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.
13 C-string manipulation and management
15 This modules provides the basic methods for creating and manipulating
16 null-terminated strings for use with FFI calls (back to C). Most C APIs require
17 that the string being passed to them is null-terminated, and by default rust's
18 string types are *not* null terminated.
20 The other problem with translating Rust strings to C strings is that Rust
21 strings can validly contain a null-byte in the middle of the string (0 is a
22 valid unicode codepoint). This means that not all Rust strings can actually be
23 translated to C strings.
25 # Creation of a C string
27 A C string is managed through the `CString` type defined in this module. It
28 "owns" the internal buffer of characters and will automatically deallocate the
29 buffer when the string is dropped. The `ToCStr` trait is implemented for `&str`
30 and `&[u8]`, but the conversions can fail due to some of the limitations
33 This also means that currently whenever a C string is created, an allocation
34 must be performed to place the data elsewhere (the lifetime of the C string is
35 not tied to the lifetime of the original string/data buffer). If C strings are
36 heavily used in applications, then caching may be advisable to prevent
37 unnecessary amounts of allocations.
39 An example of creating and using a C string would be:
45 fn puts(s: *libc::c_char);
49 let my_string = "Hello, world!";
51 // Allocate the C string with an explicit local that owns the string. The
52 // `c_buffer` pointer will be deallocated when `my_c_string` goes out of scope.
53 let my_c_string = my_string.to_c_str();
54 my_c_string.with_ref(|c_buffer| {
55 unsafe { puts(c_buffer); }
58 // Don't save off the allocation of the C string, the `c_buffer` will be
59 // deallocated when this block returns!
60 my_string.with_c_str(|c_buffer| {
61 unsafe { puts(c_buffer); }
70 use alloc::libc_heap::malloc_raw;
71 use collections::string::String;
72 use core::kinds::marker;
80 /// The representation of a C String.
82 /// This structure wraps a `*libc::c_char`, and will automatically free the
83 /// memory it is pointing to when it goes out of scope.
89 impl Clone for CString {
90 /// Clone this CString into a new, uniquely owned CString. For safety
91 /// reasons, this is always a deep clone, rather than the usual shallow
93 fn clone(&self) -> CString {
94 if self.buf.is_null() {
95 CString { buf: self.buf, owns_buffer_: self.owns_buffer_ }
97 let len = self.len() + 1;
98 let buf = unsafe { malloc_raw(len) } as *mut libc::c_char;
99 unsafe { ptr::copy_nonoverlapping_memory(buf, self.buf, len); }
100 CString { buf: buf as *libc::c_char, owns_buffer_: true }
105 impl PartialEq for CString {
106 fn eq(&self, other: &CString) -> bool {
107 if self.buf as uint == other.buf as uint {
109 } else if self.buf.is_null() || other.buf.is_null() {
113 libc::strcmp(self.buf, other.buf) == 0
120 /// Create a C String from a pointer.
121 pub unsafe fn new(buf: *libc::c_char, owns_buffer: bool) -> CString {
122 CString { buf: buf, owns_buffer_: owns_buffer }
125 /// Unwraps the wrapped `*libc::c_char` from the `CString` wrapper.
126 /// Any ownership of the buffer by the `CString` wrapper is forgotten.
127 pub unsafe fn unwrap(self) -> *libc::c_char {
128 let mut c_str = self;
129 c_str.owns_buffer_ = false;
133 /// Calls a closure with a reference to the underlying `*libc::c_char`.
137 /// Fails if the CString is null.
138 pub fn with_ref<T>(&self, f: |*libc::c_char| -> T) -> T {
139 if self.buf.is_null() { fail!("CString is null!"); }
143 /// Calls a closure with a mutable reference to the underlying `*libc::c_char`.
147 /// Fails if the CString is null.
148 pub fn with_mut_ref<T>(&mut self, f: |*mut libc::c_char| -> T) -> T {
149 if self.buf.is_null() { fail!("CString is null!"); }
150 f(self.buf as *mut libc::c_char)
153 /// Returns true if the CString is a null.
154 pub fn is_null(&self) -> bool {
158 /// Returns true if the CString is not null.
159 pub fn is_not_null(&self) -> bool {
160 self.buf.is_not_null()
163 /// Returns whether or not the `CString` owns the buffer.
