nom/bytes/

complete.rs

//! Parsers recognizing bytes streams, complete input version

use core::marker::PhantomData;

use crate::error::ParseError;
use crate::internal::{IResult, Parser};
use crate::traits::{Compare, FindSubstring, FindToken, ToUsize};
use crate::Complete;
use crate::Emit;
use crate::Input;
use crate::OutputM;

/// Recognizes a pattern
///
/// The input data will be compared to the tag combinator's argument and will return the part of
/// the input that matches the argument
///
/// It will return `Err(Err::Error((_, ErrorKind::Tag)))` if the input doesn't match the pattern
/// # Example
/// ```rust
/// # use nom::{Err, error::{Error, ErrorKind}, Needed, IResult};
/// use nom::bytes::complete::tag;
///
/// fn parser(s: &str) -> IResult<&str, &str> {
///   tag("Hello")(s)
/// }
///
/// assert_eq!(parser("Hello, World!"), Ok((", World!", "Hello")));
/// assert_eq!(parser("Something"), Err(Err::Error(Error::new("Something", ErrorKind::Tag))));
/// assert_eq!(parser(""), Err(Err::Error(Error::new("", ErrorKind::Tag))));
/// ```
pub fn tag<T, I, Error: ParseError<I>>(tag: T) -> impl Fn(I) -> IResult<I, I, Error>
where
  I: Input + Compare<T>,
  T: Input + Clone,
{
  move |i: I| {
    let mut parser = super::Tag {
      tag: tag.clone(),
      e: PhantomData,
    };

    parser.process::<OutputM<Emit, Emit, Complete>>(i)
  }
}

/// Recognizes a case insensitive pattern.
///
/// The input data will be compared to the tag combinator's argument and will return the part of
/// the input that matches the argument with no regard to case.
///
/// It will return `Err(Err::Error((_, ErrorKind::Tag)))` if the input doesn't match the pattern.
/// # Example
/// ```rust
/// # use nom::{Err, error::{Error, ErrorKind}, Needed, IResult};
/// use nom::bytes::complete::tag_no_case;
///
/// fn parser(s: &str) -> IResult<&str, &str> {
///   tag_no_case("hello")(s)
/// }
///
/// assert_eq!(parser("Hello, World!"), Ok((", World!", "Hello")));
/// assert_eq!(parser("hello, World!"), Ok((", World!", "hello")));
/// assert_eq!(parser("HeLlO, World!"), Ok((", World!", "HeLlO")));
/// assert_eq!(parser("Something"), Err(Err::Error(Error::new("Something", ErrorKind::Tag))));
/// assert_eq!(parser(""), Err(Err::Error(Error::new("", ErrorKind::Tag))));
/// ```
pub fn tag_no_case<T, I, Error: ParseError<I>>(tag: T) -> impl Fn(I) -> IResult<I, I, Error>
where
  I: Input + Compare<T>,
  T: Input + Clone,
{
  move |i: I| {
    let mut parser = super::TagNoCase {
      tag: tag.clone(),
      e: PhantomData,
    };

    parser.process::<OutputM<Emit, Emit, Complete>>(i)
  }
}

/// Parse till certain characters are met.
///
/// The parser will return the longest slice till one of the characters of the combinator's argument are met.
///
/// It doesn't consume the matched character.
///
/// It will return a `Err::Error(("", ErrorKind::IsNot))` if the pattern wasn't met.
/// # Example
/// ```rust
/// # use nom::{Err, error::{Error, ErrorKind}, Needed, IResult};
/// use nom::bytes::complete::is_not;
///
/// fn not_space(s: &str) -> IResult<&str, &str> {
///   is_not(" \t\r\n")(s)
/// }
///
/// assert_eq!(not_space("Hello, World!"), Ok((" World!", "Hello,")));
/// assert_eq!(not_space("Sometimes\t"), Ok(("\t", "Sometimes")));
/// assert_eq!(not_space("Nospace"), Ok(("", "Nospace")));
/// assert_eq!(not_space(""), Err(Err::Error(Error::new("", ErrorKind::IsNot))));
/// ```
pub fn is_not<T, I, Error: ParseError<I>>(arr: T) -> impl FnMut(I) -> IResult<I, I, Error>
where
  I: Input,
  T: FindToken<<I as Input>::Item>,
{
  let mut parser = super::is_not(arr);

