xmpp-rs/xso/src/text.rs

use std::borrow::Cow;
use std::fmt::{self, Write};
use std::marker::PhantomData;
use std::ops::Deref;

use base64::engine::{general_purpose::STANDARD as StandardBase64Engine, Engine as Base64Engine};

use crate::{error::Error, FromXmlText, IntoXmlText};

/// Represent a way to encode/decode text data into a Rust type.
///
/// This trait is used to specify how a given type is serialised into and from
/// a string. This bypasses the [`FromXmlText`] / [`IntoXmlText`] traits,
/// allowing to hide the specific implementation for transforming data from
/// text into Rust types (and back) from a container's public interface.
pub trait TextCodec<T> {
    /// Decode a string value into the type.
    fn decode(s: &str) -> Result<T, Error>;

    /// Encode the type as string value.
    ///
    /// If this returns `None`, the string value is not emitted at all.
    fn encode(value: T) -> Option<String>;
}

/// Text codec which does no transform.
pub struct Plain;

impl TextCodec<String> for Plain {
    fn decode(s: &str) -> Result<String, Error> {
        Ok(s.to_string())
    }

    fn encode(value: String) -> Option<String> {
        Some(value)
    }
}

/// Text codec which returns None instead of the empty string.
pub struct EmptyAsNone;

impl TextCodec<Option<String>> for EmptyAsNone {
    fn decode(s: &str) -> Result<Option<String>, Error> {
        if s.len() == 0 {
            Ok(None)
        } else {
            Ok(Some(s.to_owned()))
        }
    }

    fn encode(value: Option<String>) -> Option<String> {
        match value {
            Some(v) if v.len() > 0 => Some(v),
            Some(_) | None => None,
        }
    }
}

/// Text codec which trims whitespace at the beginning and the end of the
/// text.
///
/// Optionally processes the text further using a nested codec.
pub struct Trimmed<Inner = Plain>(PhantomData<Inner>);

impl<Inner: TextCodec<String>> TextCodec<String> for Trimmed<Inner> {
    fn decode(s: &str) -> Result<String, Error> {
        Inner::decode(s.trim())
    }

    fn encode(decoded: String) -> Option<String> {
        Inner::encode(decoded)
    }
}

/// Trait for preprocessing text data from XML.
///
/// This may be used by codecs to allow to customize some of their behaviour.
pub trait TextFilter {
    /// Process the incoming string and return the result of the processing.
    ///
    /// This should avoid copying wherever possible.
    fn preprocess(s: &str) -> Cow<'_, str>;
}

/// Text preprocessor which returns the input unchanged.
pub struct NoFilter;

impl TextFilter for NoFilter {
    fn preprocess(s: &str) -> Cow<'_, str> {
        Cow::Borrowed(s)
    }
}

/// Text preprocessor to remove all whitespace.
pub struct StripWhitespace;

impl TextFilter for StripWhitespace {
    fn preprocess(s: &str) -> Cow<'_, str> {
        let s: String = s
            .chars()
            .filter(|ch| *ch != ' ' && *ch != '\n' && *ch != '\t')
            .collect();
        Cow::Owned(s)
    }
}

/// Text codec transforming text to binary using standard base64.
///
/// The `Filter` type argument can be used to employ additional preprocessing
/// of incoming text data. Most interestingly, passing [`StripWhitespace`]
/// will make the implementation ignore any whitespace within the text.
pub struct Base64<Filter: TextFilter = NoFilter>(PhantomData<Filter>);

impl<Filter: TextFilter> TextCodec<Vec<u8>> for Base64<Filter> {
    fn decode(s: &str) -> Result<Vec<u8>, Error> {
        let value = Filter::preprocess(s);
        Ok(StandardBase64Engine.decode(value.as_ref())?)
    }

    fn encode(value: Vec<u8>) -> Option<String> {
        Some(StandardBase64Engine.encode(&value))
    }
}

impl<Filter: TextFilter> TextCodec<Option<Vec<u8>>> for Base64<Filter> {
    fn decode(s: &str) -> Result<Option<Vec<u8>>, Error> {
        if s.len() == 0 {
            return Ok(None);
        }
        Ok(Some(Self::decode(s)?))
    }

    fn encode(decoded: Option<Vec<u8>>) -> Option<String> {
        decoded.and_then(Self::encode)
    }
}