164 pub fn owns_buffer(&self) -> bool {
168 /// Converts the CString into a `&[u8]` without copying.
169 /// Includes the terminating NUL byte.
173 /// Fails if the CString is null.
175 pub fn as_bytes<'a>(&'a self) -> &'a [u8] {
176 if self.buf.is_null() { fail!("CString is null!"); }
178 mem::transmute(Slice { data: self.buf, len: self.len() + 1 })
182 /// Converts the CString into a `&[u8]` without copying.
183 /// Does not include the terminating NUL byte.
187 /// Fails if the CString is null.
189 pub fn as_bytes_no_nul<'a>(&'a self) -> &'a [u8] {
190 if self.buf.is_null() { fail!("CString is null!"); }
192 mem::transmute(Slice { data: self.buf, len: self.len() })
196 /// Converts the CString into a `&str` without copying.
197 /// Returns None if the CString is not UTF-8.
201 /// Fails if the CString is null.
203 pub fn as_str<'a>(&'a self) -> Option<&'a str> {
204 let buf = self.as_bytes_no_nul();
208 /// Return a CString iterator.
212 /// Fails if the CString is null.
213 pub fn iter<'a>(&'a self) -> CChars<'a> {
214 if self.buf.is_null() { fail!("CString is null!"); }
217 marker: marker::ContravariantLifetime,
222 impl Drop for CString {
224 if self.owns_buffer_ {
226 libc::free(self.buf as *mut libc::c_void)
232 impl Collection for CString {
233 /// Return the number of bytes in the CString (not including the NUL terminator).
237 /// Fails if the CString is null.
239 fn len(&self) -> uint {
240 if self.buf.is_null() { fail!("CString is null!"); }
241 let mut cur = self.buf;
253 /// A generic trait for converting a value to a CString.
255 /// Copy the receiver into a CString.
259 /// Fails the task if the receiver has an interior null.
260 fn to_c_str(&self) -> CString;
262 /// Unsafe variant of `to_c_str()` that doesn't check for nulls.
263 unsafe fn to_c_str_unchecked(&self) -> CString;
265 /// Work with a temporary CString constructed from the receiver.
266 /// The provided `*libc::c_char` will be freed immediately upon return.
271 /// extern crate libc;
274 /// let s = "PATH".with_c_str(|path| unsafe {
275 /// libc::getenv(path)
282 /// Fails the task if the receiver has an interior null.
284 fn with_c_str<T>(&self, f: |*libc::c_char| -> T) -> T {
285 self.to_c_str().with_ref(f)
288 /// Unsafe variant of `with_c_str()` that doesn't check for nulls.
290 unsafe fn with_c_str_unchecked<T>(&self, f: |*libc::c_char| -> T) -> T {
291 self.to_c_str_unchecked().with_ref(f)
295 // FIXME (#12938): Until DST lands, we cannot decompose &str into &
296 // and str, so we cannot usefully take ToCStr arguments by reference
297 // (without forcing an additional & around &str). So we are instead
298 // temporarily adding an instance for ~str and String, so that we can
299 // take ToCStr as owned. When DST lands, the string instances should
300 // be revisted, and arguments bound by ToCStr should be passed by
303 impl<'a> ToCStr for &'a str {
305 fn to_c_str(&self) -> CString {
306 self.as_bytes().to_c_str()
310 unsafe fn to_c_str_unchecked(&self) -> CString {
311 self.as_bytes().to_c_str_unchecked()
315 fn with_c_str<T>(&self, f: |*libc::c_char| -> T) -> T {
316 self.as_bytes().with_c_str(f)
320 unsafe fn with_c_str_unchecked<T>(&self, f: |*libc::c_char| -> T) -> T {
321 self.as_bytes().with_c_str_unchecked(f)
325 impl ToCStr for String {
327 fn to_c_str(&self) -> CString {
328 self.as_bytes().to_c_str()
332 unsafe fn to_c_str_unchecked(&self) -> CString {
333 self.as_bytes().to_c_str_unchecked()
337 fn with_c_str<T>(&self, f: |*libc::c_char| -> T) -> T {
338 self.as_bytes().with_c_str(f)
342 unsafe fn with_c_str_unchecked<T>(&self, f: |*libc::c_char| -> T) -> T {
343 self.as_bytes().with_c_str_unchecked(f)
347 // The length of the stack allocated buffer for `vec.with_c_str()`
348 static BUF_LEN: uint = 128;
350 impl<'a> ToCStr for &'a [u8] {
351 fn to_c_str(&self) -> CString {
352 let mut cs = unsafe { self.to_c_str_unchecked() };
353 cs.with_mut_ref(|buf| check_for_null(*self, buf));
357 unsafe fn to_c_str_unchecked(&self) -> CString {
358 let self_len = self.len();
359 let buf = malloc_raw(self_len + 1);
361 ptr::copy_memory(buf, self.as_ptr(), self_len);
362 *buf.offset(self_len as int) = 0;
364 CString::new(buf as *libc::c_char, true)
367 fn with_c_str<T>(&self, f: |*libc::c_char| -> T) -> T {
368 unsafe { with_c_str(*self, true, f) }
371 unsafe fn with_c_str_unchecked<T>(&self, f: |*libc::c_char| -> T) -> T {
372 with_c_str(*self, false, f)
376 // Unsafe function that handles possibly copying the &[u8] into a stack array.
