diff options
| author | mo khan <mo@mokhan.ca> | 2025-07-10 13:11:11 -0600 |
|---|---|---|
| committer | mo khan <mo@mokhan.ca> | 2025-07-10 13:11:11 -0600 |
| commit | 01959b16a21b22b5df5f16569c2a8e8f92beecef (patch) | |
| tree | 32afa5d747c5466345c59ec52161a7cba3d6d755 /vendor/tinystr/src | |
| parent | ff30574117a996df332e23d1fb6f65259b316b5b (diff) | |
chore: vendor dependencies
Diffstat (limited to 'vendor/tinystr/src')
| -rw-r--r-- | vendor/tinystr/src/ascii.rs | 1175 | ||||
| -rw-r--r-- | vendor/tinystr/src/asciibyte.rs | 145 | ||||
| -rw-r--r-- | vendor/tinystr/src/databake.rs | 75 | ||||
| -rw-r--r-- | vendor/tinystr/src/error.rs | 18 | ||||
| -rw-r--r-- | vendor/tinystr/src/int_ops.rs | 315 | ||||
| -rw-r--r-- | vendor/tinystr/src/lib.rs | 114 | ||||
| -rw-r--r-- | vendor/tinystr/src/macros.rs | 32 | ||||
| -rw-r--r-- | vendor/tinystr/src/serde.rs | 91 | ||||
| -rw-r--r-- | vendor/tinystr/src/ule.rs | 125 | ||||
| -rw-r--r-- | vendor/tinystr/src/unvalidated.rs | 122 |
10 files changed, 2212 insertions, 0 deletions
diff --git a/vendor/tinystr/src/ascii.rs b/vendor/tinystr/src/ascii.rs new file mode 100644 index 00000000..89ec742b --- /dev/null +++ b/vendor/tinystr/src/ascii.rs @@ -0,0 +1,1175 @@ +// This file is part of ICU4X. For terms of use, please see the file +// called LICENSE at the top level of the ICU4X source tree +// (online at: https://github.com/unicode-org/icu4x/blob/main/LICENSE ). + +use crate::asciibyte::AsciiByte; +use crate::int_ops::{Aligned4, Aligned8}; +use crate::ParseError; +use core::borrow::Borrow; +use core::fmt; +use core::ops::Deref; +use core::str::{self, FromStr}; + +#[repr(transparent)] +#[derive(PartialEq, Eq, Ord, PartialOrd, Copy, Clone, Hash)] +pub struct TinyAsciiStr<const N: usize> { + bytes: [AsciiByte; N], +} + +impl<const N: usize> TinyAsciiStr<N> { + #[inline] + pub const fn try_from_str(s: &str) -> Result<Self, ParseError> { + Self::try_from_utf8(s.as_bytes()) + } + + /// Creates a `TinyAsciiStr<N>` from the given UTF-8 slice. + /// `code_units` may contain at most `N` non-null ASCII code points. + #[inline] + pub const fn try_from_utf8(code_units: &[u8]) -> Result<Self, ParseError> { + Self::try_from_utf8_inner(code_units, false) + } + + /// Creates a `TinyAsciiStr<N>` from the given UTF-16 slice. + /// `code_units` may contain at most `N` non-null ASCII code points. + #[inline] + pub const fn try_from_utf16(code_units: &[u16]) -> Result<Self, ParseError> { + Self::try_from_utf16_inner(code_units, 0, code_units.len(), false) + } + + /// Creates a `TinyAsciiStr<N>` from a UTF-8 slice, replacing invalid code units. + /// + /// Invalid code units, as well as null or non-ASCII code points + /// (i.e. those outside the range U+0001..=U+007F`) + /// will be replaced with the replacement byte. + /// + /// The input slice will be truncated if its length exceeds `N`. + pub const fn from_utf8_lossy(code_units: &[u8], replacement: u8) -> Self { + let mut out = [0; N]; + let mut i = 0; + // Ord is not available in const, so no `.min(N)` + let len = if code_units.len() > N { + N + } else { + code_units.len() + }; + + // Indexing is protected by the len check above + #[allow(clippy::indexing_slicing)] + while i < len { + let b = code_units[i]; + if b > 0 && b < 0x80 { + out[i] = b; + } else { + out[i] = replacement; + } + i += 1; + } + + Self { + // SAFETY: `out` only contains ASCII bytes and has same size as `self.bytes` + bytes: unsafe { AsciiByte::to_ascii_byte_array(&out) }, + } + } + + /// Creates a `TinyAsciiStr<N>` from a UTF-16 slice, replacing invalid code units. + /// + /// Invalid code units, as well as null or non-ASCII code points + /// (i.e. those outside the range U+0001..