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: *const 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.
85 buf: *const libc::c_char,
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 *const 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: *const 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.
127 /// The original object is destructed after this method is called, and if
128 /// the underlying pointer was previously allocated, care must be taken to
129 /// ensure that it is deallocated properly.
130 pub unsafe fn unwrap(self) -> *const libc::c_char {
131 let mut c_str = self;
132 c_str.owns_buffer_ = false;
136 /// Return a pointer to the NUL-terminated string data.
138 /// `.as_ptr` returns an internal pointer into the `CString`, and
139 /// may be invalidated when the `CString` falls out of scope (the
140 /// destructor will run, freeing the allocation if there is
144 /// let foo = "some string";
147 /// let x = foo.to_c_str();
148 /// let p = x.as_ptr();
150 /// // wrong (the CString will be freed, invalidating `p`)
151 /// let p = foo.to_c_str().as_ptr();
156 /// Fails if the CString is null.
161 /// extern crate libc;
164 /// let c_str = "foo bar".to_c_str();
166 /// libc::puts(c_str.as_ptr());
170 pub fn as_ptr(&self) -> *const libc::c_char {
171 if self.buf.is_null() { fail!("CString is null!"); }
176 /// Return a mutable pointer to the NUL-terminated string data.
178 /// `.as_mut_ptr` returns an internal pointer into the `CString`, and
179 /// may be invalidated when the `CString` falls out of scope (the
180 /// destructor will run, freeing the allocation if there is
184 /// let foo = "some string";
187 /// let mut x = foo.to_c_str();
188 /// let p = x.as_mut_ptr();
190 /// // wrong (the CString will be freed, invalidating `p`)
191 /// let p = foo.to_c_str().as_mut_ptr();
196 /// Fails if the CString is null.
197 pub fn as_mut_ptr(&mut self) -> *mut libc::c_char {
198 if self.buf.is_null() { fail!("CString is null!") }
203 /// Calls a closure with a reference to the underlying `*libc::c_char`.
207 /// Fails if the CString is null.
208 #[deprecated="use `.as_ptr()`"]
209 pub fn with_ref<T>(&self, f: |*const libc::c_char| -> T) -> T {
210 if self.buf.is_null() { fail!("CString is null!"); }
214 /// Calls a closure with a mutable reference to the underlying `*libc::c_char`.
218 /// Fails if the CString is null.
219 #[deprecated="use `.as_mut_ptr()`"]
220 pub fn with_mut_ref<T>(&mut self, f: |*mut libc::c_char| -> T) -> T {
221 if self.buf.is_null() { fail!("CString is null!"); }
222 f(self.buf as *mut libc::c_char)
225 /// Returns true if the CString is a null.
226 pub fn is_null(&self) -> bool {
230 /// Returns true if the CString is not null.
231 pub fn is_not_null(&self) -> bool {
232 self.buf.is_not_null()
235 /// Returns whether or not the `CString` owns the buffer.
236 pub fn owns_buffer(&self) -> bool {
240 /// Converts the CString into a `&[u8]` without copying.
241 /// Includes the terminating NUL byte.
245 /// Fails if the CString is null.
247 pub fn as_bytes<'a>(&'a self) -> &'a [u8] {
248 if self.buf.is_null() { fail!("CString is null!"); }
250 mem::transmute(Slice { data: self.buf, len: self.len() + 1 })
254 /// Converts the CString into a `&[u8]` without copying.
255 /// Does not include the terminating NUL byte.
259 /// Fails if the CString is null.
261 pub fn as_bytes_no_nul<'a>(&'a self) -> &'a [u8] {
262 if self.buf.is_null() { fail!("CString is null!"); }
264 mem::transmute(Slice { data: self.buf, len: self.len() })
268 /// Converts the CString into a `&str` without copying.
269 /// Returns None if the CString is not UTF-8.
273 /// Fails if the CString is null.
275 pub fn as_str<'a>(&'a self) -> Option<&'a str> {
276 let buf = self.as_bytes_no_nul();
280 /// Return a CString iterator.