377 unsafe fn with_c_str<T>(v: &[u8], checked: bool, f: |*libc::c_char| -> T) -> T {
378 if v.len() < BUF_LEN {
379 let mut buf: [u8, .. BUF_LEN] = mem::uninitialized();
380 slice::bytes::copy_memory(buf, v);
383 let buf = buf.as_mut_ptr();
385 check_for_null(v, buf as *mut libc::c_char);
388 f(buf as *libc::c_char)
390 v.to_c_str().with_ref(f)
392 v.to_c_str_unchecked().with_ref(f)
397 fn check_for_null(v: &[u8], buf: *mut libc::c_char) {
398 for i in range(0, v.len()) {
400 let p = buf.offset(i as int);
406 /// External iterator for a CString's bytes.
408 /// Use with the `std::iter` module.
409 pub struct CChars<'a> {
411 marker: marker::ContravariantLifetime<'a>,
414 impl<'a> Iterator<libc::c_char> for CChars<'a> {
415 fn next(&mut self) -> Option<libc::c_char> {
416 let ch = unsafe { *self.ptr };
420 self.ptr = unsafe { self.ptr.offset(1) };
426 /// Parses a C "multistring", eg windows env values or
427 /// the req->ptr result in a uv_fs_readdir() call.
429 /// Optionally, a `count` can be passed in, limiting the
430 /// parsing to only being done `count`-times.
432 /// The specified closure is invoked with each string that
433 /// is found, and the number of strings found is returned.
434 pub unsafe fn from_c_multistring(buf: *libc::c_char,
436 f: |&CString|) -> uint {
438 let mut curr_ptr: uint = buf as uint;
440 let (limited_count, limit) = match count {
441 Some(limit) => (true, limit),
444 while ((limited_count && ctr < limit) || !limited_count)
445 && *(curr_ptr as *libc::c_char) != 0 as libc::c_char {
446 let cstr = CString::new(curr_ptr as *libc::c_char, false);
448 curr_ptr += cstr.len() + 1;
464 fn test_str_multistring_parsing() {
466 let input = bytes!("zero", "\x00", "one", "\x00", "\x00");
467 let ptr = input.as_ptr();
468 let expected = ["zero", "one"];
469 let mut it = expected.iter();
470 let result = from_c_multistring(ptr as *libc::c_char, None, |c| {
471 let cbytes = c.as_bytes_no_nul();
472 assert_eq!(cbytes, it.next().unwrap().as_bytes());
474 assert_eq!(result, 2);
475 assert!(it.next().is_none());
480 fn test_str_to_c_str() {
481 "".to_c_str().with_ref(|buf| {
483 assert_eq!(*buf.offset(0), 0);
487 "hello".to_c_str().with_ref(|buf| {
489 assert_eq!(*buf.offset(0), 'h' as libc::c_char);
490 assert_eq!(*buf.offset(1), 'e' as libc::c_char);
491 assert_eq!(*buf.offset(2), 'l' as libc::c_char);
492 assert_eq!(*buf.offset(3), 'l' as libc::c_char);
493 assert_eq!(*buf.offset(4), 'o' as libc::c_char);
494 assert_eq!(*buf.offset(5), 0);
500 fn test_vec_to_c_str() {
502 b.to_c_str().with_ref(|buf| {
504 assert_eq!(*buf.offset(0), 0);
508 let _ = bytes!("hello").to_c_str().with_ref(|buf| {
510 assert_eq!(*buf.offset(0), 'h' as libc::c_char);
511 assert_eq!(*buf.offset(1), 'e' as libc::c_char);
512 assert_eq!(*buf.offset(2), 'l' as libc::c_char);
513 assert_eq!(*buf.offset(3), 'l' as libc::c_char);
514 assert_eq!(*buf.offset(4), 'o' as libc::c_char);
515 assert_eq!(*buf.offset(5), 0);
