Google Git
Sign in
chromium / external / github.com / rust-lang / rust / HEAD / . / library / std / src / io / mod.rs
blob: 25a4661a0bc9ccfeb16eb1b3be4434ba6c471fe5 [file] [log] [blame]
//! Traits, helpers, and type definitions for core I/O functionality.
//!
//! The `std::io` module contains a number of common things you'll need
//! when doing input and output. The most core part of this module is
//! the [`Read`] and [`Write`] traits, which provide the
//! most general interface for reading and writing input and output.
//!
//! ## Read and Write
//!
//! Because they are traits, [`Read`] and [`Write`] are implemented by a number
//! of other types, and you can implement them for your types too. As such,
//! you'll see a few different types of I/O throughout the documentation in
//! this module: [`File`]s, [`TcpStream`]s, and sometimes even [`Vec<T>`]s. For
//! example, [`Read`] adds a [`read`][`Read::read`] method, which we can use on
//! [`File`]s:
//!
//! ```no_run
//! use std::io;
//! use std::io::prelude::*;
//! use std::fs::File;
//!
//! fn main() -> io::Result<()> {
//! let mut f = File::open("foo.txt")?;
//! let mut buffer = [0; 10];
//!
//! // read up to 10 bytes
//! let n = f.read(&mut buffer)?;
//!
//! println!("The bytes: {:?}", &buffer[..n]);
//! Ok(())
//! }
//! ```
//!
//! [`Read`] and [`Write`] are so important, implementors of the two traits have a
//! nickname: readers and writers. So you'll sometimes see 'a reader' instead
//! of 'a type that implements the [`Read`] trait'. Much easier!
//!
//! ## Seek and BufRead
//!
//! Beyond that, there are two important traits that are provided: [`Seek`]
//! and [`BufRead`]. Both of these build on top of a reader to control
//! how the reading happens. [`Seek`] lets you control where the next byte is
//! coming from:
//!
//! ```no_run
//! use std::io;
//! use std::io::prelude::*;
//! use std::io::SeekFrom;
//! use std::fs::File;
//!
//! fn main() -> io::Result<()> {
//! let mut f = File::open("foo.txt")?;
//! let mut buffer = [0; 10];
//!
//! // skip to the last 10 bytes of the file
//! f.seek(SeekFrom::End(-10))?;
//!
//! // read up to 10 bytes
//! let n = f.read(&mut buffer)?;
//!
//! println!("The bytes: {:?}", &buffer[..n]);
//! Ok(())
//! }
//! ```
//!
//! [`BufRead`] uses an internal buffer to provide a number of other ways to read, but
//! to show it off, we'll need to talk about buffers in general. Keep reading!
//!
//! ## BufReader and BufWriter
//!
//! Byte-based interfaces are unwieldy and can be inefficient, as we'd need to be
//! making near-constant calls to the operating system. To help with this,
//! `std::io` comes with two structs, [`BufReader`] and [`BufWriter`], which wrap
//! readers and writers. The wrapper uses a buffer, reducing the number of
//! calls and providing nicer methods for accessing exactly what you want.
//!
//! For example, [`BufReader`] works with the [`BufRead`] trait to add extra
//! methods to any reader:
//!
//! ```no_run
//! use std::io;
//! use std::io::prelude::*;
//! use std::io::BufReader;
//! use std::fs::File;
//!
//! fn main() -> io::Result<()> {
//! let f = File::open("foo.txt")?;
//! let mut reader = BufReader::new(f);
//! let mut buffer = String::new();
//!
//! // read a line into buffer
//! reader.read_line(&mut buffer)?;
//!
//! println!("{buffer}");
//! Ok(())
//! }
//! ```
//!
//! [`BufWriter`] doesn't add any new ways of writing; it just buffers every call
//! to [`write`][`Write::write`]:
//!
//! ```no_run
//! use std::io;
//! use std::io::prelude::*;
//! use std::io::BufWriter;
//! use std::fs::File;
//!
//! fn main() -> io::Result<()> {
//! let f = File::create("foo.txt")?;
//! {
//! let mut writer = BufWriter::new(f);
//!
//! // write a byte to the buffer
//! writer.write(&[42])?;
//!
//! } // the buffer is flushed once writer goes out of scope
//!
//! Ok(())
//! }
//! ```
//!
//! ## Standard input and output
//!
//! A very common source of input is standard input:
//!
//! ```no_run
//! use std::io;
//!
//! fn main() -> io::Result<()> {
//! let mut input = String::new();
//!
//! io::stdin().read_line(&mut input)?;
//!
//! println!("You typed: {}", input.trim());
//! Ok(())
//! }
//! ```
//!
//! Note that you cannot use the [`?` operator] in functions that do not return
//! a [`Result<T, E>`][`Result`]. Instead, you can call [`.unwrap()`]
//! or `match` on the return value to catch any possible errors:
//!
//! ```no_run
//! use std::io;
//!
//! let mut input = String::new();
//!
//! io::stdin().read_line(&mut input).unwrap();
//! ```
//!
//! And a very common source of output is standard output:
//!
//! ```no_run
//! use std::io;
//! use std::io::prelude::*;
//!
//! fn main() -> io::Result<()> {
//! io::stdout().write(&[42])?;
//! Ok(())
//! }
//! ```
//!
//! Of course, using [`io::stdout`] directly is less common than something like
//! [`println!`].
//!
//! ## Iterator types
//!
//! A large number of the structures provided by `std::io` are for various
//! ways of iterating over I/O. For example, [`Lines`] is used to split over
//! lines:
//!
//! ```no_run
//! use std::io;
//! use std::io::prelude::*;
//! use std::io::BufReader;
//! use std::fs::File;
//!
//! fn main() -> io::Result<()> {
//! let f = File::open("foo.txt")?;
//! let reader = BufReader::new(f);
//!
//! for line in reader.lines() {
//! println!("{}", line?);
//! }
//! Ok(())
//! }
//! ```
//!
//! ## Functions
//!
//! There are a number of [functions][functions-list] that offer access to various
//! features. For example, we can use three of these functions to copy everything
//! from standard input to standard output:
//!
//! ```no_run
//! use std::io;
//!
//! fn main() -> io::Result<()> {
//! io::copy(&mut io::stdin(), &mut io::stdout())?;
//! Ok(())
//! }
//! ```
//!
//! [functions-list]: #functions-1
//!
//! ## io::Result
//!
//! Last, but certainly not least, is [`io::Result`]. This type is used
//! as the return type of many `std::io` functions that can cause an error, and
//! can be returned from your own functions as well. Many of the examples in this
//! module use the [`?` operator]:
//!
//! ```
//! use std::io;
//!
//! fn read_input() -> io::Result<()> {
//! let mut input = String::new();
//!
//! io::stdin().read_line(&mut input)?;
//!
//! println!("You typed: {}", input.trim());
//!
//! Ok(())
//! }
//! ```
//!
//! The return type of `read_input()`, [`io::Result<()>`][`io::Result`], is a very
//! common type for functions which don't have a 'real' return value, but do want to
//! return errors if they happen. In this case, the only purpose of this function is
//! to read the line and print it, so we use `()`.
//!
//! ## Platform-specific behavior
//!
//! Many I/O functions throughout the standard library are documented to indicate
//! what various library or syscalls they are delegated to. This is done to help
//! applications both understand what's happening under the hood as well as investigate
//! any possibly unclear semantics. Note, however, that this is informative, not a binding
//! contract. The implementation of many of these functions are subject to change over
//! time and may call fewer or more syscalls/library functions.
//!
//! ## I/O Safety
//!
//! Rust follows an I/O safety discipline that is comparable to its memory safety discipline. This
//! means that file descriptors can be *exclusively owned*. (Here, "file descriptor" is meant to
//! subsume similar concepts that exist across a wide range of operating systems even if they might
//! use a different name, such as "handle".) An exclusively owned file descriptor is one that no
//! other code is allowed to access in any way, but the owner is allowed to access and even close
//! it any time. A type that owns its file descriptor should usually close it in its `drop`
//! function. Types like [`File`] own their file descriptor. Similarly, file descriptors
//! can be *borrowed*, granting the temporary right to perform operations on this file descriptor.
//! This indicates that the file descriptor will not be closed for the lifetime of the borrow, but
//! it does *not* imply any right to close this file descriptor, since it will likely be owned by
//! someone else.
//!
//! The platform-specific parts of the Rust standard library expose types that reflect these
//! concepts, see [`os::unix`] and [`os::windows`].
//!
//! To uphold I/O safety, it is crucial that no code acts on file descriptors it does not own or
//! borrow, and no code closes file descriptors it does not own. In other words, a safe function
//! that takes a regular integer, treats it as a file descriptor, and acts on it, is *unsound*.
//!
//! Not upholding I/O safety and acting on a file descriptor without proof of ownership can lead to
//! misbehavior and even Undefined Behavior in code that relies on ownership of its file
//! descriptors: a closed file descriptor could be re-allocated, so the original owner of that file
//! descriptor is now working on the wrong file. Some code might even rely on fully encapsulating
//! its file descriptors with no operations being performed by any other part of the program.
//!
//! Note that exclusive ownership of a file descriptor does *not* imply exclusive ownership of the
//! underlying kernel object that the file descriptor references (also called "open file description" on
//! some operating systems). File descriptors basically work like [`Arc`]: when you receive an owned
//! file descriptor, you cannot know whether there are any other file descriptors that reference the
//! same kernel object. However, when you create a new kernel object, you know that you are holding
//! the only reference to it. Just be careful not to lend it to anyone, since they can obtain a
//! clone and then you can no longer know what the reference count is! In that sense, [`OwnedFd`] is
//! like `Arc` and [`BorrowedFd<'a>`] is like `&'a Arc` (and similar for the Windows types). In
//! particular, given a `BorrowedFd<'a>`, you are not allowed to close the file descriptor -- just
//! like how, given a `&'a Arc`, you are not allowed to decrement the reference count and
//! potentially free the underlying object. There is no equivalent to `Box` for file descriptors in
//! the standard library (that would be a type that guarantees that the reference count is `1`),
//! however, it would be possible for a crate to define a type with those semantics.
//!
//! [`File`]: crate::fs::File
//! [`TcpStream`]: crate::net::TcpStream
//! [`io::stdout`]: stdout
//! [`io::Result`]: self::Result
//! [`?` operator]: ../../book/appendix-02-operators.html
//! [`Result`]: crate::result::Result
//! [`.unwrap()`]: crate::result::Result::unwrap
//! [`os::unix`]: ../os/unix/io/index.html
//! [`os::windows`]: ../os/windows/io/index.html
//! [`OwnedFd`]: ../os/fd/struct.OwnedFd.html
//! [`BorrowedFd<'a>`]: ../os/fd/struct.BorrowedFd.html
//! [`Arc`]: crate::sync::Arc
#![stable(feature = "rust1", since = "1.0.0")]
#[cfg(test)]
mod tests;
#[unstable(feature = "read_buf", issue = "78485")]
pub use core::io::{BorrowedBuf, BorrowedCursor};
use core::slice::memchr;
#[stable(feature = "bufwriter_into_parts", since = "1.56.0")]
pub use self::buffered::WriterPanicked;
#[unstable(feature = "raw_os_error_ty", issue = "107792")]
pub use self::error::RawOsError;
#[doc(hidden)]
#[unstable(feature = "io_const_error_internals", issue = "none")]
pub use self::error::SimpleMessage;
#[unstable(feature = "io_const_error", issue = "133448")]
pub use self::error::const_error;
#[stable(feature = "anonymous_pipe", since = "1.87.0")]
pub use self::pipe::{PipeReader, PipeWriter, pipe};
#[stable(feature = "is_terminal", since = "1.70.0")]
pub use self::stdio::IsTerminal;
pub(crate) use self::stdio::attempt_print_to_stderr;
#[unstable(feature = "print_internals", issue = "none")]
#[doc(hidden)]
pub use self::stdio::{_eprint, _print};
#[unstable(feature = "internal_output_capture", issue = "none")]
#[doc(no_inline, hidden)]
pub use self::stdio::{set_output_capture, try_set_output_capture};
#[stable(feature = "rust1", since = "1.0.0")]
pub use self::{
buffered::{BufReader, BufWriter, IntoInnerError, LineWriter},
copy::copy,
cursor::Cursor,
error::{Error, ErrorKind, Result},
stdio::{Stderr, StderrLock, Stdin, StdinLock, Stdout, StdoutLock, stderr, stdin, stdout},
util::{Empty, Repeat, Sink, empty, repeat, sink},
};
use crate::mem::take;
use crate::ops::{Deref, DerefMut};
use crate::{cmp, fmt, slice, str, sys};
mod buffered;
pub(crate) mod copy;
mod cursor;
mod error;
mod impls;
mod pipe;
pub mod prelude;
mod stdio;
mod util;
const DEFAULT_BUF_SIZE: usize = crate::sys::io::DEFAULT_BUF_SIZE;
pub(crate) use stdio::cleanup;
struct Guard<'a> {
buf: &'a mut Vec<u8>,
len: usize,
}
impl Drop for Guard<'_> {
fn drop(&mut self) {
unsafe {
self.buf.set_len(self.len);
}
}
}
// Several `read_to_string` and `read_line` methods in the standard library will
// append data into a `String` buffer, but we need to be pretty careful when
// doing this. The implementation will just call `.as_mut_vec()` and then
// delegate to a byte-oriented reading method, but we must ensure that when
// returning we never leave `buf` in a state such that it contains invalid UTF-8
// in its bounds.