284 /// Fails if the CString is null.
285 pub fn iter<'a>(&'a self) -> CChars<'a> {
286 if self.buf.is_null() { fail!("CString is null!"); }
289 marker: marker::ContravariantLifetime,
294 impl Drop for CString {
296 if self.owns_buffer_ {
298 libc::free(self.buf as *mut libc::c_void)
304 impl Collection for CString {
305 /// Return the number of bytes in the CString (not including the NUL terminator).
309 /// Fails if the CString is null.
311 fn len(&self) -> uint {
312 if self.buf.is_null() { fail!("CString is null!"); }
313 let mut cur = self.buf;
325 /// A generic trait for converting a value to a CString.
327 /// Copy the receiver into a CString.
331 /// Fails the task if the receiver has an interior null.
332 fn to_c_str(&self) -> CString;
334 /// Unsafe variant of `to_c_str()` that doesn't check for nulls.
335 unsafe fn to_c_str_unchecked(&self) -> CString;
337 /// Work with a temporary CString constructed from the receiver.
338 /// The provided `*libc::c_char` will be freed immediately upon return.
343 /// extern crate libc;
346 /// let s = "PATH".with_c_str(|path| unsafe {
347 /// libc::getenv(path)
354 /// Fails the task if the receiver has an interior null.
356 fn with_c_str<T>(&self, f: |*const libc::c_char| -> T) -> T {
357 self.to_c_str().with_ref(f)
360 /// Unsafe variant of `with_c_str()` that doesn't check for nulls.
362 unsafe fn with_c_str_unchecked<T>(&self, f: |*const libc::c_char| -> T) -> T {
363 self.to_c_str_unchecked().with_ref(f)
367 // FIXME (#12938): Until DST lands, we cannot decompose &str into &
368 // and str, so we cannot usefully take ToCStr arguments by reference
369 // (without forcing an additional & around &str). So we are instead
370 // temporarily adding an instance for ~str and String, so that we can
371 // take ToCStr as owned. When DST lands, the string instances should
372 // be revisted, and arguments bound by ToCStr should be passed by
375 impl<'a> ToCStr for &'a str {
377 fn to_c_str(&self) -> CString {
378 self.as_bytes().to_c_str()
382 unsafe fn to_c_str_unchecked(&self) -> CString {
383 self.as_bytes().to_c_str_unchecked()
387 fn with_c_str<T>(&self, f: |*const libc::c_char| -> T) -> T {
388 self.as_bytes().with_c_str(f)
392 unsafe fn with_c_str_unchecked<T>(&self, f: |*const libc::c_char| -> T) -> T {
393 self.as_bytes().with_c_str_unchecked(f)
397 impl ToCStr for String {
399 fn to_c_str(&self) -> CString {
400 self.as_bytes().to_c_str()
404 unsafe fn to_c_str_unchecked(&self) -> CString {
405 self.as_bytes().to_c_str_unchecked()
409 fn with_c_str<T>(&self, f: |*const libc::c_char| -> T) -> T {
410 self.as_bytes().with_c_str(f)
414 unsafe fn with_c_str_unchecked<T>(&self, f: |*const libc::c_char| -> T) -> T {
415 self.as_bytes().with_c_str_unchecked(f)
419 // The length of the stack allocated buffer for `vec.with_c_str()`
420 static BUF_LEN: uint = 128;
422 impl<'a> ToCStr for &'a [u8] {
423 fn to_c_str(&self) -> CString {
424 let mut cs = unsafe { self.to_c_str_unchecked() };
425 cs.with_mut_ref(|buf| check_for_null(*self, buf));
429 unsafe fn to_c_str_unchecked(&self) -> CString {
430 let self_len = self.len();
431 let buf = malloc_raw(self_len + 1);
433 ptr::copy_memory(buf, self.as_ptr(), self_len);
434 *buf.offset(self_len as int) = 0;
436 CString::new(buf as *const libc::c_char, true)
439 fn with_c_str<T>(&self, f: |*const libc::c_char| -> T) -> T {
440 unsafe { with_c_str(*self, true, f) }
443 unsafe fn with_c_str_unchecked<T>(&self, f: |*const libc::c_char| -> T) -> T {
444 with_c_str(*self, false, f)
448 // Unsafe function that handles possibly copying the &[u8] into a stack array.
