This is an archive of the discontinued Mercurial Phabricator instance.

Differential D8157

rust-nodemap: use proper Index API instead of using the C API
ClosedPublic

Authored by Alphare on Feb 26 2020, 12:36 PM.

Details

Reviewers
None
Group Reviewers
hg-reviewers
Commits
rHG857cc79247ac: rust-nodemap: use proper Index API instead of using the C API

Diff Detail

Repository
rHG Mercurial
Branch
default
Lint
No Linters Available
Unit
No Unit Test Coverage

Event Timeline

Alphare created this revision.Feb 26 2020, 12:36 PM
marmoute updated this revision to Diff 20376.Feb 28 2020, 1:53 PM
Alphare updated this revision to Diff 20656.Mar 10 2020, 6:51 AM
Alphare added a commit: rHG857cc79247ac: rust-nodemap: use proper Index API instead of using the C API.Mar 11 2020, 11:42 AM
This revision was not accepted when it landed; it landed in state Needs Review.
Closed by commit rHG857cc79247ac: rust-nodemap: use proper Index API instead of using the C API (authored by Alphare). · Explain Why
This revision was automatically updated to reflect the committed changes.

Revision Contents
Changeset List

PathPackages
Mrust/hg-cpython/src/revlog.rs (174 lines)

Diff 20376

rust/hg-cpython/src/revlog.rs

// revlog.rs // revlog.rs
// //
// Copyright 2019-2020 Georges Racinet <georges.racinet@octobus.net> // Copyright 2019-2020 Georges Racinet <georges.racinet@octobus.net>
// //
// This software may be used and distributed according to the terms of the // This software may be used and distributed according to the terms of the
// GNU General Public License version 2 or any later version. // GNU General Public License version 2 or any later version.
use crate::cindex; use crate::{
cindex,
utils::{node_from_py_bytes, node_from_py_object},
};
use cpython::{ use cpython::{
exc::ValueError, ObjectProtocol, PyClone, PyDict, PyErr, PyModule, exc::{IndexError, ValueError},
PyObject, PyResult, PyTuple, Python, PythonObject, ToPyObject, ObjectProtocol, PyBytes, PyClone, PyDict, PyErr, PyModule, PyObject,
PyResult, PyString, PyTuple, Python, PythonObject, ToPyObject,
};
use hg::{
nodemap::{NodeMapError, NodeTree},
revlog::{nodemap::NodeMap, RevlogIndex},
NodeError, Revision,
}; };
use hg::{nodemap::NodeMapError, NodeError, Revision};
use std::cell::RefCell; use std::cell::RefCell;
/// Return a Struct implementing the Graph trait /// Return a Struct implementing the Graph trait
pub(crate) fn pyindex_to_graph( pub(crate) fn pyindex_to_graph(
py: Python, py: Python,
index: PyObject, index: PyObject,
) -> PyResult<cindex::Index> { ) -> PyResult<cindex::Index> {
match index.extract::<MixedIndex>(py) { match index.extract::<MixedIndex>(py) {
Ok(midx) => Ok(midx.clone_cindex(py)), Ok(midx) => Ok(midx.clone_cindex(py)),
Err(_) => cindex::Index::new(py, index), Err(_) => cindex::Index::new(py, index),
} }
} }
py_class!(pub class MixedIndex |py| { py_class!(pub class MixedIndex |py| {
data cindex: RefCell<cindex::Index>; data cindex: RefCell<cindex::Index>;
data nt: RefCell<Option<NodeTree>>;
def __new__(_cls, cindex: PyObject) -> PyResult<MixedIndex> { def __new__(_cls, cindex: PyObject) -> PyResult<MixedIndex> {
Self::new(py, cindex) Self::new(py, cindex)
} }
/// Compatibility layer used for Python consumers needing access to the C index /// Compatibility layer used for Python consumers needing access to the C index