/// Text codec transforming text to binary by emitting it as colon-separated
/// hex value.
pub struct ColonSeparatedHex;

impl TextCodec<Vec<u8>> for ColonSeparatedHex {
    fn decode(s: &str) -> Result<Vec<u8>, Error> {
        let mut bytes = vec![];
        for i in 0..(1 + s.len()) / 3 {
            let byte = u8::from_str_radix(&s[3 * i..3 * i + 2], 16)?;
            if 3 * i + 2 < s.len() {
                assert_eq!(&s[3 * i + 2..3 * i + 3], ":");
            }
            bytes.push(byte);
        }
        Ok(bytes)
    }

    fn encode(decoded: Vec<u8>) -> Option<String> {
        let mut bytes = vec![];
        for byte in decoded {
            bytes.push(format!("{:02X}", byte));
        }
        Some(bytes.join(":"))
    }
}

/// Codec for bytes of lowercase hexadecimal, with a fixed length `N` (in bytes).
pub struct Hex;

impl<const N: usize> TextCodec<[u8; N]> for Hex {
    fn decode(s: &str) -> Result<[u8; N], Error> {
        if s.len() != 2 * N {
            return Err(Error::ParseError("Invalid length"));
        }

        let mut bytes = [0u8; N];
        for i in 0..N {
            bytes[i] = u8::from_str_radix(&s[2 * i..2 * i + 2], 16)?;
        }

        Ok(bytes)
    }

    fn encode(decoded: [u8; N]) -> Option<String> {
        let mut bytes = String::with_capacity(N * 2);
        for byte in decoded {
            write!(&mut bytes, "{:02x}", byte).unwrap();
        }
        Some(bytes)
    }
}

impl<const N: usize> TextCodec<Option<[u8; N]>> for Hex {
    fn decode(s: &str) -> Result<Option<[u8; N]>, Error> {
        if s.len() == 0 {
            return Ok(None);
        }
        Ok(Some(Self::decode(s)?))
    }

    fn encode(decoded: Option<[u8; N]>) -> Option<String> {
        decoded.and_then(Self::encode)
    }
}

impl TextCodec<Vec<u8>> for Hex {
    fn decode(s: &str) -> Result<Vec<u8>, Error> {
        if s.len() % 2 != 0 {
            return Err(Error::ParseError("Invalid length"));
        }
        let n = s.len() / 2;

        let mut bytes = Vec::with_capacity(n);
        bytes.resize(n, 0u8);
        for i in 0..n {
            bytes[i] = u8::from_str_radix(&s[2 * i..2 * i + 2], 16)?;
        }

        Ok(bytes)
    }

    fn encode(decoded: Vec<u8>) -> Option<String> {
        let mut bytes = String::with_capacity(decoded.len() * 2);
        for byte in decoded {
            write!(&mut bytes, "{:02x}", byte).unwrap();
        }
        Some(bytes)
    }
}

/// Shim wrapper around [`String`] which refuses to parse if the string
/// is empty
#[derive(Debug, Clone)]
pub struct NonEmptyString(String);

impl NonEmptyString {
    /// Return `Some(s)` if s is not empty, `None` otherwise.
    pub fn new(s: String) -> Option<Self> {
        if s.len() == 0 {
            return None;
        }
        Some(Self(s))
    }
}

impl Deref for NonEmptyString {
    type Target = String;

    fn deref(&self) -> &Self::Target {
        &self.0
    }
}

impl fmt::Display for NonEmptyString {
    fn fmt<'f>(&self, f: &'f mut fmt::Formatter) -> fmt::Result {
        <String as fmt::Display>::fmt(&self.0, f)
    }
}

impl FromXmlText for NonEmptyString {
    fn from_xml_text(s: &str) -> Result<Self, Error> {
        if s.len() == 0 {
            return Err(Error::ParseError("string must not be empty"));
        }
        Ok(Self(s.to_string()))
    }
}

impl IntoXmlText for NonEmptyString {
    fn into_xml_text(self) -> String {
        self.0
    }
}

/// Text codec which rejects empty strings.
impl TextCodec<String> for NonEmptyString {
    fn decode(s: &str) -> Result<String, Error> {
        if s.len() == 0 {
            return Err(Error::ParseError("string must not be empty"));
        }
        Ok(s.to_owned())
    }

    fn encode(value: String) -> Option<String> {
        Some(value)
    }
}