=U+007F`) + /// will be replaced with the replacement byte. + /// + /// The input slice will be truncated if its length exceeds `N`. + pub const fn from_utf16_lossy(code_units: &[u16], replacement: u8) -> Self { + let mut out = [0; N]; + let mut i = 0; + // Ord is not available in const, so no `.min(N)` + let len = if code_units.len() > N { + N + } else { + code_units.len() + }; + + // Indexing is protected by the len check above + #[allow(clippy::indexing_slicing)] + while i < len { + let b = code_units[i]; + if b > 0 && b < 0x80 { + out[i] = b as u8; + } else { + out[i] = replacement; + } + i += 1; + } + + Self { + // SAFETY: `out` only contains ASCII bytes and has same size as `self.bytes` + bytes: unsafe { AsciiByte::to_ascii_byte_array(&out) }, + } + } + + /// Attempts to parse a fixed-length byte array to a `TinyAsciiStr`. + /// + /// The byte array may contain trailing NUL bytes. + /// + /// # Example + /// + /// ``` + /// use tinystr::tinystr; + /// use tinystr::TinyAsciiStr; + /// + /// assert_eq!( + /// TinyAsciiStr::<3>::try_from_raw(*b"GB\0"), + /// Ok(tinystr!(3, "GB")) + /// ); + /// assert_eq!( + /// TinyAsciiStr::<3>::try_from_raw(*b"USD"), + /// Ok(tinystr!(3, "USD")) + /// ); + /// assert!(matches!(TinyAsciiStr::<3>::try_from_raw(*b"\0A\0"), Err(_))); + /// ``` + pub const fn try_from_raw(raw: [u8; N]) -> Result<Self, ParseError> { + Self::try_from_utf8_inner(&raw, true) + } + + pub(crate) const fn try_from_utf8_inner( + code_units: &[u8], + allow_trailing_null: bool, + ) -> Result<Self, ParseError> { + if code_units.len() > N { + return Err(ParseError::TooLong { + max: N, + len: code_units.len(), + }); + } + + let mut out = [0; N]; + let mut i = 0; + let mut found_null = false; + // Indexing is protected by TinyStrError::TooLarge + #[allow(clippy::indexing_slicing)] + while i < code_units.len() { + let b = code_units[i]; + + if b == 0 { + found_null = true; + } else if b >= 0x80 { + return Err(ParseError::NonAscii); + } else if found_null { + // Error if there are contentful bytes after null + return Err(ParseError::ContainsNull); + } + out[i] = b; + + i += 1; + } + + if !allow_trailing_null && found_null { + // We found some trailing nulls, error + return Err(ParseError::ContainsNull); + } + + Ok(Self { + // SAFETY: `out` only contains ASCII bytes and has same size as `self.bytes` + bytes: unsafe { AsciiByte::to_ascii_byte_array(&out) }, + }) + } + + pub(crate) const fn try_from_utf16_inner( + code_units: &[u16], + start: usize, + end: usize, + allow_trailing_null: bool, + ) -> Result<Self, ParseError> { + let len = end - start; + if len > N { + return Err(ParseError::TooLong { max: N, len }); + } + + let mut out = [0; N]; + let mut i = 0; + let mut found_null = false; + // Indexing is protected by TinyStrError::TooLarge + #[allow(clippy::indexing_slicing)] + while i < len { + let b = code_units[start + i]; + + if b == 0 { + found_null = true; + } else if b >= 0x80 { + return Err(ParseError::NonAscii); + } else