/// ///
/// Only use case so far is `scmutil.shortesthexnodeidprefix`, /// Only use case so far is `scmutil.shortesthexnodeidprefix`,
/// that may need to build a custom `nodetree`, based on a specified revset. /// that may need to build a custom `nodetree`, based on a specified revset.
/// With a Rust implementation of the nodemap, we will be able to get rid of /// With a Rust implementation of the nodemap, we will be able to get rid of
/// this, by exposing our own standalone nodemap class, /// this, by exposing our own standalone nodemap class,
/// ready to accept `MixedIndex`. /// ready to accept `MixedIndex`.
def get_cindex(&self) -> PyResult<PyObject> { def get_cindex(&self) -> PyResult<PyObject> {
Ok(self.cindex(py).borrow().inner().clone_ref(py)) Ok(self.cindex(py).borrow().inner().clone_ref(py))
} }
// Index API involving nodemap, as defined in mercurial/pure/parsers.py
/// Return Revision if found, raises a bare `error.RevlogError`
/// in case of ambiguity, same as C version does
def get_rev(&self, node: PyBytes) -> PyResult<Option<Revision>> {
let opt = self.get_nodetree(py)?.borrow();
let nt = opt.as_ref().unwrap();
let idx = &*self.cindex(py).borrow();
let node = node_from_py_bytes(py, &node)?;
nt.find_bin(idx, (&node).into()).map_err(|e| nodemap_error(py, e))
}
/// same as `get_rev()` but raises a bare `error.RevlogError` if node
/// is not found.
///
/// No need to repeat `node` in the exception, `mercurial/revlog.py`
/// will catch and rewrap with it
def rev(&self, node: PyBytes) -> PyResult<Revision> {
self.get_rev(py, node)?.ok_or_else(|| revlog_error(py))
}
/// return True if the node exist in the index
def has_node(&self, node: PyBytes) -> PyResult<bool> {
self.get_rev(py, node).map(|opt| opt.is_some())
}
/// find length of shortest hex nodeid of a binary ID
def shortest(&self, node: PyBytes) -> PyResult<usize> {
let opt = self.get_nodetree(py)?.borrow();
let nt = opt.as_ref().unwrap();
let idx = &*self.cindex(py).borrow();
match nt.unique_prefix_len_node(idx, &node_from_py_bytes(py, &node)?)
{
Ok(Some(l)) => Ok(l),
Ok(None) => Err(revlog_error(py)),
Err(e) => Err(nodemap_error(py, e)),
}
}
def partialmatch(&self, node: PyObject) -> PyResult<Option<PyBytes>> {
let opt = self.get_nodetree(py)?.borrow();
let nt = opt.as_ref().unwrap();
let idx = &*self.cindex(py).borrow();
let node_as_string = if cfg!(feature = "python3-sys") {
node.cast_as::<PyString>(py)?.to_string(py)?.to_string()
}
else {
let node = node.extract::<PyBytes>(py)?;
String::from_utf8_lossy(node.data(py)).to_string()
};
nt.find_hex(idx, &node_as_string)
// TODO make an inner API returning the node directly
.map(|opt| opt.map(
|rev| PyBytes::new(py, idx.node(rev).unwrap().as_bytes())))
.map_err(|e| nodemap_error(py, e))
}
/// append an index entry
def append(&self, tup: PyTuple) -> PyResult<PyObject> {
if tup.len(py) < 8 {
// this is better than the panic promised by tup.get_item()
return Err(
PyErr::new::<IndexError, _>(py, "tuple index out of range"))
}
let node_bytes = tup.get_item(py, 7).extract(py)?;
let node = node_from_py_object(py, &node_bytes)?;
let mut idx = self.cindex(py).borrow_mut();
let rev = idx.len() as Revision;
idx.append(py, tup)?;
self.get_nodetree(py)?.borrow_mut().as_mut().unwrap()
.insert(&*idx, &node, rev)
.map_err(|e| nodemap_error(py, e))?;
Ok(py.None())
}
