}
});
impl DirstateMap {
pub fn get_inner_mut<'a>(
&'a self,
py: Python<'a>,
) -> RefMut<'a, Box<dyn DirstateMapMethods + Send>> {
self.inner(py).borrow_mut()
}
fn translate_key(
py: Python,
res: Result<(&HgPath, DirstateEntry), DirstateV2ParseError>,
) -> PyResult<Option<PyBytes>> {
let (f, _entry) = res.map_err(|e| v2_error(py, e))?;
Ok(Some(PyBytes::new(py, f.as_bytes())))
}
fn translate_key_value(
py: Python,
res: Result<(&HgPath, DirstateEntry), DirstateV2ParseError>,
) -> PyResult<Option<(PyBytes, PyObject)>> {
let (f, entry) = res.map_err(|e| v2_error(py, e))?;
Ok(Some((
PyBytes::new(py, f.as_bytes()),
DirstateItem::new_as_pyobject(py, entry)?,
)))
}
}
py_shared_iterator!(
DirstateMapKeysIterator,
UnsafePyLeaked<StateMapIter<'static>>,
DirstateMap::translate_key,
Option<PyBytes>
);
py_shared_iterator!(
DirstateMapItemsIterator,
UnsafePyLeaked<StateMapIter<'static>>,
DirstateMap::translate_key_value,
Option<(PyBytes, PyObject)>
);
fn extract_node_id(py: Python, obj: &PyObject) -> PyResult<Node> {
let bytes = obj.extract::<PyBytes>(py)?;
match bytes.data(py).try_into() {
Ok(s) => Ok(s),
Err(e) => Err(PyErr::new::<exc::ValueError, _>(py, e.to_string())),
}
}
pub(super) fn v2_error(py: Python<'_>, _: DirstateV2ParseError) -> PyErr {
PyErr::new::<exc::ValueError, _>(py, "corrupted dirstate-v2")
}
fn dirstate_error(py: Python<'_>, e: DirstateError) -> PyErr {
PyErr::new::<exc::OSError, _>(py, format!("Dirstate error: {:?}", e))
}