if found_null { + // Error if there are contentful bytes after null + return Err(ParseError::ContainsNull); + } + out[i] = b as u8; + + i += 1; + } + + if !allow_trailing_null && found_null { + // We found some trailing nulls, error + return Err(ParseError::ContainsNull); + } + + Ok(Self { + // SAFETY: `out` only contains ASCII bytes and has same size as `self.bytes` + bytes: unsafe { AsciiByte::to_ascii_byte_array(&out) }, + }) + } + + #[inline] + pub const fn as_str(&self) -> &str { + // as_utf8 is valid utf8 + unsafe { str::from_utf8_unchecked(self.as_utf8()) } + } + + #[inline] + #[must_use] + pub const fn len(&self) -> usize { + if N <= 4 { + Aligned4::from_ascii_bytes(&self.bytes).len() + } else if N <= 8 { + Aligned8::from_ascii_bytes(&self.bytes).len() + } else { + let mut i = 0; + #[allow(clippy::indexing_slicing)] // < N is safe + while i < N && self.bytes[i] as u8 != AsciiByte::B0 as u8 { + i += 1 + } + i + } + } + + #[inline] + #[must_use] + pub const fn is_empty(&self) -> bool { + self.bytes[0] as u8 == AsciiByte::B0 as u8 + } + + #[inline] + #[must_use] + pub const fn as_utf8(&self) -> &[u8] { + // Safe because `self.bytes.as_slice()` pointer-casts to `&[u8]`, + // and changing the length of that slice to self.len() < N is safe. + unsafe { + core::slice::from_raw_parts(self.bytes.as_slice().as_ptr() as *const u8, self.len()) + } + } + + #[inline] + #[must_use] + pub const fn all_bytes(&self) -> &[u8; N] { + // SAFETY: `self.bytes` has same size as [u8; N] + unsafe { &*(self.bytes.as_ptr() as *const [u8; N]) } + } + + #[inline] + #[must_use] + /// Resizes a `TinyAsciiStr<N>` to a `TinyAsciiStr<M>`. + /// + /// If `M < len()` the string gets truncated, otherwise only the + /// memory representation changes. + pub const fn resize<const M: usize>(self) -> TinyAsciiStr<M> { + let mut bytes = [0; M]; + let mut i = 0; + // Indexing is protected by the loop guard + #[allow(clippy::indexing_slicing)] + while i < M && i < N { + bytes[i] = self.bytes[i] as u8; + i += 1; + } + // `self.bytes` only contains ASCII bytes, with no null bytes between + // ASCII characters, so this also holds for `bytes`. + unsafe { TinyAsciiStr::from_utf8_unchecked(bytes) } + } + + #[inline] + #[must_use] + /// Returns a `TinyAsciiStr<Q>` with the concatenation of this string, + /// `TinyAsciiStr<N>`, and another string, `TinyAsciiStr<M>`. + /// + /// If `Q < N + M`, the string gets truncated. + /// + /// # Examples + /// + /// ``` + /// use tinystr::tinystr; + /// use tinystr::TinyAsciiStr; + /// + /// let abc = tinystr!(6, "abc"); + /// let defg = tinystr!(6, "defg"); + /// + /// // The concatenation is successful if Q is large enough... + /// assert_eq!(abc.concat(defg), tinystr!(16, "abcdefg")); + /// assert_eq!(abc.concat(defg), tinystr!(12, "abcdefg")); + /// assert_eq!(abc.concat(defg), tinystr!(8, "abcdefg")); + /// assert_eq!(abc.concat(defg), tinystr!(7, "abcdefg")); + /// + /// /// ...but it truncates of Q is too small. + /// assert_eq!(abc.concat(defg), tinystr!(6, "abcdef")); + /// assert_eq!(abc.concat(defg), tinystr!