449 unsafe fn with_c_str<T>(v: &[u8], checked: bool,
450 f: |*const libc::c_char| -> T) -> T {
451 if v.len() < BUF_LEN {
452 let mut buf: [u8, .. BUF_LEN] = mem::uninitialized();
453 slice::bytes::copy_memory(buf, v);
456 let buf = buf.as_mut_ptr();
458 check_for_null(v, buf as *mut libc::c_char);
461 f(buf as *const libc::c_char)
463 v.to_c_str().with_ref(f)
465 v.to_c_str_unchecked().with_ref(f)
470 fn check_for_null(v: &[u8], buf: *mut libc::c_char) {
471 for i in range(0, v.len()) {
473 let p = buf.offset(i as int);
479 /// External iterator for a CString's bytes.
481 /// Use with the `std::iter` module.
482 pub struct CChars<'a> {
483 ptr: *const libc::c_char,
484 marker: marker::ContravariantLifetime<'a>,
487 impl<'a> Iterator<libc::c_char> for CChars<'a> {
488 fn next(&mut self) -> Option<libc::c_char> {
489 let ch = unsafe { *self.ptr };
493 self.ptr = unsafe { self.ptr.offset(1) };
499 /// Parses a C "multistring", eg windows env values or
500 /// the req->ptr result in a uv_fs_readdir() call.
502 /// Optionally, a `count` can be passed in, limiting the
503 /// parsing to only being done `count`-times.
505 /// The specified closure is invoked with each string that
506 /// is found, and the number of strings found is returned.
507 pub unsafe fn from_c_multistring(buf: *const libc::c_char,
509 f: |&CString|) -> uint {
511 let mut curr_ptr: uint = buf as uint;
513 let (limited_count, limit) = match count {
514 Some(limit) => (true, limit),
517 while ((limited_count && ctr < limit) || !limited_count)
518 && *(curr_ptr as *const libc::c_char) != 0 as libc::c_char {
519 let cstr = CString::new(curr_ptr as *const libc::c_char, false);
521 curr_ptr += cstr.len() + 1;
537 fn test_str_multistring_parsing() {
539 let input = b"zero\0one\0\0";
540 let ptr = input.as_ptr();
541 let expected = ["zero", "one"];
542 let mut it = expected.iter();
543 let result = from_c_multistring(ptr as *const libc::c_char, None, |c| {
544 let cbytes = c.as_bytes_no_nul();
545 assert_eq!(cbytes, it.next().unwrap().as_bytes());
547 assert_eq!(result, 2);
548 assert!(it.next().is_none());
553 fn test_str_to_c_str() {
554 "".to_c_str().with_ref(|buf| {
556 assert_eq!(*buf.offset(0), 0);
560 "hello".to_c_str().with_ref(|buf| {
562 assert_eq!(*buf.offset(0), 'h' as libc::c_char);
563 assert_eq!(*buf.offset(1), 'e' as libc::c_char);
564 assert_eq!(*buf.offset(2), 'l' as libc::c_char);
565 assert_eq!(*buf.offset(3), 'l' as libc::c_char);
566 assert_eq!(*buf.offset(4), 'o' as libc::c_char);
567 assert_eq!(*buf.offset(5), 0);
573 fn test_vec_to_c_str() {
575 b.to_c_str().with_ref(|buf| {
577 assert_eq!(*buf.offset(0), 0);
581 let _ = b"hello".to_c_str().with_ref(|buf| {
583 assert_eq!(*buf.offset(0), 'h' as libc::c_char);
584 assert_eq!(*buf.offset(1), 'e' as libc::c_char);
585 assert_eq!(*buf.offset(2), 'l' as libc::c_char);
586 assert_eq!(*buf.offset(3), 'l' as libc::c_char);
587 assert_eq!(*buf.offset(4), 'o' as libc::c_char);
