litrs/bytestr/
mod.rs

1use std::{fmt, ops::Range};
2
3use crate::{
4    Buffer, ParseError,
5    err::{perr, ParseErrorKind::*},
6    escape::{scan_raw_string, unescape_string},
7};
8
9
10/// A byte string or raw byte string literal, e.g. `b"hello"` or `br#"abc"def"#`.
11///
12/// See [the reference][ref] for more information.
13///
14/// [ref]: https://doc.rust-lang.org/reference/tokens.html#byte-string-literals
15#[derive(Debug, Clone, PartialEq, Eq)]
16pub struct ByteStringLit<B: Buffer> {
17    /// The raw input.
18    raw: B,
19
20    /// The string value (with all escaped unescaped), or `None` if there were
21    /// no escapes. In the latter case, `input` is the string value.
22    value: Option<Vec<u8>>,
23
24    /// The number of hash signs in case of a raw string literal, or `None` if
25    /// it's not a raw string literal.
26    num_hashes: Option<u32>,
27}
28
29impl<B: Buffer> ByteStringLit<B> {
30    /// Parses the input as a (raw) byte string literal. Returns an error if the
31    /// input is invalid or represents a different kind of literal.
32    pub fn parse(input: B) -> Result<Self, ParseError> {
33        if input.is_empty() {
34            return Err(perr(None, Empty));
35        }
36        if !input.starts_with(r#"b""#) && !input.starts_with("br") {
37            return Err(perr(None, InvalidByteStringLiteralStart));
38        }
39
40        Self::parse_impl(input)
41    }
42
43    /// Returns the string value this literal represents (where all escapes have
44    /// been turned into their respective values).
45    pub fn value(&self) -> &[u8] {
46        self.value.as_deref().unwrap_or(&self.raw.as_bytes()[self.inner_range()])
47    }
48
49    /// Like `value` but returns a potentially owned version of the value.
50    ///
51    /// The return value is either `Cow<'static, [u8]>` if `B = String`, or
52    /// `Cow<'a, [u8]>` if `B = &'a str`.
53    pub fn into_value(self) -> B::ByteCow {
54        let inner_range = self.inner_range();
55        let Self { raw, value, .. } = self;
56        value.map(B::ByteCow::from).unwrap_or_else(|| raw.cut(inner_range).into_byte_cow())
57    }
58
59    /// Returns whether this literal is a raw string literal (starting with
60    /// `r`).
61    pub fn is_raw_byte_string(&self) -> bool {
62        self.num_hashes.is_some()
63    }
64
65    /// The range within `self.raw` that excludes the quotes and potential `r#`.
66    fn inner_range(&self) -> Range<usize> {
67        match self.num_hashes {
68            None => 2..self.raw.len() - 1,
69            Some(n) => 2 + n as usize + 1..self.raw.len() - n as usize - 1,
70        }
71    }
72
73    /// Precondition: input has to start with either `b"` or `br`.
74    pub(crate) fn parse_impl(input: B) -> Result<Self, ParseError> {
75        if input.starts_with(r"br") {
76            let (value, num_hashes) = scan_raw_string::<u8>(&input, 2)?;
77            Ok(Self {
78                raw: input,
79                value: value.map(|s| s.into_bytes()),
80                num_hashes: Some(num_hashes),
81            })
82        } else {
83            let value = unescape_string::<u8>(&input, 2)?.map(|s| s.into_bytes());
84            Ok(Self {
85                raw: input,
86                value,
87                num_hashes: None,
88            })
89        }
90    }
91}
92
93impl ByteStringLit<&str> {
94    /// Makes a copy of the underlying buffer and returns the owned version of
95    /// `Self`.
96    pub fn into_owned(self) -> ByteStringLit<String> {
97        ByteStringLit {
98            raw: self.raw.to_owned(),
99            value: self.value,
100            num_hashes: self.num_hashes,
101        }
102    }
103}
104
105impl<B: Buffer> fmt::Display for ByteStringLit<B> {
106    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
107        f.pad(&self.raw)
108    }
109}
110
111
112#[cfg(test)]
113mod tests;
litrs/bytestr/mod.rs

litrs/bytestr/
mod.rs