  move |i: I| parser.process::<OutputM<Emit, Emit, Complete>>(i)
}

/// Returns the longest slice of the matches the pattern.
///
/// The parser will return the longest slice consisting of the characters in provided in the
/// combinator's argument.
///
/// It will return a `Err(Err::Error((_, ErrorKind::IsA)))` if the pattern wasn't met.
/// # Example
/// ```rust
/// # use nom::{Err, error::{Error, ErrorKind}, Needed, IResult};
/// use nom::bytes::complete::is_a;
///
/// fn hex(s: &str) -> IResult<&str, &str> {
///   is_a("1234567890ABCDEF")(s)
/// }
///
/// assert_eq!(hex("123 and voila"), Ok((" and voila", "123")));
/// assert_eq!(hex("DEADBEEF and others"), Ok((" and others", "DEADBEEF")));
/// assert_eq!(hex("BADBABEsomething"), Ok(("something", "BADBABE")));
/// assert_eq!(hex("D15EA5E"), Ok(("", "D15EA5E")));
/// assert_eq!(hex(""), Err(Err::Error(Error::new("", ErrorKind::IsA))));
/// ```
pub fn is_a<T, I, Error: ParseError<I>>(arr: T) -> impl FnMut(I) -> IResult<I, I, Error>
where
  I: Input,
  T: FindToken<<I as Input>::Item>,
{
  let mut parser = super::is_a(arr);

  move |i: I| parser.process::<OutputM<Emit, Emit, Complete>>(i)
}

/// Returns the longest input slice (if any) that matches the predicate.
///
/// The parser will return the longest slice that matches the given predicate *(a function that
/// takes the input and returns a bool)*.
/// # Example
/// ```rust
/// # use nom::{Err, error::ErrorKind, Needed, IResult};
/// use nom::bytes::complete::take_while;
/// use nom::AsChar;
///
/// fn alpha(s: &[u8]) -> IResult<&[u8], &[u8]> {
///   take_while(AsChar::is_alpha)(s)
/// }
///
/// assert_eq!(alpha(b"latin123"), Ok((&b"123"[..], &b"latin"[..])));
/// assert_eq!(alpha(b"12345"), Ok((&b"12345"[..], &b""[..])));
/// assert_eq!(alpha(b"latin"), Ok((&b""[..], &b"latin"[..])));
/// assert_eq!(alpha(b""), Ok((&b""[..], &b""[..])));
/// ```
pub fn take_while<F, I, Error: ParseError<I>>(cond: F) -> impl FnMut(I) -> IResult<I, I, Error>
where
  I: Input,
  F: Fn(<I as Input>::Item) -> bool,
{
  let mut parser = super::take_while(cond);

  move |i: I| parser.process::<OutputM<Emit, Emit, Complete>>(i)
}

/// Returns the longest (at least 1) input slice that matches the predicate.
///
/// The parser will return the longest slice that matches the given predicate *(a function that
/// takes the input and returns a bool)*.
///
/// It will return an `Err(Err::Error((_, ErrorKind::TakeWhile1)))` if the pattern wasn't met.
/// # Example
/// ```rust
/// # use nom::{Err, error::{Error, ErrorKind}, Needed, IResult};
/// use nom::bytes::complete::take_while1;
/// use nom::AsChar;
///
/// fn alpha(s: &[u8]) -> IResult<&[u8], &[u8]> {
///   take_while1(AsChar::is_alpha)(s)
/// }
///
/// assert_eq!(alpha(b"latin123"), Ok((&b"123"[..], &b"latin"[..])));
/// assert_eq!(alpha(b"latin"), Ok((&b""[..], &b"latin"[..])));
/// assert_eq!(alpha(b"12345"), Err(Err::Error(Error::new(&b"12345"[..], ErrorKind::TakeWhile1))));
/// ```
pub fn take_while1<F, I, Error: ParseError<I>>(cond: F) -> impl FnMut(I) -> IResult<I, I, Error>
where
  I: Input,
  F: Fn(<I as Input>::Item) -> bool,
{
  let mut parser = super::take_while1(cond);

