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
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
use std::{convert::TryInto, fmt};

use crate::{ObjectId, SIZE_OF_SHA1_DIGEST};

/// A borrowed reference to a hash identifying objects.
///
/// # Future Proofing
///
/// In case we wish to support multiple hashes with the same length we cannot discriminate
/// using the slice length anymore. To make that work, we will use the high bits of the
/// internal `bytes` slice length (a fat pointer, pointing to data and its length in bytes)
/// to encode additional information. Before accessing or returning the bytes, a new adjusted
/// slice will be constructed, while the high bits will be used to help resolving the
/// hash `[`kind()`][oid::kind()]`.
/// We expect to have quite a few bits available for such 'conflict resolution' as most hashes aren't longer
/// than 64 bytes.
#[derive(PartialEq, Eq, Hash, Ord, PartialOrd)]
#[repr(transparent)]
#[allow(non_camel_case_types)]
#[cfg_attr(feature = "serde1", derive(serde::Serialize))]
pub struct oid {
    bytes: [u8],
}

/// A utility able to format itself with the given amount of characters in hex
#[derive(PartialEq, Eq, Hash, Ord, PartialOrd)]
pub struct HexDisplay<'a> {
    inner: &'a oid,
    hex_len: usize,
}

impl<'a> fmt::Display for HexDisplay<'a> {
    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
        let mut hex = crate::Kind::hex_buf();
        let max_len = self.inner.hex_to_buf(hex.as_mut());
        let hex = std::str::from_utf8(&hex[..self.hex_len.min(max_len)]).expect("ascii only in hex");
        f.write_str(hex)
    }
}

impl fmt::Debug for oid {
    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> std::fmt::Result {
        write!(
            f,
            "{}({})",
            match self.kind() {
                crate::Kind::Sha1 => "Sha1",
            },
            self.to_hex(),
        )
    }
}

#[derive(Debug, thiserror::Error)]
pub enum Error {
    #[error("Cannot instantiate git hash from a digest of length {0}")]
    InvalidByteSliceLength(usize),
}

/// Conversion
impl oid {
    /// Try to create a shared object id from a slice of bytes representing a hash `digest`
    #[inline]
    pub fn try_from_bytes(digest: &[u8]) -> Result<&Self, Error> {
        match digest.len() {
            20 => Ok(
                #[allow(unsafe_code)]
                unsafe {
                    &*(digest as *const [u8] as *const oid)
                },
            ),
            len => Err(Error::InvalidByteSliceLength(len)),
        }
    }

    /// Create an OID from the input `value` slice without performing any safety check.
    /// Use only once sure that `value` is a hash of valid length.
    pub fn from_bytes_unchecked(value: &[u8]) -> &Self {
        Self::from_bytes(value)
    }

    /// Only from code that statically assures correct sizes using array conversions
    pub(crate) fn from_bytes(value: &[u8]) -> &Self {
        #[allow(unsafe_code)]
        unsafe {
            &*(value as *const [u8] as *const oid)
        }
    }
}

/// Access
impl oid {
    /// The kind of hash used for this Digest
    #[inline]
    pub fn kind(&self) -> crate::Kind {
        crate::Kind::from_len_in_bytes(self.bytes.len())
    }

    /// The first byte of the hash, commonly used to partition a set of `Id`s
    #[inline]
    pub fn first_byte(&self) -> u8 {
        self.bytes[0]
    }

    /// Interpret this object id as raw byte slice.
    #[inline]
    pub fn as_bytes(&self) -> &[u8] {
        &self.bytes
    }

    /// Return a type which can display itself in hexadecimal form with the `len` amount of characters.
    #[inline]
    pub fn to_hex_with_len(&self, len: usize) -> HexDisplay<'_> {
        HexDisplay {
            inner: self,
            hex_len: len,
        }
    }

    /// Return a type which displays this oid as hex in full.
    #[inline]
    pub fn to_hex(&self) -> HexDisplay<'_> {
        HexDisplay {
            inner: self,
            hex_len: self.bytes.len() * 2,
        }
    }
}