//
// To this end, we use an RAII guard (to protect against panics) which updates
// the length of the string when it is dropped. This guard initially truncates
// the string to the prior length and only after we've validated that the
// new contents are valid UTF-8 do we allow it to set a longer length.
//
// The unsafety in this function is twofold:
//
// 1. We're looking at the raw bytes of `buf`, so we take on the burden of UTF-8
// checks.
// 2. We're passing a raw buffer to the function `f`, and it is expected that
// the function only *appends* bytes to the buffer. We'll get undefined
// behavior if existing bytes are overwritten to have non-UTF-8 data.
pub(crate) unsafe fn append_to_string<F>(buf: &mut String, f: F) -> Result<usize>
where
F: FnOnce(&mut Vec<u8>) -> Result<usize>,
{
let mut g = Guard { len: buf.len(), buf: unsafe { buf.as_mut_vec() } };
let ret = f(g.buf);
// SAFETY: the caller promises to only append data to `buf`
let appended = unsafe { g.buf.get_unchecked(g.len..) };
if str::from_utf8(appended).is_err() {
ret.and_then(|_| Err(Error::INVALID_UTF8))
} else {
g.len = g.buf.len();
ret
}
}
// Here we must serve many masters with conflicting goals:
//
// - avoid allocating unless necessary
// - avoid overallocating if we know the exact size (#89165)
// - avoid passing large buffers to readers that always initialize the free capacity if they perform short reads (#23815, #23820)
// - pass large buffers to readers that do not initialize the spare capacity. this can amortize per-call overheads
// - and finally pass not-too-small and not-too-large buffers to Windows read APIs because they manage to suffer from both problems
// at the same time, i.e. small reads suffer from syscall overhead, all reads incur costs proportional to buffer size (#110650)
//
pub(crate) fn default_read_to_end<R: Read + ?Sized>(
r: &mut R,
buf: &mut Vec<u8>,
size_hint: Option<usize>,
) -> Result<usize> {
let start_len = buf.len();
let start_cap = buf.capacity();
// Optionally limit the maximum bytes read on each iteration.
// This adds an arbitrary fiddle factor to allow for more data than we expect.
let mut max_read_size = size_hint
.and_then(|s| s.checked_add(1024)?.checked_next_multiple_of(DEFAULT_BUF_SIZE))
.unwrap_or(DEFAULT_BUF_SIZE);
let mut initialized = 0; // Extra initialized bytes from previous loop iteration
const PROBE_SIZE: usize = 32;
fn small_probe_read<R: Read + ?Sized>(r: &mut R, buf: &mut Vec<u8>) -> Result<usize> {
let mut probe = [0u8; PROBE_SIZE];
loop {
match r.read(&mut probe) {
Ok(n) => {
// there is no way to recover from allocation failure here
// because the data has already been read.
buf.extend_from_slice(&probe[..n]);
return Ok(n);
}
Err(ref e) if e.is_interrupted() => continue,
Err(e) => return Err(e),
}
}
}
// avoid inflating empty/small vecs before we have determined that there's anything to read
if (size_hint.is_none() || size_hint == Some(0)) && buf.capacity() - buf.len() < PROBE_SIZE {
let read = small_probe_read(r, buf)?;
if read == 0 {
return Ok(0);
}
}
let mut consecutive_short_reads = 0;
loop {
if buf.len() == buf.capacity() && buf.capacity() == start_cap {
// The buffer might be an exact fit. Let's read into a probe buffer
// and see if it returns `Ok(0)`. If so, we've avoided an
// unnecessary doubling of the capacity. But if not, append the
// probe buffer to the primary buffer and let its capacity grow.
let read = small_probe_read(r, buf)?;
if read == 0 {
return Ok(buf.len() - start_len);
}
}
if buf.len() == buf.capacity() {
// buf is full, need more space
buf.try_reserve(PROBE_SIZE)?;
}
let mut spare = buf.spare_capacity_mut();
let buf_len = cmp::min(spare.len(), max_read_size);
spare = &mut spare[..buf_len];
let mut read_buf: BorrowedBuf<'_> = spare.into();
// SAFETY: These bytes were initialized but not filled in the previous loop
unsafe {
read_buf.set_init(initialized);
}
let mut cursor = read_buf.unfilled();
let result = loop {
match r.read_buf(cursor.reborrow()) {
Err(e) if e.is_interrupted() => continue,
// Do not stop now in case of error: we might have received both data
// and an error
res => break res,
}
};
let unfilled_but_initialized = cursor.init_mut().len();
let bytes_read = cursor.written();
let was_fully_initialized = read_buf.init_len() == buf_len;
// SAFETY: BorrowedBuf's invariants mean this much memory is initialized.
unsafe {
let new_len = bytes_read + buf.len();
buf.set_len(new_len);
}
// Now that all data is pushed to the vector, we can fail without data loss
result?;
if bytes_read == 0 {
return Ok(buf.len() - start_len);
}
if bytes_read < buf_len {
consecutive_short_reads += 1;
} else {
consecutive_short_reads = 0;
}
// store how much was initialized but not filled
initialized = unfilled_but_initialized;
// Use heuristics to determine the max read size if no initial size hint was provided
if size_hint.is_none() {
// The reader is returning short reads but it doesn't call ensure_init().
// In that case we no longer need to restrict read sizes to avoid
// initialization costs.
// When reading from disk we usually don't get any short reads except at EOF.
// So we wait for at least 2 short reads before uncapping the read buffer;
// this helps with the Windows issue.
if !was_fully_initialized && consecutive_short_reads > 1 {
max_read_size = usize::MAX;
}
// we have passed a larger buffer than previously and the
// reader still hasn't returned a short read
if buf_len >= max_read_size && bytes_read == buf_len {
max_read_size = max_read_size.saturating_mul(2);
}
}
}
}
pub(crate) fn default_read_to_string<R: Read + ?Sized>(
r: &mut R,
buf: &mut String,
size_hint: Option<usize>,
) -> Result<usize> {
// Note that we do *not* call `r.read_to_end()` here. We are passing
// `&mut Vec<u8>` (the raw contents of `buf`) into the `read_to_end`
// method to fill it up. An arbitrary implementation could overwrite the
// entire contents of the vector, not just append to it (which is what
// we are expecting).
//
// To prevent extraneously checking the UTF-8-ness of the entire buffer
// we pass it to our hardcoded `default_read_to_end` implementation which
// we know is guaranteed to only read data into the end of the buffer.
unsafe { append_to_string(buf, |b| default_read_to_end(r, b, size_hint)) }
}
pub(crate) fn default_read_vectored<F>(read: F, bufs: &mut [IoSliceMut<'_>]) -> Result<usize>
where
F: FnOnce(&mut [u8]) -> Result<usize>,
{
let buf = bufs.iter_mut().find(|b| !b.is_empty()).map_or(&mut [][..], |b| &mut **b);
read(buf)
}
pub(crate) fn default_write_vectored<F>(write: F, bufs: &[IoSlice<'_>]) -> Result<usize>
where
F: FnOnce(&[u8]) -> Result<usize>,
{
let buf = bufs.iter().find(|b| !b.is_empty()).map_or(&[][..], |b| &**b);
write(buf)
}
pub(crate) fn default_read_exact<R: Read + ?Sized>(this: &mut R, mut buf: &mut [u8]) -> Result<()> {
while !buf.is_empty() {
match this.read(buf) {
Ok(0) => break,
Ok(n) => {
buf = &mut buf[n..];
}
Err(ref e) if e.is_interrupted() => {}
Err(e) => return Err(e),
}
}
if !buf.is_empty() { Err(Error::READ_EXACT_EOF) } else { Ok(()) }
}
pub(crate) fn default_read_buf<F>(read: F, mut cursor: BorrowedCursor<'_>) -> Result<()>
where
F: FnOnce(&mut [u8]) -> Result<usize>,
{
let n = read(cursor.ensure_init().init_mut())?;
cursor.advance(n);
Ok(())
}
pub(crate) fn default_read_buf_exact<R: Read + ?Sized>(
this: &mut R,
mut cursor: BorrowedCursor<'_>,
) -> Result<()> {
while cursor.capacity() > 0 {
let prev_written = cursor.written();
match this.read_buf(cursor.reborrow()) {
Ok(()) => {}
Err(e) if e.is_interrupted() => continue,
Err(e) => return Err(e),
}
if cursor.written() == prev_written {
return Err(Error::READ_EXACT_EOF);
}
}
Ok(())
}
pub(crate) fn default_write_fmt<W: Write + ?Sized>(
this: &mut W,
args: fmt::Arguments<'_>,
) -> Result<()> {
// Create a shim which translates a `Write` to a `fmt::Write` and saves off
// I/O errors, instead of discarding them.
struct Adapter<'a, T: ?Sized + 'a> {
inner: &'a mut T,
error: Result<()>,
}
impl<T: Write + ?Sized> fmt::Write for Adapter<'_, T> {
fn write_str(&mut self, s: &str) -> fmt::Result {
match self.inner.write_all(s.as_bytes()) {
Ok(()) => Ok(()),
Err(e) => {
self.error = Err(e);
Err(fmt::Error)
}
}
}
}
let mut output = Adapter { inner: this, error: Ok(()) };
match fmt::write(&mut output, args) {
Ok(()) => Ok(()),
Err(..) => {
// Check whether the error came from the underlying `Write`.
if output.error.is_err() {
output.error
} else {
// This shouldn't happen: the underlying stream did not error,
// but somehow the formatter still errored?
panic!(
"a formatting trait implementation returned an error when the underlying stream did not"
);
}
}
}
}
/// The `Read` trait allows for reading bytes from a source.
///
/// Implementors of the `Read` trait are called 'readers'.
///
/// Readers are defined by one required method, [`read()`]. Each call to [`read()`]
/// will attempt to pull bytes from this source into a provided buffer. A
/// number of other methods are implemented in terms of [`read()`], giving
/// implementors a number of ways to read bytes while only needing to implement
/// a single method.
///
/// Readers are intended to be composable with one another. Many implementors
/// throughout [`std::io`] take and provide types which implement the `Read`
/// trait.
///
/// Please note that each call to [`read()`] may involve a system call, and
/// therefore, using something that implements [`BufRead`], such as
/// [`BufReader`], will be more efficient.
///
/// Repeated calls to the reader use the same cursor, so for example
/// calling `read_to_end` twice on a [`File`] will only return the file's
/// contents once. It's recommended to first call `rewind()` in that case.
///
/// # Examples
///
/// [`File`]s implement `Read`:
///
/// ```no_run
/// use std::io;
/// use std::io::prelude::*;
/// use std::fs::File;
///
/// fn main() -> io::Result<()> {
/// let mut f = File::open("foo.txt")?;
/// let mut buffer = [0; 10];
///
/// // read up to 10 bytes
/// f.read(&mut buffer)?;
///
/// let mut buffer = Vec::new();
/// // read the whole file
/// f.read_to_end(&mut buffer)?;
///
/// // read into a String, so that you don't need to do the conversion.
/// let mut buffer = String::new();
/// f.read_to_string(&mut buffer)?;
///
/// // and more! See the other methods for more details.
/// Ok(())
/// }
/// ```
///
/// Read from [`&str`] because [`&[u8]`][prim@slice] implements `Read`:
///
/// ```no_run
/// # use std::io;
/// use std::io::prelude::*;
///
/// fn main() -> io::Result<()> {
/// let mut b = "This string will be read".as_bytes();
/// let mut buffer = [0; 10];
///
/// // read up to 10 bytes
/// b.read(&mut buffer)?;
///
/// // etc... it works exactly as a File does!
/// Ok(())
/// }
/// ```
///
/// [`read()`]: Read::read
/// [`&str`]: prim@str
/// [`std::io`]: self
/// [`File`]: crate::fs::File
#[stable(feature = "rust1", since = "1.0.0")]
#[doc(notable_trait)]
#[cfg_attr(not(test), rustc_diagnostic_item = "IoRead")]
pub trait Read {
/// Pull some bytes from this source into the specified buffer, returning
/// how many bytes were read.
///
/// This function does not provide any guarantees about whether it blocks
/// waiting for data, but if an object needs to block for a read and cannot,
/// it will typically signal this via an [`Err`] return value.
///
/// If the return value of this method is [`Ok(n)`], then implementations must
/// guarantee that `0 <= n <= buf.len()`. A nonzero `n` value indicates
/// that the buffer `buf` has been filled in with `n` bytes of data from this
/// source. If `n` is `0`, then it can indicate one of two scenarios:
///
/// 1. This reader has reached its "end of file" and will likely no longer
/// be able to produce bytes. Note that this does not mean that the
/// reader will *always* no longer be able to produce bytes. As an example,
/// on Linux, this method will call the `recv` syscall for a [`TcpStream`],
/// where returning zero indicates the connection was shut down correctly. While
/// for [`File`], it is possible to reach the end of file and get zero as result,
/// but if more data is appended to the file, future calls to `read` will return
/// more data.
/// 2. The buffer specified was 0 bytes in length.
///
/// It is not an error if the returned value `n` is smaller than the buffer size,
/// even when the reader is not at the end of the stream yet.
/// This may happen for example because fewer bytes are actually available right now
/// (e. g. being close to end-of-file) or because read() was interrupted by a signal.
///
/// As this trait is safe to implement, callers in unsafe code cannot rely on
/// `n <= buf.len()` for safety.
/// Extra care needs to be taken when `unsafe` functions are used to access the read bytes.
/// Callers have to ensure that no unchecked out-of-bounds accesses are possible even if
/// `n > buf.len()`.