  move |i: I| parser.process::<OutputM<Emit, Emit, Complete>>(i)
}

/// Returns the longest (m <= len <= n) input slice that matches the predicate.
///
/// The parser will return the longest slice that matches the given predicate *(a function that
/// takes the input and returns a bool)*.
///
/// It will return an `Err::Error((_, ErrorKind::TakeWhileMN))` if the pattern wasn't met or is out
/// of range (m <= len <= n).
/// # Example
/// ```rust
/// # use nom::{Err, error::{Error, ErrorKind}, Needed, IResult};
/// use nom::bytes::complete::take_while_m_n;
/// use nom::AsChar;
///
/// fn short_alpha(s: &[u8]) -> IResult<&[u8], &[u8]> {
///   take_while_m_n(3, 6, AsChar::is_alpha)(s)
/// }
///
/// assert_eq!(short_alpha(b"latin123"), Ok((&b"123"[..], &b"latin"[..])));
/// assert_eq!(short_alpha(b"lengthy"), Ok((&b"y"[..], &b"length"[..])));
/// assert_eq!(short_alpha(b"latin"), Ok((&b""[..], &b"latin"[..])));
/// assert_eq!(short_alpha(b"ed"), Err(Err::Error(Error::new(&b"ed"[..], ErrorKind::TakeWhileMN))));
/// assert_eq!(short_alpha(b"12345"), Err(Err::Error(Error::new(&b"12345"[..], ErrorKind::TakeWhileMN))));
/// ```
pub fn take_while_m_n<F, I, Error: ParseError<I>>(
  m: usize,
  n: usize,
  cond: F,
) -> impl FnMut(I) -> IResult<I, I, Error>
where
  I: Input,
  F: Fn(<I as Input>::Item) -> bool,
{
  let mut parser = super::take_while_m_n(m, n, cond);

  move |i: I| parser.process::<OutputM<Emit, Emit, Complete>>(i)
}

/// Returns the longest input slice (if any) till a predicate is met.
///
/// The parser will return the longest slice till the given predicate *(a function that
/// takes the input and returns a bool)*.
/// # Example
/// ```rust
/// # use nom::{Err, error::ErrorKind, Needed, IResult};
/// use nom::bytes::complete::take_till;
///
/// fn till_colon(s: &str) -> IResult<&str, &str> {
///   take_till(|c| c == ':')(s)
/// }
///
/// assert_eq!(till_colon("latin:123"), Ok((":123", "latin")));
/// assert_eq!(till_colon(":empty matched"), Ok((":empty matched", ""))); //allowed
/// assert_eq!(till_colon("12345"), Ok(("", "12345")));
/// assert_eq!(till_colon(""), Ok(("", "")));
/// ```
#[allow(clippy::redundant_closure)]
pub fn take_till<F, I, Error: ParseError<I>>(cond: F) -> impl FnMut(I) -> IResult<I, I, Error>
where
  I: Input,
  F: Fn(<I as Input>::Item) -> bool,
{
  let mut parser = super::take_till(cond);