588 assert_eq!(*buf.offset(5), 0);
592 let _ = b"foo\xFF".to_c_str().with_ref(|buf| {
594 assert_eq!(*buf.offset(0), 'f' as libc::c_char);
595 assert_eq!(*buf.offset(1), 'o' as libc::c_char);
596 assert_eq!(*buf.offset(2), 'o' as libc::c_char);
597 assert_eq!(*buf.offset(3), 0xff as i8);
598 assert_eq!(*buf.offset(4), 0);
605 let c_str = unsafe { CString::new(ptr::null(), false) };
606 assert!(c_str.is_null());
607 assert!(!c_str.is_not_null());
612 let c_str = "hello".to_c_str();
613 unsafe { libc::free(c_str.unwrap() as *mut libc::c_void) }
618 let c_str = "hello".to_c_str();
619 let len = unsafe { c_str.with_ref(|buf| libc::strlen(buf)) };
620 assert!(!c_str.is_null());
621 assert!(c_str.is_not_null());
627 fn test_with_ref_empty_fail() {
628 let c_str = unsafe { CString::new(ptr::null(), false) };
629 c_str.with_ref(|_| ());
634 let c_str = "".to_c_str();
635 let mut iter = c_str.iter();
636 assert_eq!(iter.next(), None);
638 let c_str = "hello".to_c_str();
639 let mut iter = c_str.iter();
640 assert_eq!(iter.next(), Some('h' as libc::c_char));
641 assert_eq!(iter.next(), Some('e' as libc::c_char));
642 assert_eq!(iter.next(), Some('l' as libc::c_char));
643 assert_eq!(iter.next(), Some('l' as libc::c_char));
644 assert_eq!(iter.next(), Some('o' as libc::c_char));
645 assert_eq!(iter.next(), None);
649 fn test_to_c_str_fail() {
650 assert!(task::try(proc() { "he\x00llo".to_c_str() }).is_err());
654 fn test_to_c_str_unchecked() {
656 "he\x00llo".to_c_str_unchecked().with_ref(|buf| {
657 assert_eq!(*buf.offset(0), 'h' as libc::c_char);
658 assert_eq!(*buf.offset(1), 'e' as libc::c_char);
659 assert_eq!(*buf.offset(2), 0);
660 assert_eq!(*buf.offset(3), 'l' as libc::c_char);
661 assert_eq!(*buf.offset(4), 'l' as libc::c_char);
662 assert_eq!(*buf.offset(5), 'o' as libc::c_char);
663 assert_eq!(*buf.offset(6), 0);
670 let c_str = "hello".to_c_str();
671 assert_eq!(c_str.as_bytes(), b"hello\0");
672 let c_str = "".to_c_str();
673 assert_eq!(c_str.as_bytes(), b"\0");
674 let c_str = b"foo\xFF".to_c_str();
675 assert_eq!(c_str.as_bytes(), b"foo\xFF\0");
679 fn test_as_bytes_no_nul() {
680 let c_str = "hello".to_c_str();
681 assert_eq!(c_str.as_bytes_no_nul(), b"hello");
682 let c_str = "".to_c_str();
684 assert_eq!(c_str.as_bytes_no_nul(), exp);
685 let c_str = b"foo\xFF".to_c_str();
686 assert_eq!(c_str.as_bytes_no_nul(), b"foo\xFF");
691 fn test_as_bytes_fail() {
692 let c_str = unsafe { CString::new(ptr::null(), false) };
698 fn test_as_bytes_no_nul_fail() {
699 let c_str = unsafe { CString::new(ptr::null(), false) };
700 c_str.as_bytes_no_nul();
705 let c_str = "hello".to_c_str();
706 assert_eq!(c_str.as_str(), Some("hello"));
707 let c_str = "".to_c_str();
708 assert_eq!(c_str.as_str(), Some(""));
709 let c_str = b"foo\xFF".to_c_str();
710 assert_eq!(c_str.as_str(), None);
715 fn test_as_str_fail() {
716 let c_str = unsafe { CString::new(ptr::null(), false) };
723 let c_str = unsafe { CString::new(ptr::null(), false) };