(2, "ab")); + /// ``` + pub const fn concat<const M: usize, const Q: usize>( + self, + other: TinyAsciiStr<M>, + ) -> TinyAsciiStr<Q> { + let mut result = self.resize::<Q>(); + let mut i = self.len(); + let mut j = 0; + // Indexing is protected by the loop guard + #[allow(clippy::indexing_slicing)] + while i < Q && j < M { + result.bytes[i] = other.bytes[j]; + i += 1; + j += 1; + } + result + } + + /// # Safety + /// Must be called with a bytes array made of valid ASCII bytes, with no null bytes + /// between ASCII characters + #[must_use] + pub const unsafe fn from_utf8_unchecked(code_units: [u8; N]) -> Self { + Self { + bytes: AsciiByte::to_ascii_byte_array(&code_units), + } + } +} + +macro_rules! check_is { + ($self:ident, $check_int:ident, $check_u8:ident) => { + if N <= 4 { + Aligned4::from_ascii_bytes(&$self.bytes).$check_int() + } else if N <= 8 { + Aligned8::from_ascii_bytes(&$self.bytes).$check_int() + } else { + let mut i = 0; + // Won't panic because self.bytes has length N + #[allow(clippy::indexing_slicing)] + while i < N && $self.bytes[i] as u8 != AsciiByte::B0 as u8 { + if !($self.bytes[i] as u8).$check_u8() { + return false; + } + i += 1; + } + true + } + }; + ($self:ident, $check_int:ident, !$check_u8_0_inv:ident, !$check_u8_1_inv:ident) => { + if N <= 4 { + Aligned4::from_ascii_bytes(&$self.bytes).$check_int() + } else if N <= 8 { + Aligned8::from_ascii_bytes(&$self.bytes).$check_int() + } else { + // Won't panic because N is > 8 + if ($self.bytes[0] as u8).$check_u8_0_inv() { + return false; + } + let mut i = 1; + // Won't panic because self.bytes has length N + #[allow(clippy::indexing_slicing)] + while i < N && $self.bytes[i] as u8 != AsciiByte::B0 as u8 { + if ($self.bytes[i] as u8).$check_u8_1_inv() { + return false; + } + i += 1; + } + true + } + }; + ($self:ident, $check_int:ident, $check_u8_0_inv:ident, $check_u8_1_inv:ident) => { + if N <= 4 { + Aligned4::from_ascii_bytes(&$self.bytes).$check_int() + } else if N <= 8 { + Aligned8::from_ascii_bytes(&$self.bytes).$check_int() + } else { + // Won't panic because N is > 8 + if !($self.bytes[0] as u8).$check_u8_0_inv() { + return false; + } + let mut i = 1; + // Won't panic because self.bytes has length N + #[allow(clippy::indexing_slicing)] + while i < N && $self.bytes[i] as u8 != AsciiByte::B0 as u8 { + if !($self.bytes[i] as u8).$check_u8_1_inv() { + return false; + } + i += 1; + } + true + } + }; +} + +impl<const N: usize> TinyAsciiStr<N> { + /// Checks if the value is composed of ASCII alphabetic characters: + /// + /// * U+0041 'A' ..= U+005A 'Z', or + /// * U+0061 'a' ..= U+007A 'z'. + /// + /// # Examples + /// + /// ``` + /// use tinystr::TinyAsciiStr; + /// + /// let s1: TinyAsciiStr<4> = "Test".parse().expect("Failed to parse."); + /// let s2: TinyAsciiStr<4> = "Te3t".parse().expect("Failed to parse."); + /// + /// assert!(s1.is_ascii_alphabetic()); + /// assert!(!s2.is_ascii_alphabetic()); + /// ``` + #[inline] + #[must_use] + pub const fn is_ascii_alphabetic(&self) -> bool { + check_is!(self, is_ascii_alphabetic, is_ascii_alphabetic) + } + + /// Checks if the value is composed of ASCII alphanumeric characters: + /// + /// * U+0041 'A' ..= U+005A 'Z', or + /// * U+0061 'a' ..= U+007A 'z', or + /// * U+0030 '0' ..= U+0039 '9'. + /// + /// # Examples + /// + /// ``` + /// use tinystr::TinyAsciiStr; + /// + /// let s1: TinyAsciiStr<4> = "A15b".parse().expect("Failed to parse."); + /// let s2: TinyAsciiStr<4> = "[3@w".parse().expect("Failed to parse."); + /// + /// assert!(s1.is_ascii_alphanumeric()); + /// assert!