1 // Copyright 2012 The Rust Project Developers. See the COPYRIGHT
2 // file at the top-level directory of this distribution and at
3 // http://rust-lang.org/COPYRIGHT.
5 // Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or
6 // http://www.apache.org/licenses/LICENSE-2.0> or the MIT license
7 // <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your
8 // option. This file may not be copied, modified, or distributed
9 // except according to those terms.
12 use borrow::{Cow, ToOwned, Borrow};
14 use convert::{Into, From};
15 use cmp::{PartialEq, Eq, PartialOrd, Ord, Ordering};
17 use fmt::{self, Write};
23 use option::Option::{self, Some, None};
24 use result::Result::{self, Ok, Err};
30 /// A type representing an owned C-compatible string
32 /// This type serves the primary purpose of being able to safely generate a
33 /// C-compatible string from a Rust byte slice or vector. An instance of this
34 /// type is a static guarantee that the underlying bytes contain no interior 0
35 /// bytes and the final byte is 0.
37 /// A `CString` is created from either a byte slice or a byte vector. After
38 /// being created, a `CString` predominately inherits all of its methods from
39 /// the `Deref` implementation to `[libc::c_char]`. Note that the underlying
40 /// array is represented as an array of `libc::c_char` as opposed to `u8`. A
41 /// `u8` slice can be obtained with the `as_bytes` method. Slices produced from
42 /// a `CString` do *not* contain the trailing nul terminator unless otherwise
48 /// # #![feature(libc)]
49 /// # extern crate libc;
51 /// use std::ffi::CString;
55 /// fn my_printer(s: *const libc::c_char);
58 /// let c_to_print = CString::new("Hello, world!").unwrap();
60 /// my_printer(c_to_print.as_ptr());
64 #[derive(PartialEq, PartialOrd, Eq, Ord, Hash, Clone)]
65 #[stable(feature = "rust1", since = "1.0.0")]
70 /// Representation of a borrowed C string.
72 /// This dynamically sized type is only safely constructed via a borrowed
73 /// version of an instance of `CString`. This type can be constructed from a raw
74 /// C string as well and represents a C string borrowed from another location.
76 /// Note that this structure is **not** `repr(C)` and is not recommended to be
77 /// placed in the signatures of FFI functions. Instead safe wrappers of FFI
78 /// functions may leverage the unsafe `from_ptr` constructor to provide a safe
79 /// interface to other consumers.
83 /// Inspecting a foreign C string
86 /// # #![feature(libc)]
87 /// extern crate libc;
88 /// use std::ffi::CStr;
90 /// extern { fn my_string() -> *const libc::c_char; }
94 /// let slice = CStr::from_ptr(my_string());
95 /// println!("string length: {}", slice.to_bytes().len());
100 /// Passing a Rust-originating C string
103 /// # #![feature(libc)]
104 /// extern crate libc;
105 /// use std::ffi::{CString, CStr};
107 /// fn work(data: &CStr) {
108 /// extern { fn work_with(data: *const libc::c_char); }
110 /// unsafe { work_with(data.as_ptr()) }
114 /// let s = CString::new("data data data data").unwrap();
119 /// Converting a foreign C string into a Rust `String`
122 /// # #![feature(libc,cstr_to_str)]
123 /// extern crate libc;
124 /// use std::ffi::CStr;
126 /// extern { fn my_string() -> *const libc::c_char; }
128 /// fn my_string_safe() -> String {
130 /// CStr::from_ptr(my_string()).to_string_lossy().into_owned()
135 /// println!("string: {}", my_string_safe());
139 #[stable(feature = "rust1", since = "1.0.0")]
141 // FIXME: this should not be represented with a DST slice but rather with
142 // just a raw `libc::c_char` along with some form of marker to make
143 // this an unsized type. Essentially `sizeof(&CStr)` should be the
144 // same as `sizeof(&c_char)` but `CStr` should be an unsized type.
145 inner: [libc::c_char]
148 /// An error returned from `CString::new` to indicate that a nul byte was found
149 /// in the vector provided.
150 #[derive(Clone, PartialEq, Debug)]
151 #[stable(feature = "rust1", since = "1.0.0")]
152 pub struct NulError(usize, Vec<u8>);
155 /// Creates a new C-compatible string from a container of bytes.
157 /// This method will consume the provided data and use the underlying bytes
158 /// to construct a new string, ensuring that there is a trailing 0 byte.
