1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164

#[cfg(test)]
#[macro_use]
extern crate assert_matches;

use std::ffi::CStr;
use std::str::Utf8Error;
use std::convert::AsRef;
use std::mem::transmute;
use std::os::raw::c_char;
use std::ops::Deref;
use std::fmt;


#[derive(Debug)]
pub enum Utf8CStrError {
    NoNulTerm,
    EmbeddedNulTerm(usize),
    Utf8Error(Utf8Error)
}

/**
 * A wrapper that promises it's contents are null-terminated & validly utf-8 encoded
 *
 * Plain `std::ffi::CStr` only promises null termination, but some ffi (and other bits) require
 * strings that are both valid utf8 and null terminated.
 */
#[derive(PartialEq, Eq)]
pub struct Utf8CStr {
    inner: CStr,
}

impl fmt::Debug for Utf8CStr {
    fn fmt(&self, fmt: &mut fmt::Formatter) -> fmt::Result {
        let s : &str = self.as_ref();
        write!(fmt, "{:?}", s)
    }
}

impl fmt::Display for Utf8CStr {
    fn fmt(&self, fmt: &mut fmt::Formatter) -> fmt::Result {
        let s : &str = self.as_ref();
        write!(fmt, "{}", s)
    }
}

impl Utf8CStr {
    /// Construct a `Utf8CStr` with no checking.
    ///
    /// Unsafety:
    ///
    ///  - `'a` must be the correct lifetime
    ///  - `v` must be nul terminated
    ///  - `v` must be utf-8 encoded (for use in `&str`)
    ///  - `l` must be the length of `v` including the nul terminator
    pub unsafe fn from_raw_parts<'a>(v: *const c_char, l: usize) -> &'a Self {
        transmute((v, l))
    }

    /// Construct a `Utf8CStr` with minimal checking.
    ///
    /// This currently will scan for the terminating '\0' just to establish the length for various
    /// slices.
    ///
    /// Unsafety:
    ///
    ///  - `'a` must be the correct lifetime
    ///  - `v` must be nul terminated
    ///  - `v` must be utf-8 encoded (for use in `&str`)
    pub unsafe fn from_ptr_unchecked<'a>(v: *const c_char) -> &'a Self {
        Self::from_cstr_unchecked(CStr::from_ptr(v))
    }

    /// Failable convertion from a CStr.
    ///
    /// Verifies that the CStr is utf8 encoded.
    ///
    /// NOTE: Only handles non-mutable variants. We may want to accept &mut as well in the future.
    pub fn from_cstr(v: &CStr) -> Result<&Self, Utf8Error> {
        try!(v.to_str());
        Ok(unsafe { transmute(v)})
    }

    /// Convert from a `&CStr` without any checking
    ///
    /// Unsafety:
    ///
    ///  - `v` must be valid utf8 for use in a `&str`
    pub unsafe fn from_cstr_unchecked(v: &CStr) -> &Utf8CStr {
        Self::from_bytes_with_nul_unchecked(v.to_bytes_with_nul())
    }

    /// Convert directly from bytes
    ///
    /// NOTE: right now this scans `b` a few times over. Ideally, we'd adjust it to only scan `b`
    /// once.
    pub fn from_bytes(b: &[u8]) -> Result<&Self, Utf8CStrError> {
        // FIXME: use from_bytes_with_nul when stablized
        for (l, &v) in b[0..b.len() - 1].iter().enumerate() {
            if v == b'\0' {
                return Err(Utf8CStrError::EmbeddedNulTerm(l));
            }
        }
        if b[b.len() - 1] != b'\0' {
            return Err(Utf8CStrError::NoNulTerm);
        }

        let c : &CStr = unsafe { transmute(b) };
        Self::from_cstr(c).map_err(|e| Utf8CStrError::Utf8Error(e))
    }

    /// Raw convertion from basic data type with no checking.
    pub unsafe fn from_bytes_with_nul_unchecked(b: &[u8]) -> &Self {
        transmute(b)
    }

    pub fn as_ptr(&self) -> *const c_char {
        let v : &CStr = self.as_ref();
        v.as_ptr()
    }

}

impl AsRef<str> for Utf8CStr {
    fn as_ref(&self) -> &str {
        unsafe { transmute(self.inner.to_bytes()) }
    }
}

impl AsRef<CStr> for Utf8CStr {
    fn as_ref(&self) -> &CStr {
        unsafe { transmute(self) }
    }
}

impl Deref for Utf8CStr {
    type Target = str;
    fn deref(&self) -> &Self::Target {
        self.as_ref()
    }
}

#[cfg(test)]
mod tests {
    use super::Utf8CStr;
    use super::Utf8CStrError;
    #[test]
    fn it_works() {
        let x = Utf8CStr::from_bytes(b"hello\0").unwrap();
        assert_matches!(Utf8CStr::from_bytes(b"hell").err().unwrap(), Utf8CStrError::NoNulTerm);
        assert_matches!(Utf8CStr::from_bytes(b"hell\0d").err().unwrap(), Utf8CStrError::EmbeddedNulTerm(4));
        assert_matches!(Utf8CStr::from_bytes(&[8,1,23,4,0xff, 0]).err().unwrap(), Utf8CStrError::Utf8Error(_));


        println!("{:?}", x);
        println!("{}", x);

        assert_eq!(x, x);
        assert!(x != Utf8CStr::from_bytes(b"hell\0").unwrap());

        let v = b"hello\0";
        let b = unsafe { Utf8CStr::from_raw_parts(v.as_ptr() as *const _, v.len()) };
        assert_eq!(b, x);
    }
}