729 fn test_iter_fail() {
730 let c_str = unsafe { CString::new(ptr::null(), false) };
736 let a = "hello".to_c_str();
742 fn test_clone_noleak() {
743 fn foo(f: |c: &CString|) {
744 let s = "test".to_string();
745 let c = s.to_c_str();
746 // give the closure a non-owned CString
747 let mut c_ = c.with_ref(|c| unsafe { CString::new(c, false) } );
749 // muck with the buffer for later printing
750 c_.with_mut_ref(|c| unsafe { *c = 'X' as libc::c_char } );
753 let mut c_: Option<CString> = None;
755 c_ = Some(c.clone());
757 // force a copy, reading the memory
758 c.as_bytes().to_owned();
760 let c_ = c_.unwrap();
761 // force a copy, reading the memory
762 c_.as_bytes().to_owned();
766 fn test_clone_eq_null() {
767 let x = unsafe { CString::new(ptr::null(), false) };
780 fn check(s: &str, c_str: *const libc::c_char) {
781 let s_buf = s.as_ptr();
782 for i in range(0, s.len()) {
785 *s_buf.offset(i as int) as libc::c_char,
786 *c_str.offset(i as int));
791 static s_short: &'static str = "Mary";
792 static s_medium: &'static str = "Mary had a little lamb";
793 static s_long: &'static str = "\
794 Mary had a little lamb, Little lamb
795 Mary had a little lamb, Little lamb
796 Mary had a little lamb, Little lamb
797 Mary had a little lamb, Little lamb
798 Mary had a little lamb, Little lamb
799 Mary had a little lamb, Little lamb";
801 fn bench_to_str(b: &mut Bencher, s: &str) {
803 let c_str = s.to_c_str();
804 c_str.with_ref(|c_str_buf| check(s, c_str_buf))
809 fn bench_to_c_str_short(b: &mut Bencher) {
810 bench_to_str(b, s_short)
814 fn bench_to_c_str_medium(b: &mut Bencher) {
815 bench_to_str(b, s_medium)
819 fn bench_to_c_str_long(b: &mut Bencher) {
820 bench_to_str(b, s_long)
823 fn bench_to_c_str_unchecked(b: &mut Bencher, s: &str) {
825 let c_str = unsafe { s.to_c_str_unchecked() };
826 c_str.with_ref(|c_str_buf| check(s, c_str_buf))
831 fn bench_to_c_str_unchecked_short(b: &mut Bencher) {
832 bench_to_c_str_unchecked(b, s_short)
836 fn bench_to_c_str_unchecked_medium(b: &mut Bencher) {
837 bench_to_c_str_unchecked(b, s_medium)
841 fn bench_to_c_str_unchecked_long(b: &mut Bencher) {
842 bench_to_c_str_unchecked(b, s_long)
845 fn bench_with_c_str(b: &mut Bencher, s: &str) {
847 s.with_c_str(|c_str_buf| check(s, c_str_buf))
852 fn bench_with_c_str_short(b: &mut Bencher) {
853 bench_with_c_str(b, s_short)
857 fn bench_with_c_str_medium(b: &mut Bencher) {
858 bench_with_c_str(b, s_medium)
862 fn bench_with_c_str_long(b: &mut Bencher) {
863 bench_with_c_str(b, s_long)
866 fn bench_with_c_str_unchecked(b: &mut Bencher, s: &str) {
869 s.with_c_str_unchecked(|c_str_buf| check(s, c_str_buf))
875 fn bench_with_c_str_unchecked_short(b: &mut Bencher) {
876 bench_with_c_str_unchecked(b, s_short)
880 fn bench_with_c_str_unchecked_medium(b: &mut Bencher) {
881 bench_with_c_str_unchecked(b, s_medium)
885 fn bench_with_c_str_unchecked_long(b: &mut Bencher) {
886 bench_with_c_str_unchecked(b, s_long)