519 let _ = bytes!("foo", 0xff).to_c_str().with_ref(|buf| {
521 assert_eq!(*buf.offset(0), 'f' as libc::c_char);
522 assert_eq!(*buf.offset(1), 'o' as libc::c_char);
523 assert_eq!(*buf.offset(2), 'o' as libc::c_char);
524 assert_eq!(*buf.offset(3), 0xff as i8);
525 assert_eq!(*buf.offset(4), 0);
532 let c_str = unsafe { CString::new(ptr::null(), false) };
533 assert!(c_str.is_null());
534 assert!(!c_str.is_not_null());
539 let c_str = "hello".to_c_str();
540 unsafe { libc::free(c_str.unwrap() as *mut libc::c_void) }
545 let c_str = "hello".to_c_str();
546 let len = unsafe { c_str.with_ref(|buf| libc::strlen(buf)) };
547 assert!(!c_str.is_null());
548 assert!(c_str.is_not_null());
554 fn test_with_ref_empty_fail() {
555 let c_str = unsafe { CString::new(ptr::null(), false) };
556 c_str.with_ref(|_| ());
561 let c_str = "".to_c_str();
562 let mut iter = c_str.iter();
563 assert_eq!(iter.next(), None);
565 let c_str = "hello".to_c_str();
566 let mut iter = c_str.iter();
567 assert_eq!(iter.next(), Some('h' as libc::c_char));
568 assert_eq!(iter.next(), Some('e' as libc::c_char));
569 assert_eq!(iter.next(), Some('l' as libc::c_char));
570 assert_eq!(iter.next(), Some('l' as libc::c_char));
571 assert_eq!(iter.next(), Some('o' as libc::c_char));
572 assert_eq!(iter.next(), None);
576 fn test_to_c_str_fail() {
577 assert!(task::try(proc() { "he\x00llo".to_c_str() }).is_err());
581 fn test_to_c_str_unchecked() {
583 "he\x00llo".to_c_str_unchecked().with_ref(|buf| {
584 assert_eq!(*buf.offset(0), 'h' as libc::c_char);
585 assert_eq!(*buf.offset(1), 'e' as libc::c_char);
586 assert_eq!(*buf.offset(2), 0);
587 assert_eq!(*buf.offset(3), 'l' as libc::c_char);
588 assert_eq!(*buf.offset(4), 'l' as libc::c_char);
589 assert_eq!(*buf.offset(5), 'o' as libc::c_char);
590 assert_eq!(*buf.offset(6), 0);
597 let c_str = "hello".to_c_str();
598 assert_eq!(c_str.as_bytes(), bytes!("hello", 0));
599 let c_str = "".to_c_str();
600 assert_eq!(c_str.as_bytes(), bytes!(0));
601 let c_str = bytes!("foo", 0xff).to_c_str();
602 assert_eq!(c_str.as_bytes(), bytes!("foo", 0xff, 0));
606 fn test_as_bytes_no_nul() {
607 let c_str = "hello".to_c_str();
608 assert_eq!(c_str.as_bytes_no_nul(), bytes!("hello"));
609 let c_str = "".to_c_str();
611 assert_eq!(c_str.as_bytes_no_nul(), exp);
612 let c_str = bytes!("foo", 0xff).to_c_str();
613 assert_eq!(c_str.as_bytes_no_nul(), bytes!("foo", 0xff));
618 fn test_as_bytes_fail() {
619 let c_str = unsafe { CString::new(ptr::null(), false) };
625 fn test_as_bytes_no_nul_fail() {
626 let c_str = unsafe { CString::new(ptr::null(), false) };
627 c_str.as_bytes_no_nul();
632 let c_str = "hello".to_c_str();
633 assert_eq!(c_str.as_str(), Some("hello"));
634 let c_str = "".to_c_str();
635 assert_eq!(c_str.as_str(), Some(""));
636 let c_str = bytes!("foo", 0xff).to_c_str();
637 assert_eq!(c_str.as_str(), None);
642 fn test_as_str_fail() {
643 let c_str = unsafe { CString::new(ptr::null(), false) };
