Files
ripgrep/src/main.rs
Andrew Gallant 8bbe58d623 Add support for additional text encodings.
This includes, but is not limited to, UTF-16, latin-1, GBK, EUC-JP and
Shift_JIS. (Courtesy of the `encoding_rs` crate.)

Specifically, this feature enables ripgrep to search files that are
encoded in an encoding other than UTF-8. The list of available encodings
is tied directly to what the `encoding_rs` crate supports, which is in
turn tied to the Encoding Standard. The full list of available encodings
can be found here: https://encoding.spec.whatwg.org/#concept-encoding-get

This pull request also introduces the notion that text encodings can be
automatically detected on a best effort basis. Currently, the only
support for this is checking for a UTF-16 bom. In all other cases, a
text encoding of `auto` (the default) implies a UTF-8 or ASCII
compatible source encoding. When a text encoding is otherwise specified,
it is unconditionally used for all files searched.

Since ripgrep's regex engine is fundamentally built on top of UTF-8,
this feature works by transcoding the files to be searched from their
source encoding to UTF-8. This transcoding only happens when:

1. `auto` is specified and a non-UTF-8 encoding is detected.
2. A specific encoding is given by end users (including UTF-8).

When transcoding occurs, errors are handled by automatically inserting
the Unicode replacement character. In this case, ripgrep's output is
guaranteed to be valid UTF-8 (excluding non-UTF-8 file paths, if they
are printed).

In all other cases, the source text is searched directly, which implies
an assumption that it is at least ASCII compatible, but where UTF-8 is
most useful. In this scenario, encoding errors are not detected. In this
case, ripgrep's output will match the input exactly, byte-for-byte.

This design may not be optimal in all cases, but it has some advantages:

1. In the happy path ("UTF-8 everywhere") remains happy. I have not been
   able to witness any performance regressions.
2. In the non-UTF-8 path, implementation complexity is kept relatively
   low. The cost here is transcoding itself. A potentially superior
   implementation might build decoding of any encoding into the regex
   engine itself. In particular, the fundamental problem with
   transcoding everything first is that literal optimizations are nearly
   negated.

Future work should entail improving the user experience. For example, we
might want to auto-detect more text encodings. A more elaborate UX
experience might permit end users to specify multiple text encodings,
although this seems hard to pull off in an ergonomic way.

