datse/src/conv.rs

//! Conversion formats
use super::*;
use std::{
    ops::Deref,
    borrow::Borrow,
    fmt,
    error,
    convert::TryFrom,
};
use regex::{
    Regex,
};

//TODO: Mayb use base65536 for URL stuffs? idk..

const BASE64_VALIDATE_RE_STR: &'static str = r#"^(?:[A-Za-z0-9+/]{4})*(?:[A-Za-z0-9+/]{2}==|[A-Za-z0-9+/]{3}=)?$"#;
const MOD_BASE64_VALIDATE_RE_STR: &'static str = r#"^(?:[-A-Za-z0-9_]{4})*(?:[-_A-Za-z0-9]{2}~~|[-_A-Za-z0-9]{3}~)?$"#;

lazy_static!{
    static ref BASE64_CONV_TABLE: smallmap::Map<char, char> = smallmap![
	{'/' => '_'},
	{'+' => '-'},
	{'=' => '~'},
    ];
    static ref BASE64_CONV_TABLE_REV: smallmap::Map<char, char> = BASE64_CONV_TABLE.clone().reverse();

    static ref BASE64_VALIDATE_REGEX: Regex = Regex::new(BASE64_VALIDATE_RE_STR).expect("Failed to compile base64 validation regex");
    static ref MOD_BASE64_VALIDATE_REGEX: Regex = Regex::new(MOD_BASE64_VALIDATE_RE_STR).expect("Failed to compile modified base64 validation regex");
}

/// An error in base64 or modified base64 encoding/decoding.
#[derive(Debug)]
pub struct Base64Error<T>(T);

impl<T> Base64Error<T>
{
    /// The invalid value of this error
    pub fn value(&self) -> &T
    {
	&self.0
    }
    /// Consume into the invalid value from this instance
    #[inline] pub fn into_inner(self) -> T
    {
	self.0
    }
}

impl<T> error::Error for Base64Error<T>
where T: AsRef<str> + fmt::Debug{}
impl<T> fmt::Display for Base64Error<T>
    where T: AsRef<str>
{
    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result
    {
	write!(f, "invalid base64: {:?}", self.0.as_ref())
    }
}


/// A string of modified base64, appropriate for URL paths etc.
#[derive(Debug, Clone, PartialEq, Eq, Hash, PartialOrd, Ord, Serialize, Deserialize)]
pub struct ModifiedBase64String(String);

impl fmt::Display for ModifiedBase64String
{
    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result
    {
	write!(f, "{}", self.0)
    }
}

impl ModifiedBase64String
{
    #[inline(always)] fn from_base64_unchecked(string: &str) -> Self
    {
	Self(conv_str(&BASE64_CONV_TABLE, string).collect())
    }

    /// Consume into a normal base64 string
    pub fn into_base64(self) -> String
    {
	conv_char_iter(&BASE64_CONV_TABLE_REV, self.0.chars()).collect()
    }

    /// Consume into the inner modified base64 string
    pub fn into_string(self) -> String
    {
	self.0
    }

    /// Create an instance from a modified base64 string
    pub fn new(from: String) -> Result<Self, Base64Error<String>>
    {
	if MOD_BASE64_VALIDATE_REGEX.is_match(from.as_str()) {
	    Ok(Self(from))
	} else {
	    Err(Base64Error(from))
	}
    }

    /// Try to convert a base64 string into a modified base64 string
    pub fn try_from_base64<T: AsRef<str>>(base64: T) -> Result<Self, Base64Error<T>>
    {
	let string = base64.as_ref();
	if BASE64_VALIDATE_REGEX.is_match(string) {
	    Ok(Self::from_base64_unchecked(string))
	} else {
	    Err(Base64Error(base64))
	}
    }

    /// As a string slice
    #[inline(always)] pub fn as_str(&self) -> &str
    {
	self.0.as_str()
    }

    /// Encode from a slice
    pub fn encode(slice: impl AsRef<[u8]>) -> Self
    {
	Self::from_base64_unchecked(base64::encode(slice.as_ref()).as_str())
    }

    /// Consume into decoded bytes, write those bytes into the provided buffer
    pub fn decode(self, output: &mut [u8]) -> usize
    {
	base64::decode_config_slice(self.into_base64(), base64::STANDARD, output).expect("modified base64 string contained invalid formatted data")
    }

    /// Consume into decoded bytes, return those bytes as a new `Vec<u8>`
    pub fn decode_new(self) -> Vec<u8>
    {
	base64::decode(self.into_base64()).expect("modified base64 string contained invalid formatted data")
    }
}

impl Deref for ModifiedBase64String
{
    type Target = str;
    #[inline] fn deref(&self) -> &Self::Target {
	self.as_str()
    }
}

impl AsRef<str> for ModifiedBase64String
{
    #[inline] fn as_ref(&self) -> &str
    {
	self.as_str()
    }
}

impl TryFrom<String> for ModifiedBase64String
{
    type Error = Base64Error<String>;

    #[inline] fn try_from(from: String) -> Result<Self, Self::Error>
    {
	Self::try_from_base64(from)
    }
}

impl From<ModifiedBase64String> for String
{
    #[inline] fn from(from: ModifiedBase64String) -> Self
    {
	from.into_base64()
    }
}


/// Convert this string with a specified char map. Returns a `char` yielding iterator.
#[inline] pub fn conv_str<'a, 'b>(table: &'b smallmap::Map<char, char>, string: &'a (impl AsRef<str> + ?Sized)) -> CharSubstituteIter<'b, std::str::Chars<'a>>
where 'b: 'a
{
    conv_char_iter(table, string.as_ref().chars())
}

/// Convert this iterator of chars with this char swapping map
#[inline] pub fn conv_char_iter<T, I>(table: &smallmap::Map<char, char>, iter: I) -> CharSubstituteIter<I::IntoIter, T>
where I: IntoIterator<Item=T>,
      T: From<char> + smallmap::Collapse,
      char: Borrow<T>
{
    iter.replace_chars(table)
}

#[cfg(test)]
mod tests
{
    #[test]
    fn mod_base64_enc_dec()
    {
	let mut value = [0u8; 512];
	getrandom::getrandom(&mut value[..]).expect("setup failed");
	
	let md = super::ModifiedBase64String::encode(&value[..]);
	println!("e-md: {:?}", md);
	let ou = md.decode_new();
	assert_eq!(&ou[..], &value[..]);
    }
}