def __delitem__(&self, key: PyObject) -> PyResult<()> {
// __delitem__ is both for `del idx[r]` and `del idx[r1:r2]`
self.cindex(py).borrow().inner().del_item(py, key)?;
let mut opt = self.get_nodetree(py)?.borrow_mut();
let mut nt = opt.as_mut().unwrap();
nt.invalidate_all();
self.fill_nodemap(py, &mut nt)?;
Ok(())
}
//
// Reforwarded C index API // Reforwarded C index API
//
// index_methods (tp_methods). Same ordering as in revlog.c // index_methods (tp_methods). Same ordering as in revlog.c
/// return the gca set of the given revs /// return the gca set of the given revs
def ancestors(&self, *args, **kw) -> PyResult<PyObject> { def ancestors(&self, *args, **kw) -> PyResult<PyObject> {
self.call_cindex(py, "ancestors", args, kw) self.call_cindex(py, "ancestors", args, kw)
} }
/// return the heads of the common ancestors of the given revs /// return the heads of the common ancestors of the given revs
def commonancestorsheads(&self, *args, **kw) -> PyResult<PyObject> { def commonancestorsheads(&self, *args, **kw) -> PyResult<PyObject> {
self.call_cindex(py, "commonancestorsheads", args, kw) self.call_cindex(py, "commonancestorsheads", args, kw)
} }
/// clear the index caches /// clear the index caches
def clearcaches(&self, *args, **kw) -> PyResult<PyObject> { def clearcaches(&self, *args, **kw) -> PyResult<PyObject> {
self.call_cindex(py, "clearcaches", args, kw) self.call_cindex(py, "clearcaches", args, kw)
} }
/// get an index entry /// get an index entry
def get(&self, *args, **kw) -> PyResult<PyObject> { def get(&self, *args, **kw) -> PyResult<PyObject> {
self.call_cindex(py, "get", args, kw) self.call_cindex(py, "get", args, kw)
} }
/// return `rev` associated with a node or None
def get_rev(&self, *args, **kw) -> PyResult<PyObject> {
self.call_cindex(py, "get_rev", args, kw)
}
/// return True if the node exist in the index
def has_node(&self, *args, **kw) -> PyResult<PyObject> {
self.call_cindex(py, "has_node", args, kw)
}
/// return `rev` associated with a node or raise RevlogError
def rev(&self, *args, **kw) -> PyResult<PyObject> {
self.call_cindex(py, "rev", args, kw)
}
/// compute phases /// compute phases
def computephasesmapsets(&self, *args, **kw) -> PyResult<PyObject> { def computephasesmapsets(&self, *args, **kw) -> PyResult<PyObject> {
self.call_cindex(py, "computephasesmapsets", args, kw) self.call_cindex(py, "computephasesmapsets", args, kw)
} }
/// reachableroots /// reachableroots
def reachableroots2(&self, *args, **kw) -> PyResult<PyObject> { def reachableroots2(&self, *args, **kw) -> PyResult<PyObject> {
self.call_cindex(py, "reachableroots2", args, kw) self.call_cindex(py, "reachableroots2", args, kw)
self.call_cindex(py, "deltachain", args, kw) self.call_cindex(py, "deltachain", args, kw)
} }
/// slice planned chunk read to reach a density threshold /// slice planned chunk read to reach a density threshold
def slicechunktodensity(&self, *args, **kw) -> PyResult<PyObject> { def slicechunktodensity(&self, *args, **kw) -> PyResult<PyObject> {
self.call_cindex(py, "slicechunktodensity", args, kw) self.call_cindex(py, "slicechunktodensity", args, kw)
} }
/// append an index entry
def append(&self, *args, **kw) -> PyResult<PyObject> {
self.call_cindex(py, "append", args, kw)
}
/// match a potentially ambiguous node ID
def partialmatch(&self, *args, **kw) -> PyResult<PyObject> {