///
/// *Implementations* of this method can make no assumptions about the contents of `buf` when
/// this function is called. It is recommended that implementations only write data to `buf`
/// instead of reading its contents.
///
/// Correspondingly, however, *callers* of this method in unsafe code must not assume
/// any guarantees about how the implementation uses `buf`. The trait is safe to implement,
/// so it is possible that the code that's supposed to write to the buffer might also read
/// from it. It is your responsibility to make sure that `buf` is initialized
/// before calling `read`. Calling `read` with an uninitialized `buf` (of the kind one
/// obtains via [`MaybeUninit<T>`]) is not safe, and can lead to undefined behavior.
///
/// [`MaybeUninit<T>`]: crate::mem::MaybeUninit
///
/// # Errors
///
/// If this function encounters any form of I/O or other error, an error
/// variant will be returned. If an error is returned then it must be
/// guaranteed that no bytes were read.
///
/// An error of the [`ErrorKind::Interrupted`] kind is non-fatal and the read
/// operation should be retried if there is nothing else to do.
///
/// # Examples
///
/// [`File`]s implement `Read`:
///
/// [`Ok(n)`]: Ok
/// [`File`]: crate::fs::File
/// [`TcpStream`]: crate::net::TcpStream
///
/// ```no_run
/// use std::io;
/// use std::io::prelude::*;
/// use std::fs::File;
///
/// fn main() -> io::Result<()> {
/// let mut f = File::open("foo.txt")?;
/// let mut buffer = [0; 10];
///
/// // read up to 10 bytes
/// let n = f.read(&mut buffer[..])?;
///
/// println!("The bytes: {:?}", &buffer[..n]);
/// Ok(())
/// }
/// ```
#[stable(feature = "rust1", since = "1.0.0")]
fn read(&mut self, buf: &mut [u8]) -> Result<usize>;
/// Like `read`, except that it reads into a slice of buffers.
///
/// Data is copied to fill each buffer in order, with the final buffer
/// written to possibly being only partially filled. This method must
/// behave equivalently to a single call to `read` with concatenated
/// buffers.
///
/// The default implementation calls `read` with either the first nonempty
/// buffer provided, or an empty one if none exists.
#[stable(feature = "iovec", since = "1.36.0")]
fn read_vectored(&mut self, bufs: &mut [IoSliceMut<'_>]) -> Result<usize> {
default_read_vectored(|b| self.read(b), bufs)
}
/// Determines if this `Read`er has an efficient `read_vectored`
/// implementation.
///
/// If a `Read`er does not override the default `read_vectored`
/// implementation, code using it may want to avoid the method all together
/// and coalesce writes into a single buffer for higher performance.
///
/// The default implementation returns `false`.
#[unstable(feature = "can_vector", issue = "69941")]
fn is_read_vectored(&self) -> bool {
false
}
/// Reads all bytes until EOF in this source, placing them into `buf`.
///
/// All bytes read from this source will be appended to the specified buffer
/// `buf`. This function will continuously call [`read()`] to append more data to
/// `buf` until [`read()`] returns either [`Ok(0)`] or an error of
/// non-[`ErrorKind::Interrupted`] kind.
///
/// If successful, this function will return the total number of bytes read.
///
/// # Errors
///
/// If this function encounters an error of the kind
/// [`ErrorKind::Interrupted`] then the error is ignored and the operation
/// will continue.
///
/// If any other read error is encountered then this function immediately
/// returns. Any bytes which have already been read will be appended to
/// `buf`.
///
/// # Examples
///
/// [`File`]s implement `Read`:
///
/// [`read()`]: Read::read
/// [`Ok(0)`]: Ok
/// [`File`]: crate::fs::File
///
/// ```no_run
/// use std::io;
/// use std::io::prelude::*;
/// use std::fs::File;
///
/// fn main() -> io::Result<()> {
/// let mut f = File::open("foo.txt")?;
/// let mut buffer = Vec::new();
///
/// // read the whole file
/// f.read_to_end(&mut buffer)?;
/// Ok(())
/// }
/// ```
///
/// (See also the [`std::fs::read`] convenience function for reading from a
/// file.)
///
/// [`std::fs::read`]: crate::fs::read
///
/// ## Implementing `read_to_end`
///
/// When implementing the `io::Read` trait, it is recommended to allocate
/// memory using [`Vec::try_reserve`]. However, this behavior is not guaranteed
/// by all implementations, and `read_to_end` may not handle out-of-memory
/// situations gracefully.
///
/// ```no_run
/// # use std::io::{self, BufRead};
/// # struct Example { example_datasource: io::Empty } impl Example {
/// # fn get_some_data_for_the_example(&self) -> &'static [u8] { &[] }
/// fn read_to_end(&mut self, dest_vec: &mut Vec<u8>) -> io::Result<usize> {
/// let initial_vec_len = dest_vec.len();
/// loop {
/// let src_buf = self.example_datasource.fill_buf()?;
/// if src_buf.is_empty() {
/// break;
/// }
/// dest_vec.try_reserve(src_buf.len())?;
/// dest_vec.extend_from_slice(src_buf);
///
/// // Any irreversible side effects should happen after `try_reserve` succeeds,
/// // to avoid losing data on allocation error.
/// let read = src_buf.len();
/// self.example_datasource.consume(read);
/// }
/// Ok(dest_vec.len() - initial_vec_len)
/// }
/// # }
/// ```
///
/// # Usage Notes
///
/// `read_to_end` attempts to read a source until EOF, but many sources are continuous streams
/// that do not send EOF. In these cases, `read_to_end` will block indefinitely. Standard input
/// is one such stream which may be finite if piped, but is typically continuous. For example,
/// `cat file | my-rust-program` will correctly terminate with an `EOF` upon closure of cat.
/// Reading user input or running programs that remain open indefinitely will never terminate
/// the stream with `EOF` (e.g. `yes | my-rust-program`).
///
/// Using `.lines()` with a [`BufReader`] or using [`read`] can provide a better solution
///
///[`read`]: Read::read
///
/// [`Vec::try_reserve`]: crate::vec::Vec::try_reserve
#[stable(feature = "rust1", since = "1.0.0")]
fn read_to_end(&mut self, buf: &mut Vec<u8>) -> Result<usize> {
default_read_to_end(self, buf, None)
}
/// Reads all bytes until EOF in this source, appending them to `buf`.
///
/// If successful, this function returns the number of bytes which were read
/// and appended to `buf`.
///
/// # Errors
///
/// If the data in this stream is *not* valid UTF-8 then an error is
/// returned and `buf` is unchanged.
///
/// See [`read_to_end`] for other error semantics.
///
/// [`read_to_end`]: Read::read_to_end
///
/// # Examples
///
/// [`File`]s implement `Read`:
///
/// [`File`]: crate::fs::File
///
/// ```no_run
/// use std::io;
/// use std::io::prelude::*;
/// use std::fs::File;
///
/// fn main() -> io::Result<()> {
/// let mut f = File::open("foo.txt")?;
/// let mut buffer = String::new();
///
/// f.read_to_string(&mut buffer)?;
/// Ok(())
/// }
/// ```
///
/// (See also the [`std::fs::read_to_string`] convenience function for
/// reading from a file.)
///
/// # Usage Notes
///
/// `read_to_string` attempts to read a source until EOF, but many sources are continuous streams
/// that do not send EOF. In these cases, `read_to_string` will block indefinitely. Standard input
/// is one such stream which may be finite if piped, but is typically continuous. For example,
/// `cat file | my-rust-program` will correctly terminate with an `EOF` upon closure of cat.
/// Reading user input or running programs that remain open indefinitely will never terminate
/// the stream with `EOF` (e.g. `yes | my-rust-program`).
///
/// Using `.lines()` with a [`BufReader`] or using [`read`] can provide a better solution
///
///[`read`]: Read::read
///
/// [`std::fs::read_to_string`]: crate::fs::read_to_string
#[stable(feature = "rust1", since = "1.0.0")]
fn read_to_string(&mut self, buf: &mut String) -> Result<usize> {
default_read_to_string(self, buf, None)
}
/// Reads the exact number of bytes required to fill `buf`.
///
/// This function reads as many bytes as necessary to completely fill the
/// specified buffer `buf`.
///
/// *Implementations* of this method can make no assumptions about the contents of `buf` when
/// this function is called. It is recommended that implementations only write data to `buf`
/// instead of reading its contents. The documentation on [`read`] has a more detailed
/// explanation of this subject.
///
/// # Errors
///
/// If this function encounters an error of the kind
/// [`ErrorKind::Interrupted`] then the error is ignored and the operation
/// will continue.
///
/// If this function encounters an "end of file" before completely filling
/// the buffer, it returns an error of the kind [`ErrorKind::UnexpectedEof`].
/// The contents of `buf` are unspecified in this case.
///
/// If any other read error is encountered then this function immediately
/// returns. The contents of `buf` are unspecified in this case.
///
/// If this function returns an error, it is unspecified how many bytes it
/// has read, but it will never read more than would be necessary to
/// completely fill the buffer.
///
/// # Examples
///
/// [`File`]s implement `Read`:
///
/// [`read`]: Read::read
/// [`File`]: crate::fs::File
///
/// ```no_run
/// use std::io;
/// use std::io::prelude::*;
/// use std::fs::File;
///
/// fn main() -> io::Result<()> {
/// let mut f = File::open("foo.txt")?;
/// let mut buffer = [0; 10];
///
/// // read exactly 10 bytes
/// f.read_exact(&mut buffer)?;
/// Ok(())
/// }
/// ```
#[stable(feature = "read_exact", since = "1.6.0")]
fn read_exact(&mut self, buf: &mut [u8]) -> Result<()> {
default_read_exact(self, buf)
}
/// Pull some bytes from this source into the specified buffer.
///
/// This is equivalent to the [`read`](Read::read) method, except that it is passed a [`BorrowedCursor`] rather than `[u8]` to allow use
/// with uninitialized buffers. The new data will be appended to any existing contents of `buf`.
///
/// The default implementation delegates to `read`.
///
/// This method makes it possible to return both data and an error but it is advised against.
#[unstable(feature = "read_buf", issue = "78485")]
fn read_buf(&mut self, buf: BorrowedCursor<'_>) -> Result<()> {
default_read_buf(|b| self.read(b), buf)
}
/// Reads the exact number of bytes required to fill `cursor`.
///
/// This is similar to the [`read_exact`](Read::read_exact) method, except
/// that it is passed a [`BorrowedCursor`] rather than `[u8]` to allow use
/// with uninitialized buffers.
///
/// # Errors
///
/// If this function encounters an error of the kind [`ErrorKind::Interrupted`]
/// then the error is ignored and the operation will continue.
///
/// If this function encounters an "end of file" before completely filling
/// the buffer, it returns an error of the kind [`ErrorKind::UnexpectedEof`].
///
/// If any other read error is encountered then this function immediately
/// returns.
///
/// If this function returns an error, all bytes read will be appended to `cursor`.
#[unstable(feature = "read_buf", issue = "78485")]
fn read_buf_exact(&mut self, cursor: BorrowedCursor<'_>) -> Result<()> {
default_read_buf_exact(self, cursor)
}
/// Creates a "by reference" adapter for this instance of `Read`.
///
/// The returned adapter also implements `Read` and will simply borrow this
/// current reader.
///
/// # Examples
///
/// [`File`]s implement `Read`:
///
/// [`File`]: crate::fs::File
///
/// ```no_run
/// use std::io;
/// use std::io::Read;
/// use std::fs::File;
///
/// fn main() -> io::Result<()> {
/// let mut f = File::open("foo.txt")?;
/// let mut buffer = Vec::new();
/// let mut other_buffer = Vec::new();
///
/// {
/// let reference = f.by_ref();
///
/// // read at most 5 bytes
/// reference.take(5).read_to_end(&mut buffer)?;
///
/// } // drop our &mut reference so we can use f again
///
/// // original file still usable, read the rest
/// f.read_to_end(&mut other_buffer)?;
/// Ok(())
/// }
/// ```
#[stable(feature = "rust1", since = "1.0.0")]
fn by_ref(&mut self) -> &mut Self
where
Self: Sized,
{
self
}
/// Transforms this `Read` instance to an [`Iterator`] over its bytes.
///
/// The returned type implements [`Iterator`] where the [`Item`] is
/// <code>[Result]<[u8], [io::Error]></code>.
/// The yielded item is [`Ok`] if a byte was successfully read and [`Err`]
/// otherwise. EOF is mapped to returning [`None`] from this iterator.
///
/// The default implementation calls `read` for each byte,
/// which can be very inefficient for data that's not in memory,
/// such as [`File`]. Consider using a [`BufReader`] in such cases.
///
/// # Examples
///
/// [`File`]s implement `Read`:
///
/// [`Item`]: Iterator::Item
/// [`File`]: crate::fs::File "fs::File"
/// [Result]: crate::result::Result "Result"
/// [io::Error]: self::Error "io::Error"
///
/// ```no_run
/// use std::io;
/// use std::io::prelude::*;
/// use std::io::BufReader;
/// use std::fs::File;
///
/// fn main() -> io::Result<()> {
/// let f = BufReader::new(File::open("foo.txt")?);
///
/// for byte in f.bytes() {
/// println!("{}", byte?);
/// }
/// Ok(())
/// }
/// ```
#[stable(feature = "rust1", since = "1.0.0")]
fn bytes(self) -> Bytes<Self>
where
Self: Sized,
{
Bytes { inner: self }
}
/// Creates an adapter which will chain this stream with another.