163 /// # #![feature(libc)]
164 /// extern crate libc;
165 /// use std::ffi::CString;
167 /// extern { fn puts(s: *const libc::c_char); }
170 /// let to_print = CString::new("Hello!").unwrap();
172 /// puts(to_print.as_ptr());
179 /// This function will return an error if the bytes yielded contain an
180 /// internal 0 byte. The error returned will contain the bytes as well as
181 /// the position of the nul byte.
182 #[stable(feature = "rust1", since = "1.0.0")]
183 pub fn new<T: Into<Vec<u8>>>(t: T) -> Result<CString, NulError> {
184 let bytes = t.into();
185 match bytes.iter().position(|x| *x == 0) {
186 Some(i) => Err(NulError(i, bytes)),
187 None => Ok(unsafe { CString::from_vec_unchecked(bytes) }),
191 /// Creates a C-compatible string from a byte vector without checking for
192 /// interior 0 bytes.
194 /// This method is equivalent to `new` except that no runtime assertion
195 /// is made that `v` contains no 0 bytes, and it requires an actual
196 /// byte vector, not anything that can be converted to one with Into.
197 #[stable(feature = "rust1", since = "1.0.0")]
198 pub unsafe fn from_vec_unchecked(mut v: Vec<u8>) -> CString {
200 CString { inner: v.into_boxed_slice() }
203 /// Retakes ownership of a CString that was transferred to C.
205 /// The only appropriate argument is a pointer obtained by calling
206 /// `into_ptr`. The length of the string will be recalculated
207 /// using the pointer.
208 #[unstable(feature = "cstr_memory", reason = "recently added")]
209 // NB: may want to be called from_raw, needs to consider CStr::from_ptr,
210 // Box::from_raw (or whatever it's currently called), and
211 // slice::from_raw_parts
212 pub unsafe fn from_ptr(ptr: *const libc::c_char) -> CString {
213 let len = libc::strlen(ptr) + 1; // Including the NUL byte
214 let slice = slice::from_raw_parts(ptr, len as usize);
215 CString { inner: mem::transmute(slice) }
218 /// Transfers ownership of the string to a C caller.
220 /// The pointer must be returned to Rust and reconstituted using
221 /// `from_ptr` to be properly deallocated. Specifically, one
222 /// should *not* use the standard C `free` function to deallocate
225 /// Failure to call `from_ptr` will lead to a memory leak.
226 #[unstable(feature = "cstr_memory", reason = "recently added")]
227 // NB: may want to be called into_raw, see comments on from_ptr
228 pub fn into_ptr(self) -> *const libc::c_char {
229 // It is important that the bytes be sized to fit - we need
230 // the capacity to be determinable from the string length, and
231 // shrinking to fit is the only way to be sure.
232 Box::into_raw(self.inner) as *const libc::c_char
235 /// Returns the contents of this `CString` as a slice of bytes.
237 /// The returned slice does **not** contain the trailing nul separator and
238 /// it is guaranteed to not have any interior nul bytes.
239 #[stable(feature = "rust1", since = "1.0.0")]
240 pub fn as_bytes(&self) -> &[u8] {
241 &self.inner[..self.inner.len() - 1]
244 /// Equivalent to the `as_bytes` function except that the returned slice
245 /// includes the trailing nul byte.
246 #[stable(feature = "rust1", since = "1.0.0")]
247 pub fn as_bytes_with_nul(&self) -> &[u8] {
252 #[stable(feature = "rust1", since = "1.0.0")]
253 impl Deref for CString {
256 fn deref(&self) -> &CStr {
257 unsafe { mem::transmute(self.as_bytes_with_nul()) }
261 #[stable(feature = "rust1", since = "1.0.0")]
262 impl fmt::Debug for CString {
263 fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
264 fmt::Debug::fmt(&**self, f)
268 #[stable(feature = "cstr_debug", since = "1.3.0")]
269 impl fmt::Debug for CStr {
270 fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
271 try!(write!(f, "\""));
272 for byte in self.to_bytes().iter().flat_map(|&b| ascii::escape_default(b)) {
273 try!(f.write_char(byte as char));
279 #[stable(feature = "cstr_borrow", since = "1.3.0")]
280 impl Borrow<CStr> for CString {
281 fn borrow(&self) -> &CStr { self }
285 /// Returns the position of the nul byte in the slice that was provided to
287 #[stable(feature = "rust1", since = "1.0.0")]
288 pub fn nul_position(&self) -> usize { self.0 }
290 /// Consumes this error, returning the underlying vector of bytes which
291 /// generated the error in the first place.