  move |i: I| parser.process::<OutputM<Emit, Emit, Complete>>(i)
}

/// Returns the longest (at least 1) input slice till a predicate is met.
///
/// The parser will return the longest slice till the given predicate *(a function that
/// takes the input and returns a bool)*.
///
/// It will return `Err(Err::Error((_, ErrorKind::TakeTill1)))` if the input is empty or the
/// predicate matches the first input.
/// # Example
/// ```rust
/// # use nom::{Err, error::{Error, ErrorKind}, Needed, IResult};
/// use nom::bytes::complete::take_till1;
///
/// fn till_colon(s: &str) -> IResult<&str, &str> {
///   take_till1(|c| c == ':')(s)
/// }
///
/// assert_eq!(till_colon("latin:123"), Ok((":123", "latin")));
/// assert_eq!(till_colon(":empty matched"), Err(Err::Error(Error::new(":empty matched", ErrorKind::TakeTill1))));
/// assert_eq!(till_colon("12345"), Ok(("", "12345")));
/// assert_eq!(till_colon(""), Err(Err::Error(Error::new("", ErrorKind::TakeTill1))));
/// ```
#[allow(clippy::redundant_closure)]
pub fn take_till1<F, I, Error: ParseError<I>>(cond: F) -> impl FnMut(I) -> IResult<I, I, Error>
where
  I: Input,
  F: Fn(<I as Input>::Item) -> bool,
{
  let mut parser = super::take_till1(cond);

  move |i: I| parser.process::<OutputM<Emit, Emit, Complete>>(i)
}

/// Returns an input slice containing the first N input elements (Input[..N]).
///
/// It will return `Err(Err::Error((_, ErrorKind::Eof)))` if the input is shorter than the argument.
/// # Example
/// ```rust
/// # use nom::{Err, error::{Error, ErrorKind}, Needed, IResult};
/// use nom::bytes::complete::take;
///
/// fn take6(s: &str) -> IResult<&str, &str> {
///   take(6usize)(s)
/// }
///
/// assert_eq!(take6("1234567"), Ok(("7", "123456")));
/// assert_eq!(take6("things"), Ok(("", "things")));
/// assert_eq!(take6("short"), Err(Err::Error(Error::new("short", ErrorKind::Eof))));
/// assert_eq!(take6(""), Err(Err::Error(Error::new("", ErrorKind::Eof))));
/// ```
///
/// The units that are taken will depend on the input type. For example, for a
/// `&str` it will take a number of `char`'s, whereas for a `&[u8]` it will
/// take that many `u8`'s:
///
/// ```rust
/// use nom::error::Error;
/// use nom::bytes::complete::take;
///
/// assert_eq!(take::<_, _, Error<_>>(1usize)("💙"), Ok(("", "💙")));
/// assert_eq!(take::<_, _, Error<_>>(1usize)("💙".as_bytes()), Ok((b"\x9F\x92\x99".as_ref(), b"\xF0".as_ref())));
/// ```
pub fn take<C, I, Error: ParseError<I>>(count: C) -> impl FnMut(I) -> IResult<I, I, Error>
where
  I: Input,
  C: ToUsize,
{
  let mut parser = super::take(count);

  move |i: I| parser.process::<OutputM<Emit, Emit, Complete>>(i)
}

/// Returns the input slice up to the first occurrence of the pattern.
///
/// It doesn't consume the pattern. It will return `Err(Err::Error((_, ErrorKind::TakeUntil)))`
/// if the pattern wasn't met.
/// # Example
/// ```rust
/// # use nom::{Err, error::{Error, ErrorKind}, Needed, IResult};
/// use nom::bytes::complete::take_until;
///
/// fn until_eof(s: &str) -> IResult<&str, &str> {
///   take_until("eof")(s)
/// }
///
/// assert_eq!(until_eof("hello, worldeof"), Ok(("eof", "hello, world")));
/// assert_eq!(until_eof("hello, world"), Err(Err::Error(Error::new("hello, world", ErrorKind::TakeUntil))));
/// assert_eq!(until_eof(""), Err(Err::Error(Error::new("", ErrorKind::TakeUntil))));
/// assert_eq!(until_eof("1eof2eof"), Ok(("eof2eof", "1")));
/// ```
pub fn take_until<T, I, Error: ParseError<I>>(tag: T) -> impl FnMut(I) -> IResult<I, I, Error>
where
  I: Input + FindSubstring<T>,
  T: Input + Clone,
{
  let mut parser = super::take_until(tag);