(!s2.is_ascii_alphanumeric()); + /// ``` + #[inline] + #[must_use] + pub const fn is_ascii_alphanumeric(&self) -> bool { + check_is!(self, is_ascii_alphanumeric, is_ascii_alphanumeric) + } + + /// Checks if the value is composed of ASCII decimal digits: + /// + /// * U+0030 '0' ..= U+0039 '9'. + /// + /// # Examples + /// + /// ``` + /// use tinystr::TinyAsciiStr; + /// + /// let s1: TinyAsciiStr<4> = "312".parse().expect("Failed to parse."); + /// let s2: TinyAsciiStr<4> = "3d".parse().expect("Failed to parse."); + /// + /// assert!(s1.is_ascii_numeric()); + /// assert!(!s2.is_ascii_numeric()); + /// ``` + #[inline] + #[must_use] + pub const fn is_ascii_numeric(&self) -> bool { + check_is!(self, is_ascii_numeric, is_ascii_digit) + } + + /// Checks if the value is in ASCII lower case. + /// + /// All letter characters are checked for case. Non-letter characters are ignored. + /// + /// # Examples + /// + /// ``` + /// use tinystr::TinyAsciiStr; + /// + /// let s1: TinyAsciiStr<4> = "teSt".parse().expect("Failed to parse."); + /// let s2: TinyAsciiStr<4> = "test".parse().expect("Failed to parse."); + /// let s3: TinyAsciiStr<4> = "001z".parse().expect("Failed to parse."); + /// + /// assert!(!s1.is_ascii_lowercase()); + /// assert!(s2.is_ascii_lowercase()); + /// assert!(s3.is_ascii_lowercase()); + /// ``` + #[inline] + #[must_use] + pub const fn is_ascii_lowercase(&self) -> bool { + check_is!( + self, + is_ascii_lowercase, + !is_ascii_uppercase, + !is_ascii_uppercase + ) + } + + /// Checks if the value is in ASCII title case. + /// + /// This verifies that the first character is ASCII uppercase and all others ASCII lowercase. + /// Non-letter characters are ignored. + /// + /// # Examples + /// + /// ``` + /// use tinystr::TinyAsciiStr; + /// + /// let s1: TinyAsciiStr<4> = "teSt".parse().expect("Failed to parse."); + /// let s2: TinyAsciiStr<4> = "Test".parse().expect("Failed to parse."); + /// let s3: TinyAsciiStr<4> = "001z".parse().expect("Failed to parse."); + /// + /// assert!(!s1.is_ascii_titlecase()); + /// assert!(s2.is_ascii_titlecase()); + /// assert!(s3.is_ascii_titlecase()); + /// ``` + #[inline] + #[must_use] + pub const fn is_ascii_titlecase(&self) -> bool { + check_is!( + self, + is_ascii_titlecase, + !is_ascii_lowercase, + !is_ascii_uppercase + ) + } + + /// Checks if the value is in ASCII upper case. + /// + /// All letter characters are checked for case. Non-letter characters are ignored. + /// + /// # Examples + /// + /// ``` + /// use tinystr::TinyAsciiStr; + /// + /// let s1: TinyAsciiStr<4> = "teSt".parse().expect("Failed to parse."); + /// let s2: TinyAsciiStr<4> = "TEST".parse().expect("Failed to parse."); + /// let s3: TinyAsciiStr<4> = "001z".parse().expect("Failed to parse."); + /// + /// assert!(!s1.is_ascii_uppercase()); + /// assert!(s2.is_ascii_uppercase()); + /// assert!(!s3.is_ascii_uppercase()); + /// ``` + #[inline] + #[must_use] + pub const fn is_ascii_uppercase(&self) -> bool { + check_is!( + self, + is_ascii_uppercase, + !is_ascii_lowercase, + !is_ascii_lowercase + ) + } + + /// Checks if the value is composed of ASCII alphabetic lower case characters: + /// + /// * U+0061 'a' ..