##// END OF EJS Templates
revlog: add a mechanism to verify expected file position before appending...
revlog: add a mechanism to verify expected file position before appending If someone uses `hg debuglocks`, or some non-hg process writes to the .hg directory without respecting the locks, or if the repo's on a networked filesystem, it's possible for the revlog code to write out corrupted data. The form of this corruption can vary depending on what data was written and how that happened. We are in the "networked filesystem" case (though I've had users also do this to themselves with the "`hg debuglocks`" scenario), and most often see this with the changelog. What ends up happening is we produce two items (let's call them rev1 and rev2) in the .i file that have the same linkrev, baserev, and offset into the .d file, while the data in the .d file is appended properly. rev2's compressed_size is accurate for rev2, but when we go to decompress the data in the .d file, we use the offset that's recorded in the index file, which is the same as rev1, and attempt to decompress rev2.compressed_size bytes of rev1's data. This usually does not succeed. :) When using inline data, this also fails, though I haven't investigated why too closely. This shows up as a "patch decode" error. I believe what's happening there is that we're basically ignoring the offset field, getting the data properly, but since baserev != rev, it thinks this is a delta based on rev (instead of a full text) and can't actually apply it as such. For now, I'm going to make this an optional component and default it to entirely off. I may increase the default severity of this in the future, once I've enabled it for my users and we gain more experience with it. Luckily, most of my users have a versioned filesystem and can roll back to before the corruption has been written, it's just a hassle to do so and not everyone knows how (so it's a support burden). Users on other filesystems will not have that luxury, and this can cause them to have a corrupted repository that they are unlikely to know how to resolve, and they'll see this as a data-loss event. Refusing to create the corruption is a much better user experience. This mechanism is not perfect. There may be false-negatives (racy writes that are not detected). There should not be any false-positives (non-racy writes that are detected as such). This is not a mechanism that makes putting a repo on a networked filesystem "safe" or "supported", just *less* likely to cause corruption. Differential Revision: https://phab.mercurial-scm.org/D9952

File last commit:

r47341:1f55cd5b default
r47349:e9901d01 default
Show More
errors.rs
192 lines | 6.3 KiB | application/rls-services+xml | RustLexer
use crate::config::ConfigValueParseError;
use std::fmt;
/// Common error cases that can happen in many different APIs
#[derive(Debug, derive_more::From)]
pub enum HgError {
IoError {
error: std::io::Error,
context: IoErrorContext,
},
/// A file under `.hg/` normally only written by Mercurial is not in the
/// expected format. This indicates a bug in Mercurial, filesystem
/// corruption, or hardware failure.
///
/// The given string is a short explanation for users, not intended to be
/// machine-readable.
CorruptedRepository(String),
/// The respository or requested operation involves a feature not
/// supported by the Rust implementation. Falling back to the Python
/// implementation may or may not work.
///
/// The given string is a short explanation for users, not intended to be
/// machine-readable.
UnsupportedFeature(String),
/// Operation cannot proceed for some other reason.
///
/// The given string is a short explanation for users, not intended to be
/// machine-readable.
Abort(String),
/// A configuration value is not in the expected syntax.
///
/// These errors can happen in many places in the code because values are
/// parsed lazily as the file-level parser does not know the expected type
/// and syntax of each value.
#[from]
ConfigValueParseError(ConfigValueParseError),
}
/// Details about where an I/O error happened
#[derive(Debug)]
pub enum IoErrorContext {
ReadingFile(std::path::PathBuf),
WritingFile(std::path::PathBuf),
RemovingFile(std::path::PathBuf),
RenamingFile {
from: std::path::PathBuf,
to: std::path::PathBuf,
},
/// `std::env::current_dir`
CurrentDir,
/// `std::env::current_exe`
CurrentExe,
}
impl HgError {
pub fn corrupted(explanation: impl Into<String>) -> Self {
// TODO: capture a backtrace here and keep it in the error value
// to aid debugging?
// https://doc.rust-lang.org/std/backtrace/struct.Backtrace.html
HgError::CorruptedRepository(explanation.into())
}
pub fn unsupported(explanation: impl Into<String>) -> Self {
HgError::UnsupportedFeature(explanation.into())
}
pub fn abort(explanation: impl Into<String>) -> Self {
HgError::Abort(explanation.into())
}
}
// TODO: use `DisplayBytes` instead to show non-Unicode filenames losslessly?
impl fmt::Display for HgError {
fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
match self {
HgError::Abort(explanation) => write!(f, "{}", explanation),
HgError::IoError { error, context } => {
write!(f, "{}: {}", error, context)
}
HgError::CorruptedRepository(explanation) => {
write!(f, "corrupted repository: {}", explanation)
}
HgError::UnsupportedFeature(explanation) => {
write!(f, "unsupported feature: {}", explanation)
}
HgError::ConfigValueParseError(ConfigValueParseError {
origin: _,
line: _,
section,
item,
value,
expected_type,
}) => {
// TODO: add origin and line number information, here and in
// corresponding python code
write!(
f,
"config error: {}.{} is not a {} ('{}')",
String::from_utf8_lossy(section),
String::from_utf8_lossy(item),
expected_type,
String::from_utf8_lossy(value)
)
}
}
}
}
// TODO: use `DisplayBytes` instead to show non-Unicode filenames losslessly?
impl fmt::Display for IoErrorContext {
fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
match self {
IoErrorContext::ReadingFile(path) => {
write!(f, "when reading {}", path.display())
}
IoErrorContext::WritingFile(path) => {
write!(f, "when writing {}", path.display())
}
IoErrorContext::RemovingFile(path) => {
write!(f, "when removing {}", path.display())
}
IoErrorContext::RenamingFile { from, to } => write!(
f,
"when renaming {} to {}",
from.display(),
to.display()
),
IoErrorContext::CurrentDir => write!(f, "current directory"),
IoErrorContext::CurrentExe => write!(f, "current executable"),
}
}
}
pub trait IoResultExt<T> {
/// Annotate a possible I/O error as related to a reading a file at the
/// given path.
///
/// This allows printing something like “File not found when reading
/// example.txt” instead of just “File not found”.
///
/// Converts a `Result` with `std::io::Error` into one with `HgError`.
fn when_reading_file(self, path: &std::path::Path) -> Result<T, HgError>;
fn with_context(
self,
context: impl FnOnce() -> IoErrorContext,
) -> Result<T, HgError>;
}
impl<T> IoResultExt<T> for std::io::Result<T> {
fn when_reading_file(self, path: &std::path::Path) -> Result<T, HgError> {
self.with_context(|| IoErrorContext::ReadingFile(path.to_owned()))
}
fn with_context(
self,
context: impl FnOnce() -> IoErrorContext,
) -> Result<T, HgError> {
self.map_err(|error| HgError::IoError {
error,
context: context(),
})
}
}
pub trait HgResultExt<T> {
/// Handle missing files separately from other I/O error cases.
///
/// Wraps the `Ok` type in an `Option`:
///
/// * `Ok(x)` becomes `Ok(Some(x))`
/// * An I/O "not found" error becomes `Ok(None)`
/// * Other errors are unchanged
fn io_not_found_as_none(self) -> Result<Option<T>, HgError>;
}
impl<T> HgResultExt<T> for Result<T, HgError> {
fn io_not_found_as_none(self) -> Result<Option<T>, HgError> {
match self {
Ok(x) => Ok(Some(x)),
Err(HgError::IoError { error, .. })
if error.kind() == std::io::ErrorKind::NotFound =>
{
Ok(None)
}
Err(other_error) => Err(other_error),
}
}
}