##// END OF EJS Templates
exchange: ensure all outgoing subrepo references are present before pushing...
exchange: ensure all outgoing subrepo references are present before pushing We've run into occasional problems with people committing a repo, and then amending or rebasing in the subrepo. That makes it so that the revision in the parent can't be checked out, and the problem gets propagated on push. Mercurial already tries to defend against this sort of dangling reference by pushing *all* subrepo revisions first. This reuses the checks that trigger warnings in `hg verify` to bail on the push unless using `--force`. I thought about putting this on the server side, but at that point, all of the data has been transferred, only to bail out. Additionally, SCM Manager hosts subrepos in a location that isn't nested in the parent, so normal subrepo code would complain that the subrepo is missing when run on the server. Because the push command pushes subrepos before calling this exchange code, a subrepo will be pushed before the parent is verified. Not great, but no dangling references are exchanged, so it solves the problem. This code isn't in the loop that pushes the subrepos because: 1) the list of outgoing revisions is needed to limit the scope of the check 2) the loop only accesses the current revision, and therefore can miss subrepos that were dropped in previous commits 3) this code is called when pushing a subrepo, so the protection is recursive I'm not sure if there's a cheap check for the list of files in the outgoing bundle. If there is, that would provide a fast path to bypass this check for people not using subrepos (or if no subrepo changes were made). There's probably also room for verifying other references like tags. But since that doesn't break checkouts, it's much less of a problem. Differential Revision: https://phab.mercurial-scm.org/D7616

File last commit:

r43484:0246bbe1 default
r44365:4b7d5d10 default
Show More
cindex.rs
102 lines | 3.3 KiB | application/rls-services+xml | RustLexer
// cindex.rs
//
// Copyright 2018 Georges Racinet <gracinet@anybox.fr>
//
// This software may be used and distributed according to the terms of the
// GNU General Public License version 2 or any later version.
//! Bindings to use the Index defined by the parsers C extension
//!
//! Ideally, we should use an Index entirely implemented in Rust,
//! but this will take some time to get there.
use cpython::{PyClone, PyObject, PyResult, Python};
use hg::{Graph, GraphError, Revision, WORKING_DIRECTORY_REVISION};
use libc::c_int;
py_capsule_fn!(
from mercurial.cext.parsers import index_get_parents_CAPI
as get_parents_capi
signature (
index: *mut RawPyObject,
rev: c_int,
ps: *mut [c_int; 2],
) -> c_int
);
/// A `Graph` backed up by objects and functions from revlog.c
///
/// This implementation of the `Graph` trait, relies on (pointers to)
/// - the C index object (`index` member)
/// - the `index_get_parents()` function (`parents` member)
///
/// # Safety
///
/// The C index itself is mutable, and this Rust exposition is **not
/// protected by the GIL**, meaning that this construct isn't safe with respect
/// to Python threads.
///
/// All callers of this `Index` must acquire the GIL and must not release it
/// while working.
///
/// # TODO find a solution to make it GIL safe again.
///
/// This is non trivial, and can wait until we have a clearer picture with
/// more Rust Mercurial constructs.
///
/// One possibility would be to a `GILProtectedIndex` wrapper enclosing
/// a `Python<'p>` marker and have it be the one implementing the
/// `Graph` trait, but this would mean the `Graph` implementor would become
/// likely to change between subsequent method invocations of the `hg-core`
/// objects (a serious change of the `hg-core` API):
/// either exposing ways to mutate the `Graph`, or making it a non persistent
/// parameter in the relevant methods that need one.
///
/// Another possibility would be to introduce an abstract lock handle into
/// the core API, that would be tied to `GILGuard` / `Python<'p>`
/// in the case of the `cpython` crate bindings yet could leave room for other
/// mechanisms in other contexts.
pub struct Index {
index: PyObject,
parents: get_parents_capi::CapsuleFn,
}
impl Index {
pub fn new(py: Python, index: PyObject) -> PyResult<Self> {
Ok(Index {
index: index,
parents: get_parents_capi::retrieve(py)?,
})
}
}
impl Clone for Index {
fn clone(&self) -> Self {
let guard = Python::acquire_gil();
Index {
index: self.index.clone_ref(guard.python()),
parents: self.parents.clone(),
}
}
}
impl Graph for Index {
/// wrap a call to the C extern parents function
fn parents(&self, rev: Revision) -> Result<[Revision; 2], GraphError> {
if rev == WORKING_DIRECTORY_REVISION {
return Err(GraphError::WorkingDirectoryUnsupported);
}
let mut res: [c_int; 2] = [0; 2];
let code = unsafe {
(self.parents)(
self.index.as_ptr(),
rev as c_int,
&mut res as *mut [c_int; 2],
)
};
match code {
0 => Ok(res),
_ => Err(GraphError::ParentOutOfRange(rev)),
}
}
}