650 let c_str = unsafe { CString::new(ptr::null(), false) };
656 fn test_iter_fail() {
657 let c_str = unsafe { CString::new(ptr::null(), false) };
663 let a = "hello".to_c_str();
669 fn test_clone_noleak() {
670 fn foo(f: |c: &CString|) {
671 let s = "test".to_string();
672 let c = s.to_c_str();
673 // give the closure a non-owned CString
674 let mut c_ = c.with_ref(|c| unsafe { CString::new(c, false) } );
676 // muck with the buffer for later printing
677 c_.with_mut_ref(|c| unsafe { *c = 'X' as libc::c_char } );
680 let mut c_: Option<CString> = None;
682 c_ = Some(c.clone());
684 // force a copy, reading the memory
685 c.as_bytes().to_owned();
687 let c_ = c_.unwrap();
688 // force a copy, reading the memory
689 c_.as_bytes().to_owned();
693 fn test_clone_eq_null() {
694 let x = unsafe { CString::new(ptr::null(), false) };
707 fn check(s: &str, c_str: *libc::c_char) {
708 let s_buf = s.as_ptr();
709 for i in range(0, s.len()) {
712 *s_buf.offset(i as int) as libc::c_char,
713 *c_str.offset(i as int));
718 static s_short: &'static str = "Mary";
719 static s_medium: &'static str = "Mary had a little lamb";
720 static s_long: &'static str = "\
721 Mary had a little lamb, Little lamb
722 Mary had a little lamb, Little lamb
723 Mary had a little lamb, Little lamb
724 Mary had a little lamb, Little lamb
725 Mary had a little lamb, Little lamb
726 Mary had a little lamb, Little lamb";
728 fn bench_to_str(b: &mut Bencher, s: &str) {
730 let c_str = s.to_c_str();
731 c_str.with_ref(|c_str_buf| check(s, c_str_buf))
736 fn bench_to_c_str_short(b: &mut Bencher) {
737 bench_to_str(b, s_short)
741 fn bench_to_c_str_medium(b: &mut Bencher) {
742 bench_to_str(b, s_medium)
746 fn bench_to_c_str_long(b: &mut Bencher) {
747 bench_to_str(b, s_long)
750 fn bench_to_c_str_unchecked(b: &mut Bencher, s: &str) {
752 let c_str = unsafe { s.to_c_str_unchecked() };
753 c_str.with_ref(|c_str_buf| check(s, c_str_buf))
758 fn bench_to_c_str_unchecked_short(b: &mut Bencher) {
759 bench_to_c_str_unchecked(b, s_short)
763 fn bench_to_c_str_unchecked_medium(b: &mut Bencher) {
764 bench_to_c_str_unchecked(b, s_medium)
768 fn bench_to_c_str_unchecked_long(b: &mut Bencher) {
769 bench_to_c_str_unchecked(b, s_long)
772 fn bench_with_c_str(b: &mut Bencher, s: &str) {
774 s.with_c_str(|c_str_buf| check(s, c_str_buf))
779 fn bench_with_c_str_short(b: &mut Bencher) {
780 bench_with_c_str(b, s_short)
784 fn bench_with_c_str_medium(b: &mut Bencher) {
785 bench_with_c_str(b, s_medium)
789 fn bench_with_c_str_long(b: &mut Bencher) {
790 bench_with_c_str(b, s_long)
793 fn bench_with_c_str_unchecked(b: &mut Bencher, s: &str) {
796 s.with_c_str_unchecked(|c_str_buf| check(s, c_str_buf))
802 fn bench_with_c_str_unchecked_short(b: &mut Bencher) {
803 bench_with_c_str_unchecked(b, s_short)
807 fn bench_with_c_str_unchecked_medium(b: &mut Bencher) {
808 bench_with_c_str_unchecked(b, s_medium)
812 fn bench_with_c_str_unchecked_long(b: &mut Bencher) {
813 bench_with_c_str_unchecked(b, s_long)