///
/// The returned `Read` instance will first read all bytes from this object
/// until EOF is encountered. Afterwards the output is equivalent to the
/// output of `next`.
///
/// # Examples
///
/// [`File`]s implement `Read`:
///
/// [`File`]: crate::fs::File
///
/// ```no_run
/// use std::io;
/// use std::io::prelude::*;
/// use std::fs::File;
///
/// fn main() -> io::Result<()> {
/// let f1 = File::open("foo.txt")?;
/// let f2 = File::open("bar.txt")?;
///
/// let mut handle = f1.chain(f2);
/// let mut buffer = String::new();
///
/// // read the value into a String. We could use any Read method here,
/// // this is just one example.
/// handle.read_to_string(&mut buffer)?;
/// Ok(())
/// }
/// ```
#[stable(feature = "rust1", since = "1.0.0")]
fn chain<R: Read>(self, next: R) -> Chain<Self, R>
where
Self: Sized,
{
Chain { first: self, second: next, done_first: false }
}
/// Creates an adapter which will read at most `limit` bytes from it.
///
/// This function returns a new instance of `Read` which will read at most
/// `limit` bytes, after which it will always return EOF ([`Ok(0)`]). Any
/// read errors will not count towards the number of bytes read and future
/// calls to [`read()`] may succeed.
///
/// # Examples
///
/// [`File`]s implement `Read`:
///
/// [`File`]: crate::fs::File
/// [`Ok(0)`]: Ok
/// [`read()`]: Read::read
///
/// ```no_run
/// use std::io;
/// use std::io::prelude::*;
/// use std::fs::File;
///
/// fn main() -> io::Result<()> {
/// let f = File::open("foo.txt")?;
/// let mut buffer = [0; 5];
///
/// // read at most five bytes
/// let mut handle = f.take(5);
///
/// handle.read(&mut buffer)?;
/// Ok(())
/// }
/// ```
#[stable(feature = "rust1", since = "1.0.0")]
fn take(self, limit: u64) -> Take<Self>
where
Self: Sized,
{
Take { inner: self, len: limit, limit }
}
}
/// Reads all bytes from a [reader][Read] into a new [`String`].
///
/// This is a convenience function for [`Read::read_to_string`]. Using this
/// function avoids having to create a variable first and provides more type
/// safety since you can only get the buffer out if there were no errors. (If you
/// use [`Read::read_to_string`] you have to remember to check whether the read
/// succeeded because otherwise your buffer will be empty or only partially full.)
///
/// # Performance
///
/// The downside of this function's increased ease of use and type safety is
/// that it gives you less control over performance. For example, you can't
/// pre-allocate memory like you can using [`String::with_capacity`] and
/// [`Read::read_to_string`]. Also, you can't re-use the buffer if an error
/// occurs while reading.
///
/// In many cases, this function's performance will be adequate and the ease of use
/// and type safety tradeoffs will be worth it. However, there are cases where you
/// need more control over performance, and in those cases you should definitely use
/// [`Read::read_to_string`] directly.
///
/// Note that in some special cases, such as when reading files, this function will
/// pre-allocate memory based on the size of the input it is reading. In those
/// cases, the performance should be as good as if you had used
/// [`Read::read_to_string`] with a manually pre-allocated buffer.
///
/// # Errors
///
/// This function forces you to handle errors because the output (the `String`)
/// is wrapped in a [`Result`]. See [`Read::read_to_string`] for the errors
/// that can occur. If any error occurs, you will get an [`Err`], so you
/// don't have to worry about your buffer being empty or partially full.
///
/// # Examples
///
/// ```no_run
/// # use std::io;
/// fn main() -> io::Result<()> {
/// let stdin = io::read_to_string(io::stdin())?;
/// println!("Stdin was:");
/// println!("{stdin}");
/// Ok(())
/// }
/// ```
///
/// # Usage Notes
///
/// `read_to_string` attempts to read a source until EOF, but many sources are continuous streams
/// that do not send EOF. In these cases, `read_to_string` will block indefinitely. Standard input
/// is one such stream which may be finite if piped, but is typically continuous. For example,
/// `cat file | my-rust-program` will correctly terminate with an `EOF` upon closure of cat.
/// Reading user input or running programs that remain open indefinitely will never terminate
/// the stream with `EOF` (e.g. `yes | my-rust-program`).
///
/// Using `.lines()` with a [`BufReader`] or using [`read`] can provide a better solution
///
///[`read`]: Read::read
///
#[stable(feature = "io_read_to_string", since = "1.65.0")]
pub fn read_to_string<R: Read>(mut reader: R) -> Result<String> {
let mut buf = String::new();
reader.read_to_string(&mut buf)?;
Ok(buf)
}
/// A buffer type used with `Read::read_vectored`.
///
/// It is semantically a wrapper around a `&mut [u8]`, but is guaranteed to be
/// ABI compatible with the `iovec` type on Unix platforms and `WSABUF` on
/// Windows.
#[stable(feature = "iovec", since = "1.36.0")]
#[repr(transparent)]
pub struct IoSliceMut<'a>(sys::io::IoSliceMut<'a>);
#[stable(feature = "iovec_send_sync", since = "1.44.0")]
unsafe impl<'a> Send for IoSliceMut<'a> {}
#[stable(feature = "iovec_send_sync", since = "1.44.0")]
unsafe impl<'a> Sync for IoSliceMut<'a> {}
#[stable(feature = "iovec", since = "1.36.0")]
impl<'a> fmt::Debug for IoSliceMut<'a> {
fn fmt(&self, fmt: &mut fmt::Formatter<'_>) -> fmt::Result {
fmt::Debug::fmt(self.0.as_slice(), fmt)
}
}
impl<'a> IoSliceMut<'a> {
/// Creates a new `IoSliceMut` wrapping a byte slice.
///
/// # Panics
///
/// Panics on Windows if the slice is larger than 4GB.
#[stable(feature = "iovec", since = "1.36.0")]
#[inline]
pub fn new(buf: &'a mut [u8]) -> IoSliceMut<'a> {
IoSliceMut(sys::io::IoSliceMut::new(buf))
}
/// Advance the internal cursor of the slice.
///
/// Also see [`IoSliceMut::advance_slices`] to advance the cursors of
/// multiple buffers.
///
/// # Panics
///
/// Panics when trying to advance beyond the end of the slice.
///
/// # Examples
///
/// ```
/// use std::io::IoSliceMut;
/// use std::ops::Deref;
///
/// let mut data = [1; 8];
/// let mut buf = IoSliceMut::new(&mut data);
///
/// // Mark 3 bytes as read.
/// buf.advance(3);
/// assert_eq!(buf.deref(), [1; 5].as_ref());
/// ```
#[stable(feature = "io_slice_advance", since = "1.81.0")]
#[inline]
pub fn advance(&mut self, n: usize) {
self.0.advance(n)
}
/// Advance a slice of slices.
///
/// Shrinks the slice to remove any `IoSliceMut`s that are fully advanced over.
/// If the cursor ends up in the middle of an `IoSliceMut`, it is modified
/// to start at that cursor.
///
/// For example, if we have a slice of two 8-byte `IoSliceMut`s, and we advance by 10 bytes,
/// the result will only include the second `IoSliceMut`, advanced by 2 bytes.
///
/// # Panics
///
/// Panics when trying to advance beyond the end of the slices.
///
/// # Examples
///
/// ```
/// use std::io::IoSliceMut;
/// use std::ops::Deref;
///
/// let mut buf1 = [1; 8];
/// let mut buf2 = [2; 16];
/// let mut buf3 = [3; 8];
/// let mut bufs = &mut [
/// IoSliceMut::new(&mut buf1),
/// IoSliceMut::new(&mut buf2),
/// IoSliceMut::new(&mut buf3),
/// ][..];
///
/// // Mark 10 bytes as read.
/// IoSliceMut::advance_slices(&mut bufs, 10);
/// assert_eq!(bufs[0].deref(), [2; 14].as_ref());
/// assert_eq!(bufs[1].deref(), [3; 8].as_ref());
/// ```
#[stable(feature = "io_slice_advance", since = "1.81.0")]
#[inline]
pub fn advance_slices(bufs: &mut &mut [IoSliceMut<'a>], n: usize) {
// Number of buffers to remove.
let mut remove = 0;
// Remaining length before reaching n.
let mut left = n;
for buf in bufs.iter() {
if let Some(remainder) = left.checked_sub(buf.len()) {
left = remainder;
remove += 1;
} else {
break;
}
}
*bufs = &mut take(bufs)[remove..];
if bufs.is_empty() {
assert!(left == 0, "advancing io slices beyond their length");
} else {
bufs[0].advance(left);
}
}
/// Get the underlying bytes as a mutable slice with the original lifetime.
///
/// # Examples
///
/// ```
/// #![feature(io_slice_as_bytes)]
/// use std::io::IoSliceMut;
///
/// let mut data = *b"abcdef";
/// let io_slice = IoSliceMut::new(&mut data);
/// io_slice.into_slice()[0] = b'A';
///
/// assert_eq!(&data, b"Abcdef");
/// ```
#[unstable(feature = "io_slice_as_bytes", issue = "132818")]
pub const fn into_slice(self) -> &'a mut [u8] {
self.0.into_slice()
}
}
#[stable(feature = "iovec", since = "1.36.0")]
impl<'a> Deref for IoSliceMut<'a> {
type Target = [u8];
#[inline]
fn deref(&self) -> &[u8] {
self.0.as_slice()
}
}
#[stable(feature = "iovec", since = "1.36.0")]
impl<'a> DerefMut for IoSliceMut<'a> {
#[inline]
fn deref_mut(&mut self) -> &mut [u8] {
self.0.as_mut_slice()
}
}
/// A buffer type used with `Write::write_vectored`.
///
/// It is semantically a wrapper around a `&[u8]`, but is guaranteed to be
/// ABI compatible with the `iovec` type on Unix platforms and `WSABUF` on
/// Windows.
#[stable(feature = "iovec", since = "1.36.0")]
#[derive(Copy, Clone)]
#[repr(transparent)]
pub struct IoSlice<'a>(sys::io::IoSlice<'a>);
#[stable(feature = "iovec_send_sync", since = "1.44.0")]
unsafe impl<'a> Send for IoSlice<'a> {}
#[stable(feature = "iovec_send_sync", since = "1.44.0")]
unsafe impl<'a> Sync for IoSlice<'a> {}
#[stable(feature = "iovec", since = "1.36.0")]
impl<'a> fmt::Debug for IoSlice<'a> {
fn fmt(&self, fmt: &mut fmt::Formatter<'_>) -> fmt::Result {
fmt::Debug::fmt(self.0.as_slice(), fmt)
}
}
impl<'a> IoSlice<'a> {
/// Creates a new `IoSlice` wrapping a byte slice.
///
/// # Panics
///
/// Panics on Windows if the slice is larger than 4GB.
#[stable(feature = "iovec", since = "1.36.0")]
#[must_use]
#[inline]
pub fn new(buf: &'a [u8]) -> IoSlice<'a> {
IoSlice(sys::io::IoSlice::new(buf))
}
/// Advance the internal cursor of the slice.
///
/// Also see [`IoSlice::advance_slices`] to advance the cursors of multiple
/// buffers.
///
/// # Panics
///
/// Panics when trying to advance beyond the end of the slice.
///
/// # Examples
///
/// ```
/// use std::io::IoSlice;
/// use std::ops::Deref;
///
/// let data = [1; 8];
/// let mut buf = IoSlice::new(&data);
///
/// // Mark 3 bytes as read.
/// buf.advance(3);
/// assert_eq!(buf.deref(), [1; 5].as_ref());
/// ```
#[stable(feature = "io_slice_advance", since = "1.81.0")]
#[inline]
pub fn advance(&mut self, n: usize) {
self.0.advance(n)
}
/// Advance a slice of slices.
///
/// Shrinks the slice to remove any `IoSlice`s that are fully advanced over.
/// If the cursor ends up in the middle of an `IoSlice`, it is modified
/// to start at that cursor.
///
/// For example, if we have a slice of two 8-byte `IoSlice`s, and we advance by 10 bytes,
/// the result will only include the second `IoSlice`, advanced by 2 bytes.
///
/// # Panics
///
/// Panics when trying to advance beyond the end of the slices.
///
/// # Examples
///
/// ```
/// use std::io::IoSlice;
/// use std::ops::Deref;
///
/// let buf1 = [1; 8];
/// let buf2 = [2; 16];
/// let buf3 = [3; 8];
/// let mut bufs = &mut [
/// IoSlice::new(&buf1),
/// IoSlice::new(&buf2),
/// IoSlice::new(&buf3),
/// ][..];
///
/// // Mark 10 bytes as written.
/// IoSlice::advance_slices(&mut bufs, 10);
/// assert_eq!(bufs[0].deref(), [2; 14].as_ref());
/// assert_eq!(bufs[1].deref(), [3; 8].as_ref());
#[stable(feature = "io_slice_advance", since = "1.81.0")]
#[inline]
pub fn advance_slices(bufs: &mut &mut [IoSlice<'a>], n: usize) {
// Number of buffers to remove.
let mut remove = 0;
// Remaining length before reaching n. This prevents overflow
// that could happen if the length of slices in `bufs` were instead
// accumulated. Those slice may be aliased and, if they are large
// enough, their added length may overflow a `usize`.
let mut left = n;
for buf in bufs.iter() {
if let Some(remainder) = left.checked_sub(buf.len()) {
left = remainder;
remove += 1;
} else {
break;
}
}
*bufs = &mut take(bufs)[remove..];
if bufs.is_empty() {
assert!(left == 0, "advancing io slices beyond their length");
} else {
bufs[0].advance(left);
}
}
/// Get the underlying bytes as a slice with the original lifetime.
