##// 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:

r47343:755c31a1 default
r47349:e9901d01 default
Show More
main.rs
191 lines | 5.8 KiB | application/rls-services+xml | RustLexer
extern crate log;
use crate::ui::Ui;
use clap::App;
use clap::AppSettings;
use clap::Arg;
use clap::ArgMatches;
use format_bytes::format_bytes;
use hg::config::Config;
use hg::repo::{Repo, RepoError};
use std::path::{Path, PathBuf};
mod blackbox;
mod error;
mod exitcode;
mod ui;
use error::CommandError;
fn add_global_args<'a, 'b>(app: App<'a, 'b>) -> App<'a, 'b> {
app.arg(
Arg::with_name("repository")
.help("repository root directory")
.short("-R")
.long("--repository")
.value_name("REPO")
.takes_value(true),
)
.arg(
Arg::with_name("config")
.help("set/override config option (use 'section.name=value')")
.long("--config")
.value_name("CONFIG")
.takes_value(true)
// Ok: `--config section.key1=val --config section.key2=val2`
.multiple(true)
// Not ok: `--config section.key1=val section.key2=val2`
.number_of_values(1),
)
}
fn main_with_result(
ui: &ui::Ui,
process_start_time: &blackbox::ProcessStartTime,
) -> Result<(), CommandError> {
env_logger::init();
let app = App::new("rhg")
.setting(AppSettings::AllowInvalidUtf8)
.setting(AppSettings::SubcommandRequired)
.setting(AppSettings::VersionlessSubcommands)
.version("0.0.1");
let app = add_global_args(app);
let app = add_subcommand_args(app);
let matches = app.clone().get_matches_safe()?;
let (subcommand_name, subcommand_matches) = matches.subcommand();
let run = subcommand_run_fn(subcommand_name)
.expect("unknown subcommand name from clap despite AppSettings::SubcommandRequired");
let subcommand_args = subcommand_matches
.expect("no subcommand arguments from clap despite AppSettings::SubcommandRequired");
// Global arguments can be in either based on e.g. `hg -R ./foo log` v.s.
// `hg log -R ./foo`
let value_of_global_arg = |name| {
subcommand_args
.value_of_os(name)
.or_else(|| matches.value_of_os(name))
};
// For arguments where multiple occurences are allowed, return a
// possibly-iterator of all values.
let values_of_global_arg = |name: &str| {
let a = matches.values_of_os(name).into_iter().flatten();
let b = subcommand_args.values_of_os(name).into_iter().flatten();
a.chain(b)
};
let config_args = values_of_global_arg("config")
.map(hg::utils::files::get_bytes_from_os_str);
let non_repo_config = &hg::config::Config::load(config_args)?;
let repo_path = value_of_global_arg("repository").map(Path::new);
let repo = match Repo::find(non_repo_config, repo_path) {
Ok(repo) => Ok(repo),
Err(RepoError::NotFound { at }) if repo_path.is_none() => {
// Not finding a repo is not fatal yet, if `-R` was not given
Err(NoRepoInCwdError { cwd: at })
}
Err(error) => return Err(error.into()),
};
let invocation = CliInvocation {
ui,
subcommand_args,
non_repo_config,
repo: repo.as_ref(),
};
let blackbox = blackbox::Blackbox::new(&invocation, process_start_time)?;
blackbox.log_command_start();
let result = run(&invocation);
blackbox.log_command_end(exit_code(&result));
result
}
fn main() {
// Run this first, before we find out if the blackbox extension is even
// enabled, in order to include everything in-between in the duration
// measurements. Reading config files can be slow if they’re on NFS.
let process_start_time = blackbox::ProcessStartTime::now();
let ui = ui::Ui::new();
let result = main_with_result(&ui, &process_start_time);
if let Err(CommandError::Abort { message }) = &result {
if !message.is_empty() {
// Ignore errors when writing to stderr, we’re already exiting
// with failure code so there’s not much more we can do.
let _ = ui.write_stderr(&format_bytes!(b"abort: {}\n", message));
}
}
std::process::exit(exit_code(&result))
}
fn exit_code(result: &Result<(), CommandError>) -> i32 {
match result {
Ok(()) => exitcode::OK,
Err(CommandError::Abort { .. }) => exitcode::ABORT,
// Exit with a specific code and no error message to let a potential
// wrapper script fallback to Python-based Mercurial.
Err(CommandError::Unimplemented) => exitcode::UNIMPLEMENTED,
}
}
macro_rules! subcommands {
($( $command: ident )+) => {
mod commands {
$(
pub mod $command;
)+
}
fn add_subcommand_args<'a, 'b>(app: App<'a, 'b>) -> App<'a, 'b> {
app
$(
.subcommand(add_global_args(commands::$command::args()))
)+
}
pub type RunFn = fn(&CliInvocation) -> Result<(), CommandError>;
fn subcommand_run_fn(name: &str) -> Option<RunFn> {
match name {
$(
stringify!($command) => Some(commands::$command::run),
)+
_ => None,
}
}
};
}
subcommands! {
cat
debugdata
debugrequirements
files
root
config
}
pub struct CliInvocation<'a> {
ui: &'a Ui,
subcommand_args: &'a ArgMatches<'a>,
non_repo_config: &'a Config,
/// References inside `Result` is a bit peculiar but allow
/// `invocation.repo?` to work out with `&CliInvocation` since this
/// `Result` type is `Copy`.
repo: Result<&'a Repo, &'a NoRepoInCwdError>,
}
struct NoRepoInCwdError {
cwd: PathBuf,
}
impl CliInvocation<'_> {
fn config(&self) -> &Config {
if let Ok(repo) = self.repo {
repo.config()
} else {
self.non_repo_config
}
}
}