292 #[stable(feature = "rust1", since = "1.0.0")]
293 pub fn into_vec(self) -> Vec<u8> { self.1 }
296 #[stable(feature = "rust1", since = "1.0.0")]
297 impl Error for NulError {
298 fn description(&self) -> &str { "nul byte found in data" }
301 #[stable(feature = "rust1", since = "1.0.0")]
302 impl fmt::Display for NulError {
303 fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
304 write!(f, "nul byte found in provided data at position: {}", self.0)
308 #[stable(feature = "rust1", since = "1.0.0")]
309 impl From<NulError> for io::Error {
310 fn from(_: NulError) -> io::Error {
311 io::Error::new(io::ErrorKind::InvalidInput,
312 "data provided contains a nul byte")
317 /// Casts a raw C string to a safe C string wrapper.
319 /// This function will cast the provided `ptr` to the `CStr` wrapper which
320 /// allows inspection and interoperation of non-owned C strings. This method
321 /// is unsafe for a number of reasons:
323 /// * There is no guarantee to the validity of `ptr`
324 /// * The returned lifetime is not guaranteed to be the actual lifetime of
326 /// * There is no guarantee that the memory pointed to by `ptr` contains a
327 /// valid nul terminator byte at the end of the string.
329 /// > **Note**: This operation is intended to be a 0-cost cast but it is
330 /// > currently implemented with an up-front calculation of the length of
331 /// > the string. This is not guaranteed to always be the case.
336 /// # #![feature(libc)]
337 /// # extern crate libc;
339 /// use std::ffi::CStr;
344 /// fn my_string() -> *const libc::c_char;
348 /// let slice = CStr::from_ptr(my_string());
349 /// println!("string returned: {}",
350 /// str::from_utf8(slice.to_bytes()).unwrap());
354 #[stable(feature = "rust1", since = "1.0.0")]
355 pub unsafe fn from_ptr<'a>(ptr: *const libc::c_char) -> &'a CStr {
356 let len = libc::strlen(ptr);
357 mem::transmute(slice::from_raw_parts(ptr, len as usize + 1))
360 /// Returns the inner pointer to this C string.
362 /// The returned pointer will be valid for as long as `self` is and points
363 /// to a contiguous region of memory terminated with a 0 byte to represent
364 /// the end of the string.
365 #[stable(feature = "rust1", since = "1.0.0")]
366 pub fn as_ptr(&self) -> *const libc::c_char {
370 /// Converts this C string to a byte slice.
372 /// This function will calculate the length of this string (which normally
373 /// requires a linear amount of work to be done) and then return the
374 /// resulting slice of `u8` elements.
376 /// The returned slice will **not** contain the trailing nul that this C
379 /// > **Note**: This method is currently implemented as a 0-cost cast, but
380 /// > it is planned to alter its definition in the future to perform the
381 /// > length calculation whenever this method is called.
382 #[stable(feature = "rust1", since = "1.0.0")]
383 pub fn to_bytes(&self) -> &[u8] {
384 let bytes = self.to_bytes_with_nul();
385 &bytes[..bytes.len() - 1]
388 /// Converts this C string to a byte slice containing the trailing 0 byte.
390 /// This function is the equivalent of `to_bytes` except that it will retain
391 /// the trailing nul instead of chopping it off.
393 /// > **Note**: This method is currently implemented as a 0-cost cast, but
394 /// > it is planned to alter its definition in the future to perform the
395 /// > length calculation whenever this method is called.
396 #[stable(feature = "rust1", since = "1.0.0")]
397 pub fn to_bytes_with_nul(&self) -> &[u8] {
398 unsafe { mem::transmute(&self.inner) }
401 /// Yields a `&str` slice if the `CStr` contains valid UTF-8.
403 /// This function will calculate the length of this string and check for
404 /// UTF-8 validity, and then return the `&str` if it's valid.
406 /// > **Note**: This method is currently implemented to check for validity
407 /// > after a 0-cost cast, but it is planned to alter its definition in the
408 /// > future to perform the length calculation in addition to the UTF-8
409 /// > check whenever this method is called.
410 #[unstable(feature = "cstr_to_str", reason = "recently added")]
411 pub fn to_str(&self) -> Result<&str, str::Utf8Error> {
412 // NB: When CStr is changed to perform the length check in .to_bytes() instead of in
413 // from_ptr(), it may be worth considering if this should be rewritten to do the UTF-8
414 // check inline with the length calculation instead of doing it afterwards.
415 str::from_utf8(self.to_bytes())
418 /// Converts a `CStr` into a `Cow<str>`.