///
/// This doesn't borrow from `self`, so is less restrictive than calling
/// `.deref()`, which does.
///
/// # Examples
///
/// ```
/// #![feature(io_slice_as_bytes)]
/// use std::io::IoSlice;
///
/// let data = b"abcdef";
///
/// let mut io_slice = IoSlice::new(data);
/// let tail = &io_slice.as_slice()[3..];
///
/// // This works because `tail` doesn't borrow `io_slice`
/// io_slice = IoSlice::new(tail);
///
/// assert_eq!(io_slice.as_slice(), b"def");
/// ```
#[unstable(feature = "io_slice_as_bytes", issue = "132818")]
pub const fn as_slice(self) -> &'a [u8] {
self.0.as_slice()
}
}
#[stable(feature = "iovec", since = "1.36.0")]
impl<'a> Deref for IoSlice<'a> {
type Target = [u8];
#[inline]
fn deref(&self) -> &[u8] {
self.0.as_slice()
}
}
/// A trait for objects which are byte-oriented sinks.
///
/// Implementors of the `Write` trait are sometimes called 'writers'.
///
/// Writers are defined by two required methods, [`write`] and [`flush`]:
///
/// * The [`write`] method will attempt to write some data into the object,
/// returning how many bytes were successfully written.
///
/// * The [`flush`] method is useful for adapters and explicit buffers
/// themselves for ensuring that all buffered data has been pushed out to the
/// 'true sink'.
///
/// Writers are intended to be composable with one another. Many implementors
/// throughout [`std::io`] take and provide types which implement the `Write`
/// trait.
///
/// [`write`]: Write::write
/// [`flush`]: Write::flush
/// [`std::io`]: self
///
/// # Examples
///
/// ```no_run
/// use std::io::prelude::*;
/// use std::fs::File;
///
/// fn main() -> std::io::Result<()> {
/// let data = b"some bytes";
///
/// let mut pos = 0;
/// let mut buffer = File::create("foo.txt")?;
///
/// while pos < data.len() {
/// let bytes_written = buffer.write(&data[pos..])?;
/// pos += bytes_written;
/// }
/// Ok(())
/// }
/// ```
///
/// The trait also provides convenience methods like [`write_all`], which calls
/// `write` in a loop until its entire input has been written.
///
/// [`write_all`]: Write::write_all
#[stable(feature = "rust1", since = "1.0.0")]
#[doc(notable_trait)]
#[cfg_attr(not(test), rustc_diagnostic_item = "IoWrite")]
pub trait Write {
/// Writes a buffer into this writer, returning how many bytes were written.
///
/// This function will attempt to write the entire contents of `buf`, but
/// the entire write might not succeed, or the write may also generate an
/// error. Typically, a call to `write` represents one attempt to write to
/// any wrapped object.
///
/// Calls to `write` are not guaranteed to block waiting for data to be
/// written, and a write which would otherwise block can be indicated through
/// an [`Err`] variant.
///
/// If this method consumed `n > 0` bytes of `buf` it must return [`Ok(n)`].
/// If the return value is `Ok(n)` then `n` must satisfy `n <= buf.len()`.
/// A return value of `Ok(0)` typically means that the underlying object is
/// no longer able to accept bytes and will likely not be able to in the
/// future as well, or that the buffer provided is empty.
///
/// # Errors
///
/// Each call to `write` may generate an I/O error indicating that the
/// operation could not be completed. If an error is returned then no bytes
/// in the buffer were written to this writer.
///
/// It is **not** considered an error if the entire buffer could not be
/// written to this writer.
///
/// An error of the [`ErrorKind::Interrupted`] kind is non-fatal and the
/// write operation should be retried if there is nothing else to do.
///
/// # Examples
///
/// ```no_run
/// use std::io::prelude::*;
/// use std::fs::File;
///
/// fn main() -> std::io::Result<()> {
/// let mut buffer = File::create("foo.txt")?;
///
/// // Writes some prefix of the byte string, not necessarily all of it.
/// buffer.write(b"some bytes")?;
/// Ok(())
/// }
/// ```
///
/// [`Ok(n)`]: Ok
#[stable(feature = "rust1", since = "1.0.0")]
fn write(&mut self, buf: &[u8]) -> Result<usize>;
/// Like [`write`], except that it writes from a slice of buffers.
///
/// Data is copied from each buffer in order, with the final buffer
/// read from possibly being only partially consumed. This method must
/// behave as a call to [`write`] with the buffers concatenated would.
///
/// The default implementation calls [`write`] with either the first nonempty
/// buffer provided, or an empty one if none exists.
///
/// # Examples
///
/// ```no_run
/// use std::io::IoSlice;
/// use std::io::prelude::*;
/// use std::fs::File;
///
/// fn main() -> std::io::Result<()> {
/// let data1 = [1; 8];
/// let data2 = [15; 8];
/// let io_slice1 = IoSlice::new(&data1);
/// let io_slice2 = IoSlice::new(&data2);
///
/// let mut buffer = File::create("foo.txt")?;
///
/// // Writes some prefix of the byte string, not necessarily all of it.
/// buffer.write_vectored(&[io_slice1, io_slice2])?;
/// Ok(())
/// }
/// ```
///
/// [`write`]: Write::write
#[stable(feature = "iovec", since = "1.36.0")]
fn write_vectored(&mut self, bufs: &[IoSlice<'_>]) -> Result<usize> {
default_write_vectored(|b| self.write(b), bufs)
}
/// Determines if this `Write`r has an efficient [`write_vectored`]
/// implementation.
///
/// If a `Write`r does not override the default [`write_vectored`]
/// implementation, code using it may want to avoid the method all together
/// and coalesce writes into a single buffer for higher performance.
///
/// The default implementation returns `false`.
///
/// [`write_vectored`]: Write::write_vectored
#[unstable(feature = "can_vector", issue = "69941")]
fn is_write_vectored(&self) -> bool {
false
}
/// Flushes this output stream, ensuring that all intermediately buffered
/// contents reach their destination.
///
/// # Errors
///
/// It is considered an error if not all bytes could be written due to
/// I/O errors or EOF being reached.
///
/// # Examples
///
/// ```no_run
/// use std::io::prelude::*;
/// use std::io::BufWriter;
/// use std::fs::File;
///
/// fn main() -> std::io::Result<()> {
/// let mut buffer = BufWriter::new(File::create("foo.txt")?);
///
/// buffer.write_all(b"some bytes")?;
/// buffer.flush()?;
/// Ok(())
/// }
/// ```
#[stable(feature = "rust1", since = "1.0.0")]
fn flush(&mut self) -> Result<()>;
/// Attempts to write an entire buffer into this writer.
///
/// This method will continuously call [`write`] until there is no more data
/// to be written or an error of non-[`ErrorKind::Interrupted`] kind is
/// returned. This method will not return until the entire buffer has been
/// successfully written or such an error occurs. The first error that is
/// not of [`ErrorKind::Interrupted`] kind generated from this method will be
/// returned.
///
/// If the buffer contains no data, this will never call [`write`].
///
/// # Errors
///
/// This function will return the first error of
/// non-[`ErrorKind::Interrupted`] kind that [`write`] returns.
///
/// [`write`]: Write::write
///
/// # Examples
///
/// ```no_run
/// use std::io::prelude::*;
/// use std::fs::File;
///
/// fn main() -> std::io::Result<()> {
/// let mut buffer = File::create("foo.txt")?;
///
/// buffer.write_all(b"some bytes")?;
/// Ok(())
/// }
/// ```
#[stable(feature = "rust1", since = "1.0.0")]
fn write_all(&mut self, mut buf: &[u8]) -> Result<()> {
while !buf.is_empty() {
match self.write(buf) {
Ok(0) => {
return Err(Error::WRITE_ALL_EOF);
}
Ok(n) => buf = &buf[n..],
Err(ref e) if e.is_interrupted() => {}
Err(e) => return Err(e),
}
}
Ok(())
}
/// Attempts to write multiple buffers into this writer.
///
/// This method will continuously call [`write_vectored`] until there is no
/// more data to be written or an error of non-[`ErrorKind::Interrupted`]
/// kind is returned. This method will not return until all buffers have
/// been successfully written or such an error occurs. The first error that
/// is not of [`ErrorKind::Interrupted`] kind generated from this method
/// will be returned.
///
/// If the buffer contains no data, this will never call [`write_vectored`].
///
/// # Notes
///
/// Unlike [`write_vectored`], this takes a *mutable* reference to
/// a slice of [`IoSlice`]s, not an immutable one. That's because we need to
/// modify the slice to keep track of the bytes already written.
///
/// Once this function returns, the contents of `bufs` are unspecified, as
/// this depends on how many calls to [`write_vectored`] were necessary. It is
/// best to understand this function as taking ownership of `bufs` and to
/// not use `bufs` afterwards. The underlying buffers, to which the
/// [`IoSlice`]s point (but not the [`IoSlice`]s themselves), are unchanged and
/// can be reused.
///
/// [`write_vectored`]: Write::write_vectored
///
/// # Examples
///
/// ```
/// #![feature(write_all_vectored)]
/// # fn main() -> std::io::Result<()> {
///
/// use std::io::{Write, IoSlice};
///
/// let mut writer = Vec::new();
/// let bufs = &mut [
/// IoSlice::new(&[1]),
/// IoSlice::new(&[2, 3]),
/// IoSlice::new(&[4, 5, 6]),
/// ];
///
/// writer.write_all_vectored(bufs)?;
/// // Note: the contents of `bufs` is now undefined, see the Notes section.
///
/// assert_eq!(writer, &[1, 2, 3, 4, 5, 6]);
/// # Ok(()) }
/// ```
#[unstable(feature = "write_all_vectored", issue = "70436")]
fn write_all_vectored(&mut self, mut bufs: &mut [IoSlice<'_>]) -> Result<()> {
// Guarantee that bufs is empty if it contains no data,
// to avoid calling write_vectored if there is no data to be written.
IoSlice::advance_slices(&mut bufs, 0);
while !bufs.is_empty() {
match self.write_vectored(bufs) {
Ok(0) => {
return Err(Error::WRITE_ALL_EOF);
}
Ok(n) => IoSlice::advance_slices(&mut bufs, n),
Err(ref e) if e.is_interrupted() => {}
Err(e) => return Err(e),
}
}
Ok(())
}
/// Writes a formatted string into this writer, returning any error
/// encountered.
///
/// This method is primarily used to interface with the
/// [`format_args!()`] macro, and it is rare that this should
/// explicitly be called. The [`write!()`] macro should be favored to
/// invoke this method instead.
///
/// This function internally uses the [`write_all`] method on
/// this trait and hence will continuously write data so long as no errors
/// are received. This also means that partial writes are not indicated in
/// this signature.
///
/// [`write_all`]: Write::write_all
///
/// # Errors
///
/// This function will return any I/O error reported while formatting.
///
/// # Examples
///
/// ```no_run
/// use std::io::prelude::*;
/// use std::fs::File;
///
/// fn main() -> std::io::Result<()> {
/// let mut buffer = File::create("foo.txt")?;
///
/// // this call
/// write!(buffer, "{:.*}", 2, 1.234567)?;
/// // turns into this:
/// buffer.write_fmt(format_args!("{:.*}", 2, 1.234567))?;
/// Ok(())
/// }
/// ```
#[stable(feature = "rust1", since = "1.0.0")]
fn write_fmt(&mut self, args: fmt::Arguments<'_>) -> Result<()> {
if let Some(s) = args.as_statically_known_str() {
self.write_all(s.as_bytes())
} else {
default_write_fmt(self, args)
}
}
/// Creates a "by reference" adapter for this instance of `Write`.
///
/// The returned adapter also implements `Write` and will simply borrow this
/// current writer.
///
/// # Examples
///
/// ```no_run
/// use std::io::Write;
/// use std::fs::File;
///
/// fn main() -> std::io::Result<()> {
/// let mut buffer = File::create("foo.txt")?;
///
/// let reference = buffer.by_ref();
///
/// // we can use reference just like our original buffer
/// reference.write_all(b"some bytes")?;
/// Ok(())
/// }
/// ```
#[stable(feature = "rust1", since = "1.0.0")]
fn by_ref(&mut self) -> &mut Self
where
Self: Sized,
{
self
}
}
/// The `Seek` trait provides a cursor which can be moved within a stream of
/// bytes.
///
/// The stream typically has a fixed size, allowing seeking relative to either
/// end or the current offset.
///
/// # Examples
///
/// [`File`]s implement `Seek`:
///
/// [`File`]: crate::fs::File
///
/// ```no_run
/// use std::io;
/// use std::io::prelude::*;
/// use std::fs::File;
/// use std::io::SeekFrom;
///
/// fn main() -> io::Result<()> {
/// let mut f = File::open("foo.txt")?;
///
/// // move the cursor 42 bytes from the start of the file
/// f.seek(SeekFrom::Start(42))?;
/// Ok(())
/// }
/// ```
#[stable(feature = "rust1", since = "1.0.0")]
#[cfg_attr(not(test), rustc_diagnostic_item = "IoSeek")]
pub trait Seek {
/// Seek to an offset, in bytes, in a stream.
///
/// A seek beyond the end of a stream is allowed, but behavior is defined
/// by the implementation.
///
/// If the seek operation completed successfully,
/// this method returns the new position from the start of the stream.