Fixes #1
2017-03-12 19:54:48 -04:00

329 lines
9.0 KiB
Rust

extern crate atty;
extern crate bytecount;
#[macro_use]
extern crate clap;
extern crate encoding_rs;
extern crate env_logger;
extern crate grep;
extern crate ignore;
#[macro_use]
extern crate lazy_static;
extern crate libc;
#[macro_use]
extern crate log;
extern crate memchr;
extern crate memmap;
extern crate num_cpus;
extern crate regex;
extern crate same_file;
extern crate termcolor;
use std::error::Error;
use std::process;
use std::result;
use std::sync::Arc;
use std::sync::atomic::{AtomicUsize, Ordering};
use std::sync::mpsc;
use std::thread;
use args::Args;
use worker::Work;
macro_rules! errored {
($($tt:tt)*) => {
return Err(From::from(format!($($tt)*)));
}
}
macro_rules! eprintln {
($($tt:tt)*) => {{
use std::io::Write;
let _ = writeln!(&mut ::std::io::stderr(), $($tt)*);
}}
}
mod app;
mod args;
mod decoder;
mod pathutil;
mod printer;
mod search_buffer;
mod search_stream;
mod unescape;
mod worker;
pub type Result<T> = result::Result<T, Box<Error + Send + Sync>>;
fn main() {
match Args::parse().map(Arc::new).and_then(run) {
Ok(0) => process::exit(1),
Ok(_) => process::exit(0),
Err(err) => {
eprintln!("{}", err);
process::exit(1);
}
}
}
fn run(args: Arc<Args>) -> Result<u64> {
if args.never_match() {
return Ok(0);
}
let threads = args.threads();
if args.files() {
if threads == 1 || args.is_one_path() {
run_files_one_thread(args)
} else {
run_files_parallel(args)
}
} else if args.type_list() {
run_types(args)
} else if threads == 1 || args.is_one_path() {
run_one_thread(args)
} else {
run_parallel(args)
}
}
fn run_parallel(args: Arc<Args>) -> Result<u64> {
let bufwtr = Arc::new(args.buffer_writer());
let quiet_matched = args.quiet_matched();
let paths_searched = Arc::new(AtomicUsize::new(0));
let match_count = Arc::new(AtomicUsize::new(0));
args.walker_parallel().run(|| {
let args = args.clone();
let quiet_matched = quiet_matched.clone();
let paths_searched = paths_searched.clone();
let match_count = match_count.clone();
let bufwtr = bufwtr.clone();
let mut buf = bufwtr.buffer();
let mut worker = args.worker();
Box::new(move |result| {
use ignore::WalkState::*;
if quiet_matched.has_match() {
return Quit;
}
let dent = match get_or_log_dir_entry(
result,
args.stdout_handle(),
args.no_messages(),
) {
None => return Continue,
Some(dent) => dent,
};
paths_searched.fetch_add(1, Ordering::SeqCst);
buf.clear();
{
// This block actually executes the search and prints the
// results into outbuf.
let mut printer = args.printer(&mut buf);
let count =
if dent.is_stdin() {
worker.run(&mut printer, Work::Stdin)
} else {
worker.run(&mut printer, Work::DirEntry(dent))
};
match_count.fetch_add(count as usize, Ordering::SeqCst);
if quiet_matched.set_match(count > 0) {
return Quit;
}
}
// BUG(burntsushi): We should handle this error instead of ignoring
// it. See: https://github.com/BurntSushi/ripgrep/issues/200
let _ = bufwtr.print(&buf);
Continue
})
});
if !args.paths().is_empty() && paths_searched.load(Ordering::SeqCst) == 0 {
if !args.no_messages() {
eprint_nothing_searched();
}
}
Ok(match_count.load(Ordering::SeqCst) as u64)
}
fn run_one_thread(args: Arc<Args>) -> Result<u64> {
let stdout = args.stdout();
let mut stdout = stdout.lock();
let mut worker = args.worker();
let mut paths_searched: u64 = 0;
let mut match_count = 0;
for result in args.walker() {
let dent = match get_or_log_dir_entry(
result,
args.stdout_handle(),
args.no_messages(),
) {
None => continue,
Some(dent) => dent,
};
let mut printer = args.printer(&mut stdout);
if match_count > 0 {
if args.quiet() {
break;
}
if let Some(sep) = args.file_separator() {
printer = printer.file_separator(sep);
}
}
paths_searched += 1;
match_count +=
if dent.is_stdin() {
worker.run(&mut printer, Work::Stdin)
} else {
worker.run(&mut printer, Work::DirEntry(dent))
};
}
if !args.paths().is_empty() && paths_searched == 0 {
if !args.no_messages() {
eprint_nothing_searched();
}
}
Ok(match_count)
}
fn run_files_parallel(args: Arc<Args>) -> Result<u64> {
let print_args = args.clone();
let (tx, rx) = mpsc::channel::<ignore::DirEntry>();
let print_thread = thread::spawn(move || {
let stdout = print_args.stdout();
let mut printer = print_args.printer(stdout.lock());
let mut file_count = 0;
for dent in rx.iter() {
printer.path(dent.path());
file_count += 1;
}
file_count
});
args.walker_parallel().run(move || {
let args = args.clone();
let tx = tx.clone();
Box::new(move |result| {
if let Some(dent) = get_or_log_dir_entry(
result,
args.stdout_handle(),
args.no_messages(),
) {
tx.send(dent).unwrap();
}
ignore::WalkState::Continue
})
});
Ok(print_thread.join().unwrap())
}
fn run_files_one_thread(args: Arc<Args>) -> Result<u64> {
let stdout = args.stdout();
let mut printer = args.printer(stdout.lock());
let mut file_count = 0;
for result in args.walker() {
let dent = match get_or_log_dir_entry(
result,
args.stdout_handle(),
args.no_messages(),
) {
None => continue,
Some(dent) => dent,
};
printer.path(dent.path());
file_count += 1;
}
Ok(file_count)
}
fn run_types(args: Arc<Args>) -> Result<u64> {
let stdout = args.stdout();
let mut printer = args.printer(stdout.lock());
let mut ty_count = 0;
for def in args.type_defs() {
printer.type_def(def);
ty_count += 1;
}
Ok(ty_count)
}
fn get_or_log_dir_entry(
result: result::Result<ignore::DirEntry, ignore::Error>,
stdout_handle: Option<&same_file::Handle>,
no_messages: bool,
) -> Option<ignore::DirEntry> {
match result {
Err(err) => {
if !no_messages {
eprintln!("{}", err);
}
None
}
Ok(dent) => {
if let Some(err) = dent.error() {
if !no_messages {
eprintln!("{}", err);
}
}
let ft = match dent.file_type() {
None => return Some(dent), // entry is stdin
Some(ft) => ft,
};
// A depth of 0 means the user gave the path explicitly, so we
// should always try to search it.
if dent.depth() == 0 && !ft.is_dir() {
return Some(dent);
} else if !ft.is_file() {
return None;
}
// If we are redirecting stdout to a file, then don't search that
// file.
if is_stdout_file(&dent, stdout_handle, no_messages) {
return None;
}
Some(dent)
}
}
}
fn is_stdout_file(
dent: &ignore::DirEntry,
stdout_handle: Option<&same_file::Handle>,
no_messages: bool,
) -> bool {
let stdout_handle = match stdout_handle {
None => return false,
Some(stdout_handle) => stdout_handle,
};
// If we know for sure that these two things aren't equal, then avoid
// the costly extra stat call to determine equality.
if !maybe_dent_eq_handle(dent, stdout_handle) {
return false;
}
match same_file::Handle::from_path(dent.path()) {
Ok(h) => stdout_handle == &h,
Err(err) => {
if !no_messages {
eprintln!("{}: {}", dent.path().display(), err);
}
false
}
}
}
#[cfg(unix)]
fn maybe_dent_eq_handle(
dent: &ignore::DirEntry,
handle: &same_file::Handle,
) -> bool {
dent.ino() == Some(handle.ino())
}
#[cfg(not(unix))]
fn maybe_dent_eq_handle(_: &ignore::DirEntry, _: &same_file::Handle) -> bool {
true
}
fn eprint_nothing_searched() {
eprintln!("No files were searched, which means ripgrep probably \
applied a filter you didn't expect. \
Try running again with --debug.");
}