= U+007A 'z', + /// + /// # Examples + /// + /// ``` + /// use tinystr::TinyAsciiStr; + /// + /// let s1: TinyAsciiStr<4> = "Test".parse().expect("Failed to parse."); + /// let s2: TinyAsciiStr<4> = "Te3t".parse().expect("Failed to parse."); + /// let s3: TinyAsciiStr<4> = "teSt".parse().expect("Failed to parse."); + /// let s4: TinyAsciiStr<4> = "test".parse().expect("Failed to parse."); + /// let s5: TinyAsciiStr<4> = "001z".parse().expect("Failed to parse."); + /// + /// assert!(!s1.is_ascii_alphabetic_lowercase()); + /// assert!(!s2.is_ascii_alphabetic_lowercase()); + /// assert!(!s3.is_ascii_alphabetic_lowercase()); + /// assert!(s4.is_ascii_alphabetic_lowercase()); + /// assert!(!s5.is_ascii_alphabetic_lowercase()); + /// ``` + #[inline] + #[must_use] + pub const fn is_ascii_alphabetic_lowercase(&self) -> bool { + check_is!( + self, + is_ascii_alphabetic_lowercase, + is_ascii_lowercase, + is_ascii_lowercase + ) + } + + /// Checks if the value is composed of ASCII alphabetic, with the first character being ASCII uppercase, and all others ASCII lowercase. + /// + /// # Examples + /// + /// ``` + /// use tinystr::TinyAsciiStr; + /// + /// let s1: TinyAsciiStr<4> = "Test".parse().expect("Failed to parse."); + /// let s2: TinyAsciiStr<4> = "Te3t".parse().expect("Failed to parse."); + /// let s3: TinyAsciiStr<4> = "teSt".parse().expect("Failed to parse."); + /// let s4: TinyAsciiStr<4> = "test".parse().expect("Failed to parse."); + /// let s5: TinyAsciiStr<4> = "001z".parse().expect("Failed to parse."); + /// + /// assert!(s1.is_ascii_alphabetic_titlecase()); + /// assert!(!s2.is_ascii_alphabetic_titlecase()); + /// assert!(!s3.is_ascii_alphabetic_titlecase()); + /// assert!(!s4.is_ascii_alphabetic_titlecase()); + /// assert!(!s5.is_ascii_alphabetic_titlecase()); + /// ``` + #[inline] + #[must_use] + pub const fn is_ascii_alphabetic_titlecase(&self) -> bool { + check_is!( + self, + is_ascii_alphabetic_titlecase, + is_ascii_uppercase, + is_ascii_lowercase + ) + } + + /// Checks if the value is composed of ASCII alphabetic upper case characters: + /// + /// * U+0041 'A' ..= U+005A 'Z', + /// + /// # Examples + /// + /// ``` + /// use tinystr::TinyAsciiStr; + /// + /// let s1: TinyAsciiStr<4> = "Test".parse().expect("Failed to parse."); + /// let s2: TinyAsciiStr<4> = "Te3t".parse().expect("Failed to parse."); + /// let s3: TinyAsciiStr<4> = "teSt".parse().expect("Failed to parse."); + /// let s4: TinyAsciiStr<4> = "TEST".parse().expect("Failed to parse."); + /// let s5: TinyAsciiStr<4> = "001z".parse().expect("Failed to parse."); + /// + /// assert!(!s1.is_ascii_alphabetic_uppercase()); + /// assert!(!s2.is_ascii_alphabetic_uppercase()); + /// assert!(!s3.is_ascii_alphabetic_uppercase()); + /// assert!(s4.is_ascii_alphabetic_uppercase()); + /// assert!(!s5.is_ascii_alphabetic_uppercase()); + /// ``` + #[inline] + #[must_use] + pub const fn is_ascii_alphabetic_uppercase(&self) -> bool { + check_is!( + self, + is_ascii_alphabetic_uppercase, + is_ascii_uppercase, + is_ascii_uppercase + ) + } +} + +macro_rules! to { + ($self:ident, $to:ident, $later_char_to:ident $(,$first_char_to:ident)?) => {{ + let mut i = 0; + if N <= 4 { + let aligned = Aligned4::from_ascii_bytes(&$self.bytes).$to().to_ascii_bytes(); + // Won't panic because self.bytes has length N and aligned has length >= N + #[allow(clippy::indexing_slicing)] + while i < N { + $self.bytes[i] = aligned[i]; + i += 1; + } + } else if N <= 8 { + let aligned = Aligned8::from_ascii_bytes(&$self.bytes).