/// That position can be used later with [`SeekFrom::Start`].
///
/// # Errors
///
/// Seeking can fail, for example because it might involve flushing a buffer.
///
/// Seeking to a negative offset is considered an error.
#[stable(feature = "rust1", since = "1.0.0")]
fn seek(&mut self, pos: SeekFrom) -> Result<u64>;
/// Rewind to the beginning of a stream.
///
/// This is a convenience method, equivalent to `seek(SeekFrom::Start(0))`.
///
/// # Errors
///
/// Rewinding can fail, for example because it might involve flushing a buffer.
///
/// # Example
///
/// ```no_run
/// use std::io::{Read, Seek, Write};
/// use std::fs::OpenOptions;
///
/// let mut f = OpenOptions::new()
/// .write(true)
/// .read(true)
/// .create(true)
/// .open("foo.txt")?;
///
/// let hello = "Hello!\n";
/// write!(f, "{hello}")?;
/// f.rewind()?;
///
/// let mut buf = String::new();
/// f.read_to_string(&mut buf)?;
/// assert_eq!(&buf, hello);
/// # std::io::Result::Ok(())
/// ```
#[stable(feature = "seek_rewind", since = "1.55.0")]
fn rewind(&mut self) -> Result<()> {
self.seek(SeekFrom::Start(0))?;
Ok(())
}
/// Returns the length of this stream (in bytes).
///
/// The default implementation uses up to three seek operations. If this
/// method returns successfully, the seek position is unchanged (i.e. the
/// position before calling this method is the same as afterwards).
/// However, if this method returns an error, the seek position is
/// unspecified.
///
/// If you need to obtain the length of *many* streams and you don't care
/// about the seek position afterwards, you can reduce the number of seek
/// operations by simply calling `seek(SeekFrom::End(0))` and using its
/// return value (it is also the stream length).
///
/// Note that length of a stream can change over time (for example, when
/// data is appended to a file). So calling this method multiple times does
/// not necessarily return the same length each time.
///
/// # Example
///
/// ```no_run
/// #![feature(seek_stream_len)]
/// use std::{
/// io::{self, Seek},
/// fs::File,
/// };
///
/// fn main() -> io::Result<()> {
/// let mut f = File::open("foo.txt")?;
///
/// let len = f.stream_len()?;
/// println!("The file is currently {len} bytes long");
/// Ok(())
/// }
/// ```
#[unstable(feature = "seek_stream_len", issue = "59359")]
fn stream_len(&mut self) -> Result<u64> {
stream_len_default(self)
}
/// Returns the current seek position from the start of the stream.
///
/// This is equivalent to `self.seek(SeekFrom::Current(0))`.
///
/// # Example
///
/// ```no_run
/// use std::{
/// io::{self, BufRead, BufReader, Seek},
/// fs::File,
/// };
///
/// fn main() -> io::Result<()> {
/// let mut f = BufReader::new(File::open("foo.txt")?);
///
/// let before = f.stream_position()?;
/// f.read_line(&mut String::new())?;
/// let after = f.stream_position()?;
///
/// println!("The first line was {} bytes long", after - before);
/// Ok(())
/// }
/// ```
#[stable(feature = "seek_convenience", since = "1.51.0")]
fn stream_position(&mut self) -> Result<u64> {
self.seek(SeekFrom::Current(0))
}
/// Seeks relative to the current position.
///
/// This is equivalent to `self.seek(SeekFrom::Current(offset))` but
/// doesn't return the new position which can allow some implementations
/// such as [`BufReader`] to perform more efficient seeks.
///
/// # Example
///
/// ```no_run
/// use std::{
/// io::{self, Seek},
/// fs::File,
/// };
///
/// fn main() -> io::Result<()> {
/// let mut f = File::open("foo.txt")?;
/// f.seek_relative(10)?;
/// assert_eq!(f.stream_position()?, 10);
/// Ok(())
/// }
/// ```
///
/// [`BufReader`]: crate::io::BufReader
#[stable(feature = "seek_seek_relative", since = "1.80.0")]
fn seek_relative(&mut self, offset: i64) -> Result<()> {
self.seek(SeekFrom::Current(offset))?;
Ok(())
}
}
pub(crate) fn stream_len_default<T: Seek + ?Sized>(self_: &mut T) -> Result<u64> {
let old_pos = self_.stream_position()?;
let len = self_.seek(SeekFrom::End(0))?;
// Avoid seeking a third time when we were already at the end of the
// stream. The branch is usually way cheaper than a seek operation.
if old_pos != len {
self_.seek(SeekFrom::Start(old_pos))?;
}
Ok(len)
}
/// Enumeration of possible methods to seek within an I/O object.
///
/// It is used by the [`Seek`] trait.
#[derive(Copy, PartialEq, Eq, Clone, Debug)]
#[stable(feature = "rust1", since = "1.0.0")]
#[cfg_attr(not(test), rustc_diagnostic_item = "SeekFrom")]
pub enum SeekFrom {
/// Sets the offset to the provided number of bytes.
#[stable(feature = "rust1", since = "1.0.0")]
Start(#[stable(feature = "rust1", since = "1.0.0")] u64),
/// Sets the offset to the size of this object plus the specified number of
/// bytes.
///
/// It is possible to seek beyond the end of an object, but it's an error to
/// seek before byte 0.
#[stable(feature = "rust1", since = "1.0.0")]
End(#[stable(feature = "rust1", since = "1.0.0")] i64),
/// Sets the offset to the current position plus the specified number of
/// bytes.
///
/// It is possible to seek beyond the end of an object, but it's an error to
/// seek before byte 0.
#[stable(feature = "rust1", since = "1.0.0")]
Current(#[stable(feature = "rust1", since = "1.0.0")] i64),
}
fn read_until<R: BufRead + ?Sized>(r: &mut R, delim: u8, buf: &mut Vec<u8>) -> Result<usize> {
let mut read = 0;
loop {
let (done, used) = {
let available = match r.fill_buf() {
Ok(n) => n,
Err(ref e) if e.is_interrupted() => continue,
Err(e) => return Err(e),
};
match memchr::memchr(delim, available) {
Some(i) => {
buf.extend_from_slice(&available[..=i]);
(true, i + 1)
}
None => {
buf.extend_from_slice(available);
(false, available.len())
}
}
};
r.consume(used);
read += used;
if done || used == 0 {
return Ok(read);
}
}
}
fn skip_until<R: BufRead + ?Sized>(r: &mut R, delim: u8) -> Result<usize> {
let mut read = 0;
loop {
let (done, used) = {
let available = match r.fill_buf() {
Ok(n) => n,
Err(ref e) if e.kind() == ErrorKind::Interrupted => continue,
Err(e) => return Err(e),
};
match memchr::memchr(delim, available) {
Some(i) => (true, i + 1),
None => (false, available.len()),
}
};
r.consume(used);
read += used;
if done || used == 0 {
return Ok(read);
}
}
}
/// A `BufRead` is a type of `Read`er which has an internal buffer, allowing it
/// to perform extra ways of reading.
///
/// For example, reading line-by-line is inefficient without using a buffer, so
/// if you want to read by line, you'll need `BufRead`, which includes a
/// [`read_line`] method as well as a [`lines`] iterator.
///
/// # Examples
///
/// A locked standard input implements `BufRead`:
///
/// ```no_run
/// use std::io;
/// use std::io::prelude::*;
///
/// let stdin = io::stdin();
/// for line in stdin.lock().lines() {
/// println!("{}", line?);
/// }
/// # std::io::Result::Ok(())
/// ```
///
/// If you have something that implements [`Read`], you can use the [`BufReader`
/// type][`BufReader`] to turn it into a `BufRead`.
///
/// For example, [`File`] implements [`Read`], but not `BufRead`.
/// [`BufReader`] to the rescue!
///
/// [`File`]: crate::fs::File
/// [`read_line`]: BufRead::read_line
/// [`lines`]: BufRead::lines
///
/// ```no_run
/// use std::io::{self, BufReader};
/// use std::io::prelude::*;
/// use std::fs::File;
///
/// fn main() -> io::Result<()> {
/// let f = File::open("foo.txt")?;
/// let f = BufReader::new(f);
///
/// for line in f.lines() {
/// let line = line?;
/// println!("{line}");
/// }
///
/// Ok(())
/// }
/// ```
#[stable(feature = "rust1", since = "1.0.0")]
#[cfg_attr(not(test), rustc_diagnostic_item = "IoBufRead")]
pub trait BufRead: Read {
/// Returns the contents of the internal buffer, filling it with more data, via `Read` methods, if empty.
///
/// This is a lower-level method and is meant to be used together with [`consume`],
/// which can be used to mark bytes that should not be returned by subsequent calls to `read`.
///
/// [`consume`]: BufRead::consume
///
/// Returns an empty buffer when the stream has reached EOF.
///
/// # Errors
///
/// This function will return an I/O error if a `Read` method was called, but returned an error.
///
/// # Examples
///
/// A locked standard input implements `BufRead`:
///
/// ```no_run
/// use std::io;
/// use std::io::prelude::*;
///
/// let stdin = io::stdin();
/// let mut stdin = stdin.lock();
///
/// let buffer = stdin.fill_buf()?;
///
/// // work with buffer
/// println!("{buffer:?}");
///
/// // mark the bytes we worked with as read
/// let length = buffer.len();
/// stdin.consume(length);
/// # std::io::Result::Ok(())
/// ```
#[stable(feature = "rust1", since = "1.0.0")]
fn fill_buf(&mut self) -> Result<&[u8]>;
/// Marks the given `amount` of additional bytes from the internal buffer as having been read.
/// Subsequent calls to `read` only return bytes that have not been marked as read.
///
/// This is a lower-level method and is meant to be used together with [`fill_buf`],
/// which can be used to fill the internal buffer via `Read` methods.
///
/// It is a logic error if `amount` exceeds the number of unread bytes in the internal buffer, which is returned by [`fill_buf`].
///
/// # Examples
///
/// Since `consume()` is meant to be used with [`fill_buf`],
/// that method's example includes an example of `consume()`.
///
/// [`fill_buf`]: BufRead::fill_buf
#[stable(feature = "rust1", since = "1.0.0")]
fn consume(&mut self, amount: usize);
/// Checks if there is any data left to be `read`.
///
/// This function may fill the buffer to check for data,
/// so this function returns `Result<bool>`, not `bool`.
///
/// The default implementation calls `fill_buf` and checks that the
/// returned slice is empty (which means that there is no data left,
/// since EOF is reached).
///
/// # Errors
///
/// This function will return an I/O error if a `Read` method was called, but returned an error.
///
/// Examples
///
/// ```
/// #![feature(buf_read_has_data_left)]
/// use std::io;
/// use std::io::prelude::*;
///
/// let stdin = io::stdin();
/// let mut stdin = stdin.lock();
///
/// while stdin.has_data_left()? {
/// let mut line = String::new();
/// stdin.read_line(&mut line)?;
/// // work with line
/// println!("{line:?}");
/// }
/// # std::io::Result::Ok(())
/// ```
#[unstable(feature = "buf_read_has_data_left", reason = "recently added", issue = "86423")]
fn has_data_left(&mut self) -> Result<bool> {
self.fill_buf().map(|b| !b.is_empty())
}
/// Reads all bytes into `buf` until the delimiter `byte` or EOF is reached.
///
/// This function will read bytes from the underlying stream until the
/// delimiter or EOF is found. Once found, all bytes up to, and including,
/// the delimiter (if found) will be appended to `buf`.
///
/// If successful, this function will return the total number of bytes read.
///
/// This function is blocking and should be used carefully: it is possible for
/// an attacker to continuously send bytes without ever sending the delimiter
/// or EOF.
///
/// # Errors
///
/// This function will ignore all instances of [`ErrorKind::Interrupted`] and
/// will otherwise return any errors returned by [`fill_buf`].
///
/// If an I/O error is encountered then all bytes read so far will be
/// present in `buf` and its length will have been adjusted appropriately.
///
/// [`fill_buf`]: BufRead::fill_buf
///
/// # Examples
///
/// [`std::io::Cursor`][`Cursor`] is a type that implements `BufRead`. In
/// this example, we use [`Cursor`] to read all the bytes in a byte slice
/// in hyphen delimited segments:
///
/// ```
/// use std::io::{self, BufRead};
///
/// let mut cursor = io::Cursor::new(b"lorem-ipsum");
/// let mut buf = vec![];
///
/// // cursor is at 'l'
/// let num_bytes = cursor.read_until(b'-', &mut buf)
/// .expect("reading from cursor won't fail");
/// assert_eq!(num_bytes, 6);
/// assert_eq!(buf, b"lorem-");
/// buf.clear();
///
/// // cursor is at 'i'
/// let num_bytes = cursor.read_until(b'-', &mut buf)
/// .expect("reading from cursor won't fail");
/// assert_eq!(num_bytes, 5);
/// assert_eq!(buf, b"ipsum");
/// buf.clear();
///
/// // cursor is at EOF
/// let num_bytes = cursor.read_until(b'-', &mut buf)
/// .expect("reading from cursor won't fail");
/// assert_eq!(num_bytes, 0);
/// assert_eq!(buf, b"");
/// ```
#[stable(feature = "rust1", since = "1.0.0")]
fn read_until(&mut self, byte: u8, buf: &mut Vec<u8>) -> Result<usize> {
read_until(self, byte, buf)
}
/// Skips all bytes until the delimiter `byte` or EOF is reached.
///
/// This function will read (and discard) bytes from the underlying stream until the
/// delimiter or EOF is found.