  move |i: I| parser.process::<OutputM<Emit, Emit, Complete>>(i)
}

/// Returns the non empty input slice up to the first occurrence of the pattern.
///
/// It doesn't consume the pattern. It will return `Err(Err::Error((_, ErrorKind::TakeUntil)))`
/// if the pattern wasn't met.
/// # Example
/// ```rust
/// # use nom::{Err, error::{Error, ErrorKind}, Needed, IResult};
/// use nom::bytes::complete::take_until1;
///
/// fn until_eof(s: &str) -> IResult<&str, &str> {
///   take_until1("eof")(s)
/// }
///
/// assert_eq!(until_eof("hello, worldeof"), Ok(("eof", "hello, world")));
/// assert_eq!(until_eof("hello, world"), Err(Err::Error(Error::new("hello, world", ErrorKind::TakeUntil))));
/// assert_eq!(until_eof(""), Err(Err::Error(Error::new("", ErrorKind::TakeUntil))));
/// assert_eq!(until_eof("1eof2eof"), Ok(("eof2eof", "1")));
/// assert_eq!(until_eof("eof"), Err(Err::Error(Error::new("eof", ErrorKind::TakeUntil))));
/// ```
pub fn take_until1<T, I, Error: ParseError<I>>(tag: T) -> impl FnMut(I) -> IResult<I, I, Error>
where
  I: Input + FindSubstring<T>,
  T: Input + Clone,
{
  let mut parser = super::take_until1(tag);

  move |i: I| parser.process::<OutputM<Emit, Emit, Complete>>(i)
}

/// Matches a byte string with escaped characters.
///
/// * The first argument matches the normal characters (it must not accept the control character)
/// * The second argument is the control character (like `\` in most languages)
/// * The third argument matches the escaped characters
/// # Example
/// ```
/// # use nom::{Err, error::ErrorKind, Needed, IResult};
/// # use nom::character::complete::digit1;
/// use nom::bytes::complete::escaped;
/// use nom::character::complete::one_of;
///
/// fn esc(s: &str) -> IResult<&str, &str> {
///   escaped(digit1, '\\', one_of(r#""n\"#))(s)
/// }
///
/// assert_eq!(esc("123;"), Ok((";", "123")));
/// assert_eq!(esc(r#"12\"34;"#), Ok((";", r#"12\"34"#)));
/// ```
///
pub fn escaped<'a, I, Error, F, G>(
  normal: F,
  control_char: char,
  escapable: G,
) -> impl FnMut(I) -> IResult<I, I, Error>
where
  I: Clone + crate::traits::Offset + Input + 'a,
  <I as Input>::Item: crate::traits::AsChar,
  F: Parser<I, Error = Error>,
  G: Parser<I, Error = Error>,
  Error: ParseError<I>,
{
  let mut parser = super::escaped(normal, control_char, escapable);

  move |i: I| parser.process::<OutputM<Emit, Emit, Complete>>(i)
}

/// Matches a byte string with escaped characters.
///
/// * The first argument matches the normal characters (it must not match the control character)
/// * The second argument is the control character (like `\` in most languages)
/// * The third argument matches the escaped characters and transforms them
///
/// As an example, the chain `abc\tdef` could be `abc    def` (it also consumes the control character)
///
/// ```
/// # use nom::{Err, error::ErrorKind, Needed, IResult};
/// # use std::str::from_utf8;
/// use nom::bytes::complete::{escaped_transform, tag};
/// use nom::character::complete::alpha1;
/// use nom::branch::alt;
/// use nom::combinator::value;
///
/// fn parser(input: &str) -> IResult<&str, String> {
///   escaped_transform(
///     alpha1,
///     '\\',
///     alt((
///       value("\\", tag("\\")),
///       value("\"", tag("\"")),
///       value("\n", tag("n")),
///     ))
///   )(input)
/// }
///
/// assert_eq!(parser("ab\\\"cd"), Ok(("", String::from("ab\"cd"))));
/// assert_eq!(parser("ab\\ncd"), Ok(("", String::from("ab\ncd"))));
/// ```
#[cfg(feature = "alloc")]
#[cfg_attr(feature = "docsrs", doc(cfg(feature = "alloc")))]
pub fn escaped_transform<I, Error, F, G, O1, O2, ExtendItem, Output>(
  normal: F,
  control_char: char,
  transform: G,
) -> impl FnMut(I) -> IResult<I, Output, Error>
where
  I: Clone + crate::traits::Offset + Input,
  I: crate::traits::ExtendInto<Item = ExtendItem, Extender = Output>,
  O1: crate::traits::ExtendInto<Item = ExtendItem, Extender = Output>,
  O2: crate::traits::ExtendInto<Item = ExtendItem, Extender = Output>,
  <I as Input>::Item: crate::traits::AsChar,
  F: Parser<I, Output = O1, Error = Error>,
  G: Parser<I, Output = O2, Error = Error>,
  Error: ParseError<I>,
{
  let mut parser = super::escaped_transform(normal, control_char, transform);