self.call_cindex(py, "partialmatch", args, kw)
}
/// find length of shortest hex nodeid of a binary ID
def shortest(&self, *args, **kw) -> PyResult<PyObject> {
self.call_cindex(py, "shortest", args, kw)
}
/// stats for the index /// stats for the index
def stats(&self, *args, **kw) -> PyResult<PyObject> { def stats(&self, *args, **kw) -> PyResult<PyObject> {
self.call_cindex(py, "stats", args, kw) self.call_cindex(py, "stats", args, kw)
} }
// index_sequence_methods and index_mapping_methods. // index_sequence_methods and index_mapping_methods.
// //
// Since we call back through the high level Python API, // Since we call back through the high level Python API,
}; };
self.cindex(py).borrow().inner().get_item(py, key) self.cindex(py).borrow().inner().get_item(py, key)
} }
def __setitem__(&self, key: PyObject, value: PyObject) -> PyResult<()> { def __setitem__(&self, key: PyObject, value: PyObject) -> PyResult<()> {
self.cindex(py).borrow().inner().set_item(py, key, value) self.cindex(py).borrow().inner().set_item(py, key, value)
} }
def __delitem__(&self, key: PyObject) -> PyResult<()> {
self.cindex(py).borrow().inner().del_item(py, key)
}
def __contains__(&self, item: PyObject) -> PyResult<bool> { def __contains__(&self, item: PyObject) -> PyResult<bool> {
// ObjectProtocol does not seem to provide contains(), so // ObjectProtocol does not seem to provide contains(), so
// this is an equivalent implementation of the index_contains() // this is an equivalent implementation of the index_contains()
// defined in revlog.c // defined in revlog.c
let cindex = self.cindex(py).borrow(); let cindex = self.cindex(py).borrow();
match item.extract::<Revision>(py) { match item.extract::<Revision>(py) {
Ok(rev) => { Ok(rev) => {
Ok(rev >= -1 && rev < cindex.inner().len(py)? as Revision) Ok(rev >= -1 && rev < cindex.inner().len(py)? as Revision)
}); });
impl MixedIndex { impl MixedIndex {
fn new(py: Python, cindex: PyObject) -> PyResult<MixedIndex> { fn new(py: Python, cindex: PyObject) -> PyResult<MixedIndex> {
Self::create_instance( Self::create_instance(
py, py,
RefCell::new(cindex::Index::new(py, cindex)?), RefCell::new(cindex::Index::new(py, cindex)?),
RefCell::new(None),
) )
} }
/// This is scaffolding at this point, but it could also become
/// a way to start a persistent nodemap or perform a
/// vacuum / repack operation
fn fill_nodemap(
&self,
py: Python,
nt: &mut NodeTree,
) -> PyResult<PyObject> {
let index = self.cindex(py).borrow();
for r in 0..index.len() {
let rev = r as Revision;
// in this case node() won't ever return None
nt.insert(&*index, index.node(rev).unwrap(), rev)
.map_err(|e| nodemap_error(py, e))?
}
Ok(py.None())
}
fn get_nodetree<'a>(
&'a self,
py: Python<'a>,
) -> PyResult<&'a RefCell<Option<NodeTree>>> {
if self.nt(py).borrow().is_none() {
let readonly = Box::new(Vec::new());
let mut nt = NodeTree::load_bytes(readonly, 0);
self.fill_nodemap(py, &mut nt)?;
self.nt(py).borrow_mut().replace(nt);
}
Ok(self.nt(py))
}
/// forward a method call to the underlying C index /// forward a method call to the underlying C index
fn call_cindex( fn call_cindex(
&self, &self,
py: Python, py: Python,
name: &str, name: &str,
args: &PyTuple, args: &PyTuple,
kwargs: Option<&PyDict>, kwargs: Option<&PyDict>,
) -> PyResult<PyObject> { ) -> PyResult<PyObject> {