///
/// If successful, this function will return the total number of bytes read,
/// including the delimiter byte if found.
///
/// This is useful for efficiently skipping data such as NUL-terminated strings
/// in binary file formats without buffering.
///
/// This function is blocking and should be used carefully: it is possible for
/// an attacker to continuously send bytes without ever sending the delimiter
/// or EOF.
///
/// # Errors
///
/// This function will ignore all instances of [`ErrorKind::Interrupted`] and
/// will otherwise return any errors returned by [`fill_buf`].
///
/// If an I/O error is encountered then all bytes read so far will be
/// present in `buf` and its length will have been adjusted appropriately.
///
/// [`fill_buf`]: BufRead::fill_buf
///
/// # Examples
///
/// [`std::io::Cursor`][`Cursor`] is a type that implements `BufRead`. In
/// this example, we use [`Cursor`] to read some NUL-terminated information
/// about Ferris from a binary string, skipping the fun fact:
///
/// ```
/// use std::io::{self, BufRead};
///
/// let mut cursor = io::Cursor::new(b"Ferris\0Likes long walks on the beach\0Crustacean\0!");
///
/// // read name
/// let mut name = Vec::new();
/// let num_bytes = cursor.read_until(b'\0', &mut name)
/// .expect("reading from cursor won't fail");
/// assert_eq!(num_bytes, 7);
/// assert_eq!(name, b"Ferris\0");
///
/// // skip fun fact
/// let num_bytes = cursor.skip_until(b'\0')
/// .expect("reading from cursor won't fail");
/// assert_eq!(num_bytes, 30);
///
/// // read animal type
/// let mut animal = Vec::new();
/// let num_bytes = cursor.read_until(b'\0', &mut animal)
/// .expect("reading from cursor won't fail");
/// assert_eq!(num_bytes, 11);
/// assert_eq!(animal, b"Crustacean\0");
///
/// // reach EOF
/// let num_bytes = cursor.skip_until(b'\0')
/// .expect("reading from cursor won't fail");
/// assert_eq!(num_bytes, 1);
/// ```
#[stable(feature = "bufread_skip_until", since = "1.83.0")]
fn skip_until(&mut self, byte: u8) -> Result<usize> {
skip_until(self, byte)
}
/// Reads all bytes until a newline (the `0xA` byte) is reached, and append
/// them to the provided `String` buffer.
///
/// Previous content of the buffer will be preserved. To avoid appending to
/// the buffer, you need to [`clear`] it first.
///
/// This function will read bytes from the underlying stream until the
/// newline delimiter (the `0xA` byte) or EOF is found. Once found, all bytes
/// up to, and including, the delimiter (if found) will be appended to
/// `buf`.
///
/// If successful, this function will return the total number of bytes read.
///
/// If this function returns [`Ok(0)`], the stream has reached EOF.
///
/// This function is blocking and should be used carefully: it is possible for
/// an attacker to continuously send bytes without ever sending a newline
/// or EOF. You can use [`take`] to limit the maximum number of bytes read.
///
/// [`Ok(0)`]: Ok
/// [`clear`]: String::clear
/// [`take`]: crate::io::Read::take
///
/// # Errors
///
/// This function has the same error semantics as [`read_until`] and will
/// also return an error if the read bytes are not valid UTF-8. If an I/O
/// error is encountered then `buf` may contain some bytes already read in
/// the event that all data read so far was valid UTF-8.
///
/// [`read_until`]: BufRead::read_until
///
/// # Examples
///
/// [`std::io::Cursor`][`Cursor`] is a type that implements `BufRead`. In
/// this example, we use [`Cursor`] to read all the lines in a byte slice:
///
/// ```
/// use std::io::{self, BufRead};
///
/// let mut cursor = io::Cursor::new(b"foo\nbar");
/// let mut buf = String::new();
///
/// // cursor is at 'f'
/// let num_bytes = cursor.read_line(&mut buf)
/// .expect("reading from cursor won't fail");
/// assert_eq!(num_bytes, 4);
/// assert_eq!(buf, "foo\n");
/// buf.clear();
///
/// // cursor is at 'b'
/// let num_bytes = cursor.read_line(&mut buf)
/// .expect("reading from cursor won't fail");
/// assert_eq!(num_bytes, 3);
/// assert_eq!(buf, "bar");
/// buf.clear();
///
/// // cursor is at EOF
/// let num_bytes = cursor.read_line(&mut buf)
/// .expect("reading from cursor won't fail");
/// assert_eq!(num_bytes, 0);
/// assert_eq!(buf, "");
/// ```
#[stable(feature = "rust1", since = "1.0.0")]
fn read_line(&mut self, buf: &mut String) -> Result<usize> {
// Note that we are not calling the `.read_until` method here, but
// rather our hardcoded implementation. For more details as to why, see
// the comments in `default_read_to_string`.
unsafe { append_to_string(buf, |b| read_until(self, b'\n', b)) }
}
/// Returns an iterator over the contents of this reader split on the byte
/// `byte`.
///
/// The iterator returned from this function will return instances of
/// <code>[io::Result]<[Vec]\<u8>></code>. Each vector returned will *not* have
/// the delimiter byte at the end.
///
/// This function will yield errors whenever [`read_until`] would have
/// also yielded an error.
///
/// [io::Result]: self::Result "io::Result"
/// [`read_until`]: BufRead::read_until
///
/// # Examples
///
/// [`std::io::Cursor`][`Cursor`] is a type that implements `BufRead`. In
/// this example, we use [`Cursor`] to iterate over all hyphen delimited
/// segments in a byte slice
///
/// ```
/// use std::io::{self, BufRead};
///
/// let cursor = io::Cursor::new(b"lorem-ipsum-dolor");
///
/// let mut split_iter = cursor.split(b'-').map(|l| l.unwrap());
/// assert_eq!(split_iter.next(), Some(b"lorem".to_vec()));
/// assert_eq!(split_iter.next(), Some(b"ipsum".to_vec()));
/// assert_eq!(split_iter.next(), Some(b"dolor".to_vec()));
/// assert_eq!(split_iter.next(), None);
/// ```
#[stable(feature = "rust1", since = "1.0.0")]
fn split(self, byte: u8) -> Split<Self>
where
Self: Sized,
{
Split { buf: self, delim: byte }
}
/// Returns an iterator over the lines of this reader.
///
/// The iterator returned from this function will yield instances of
/// <code>[io::Result]<[String]></code>. Each string returned will *not* have a newline
/// byte (the `0xA` byte) or `CRLF` (`0xD`, `0xA` bytes) at the end.
///
/// [io::Result]: self::Result "io::Result"
///
/// # Examples
///
/// [`std::io::Cursor`][`Cursor`] is a type that implements `BufRead`. In
/// this example, we use [`Cursor`] to iterate over all the lines in a byte
/// slice.
///
/// ```
/// use std::io::{self, BufRead};
///
/// let cursor = io::Cursor::new(b"lorem\nipsum\r\ndolor");
///
/// let mut lines_iter = cursor.lines().map(|l| l.unwrap());
/// assert_eq!(lines_iter.next(), Some(String::from("lorem")));
/// assert_eq!(lines_iter.next(), Some(String::from("ipsum")));
/// assert_eq!(lines_iter.next(), Some(String::from("dolor")));
/// assert_eq!(lines_iter.next(), None);
/// ```
///
/// # Errors
///
/// Each line of the iterator has the same error semantics as [`BufRead::read_line`].
#[stable(feature = "rust1", since = "1.0.0")]
fn lines(self) -> Lines<Self>
where
Self: Sized,
{
Lines { buf: self }
}
}
/// Adapter to chain together two readers.
///
/// This struct is generally created by calling [`chain`] on a reader.
/// Please see the documentation of [`chain`] for more details.
///
/// [`chain`]: Read::chain
#[stable(feature = "rust1", since = "1.0.0")]
#[derive(Debug)]
pub struct Chain<T, U> {
first: T,
second: U,
done_first: bool,
}
impl<T, U> Chain<T, U> {
/// Consumes the `Chain`, returning the wrapped readers.
///
/// # Examples
///
/// ```no_run
/// use std::io;
/// use std::io::prelude::*;
/// use std::fs::File;
///
/// fn main() -> io::Result<()> {
/// let mut foo_file = File::open("foo.txt")?;
/// let mut bar_file = File::open("bar.txt")?;
///
/// let chain = foo_file.chain(bar_file);
/// let (foo_file, bar_file) = chain.into_inner();
/// Ok(())
/// }
/// ```
#[stable(feature = "more_io_inner_methods", since = "1.20.0")]
pub fn into_inner(self) -> (T, U) {
(self.first, self.second)
}
/// Gets references to the underlying readers in this `Chain`.
///
/// Care should be taken to avoid modifying the internal I/O state of the
/// underlying readers as doing so may corrupt the internal state of this
/// `Chain`.
///
/// # Examples
///
/// ```no_run
/// use std::io;
/// use std::io::prelude::*;
/// use std::fs::File;
///
/// fn main() -> io::Result<()> {
/// let mut foo_file = File::open("foo.txt")?;
/// let mut bar_file = File::open("bar.txt")?;
///
/// let chain = foo_file.chain(bar_file);
/// let (foo_file, bar_file) = chain.get_ref();
/// Ok(())
/// }
/// ```
#[stable(feature = "more_io_inner_methods", since = "1.20.0")]
pub fn get_ref(&self) -> (&T, &U) {
(&self.first, &self.second)
}
/// Gets mutable references to the underlying readers in this `Chain`.
///
/// Care should be taken to avoid modifying the internal I/O state of the
/// underlying readers as doing so may corrupt the internal state of this
/// `Chain`.
///
/// # Examples
///
/// ```no_run
/// use std::io;
/// use std::io::prelude::*;
/// use std::fs::File;
///
/// fn main() -> io::Result<()> {
/// let mut foo_file = File::open("foo.txt")?;
/// let mut bar_file = File::open("bar.txt")?;
///
/// let mut chain = foo_file.chain(bar_file);
/// let (foo_file, bar_file) = chain.get_mut();
/// Ok(())
/// }
/// ```
#[stable(feature = "more_io_inner_methods", since = "1.20.0")]
pub fn get_mut(&mut self) -> (&mut T, &mut U) {
(&mut self.first, &mut self.second)
}
}
#[stable(feature = "rust1", since = "1.0.0")]
impl<T: Read, U: Read> Read for Chain<T, U> {
fn read(&mut self, buf: &mut [u8]) -> Result<usize> {
if !self.done_first {
match self.first.read(buf)? {
0 if !buf.is_empty() => self.done_first = true,
n => return Ok(n),
}
}
self.second.read(buf)
}
fn read_vectored(&mut self, bufs: &mut [IoSliceMut<'_>]) -> Result<usize> {
if !self.done_first {
match self.first.read_vectored(bufs)? {
0 if bufs.iter().any(|b| !b.is_empty()) => self.done_first = true,
n => return Ok(n),
}
}
self.second.read_vectored(bufs)
}
#[inline]
fn is_read_vectored(&self) -> bool {
self.first.is_read_vectored() || self.second.is_read_vectored()
}
fn read_to_end(&mut self, buf: &mut Vec<u8>) -> Result<usize> {
let mut read = 0;
if !self.done_first {
read += self.first.read_to_end(buf)?;
self.done_first = true;
}
read += self.second.read_to_end(buf)?;
Ok(read)
}
// We don't override `read_to_string` here because an UTF-8 sequence could
// be split between the two parts of the chain
fn read_buf(&mut self, mut buf: BorrowedCursor<'_>) -> Result<()> {
if buf.capacity() == 0 {
return Ok(());
}
if !self.done_first {
let old_len = buf.written();
self.first.read_buf(buf.reborrow())?;
if buf.written() != old_len {
return Ok(());
} else {
self.done_first = true;
}
}
self.second.read_buf(buf)
}
}
#[stable(feature = "chain_bufread", since = "1.9.0")]
impl<T: BufRead, U: BufRead> BufRead for Chain<T, U> {
fn fill_buf(&mut self) -> Result<&[u8]> {
if !self.done_first {
match self.first.fill_buf()? {
buf if buf.is_empty() => self.done_first = true,
buf => return Ok(buf),
}
}
self.second.fill_buf()
}
fn consume(&mut self, amt: usize) {
if !self.done_first { self.first.consume(amt) } else { self.second.consume(amt) }
}
fn read_until(&mut self, byte: u8, buf: &mut Vec<u8>) -> Result<usize> {
let mut read = 0;
if !self.done_first {
let n = self.first.read_until(byte, buf)?;
read += n;
match buf.last() {
Some(b) if *b == byte && n != 0 => return Ok(read),
_ => self.done_first = true,
}
}
read += self.second.read_until(byte, buf)?;
Ok(read)
}
// We don't override `read_line` here because an UTF-8 sequence could be
// split between the two parts of the chain
}
impl<T, U> SizeHint for Chain<T, U> {
#[inline]
fn lower_bound(&self) -> usize {
SizeHint::lower_bound(&self.first) + SizeHint::lower_bound(&self.second)
}
#[inline]
fn upper_bound(&self) -> Option<usize> {
match (SizeHint::upper_bound(&self.first), SizeHint::upper_bound(&self.second)) {
(Some(first), Some(second)) => first.checked_add(second),
_ => None,
}
}
}
/// Reader adapter which limits the bytes read from an underlying reader.
///
/// This struct is generally created by calling [`take`] on a reader.
/// Please see the documentation of [`take`] for more details.
