Haibo Huang | 914311b | 2021-01-07 18:06:15 -0800 | [diff] [blame] | 1 | use super::lexer_impl::Lexer; |
| 2 | use super::lexer_impl::LexerError; |
| 3 | use crate::text_format::lexer::ParserLanguage; |
| 4 | use std::fmt; |
| 5 | use std::string::FromUtf8Error; |
| 6 | |
| 7 | #[derive(Debug)] |
| 8 | pub enum StrLitDecodeError { |
| 9 | FromUtf8Error(FromUtf8Error), |
| 10 | // TODO: be more specific |
| 11 | OtherError, |
| 12 | } |
| 13 | |
| 14 | impl fmt::Display for StrLitDecodeError { |
| 15 | fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { |
| 16 | match self { |
| 17 | StrLitDecodeError::FromUtf8Error(e) => write!(f, "{}", e), |
| 18 | StrLitDecodeError::OtherError => write!(f, "String literal decode error"), |
| 19 | } |
| 20 | } |
| 21 | } |
| 22 | |
| 23 | impl std::error::Error for StrLitDecodeError {} |
| 24 | |
| 25 | impl From<LexerError> for StrLitDecodeError { |
| 26 | fn from(_: LexerError) -> Self { |
| 27 | StrLitDecodeError::OtherError |
| 28 | } |
| 29 | } |
| 30 | |
| 31 | impl From<FromUtf8Error> for StrLitDecodeError { |
| 32 | fn from(e: FromUtf8Error) -> Self { |
| 33 | StrLitDecodeError::FromUtf8Error(e) |
| 34 | } |
| 35 | } |
| 36 | |
| 37 | pub type StrLitDecodeResult<T> = Result<T, StrLitDecodeError>; |
| 38 | |
| 39 | /// String literal, both `string` and `bytes`. |
| 40 | #[derive(Clone, Eq, PartialEq, Debug)] |
| 41 | pub struct StrLit { |
| 42 | pub escaped: String, |
| 43 | } |
| 44 | |
| 45 | impl StrLit { |
| 46 | /// May fail if not valid UTF8 |
| 47 | pub fn decode_utf8(&self) -> StrLitDecodeResult<String> { |
| 48 | let mut lexer = Lexer::new(&self.escaped, ParserLanguage::Json); |
| 49 | let mut r = Vec::new(); |
| 50 | while !lexer.eof() { |
| 51 | r.push(lexer.next_byte_value()?); |
| 52 | } |
| 53 | Ok(String::from_utf8(r)?) |
| 54 | } |
| 55 | |
| 56 | pub fn decode_bytes(&self) -> StrLitDecodeResult<Vec<u8>> { |
| 57 | let mut lexer = Lexer::new(&self.escaped, ParserLanguage::Json); |
| 58 | let mut r = Vec::new(); |
| 59 | while !lexer.eof() { |
| 60 | r.push(lexer.next_byte_value()?); |
| 61 | } |
| 62 | Ok(r) |
| 63 | } |
| 64 | |
| 65 | pub fn quoted(&self) -> String { |
| 66 | format!("\"{}\"", self.escaped) |
| 67 | } |
| 68 | } |
| 69 | |
| 70 | #[cfg(test)] |
| 71 | mod test { |
| 72 | use crate::text_format::lexer::StrLit; |
| 73 | |
| 74 | #[test] |
| 75 | fn decode_utf8() { |
| 76 | assert_eq!( |
| 77 | "\u{1234}".to_owned(), |
| 78 | StrLit { |
| 79 | escaped: "\\341\\210\\264".to_owned() |
| 80 | } |
| 81 | .decode_utf8() |
| 82 | .unwrap() |
| 83 | ) |
| 84 | } |
| 85 | } |