420 /// This function will calculate the length of this string (which normally
421 /// requires a linear amount of work to be done) and then return the
422 /// resulting slice as a `Cow<str>`, replacing any invalid UTF-8 sequences
423 /// with `U+FFFD REPLACEMENT CHARACTER`.
425 /// > **Note**: This method is currently implemented to check for validity
426 /// > after a 0-cost cast, but it is planned to alter its definition in the
427 /// > future to perform the length calculation in addition to the UTF-8
428 /// > check whenever this method is called.
429 #[unstable(feature = "cstr_to_str", reason = "recently added")]
430 pub fn to_string_lossy(&self) -> Cow<str> {
431 String::from_utf8_lossy(self.to_bytes())
435 #[stable(feature = "rust1", since = "1.0.0")]
436 impl PartialEq for CStr {
437 fn eq(&self, other: &CStr) -> bool {
438 self.to_bytes().eq(other.to_bytes())
441 #[stable(feature = "rust1", since = "1.0.0")]
443 #[stable(feature = "rust1", since = "1.0.0")]
444 impl PartialOrd for CStr {
445 fn partial_cmp(&self, other: &CStr) -> Option<Ordering> {
446 self.to_bytes().partial_cmp(&other.to_bytes())
449 #[stable(feature = "rust1", since = "1.0.0")]
451 fn cmp(&self, other: &CStr) -> Ordering {
452 self.to_bytes().cmp(&other.to_bytes())
456 #[stable(feature = "cstr_borrow", since = "1.3.0")]
457 impl ToOwned for CStr {
458 type Owned = CString;
460 fn to_owned(&self) -> CString {
461 unsafe { CString::from_vec_unchecked(self.to_bytes().to_vec()) }
470 use borrow::Cow::{Borrowed, Owned};
471 use hash::{SipHasher, Hash, Hasher};
476 let ptr = data.as_ptr() as *const libc::c_char;
478 assert_eq!(CStr::from_ptr(ptr).to_bytes(), b"123");
479 assert_eq!(CStr::from_ptr(ptr).to_bytes_with_nul(), b"123\0");
485 let s = CString::new("1234").unwrap();
486 assert_eq!(s.as_bytes(), b"1234");
487 assert_eq!(s.as_bytes_with_nul(), b"1234\0");
491 fn build_with_zero1() {
492 assert!(CString::new(&b"\0"[..]).is_err());
495 fn build_with_zero2() {
496 assert!(CString::new(vec![0]).is_err());
500 fn build_with_zero3() {
502 let s = CString::from_vec_unchecked(vec![0]);
503 assert_eq!(s.as_bytes(), b"\0");
509 let s = CString::new(&b"abc\x01\x02\n\xE2\x80\xA6\xFF"[..]).unwrap();
510 assert_eq!(format!("{:?}", s), r#""abc\x01\x02\n\xe2\x80\xa6\xff""#);
516 let s = CStr::from_ptr(b"12\0".as_ptr() as *const _);
517 assert_eq!(s.to_bytes(), b"12");
518 assert_eq!(s.to_bytes_with_nul(), b"12\0");
524 let data = b"123\xE2\x80\xA6\0";
525 let ptr = data.as_ptr() as *const libc::c_char;
527 assert_eq!(CStr::from_ptr(ptr).to_str(), Ok("123…"));
528 assert_eq!(CStr::from_ptr(ptr).to_string_lossy(), Borrowed("123…"));
530 let data = b"123\xE2\0";
531 let ptr = data.as_ptr() as *const libc::c_char;
533 assert!(CStr::from_ptr(ptr).to_str().is_err());
534 assert_eq!(CStr::from_ptr(ptr).to_string_lossy(), Owned::<str>(format!("123\u{FFFD}")));
541 let ptr = data.as_ptr() as *const libc::c_char;
543 let owned = unsafe { CStr::from_ptr(ptr).to_owned() };
544 assert_eq!(owned.as_bytes_with_nul(), data);
549 let data = b"123\xE2\xFA\xA6\0";
550 let ptr = data.as_ptr() as *const libc::c_char;
551 let cstr: &'static CStr = unsafe { CStr::from_ptr(ptr) };
553 let mut s = SipHasher::new_with_keys(0, 0);
555 let cstr_hash = s.finish();
556 let mut s = SipHasher::new_with_keys(0, 0);
557 CString::new(&data[..data.len() - 1]).unwrap().hash(&mut s);
558 let cstring_hash = s.finish();
560 assert_eq!(cstr_hash, cstring_hash);