///
/// [`take`]: Read::take
#[stable(feature = "rust1", since = "1.0.0")]
#[derive(Debug)]
pub struct Take<T> {
inner: T,
len: u64,
limit: u64,
}
impl<T> Take<T> {
/// Returns the number of bytes that can be read before this instance will
/// return EOF.
///
/// # Note
///
/// This instance may reach `EOF` after reading fewer bytes than indicated by
/// this method if the underlying [`Read`] instance reaches EOF.
///
/// # Examples
///
/// ```no_run
/// use std::io;
/// use std::io::prelude::*;
/// use std::fs::File;
///
/// fn main() -> io::Result<()> {
/// let f = File::open("foo.txt")?;
///
/// // read at most five bytes
/// let handle = f.take(5);
///
/// println!("limit: {}", handle.limit());
/// Ok(())
/// }
/// ```
#[stable(feature = "rust1", since = "1.0.0")]
pub fn limit(&self) -> u64 {
self.limit
}
/// Returns the number of bytes read so far.
#[unstable(feature = "seek_io_take_position", issue = "97227")]
pub fn position(&self) -> u64 {
self.len - self.limit
}
/// Sets the number of bytes that can be read before this instance will
/// return EOF. This is the same as constructing a new `Take` instance, so
/// the amount of bytes read and the previous limit value don't matter when
/// calling this method.
///
/// # Examples
///
/// ```no_run
/// use std::io;
/// use std::io::prelude::*;
/// use std::fs::File;
///
/// fn main() -> io::Result<()> {
/// let f = File::open("foo.txt")?;
///
/// // read at most five bytes
/// let mut handle = f.take(5);
/// handle.set_limit(10);
///
/// assert_eq!(handle.limit(), 10);
/// Ok(())
/// }
/// ```
#[stable(feature = "take_set_limit", since = "1.27.0")]
pub fn set_limit(&mut self, limit: u64) {
self.len = limit;
self.limit = limit;
}
/// Consumes the `Take`, returning the wrapped reader.
///
/// # Examples
///
/// ```no_run
/// use std::io;
/// use std::io::prelude::*;
/// use std::fs::File;
///
/// fn main() -> io::Result<()> {
/// let mut file = File::open("foo.txt")?;
///
/// let mut buffer = [0; 5];
/// let mut handle = file.take(5);
/// handle.read(&mut buffer)?;
///
/// let file = handle.into_inner();
/// Ok(())
/// }
/// ```
#[stable(feature = "io_take_into_inner", since = "1.15.0")]
pub fn into_inner(self) -> T {
self.inner
}
/// Gets a reference to the underlying reader.
///
/// Care should be taken to avoid modifying the internal I/O state of the
/// underlying reader as doing so may corrupt the internal limit of this
/// `Take`.
///
/// # Examples
///
/// ```no_run
/// use std::io;
/// use std::io::prelude::*;
/// use std::fs::File;
///
/// fn main() -> io::Result<()> {
/// let mut file = File::open("foo.txt")?;
///
/// let mut buffer = [0; 5];
/// let mut handle = file.take(5);
/// handle.read(&mut buffer)?;
///
/// let file = handle.get_ref();
/// Ok(())
/// }
/// ```
#[stable(feature = "more_io_inner_methods", since = "1.20.0")]
pub fn get_ref(&self) -> &T {
&self.inner
}
/// Gets a mutable reference to the underlying reader.
///
/// Care should be taken to avoid modifying the internal I/O state of the
/// underlying reader as doing so may corrupt the internal limit of this
/// `Take`.
///
/// # Examples
///
/// ```no_run
/// use std::io;
/// use std::io::prelude::*;
/// use std::fs::File;
///
/// fn main() -> io::Result<()> {
/// let mut file = File::open("foo.txt")?;
///
/// let mut buffer = [0; 5];
/// let mut handle = file.take(5);
/// handle.read(&mut buffer)?;
///
/// let file = handle.get_mut();
/// Ok(())
/// }
/// ```
#[stable(feature = "more_io_inner_methods", since = "1.20.0")]
pub fn get_mut(&mut self) -> &mut T {
&mut self.inner
}
}
#[stable(feature = "rust1", since = "1.0.0")]
impl<T: Read> Read for Take<T> {
fn read(&mut self, buf: &mut [u8]) -> Result<usize> {
// Don't call into inner reader at all at EOF because it may still block
if self.limit == 0 {
return Ok(0);
}
let max = cmp::min(buf.len() as u64, self.limit) as usize;
let n = self.inner.read(&mut buf[..max])?;
assert!(n as u64 <= self.limit, "number of read bytes exceeds limit");
self.limit -= n as u64;
Ok(n)
}
fn read_buf(&mut self, mut buf: BorrowedCursor<'_>) -> Result<()> {
// Don't call into inner reader at all at EOF because it may still block
if self.limit == 0 {
return Ok(());
}
if self.limit < buf.capacity() as u64 {
// The condition above guarantees that `self.limit` fits in `usize`.
let limit = self.limit as usize;
let extra_init = cmp::min(limit, buf.init_mut().len());
// SAFETY: no uninit data is written to ibuf
let ibuf = unsafe { &mut buf.as_mut()[..limit] };
let mut sliced_buf: BorrowedBuf<'_> = ibuf.into();
// SAFETY: extra_init bytes of ibuf are known to be initialized
unsafe {
sliced_buf.set_init(extra_init);
}
let mut cursor = sliced_buf.unfilled();
let result = self.inner.read_buf(cursor.reborrow());
let new_init = cursor.init_mut().len();
let filled = sliced_buf.len();
// cursor / sliced_buf / ibuf must drop here
unsafe {
// SAFETY: filled bytes have been filled and therefore initialized
buf.advance_unchecked(filled);
// SAFETY: new_init bytes of buf's unfilled buffer have been initialized
buf.set_init(new_init);
}
self.limit -= filled as u64;
result
} else {
let written = buf.written();
let result = self.inner.read_buf(buf.reborrow());
self.limit -= (buf.written() - written) as u64;
result
}
}
}
#[stable(feature = "rust1", since = "1.0.0")]
impl<T: BufRead> BufRead for Take<T> {
fn fill_buf(&mut self) -> Result<&[u8]> {
// Don't call into inner reader at all at EOF because it may still block
if self.limit == 0 {
return Ok(&[]);
}
let buf = self.inner.fill_buf()?;
let cap = cmp::min(buf.len() as u64, self.limit) as usize;
Ok(&buf[..cap])
}
fn consume(&mut self, amt: usize) {
// Don't let callers reset the limit by passing an overlarge value
let amt = cmp::min(amt as u64, self.limit) as usize;
self.limit -= amt as u64;
self.inner.consume(amt);
}
}
impl<T> SizeHint for Take<T> {
#[inline]
fn lower_bound(&self) -> usize {
cmp::min(SizeHint::lower_bound(&self.inner) as u64, self.limit) as usize
}
#[inline]
fn upper_bound(&self) -> Option<usize> {
match SizeHint::upper_bound(&self.inner) {
Some(upper_bound) => Some(cmp::min(upper_bound as u64, self.limit) as usize),
None => self.limit.try_into().ok(),
}
}
}
#[stable(feature = "seek_io_take", since = "1.89.0")]
impl<T: Seek> Seek for Take<T> {
fn seek(&mut self, pos: SeekFrom) -> Result<u64> {
let new_position = match pos {
SeekFrom::Start(v) => Some(v),
SeekFrom::Current(v) => self.position().checked_add_signed(v),
SeekFrom::End(v) => self.len.checked_add_signed(v),
};
let new_position = match new_position {
Some(v) if v <= self.len => v,
_ => return Err(ErrorKind::InvalidInput.into()),
};
while new_position != self.position() {
if let Some(offset) = new_position.checked_signed_diff(self.position()) {
self.inner.seek_relative(offset)?;
self.limit = self.limit.wrapping_sub(offset as u64);
break;
}
let offset = if new_position > self.position() { i64::MAX } else { i64::MIN };
self.inner.seek_relative(offset)?;
self.limit = self.limit.wrapping_sub(offset as u64);
}
Ok(new_position)
}
fn stream_len(&mut self) -> Result<u64> {
Ok(self.len)
}
fn stream_position(&mut self) -> Result<u64> {
Ok(self.position())
}
fn seek_relative(&mut self, offset: i64) -> Result<()> {
if !self.position().checked_add_signed(offset).is_some_and(|p| p <= self.len) {
return Err(ErrorKind::InvalidInput.into());
}
self.inner.seek_relative(offset)?;
self.limit = self.limit.wrapping_sub(offset as u64);
Ok(())
}
}
/// An iterator over `u8` values of a reader.
///
/// This struct is generally created by calling [`bytes`] on a reader.
/// Please see the documentation of [`bytes`] for more details.
///
/// [`bytes`]: Read::bytes
#[stable(feature = "rust1", since = "1.0.0")]
#[derive(Debug)]
pub struct Bytes<R> {
inner: R,
}
#[stable(feature = "rust1", since = "1.0.0")]
impl<R: Read> Iterator for Bytes<R> {
type Item = Result<u8>;
// Not `#[inline]`. This function gets inlined even without it, but having
// the inline annotation can result in worse code generation. See #116785.
fn next(&mut self) -> Option<Result<u8>> {
SpecReadByte::spec_read_byte(&mut self.inner)
}
#[inline]
fn size_hint(&self) -> (usize, Option<usize>) {
SizeHint::size_hint(&self.inner)
}
}
/// For the specialization of `Bytes::next`.
trait SpecReadByte {
fn spec_read_byte(&mut self) -> Option<Result<u8>>;
}
impl<R> SpecReadByte for R
where
Self: Read,
{
#[inline]
default fn spec_read_byte(&mut self) -> Option<Result<u8>> {
inlined_slow_read_byte(self)
}
}
/// Reads a single byte in a slow, generic way. This is used by the default
/// `spec_read_byte`.
#[inline]
fn inlined_slow_read_byte<R: Read>(reader: &mut R) -> Option<Result<u8>> {
let mut byte = 0;
loop {
return match reader.read(slice::from_mut(&mut byte)) {
Ok(0) => None,
Ok(..) => Some(Ok(byte)),
Err(ref e) if e.is_interrupted() => continue,
Err(e) => Some(Err(e)),
};
}
}
// Used by `BufReader::spec_read_byte`, for which the `inline(never)` is
// important.
#[inline(never)]
fn uninlined_slow_read_byte<R: Read>(reader: &mut R) -> Option<Result<u8>> {
inlined_slow_read_byte(reader)
}
trait SizeHint {
fn lower_bound(&self) -> usize;
fn upper_bound(&self) -> Option<usize>;
fn size_hint(&self) -> (usize, Option<usize>) {
(self.lower_bound(), self.upper_bound())
}
}
impl<T: ?Sized> SizeHint for T {
#[inline]
default fn lower_bound(&self) -> usize {
0
}
#[inline]
default fn upper_bound(&self) -> Option<usize> {
None
}
}
impl<T> SizeHint for &mut T {
#[inline]
fn lower_bound(&self) -> usize {
SizeHint::lower_bound(*self)
}
#[inline]
fn upper_bound(&self) -> Option<usize> {
SizeHint::upper_bound(*self)
}
}
impl<T> SizeHint for Box<T> {
#[inline]
fn lower_bound(&self) -> usize {
SizeHint::lower_bound(&**self)
}
#[inline]
fn upper_bound(&self) -> Option<usize> {
SizeHint::upper_bound(&**self)
}
}
impl SizeHint for &[u8] {
#[inline]
fn lower_bound(&self) -> usize {
self.len()
}
#[inline]
fn upper_bound(&self) -> Option<usize> {
Some(self.len())
}
}
/// An iterator over the contents of an instance of `BufRead` split on a
/// particular byte.
///
/// This struct is generally created by calling [`split`] on a `BufRead`.
/// Please see the documentation of [`split`] for more details.
///
/// [`split`]: BufRead::split
#[stable(feature = "rust1", since = "1.0.0")]
#[derive(Debug)]
pub struct Split<B> {
buf: B,
delim: u8,
}
#[stable(feature = "rust1", since = "1.0.0")]
impl<B: BufRead> Iterator for Split<B> {
type Item = Result<Vec<u8>>;
fn next(&mut self) -> Option<Result<Vec<u8>>> {
let mut buf = Vec::new();
match self.buf.read_until(self.delim, &mut buf) {
Ok(0) => None,
Ok(_n) => {
if buf[buf.len() - 1] == self.delim {
buf.pop();
}
Some(Ok(buf))
}
Err(e) => Some(Err(e)),
}
}
}
/// An iterator over the lines of an instance of `BufRead`.
///
/// This struct is generally created by calling [`lines`] on a `BufRead`.
/// Please see the documentation of [`lines`] for more details.
///
/// [`lines`]: BufRead::lines
#[stable(feature = "rust1", since = "1.0.0")]
#[derive(Debug)]
#[cfg_attr(not(test), rustc_diagnostic_item = "IoLines")]
pub struct Lines<B> {
buf: B,
}
#[stable(feature = "rust1", since = "1.0.0")]
impl<B: BufRead> Iterator for Lines<B> {
type Item = Result<String>;
fn next(&mut self) -> Option<Result<String>> {
let mut buf = String::new();
match self.buf.read_line(&mut buf) {
Ok(0) => None,
Ok(_n) => {
if buf.ends_with('\n') {
buf.pop();
if buf.ends_with('\r') {
buf.pop();
}
}
Some(Ok(buf))
}
Err(e) => Some(Err(e)),
}
}
}