  move |i: I| parser.process::<OutputM<Emit, Emit, Complete>>(i)
}

#[cfg(test)]
mod tests {
  use super::*;
  use crate::error::ErrorKind;
  use crate::AsChar;
  use crate::Err;

  #[test]
  fn complete_take_while_m_n_utf8_all_matching() {
    let result: IResult<&str, &str> =
      super::take_while_m_n(1, 4, |c: char| c.is_alphabetic())("øn");
    assert_eq!(result, Ok(("", "øn")));
  }

  #[test]
  fn complete_take_while_m_n_utf8_all_matching_substring() {
    let result: IResult<&str, &str> =
      super::take_while_m_n(1, 1, |c: char| c.is_alphabetic())("øn");
    assert_eq!(result, Ok(("n", "ø")));
  }

  // issue #1336 "escaped hangs if normal parser accepts empty"
  fn escaped_string(input: &str) -> IResult<&str, &str> {
    use crate::character::complete::{alpha0, one_of};
    escaped(alpha0, '\\', one_of("n"))(input)
  }

  // issue #1336 "escaped hangs if normal parser accepts empty"
  #[test]
  fn escaped_hang() {
    escaped_string("7").unwrap();
    escaped_string("a7").unwrap();
  }

  // issue ##1118 escaped does not work with empty string
  fn unquote(input: &str) -> IResult<&str, &str> {
    use crate::bytes::complete::*;
    use crate::character::complete::*;
    use crate::combinator::opt;
    use crate::sequence::delimited;

    delimited(
      char('"'),
      escaped(opt(none_of(r#"\""#)), '\\', one_of(r#"\"rnt"#)),
      char('"'),
    )
    .parse(input)
  }

  #[test]
  fn escaped_hang_1118() {
    assert_eq!(unquote(r#""""#), Ok(("", "")));
  }

  // issue #1630 take_while_m_n is invalid for multi-byte UTF-8 characters
  #[test]
  fn complete_take_while_m_n_multibyte() {
    use crate::error::Error;

    fn multi_byte_chars(s: &str, m: usize, n: usize) -> IResult<&str, &str> {
      take_while_m_n(m, n, |c: char| c.len() > 1)(s)
    }

    assert_eq!(multi_byte_chars("€ latin", 0, 64), Ok((" latin", "€")));
    assert_eq!(multi_byte_chars("𝄠 latin", 0, 1), Ok((" latin", "𝄠")));
    assert_eq!(multi_byte_chars("باب latin", 0, 64), Ok((" latin", "باب")));
    assert_eq!(
      multi_byte_chars("💣💢ᾠ latin", 3, 3),
      Ok((" latin", "💣💢ᾠ"))
    );
    assert_eq!(multi_byte_chars("latin", 0, 64), Ok(("latin", "")));
    assert_eq!(multi_byte_chars("باب", 1, 3), Ok(("", "باب")));
    assert_eq!(multi_byte_chars("باب", 1, 2), Ok(("ب", "با")));
    assert_eq!(
      multi_byte_chars("latin", 1, 64),
      Err(Err::Error(Error::new("latin", ErrorKind::TakeWhileMN)))
    );
  }
}