/// Sha1 specific methods
impl oid {
    /// Write ourselves to the `out` in hexadecimal notation, returning the amount of written bytes.
    ///
    /// **Panics** if the buffer isn't big enough to hold twice as many bytes as the current binary size.
    #[inline]
    #[must_use]
    pub fn hex_to_buf(&self, buf: &mut [u8]) -> usize {
        let num_hex_bytes = self.bytes.len() * 2;
        hex::encode_to_slice(&self.bytes, &mut buf[..num_hex_bytes]).expect("to count correctly");
        num_hex_bytes
    }

    /// Write ourselves to `out` in hexadecimal notation
    #[inline]
    pub fn write_hex_to(&self, mut out: impl std::io::Write) -> std::io::Result<()> {
        let mut hex = crate::Kind::hex_buf();
        let hex_len = self.hex_to_buf(&mut hex);
        out.write_all(&hex[..hex_len])
    }

    /// Returns a Sha1 digest with all bytes being initialized to zero.
    #[inline]
    pub(crate) fn null_sha1() -> &'static Self {
        oid::from_bytes([0u8; SIZE_OF_SHA1_DIGEST].as_ref())
    }
}

impl AsRef<oid> for &oid {
    fn as_ref(&self) -> &oid {
        self
    }
}

impl ToOwned for oid {
    type Owned = crate::ObjectId;

    fn to_owned(&self) -> Self::Owned {
        match self.kind() {
            crate::Kind::Sha1 => crate::ObjectId::Sha1(self.bytes.try_into().expect("no bug in hash detection")),
        }
    }
}

impl<'a> From<&'a [u8; SIZE_OF_SHA1_DIGEST]> for &'a oid {
    fn from(v: &'a [u8; SIZE_OF_SHA1_DIGEST]) -> Self {
        oid::from_bytes(v.as_ref())
    }
}

impl fmt::Display for &oid {
    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
        for b in self.as_bytes() {
            write!(f, "{b:02x}")?;
        }
        Ok(())
    }
}

impl PartialEq<crate::ObjectId> for &oid {
    fn eq(&self, other: &ObjectId) -> bool {
        *self == other.as_ref()
    }
}

/// Manually created from a version that uses a slice, and we forcefully try to convert it into a borrowed array of the desired size
/// Could be improved by fitting this into serde
/// Unfortunately the serde::Deserialize derive wouldn't work for borrowed arrays.
#[cfg(feature = "serde1")]
impl<'de: 'a, 'a> serde::Deserialize<'de> for &'a oid {
    fn deserialize<D>(deserializer: D) -> Result<Self, <D as serde::Deserializer<'de>>::Error>
    where
        D: serde::Deserializer<'de>,
    {
        struct __Visitor<'de: 'a, 'a> {
            marker: std::marker::PhantomData<&'a oid>,
            lifetime: std::marker::PhantomData<&'de ()>,
        }
        impl<'de: 'a, 'a> serde::de::Visitor<'de> for __Visitor<'de, 'a> {
            type Value = &'a oid;
            fn expecting(&self, __formatter: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
                std::fmt::Formatter::write_str(__formatter, "tuple struct Digest")
            }
            #[inline]
            fn visit_newtype_struct<__E>(self, __e: __E) -> std::result::Result<Self::Value, __E::Error>
            where
                __E: serde::Deserializer<'de>,
            {
                let __field0: &'a [u8] = match <&'a [u8] as serde::Deserialize>::deserialize(__e) {
                    Ok(__val) => __val,
                    Err(__err) => {
                        return Err(__err);
                    }
                };
                Ok(oid::try_from_bytes(__field0).expect("hash of known length"))
            }
            #[inline]
            fn visit_seq<__A>(self, mut __seq: __A) -> std::result::Result<Self::Value, __A::Error>
            where
                __A: serde::de::SeqAccess<'de>,
            {
                let __field0 = match match serde::de::SeqAccess::next_element::<&'a [u8]>(&mut __seq) {
                    Ok(__val) => __val,
                    Err(__err) => {
                        return Err(__err);
                    }
                } {
                    Some(__value) => __value,
                    None => {
                        return Err(serde::de::Error::invalid_length(
                            0usize,
                            &"tuple struct Digest with 1 element",
                        ));
                    }
                };
                Ok(oid::try_from_bytes(__field0).expect("hash of known length"))
            }
        }
        serde::Deserializer::deserialize_newtype_struct(
            deserializer,
            "Digest",
            __Visitor {
                marker: std::marker::PhantomData::<&'a oid>,
                lifetime: std::marker::PhantomData,
            },
        )
    }
}