$to().to_ascii_bytes(); + // Won't panic because self.bytes has length N and aligned has length >= N + #[allow(clippy::indexing_slicing)] + while i < N { + $self.bytes[i] = aligned[i]; + i += 1; + } + } else { + // Won't panic because self.bytes has length N + #[allow(clippy::indexing_slicing)] + while i < N && $self.bytes[i] as u8 != AsciiByte::B0 as u8 { + // SAFETY: AsciiByte is repr(u8) and has same size as u8 + unsafe { + $self.bytes[i] = core::mem::transmute::<u8, AsciiByte>( + ($self.bytes[i] as u8).$later_char_to() + ); + } + i += 1; + } + // SAFETY: AsciiByte is repr(u8) and has same size as u8 + $( + $self.bytes[0] = unsafe { + core::mem::transmute::<u8, AsciiByte>(($self.bytes[0] as u8).$first_char_to()) + }; + )? + } + $self + }}; +} + +impl<const N: usize> TinyAsciiStr<N> { + /// Converts this type to its ASCII lower case equivalent in-place. + /// + /// ASCII letters 'A' to 'Z' are mapped to 'a' to 'z', other characters are unchanged. + /// + /// # Examples + /// + /// ``` + /// use tinystr::TinyAsciiStr; + /// + /// let s1: TinyAsciiStr<4> = "TeS3".parse().expect("Failed to parse."); + /// + /// assert_eq!(&*s1.to_ascii_lowercase(), "tes3"); + /// ``` + #[inline] + #[must_use] + pub const fn to_ascii_lowercase(mut self) -> Self { + to!(self, to_ascii_lowercase, to_ascii_lowercase) + } + + /// Converts this type to its ASCII title case equivalent in-place. + /// + /// The first character is converted to ASCII uppercase; the remaining characters + /// are converted to ASCII lowercase. + /// + /// # Examples + /// + /// ``` + /// use tinystr::TinyAsciiStr; + /// + /// let s1: TinyAsciiStr<4> = "teSt".parse().expect("Failed to parse."); + /// + /// assert_eq!(&*s1.to_ascii_titlecase(), "Test"); + /// ``` + #[inline] + #[must_use] + pub const fn to_ascii_titlecase(mut self) -> Self { + to!( + self, + to_ascii_titlecase, + to_ascii_lowercase, + to_ascii_uppercase + ) + } + + /// Converts this type to its ASCII upper case equivalent in-place. + /// + /// ASCII letters 'a' to 'z' are mapped to 'A' to 'Z', other characters are unchanged. + /// + /// # Examples + /// + /// ``` + /// use tinystr::TinyAsciiStr; + /// + /// let s1: TinyAsciiStr<4> = "Tes3".parse().expect("Failed to parse."); + /// + /// assert_eq!(&*s1.to_ascii_uppercase(), "TES3"); + /// ``` + #[inline] + #[must_use] + pub const fn to_ascii_uppercase(mut self) -> Self { + to!(self, to_ascii_uppercase, to_ascii_uppercase) + } +} + +impl<const N: usize> fmt::Debug for TinyAsciiStr<N> { + fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result { + fmt::Debug::fmt(self.as_str(), f) + } +} + +impl<const N: usize> fmt::Display for TinyAsciiStr<N> { + fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result { + fmt::Display::fmt(self.as_str(), f) + } +} + +impl<const N: usize> Deref for TinyAsciiStr<N> { + type Target = str; + #[inline] + fn deref(&self) -> &str { + self.as_str() + } +} + +impl<const N: usize> Borrow<str> for TinyAsciiStr<N> { + #[inline] + fn borrow(&self) -> &str { + self.as_str() + } +} + +impl<const N: usize> FromStr for TinyAsciiStr<N> { + type Err = ParseError; + #[inline] + fn from_str(s: &str) -> Result<Self, Self::Err> { + Self::try_from_str(s) + } +} + +impl<const N: usize> PartialEq<str> for TinyAsciiStr<N> { + fn eq(&self, other: &str) -> bool { + self.deref() == other + } +} + +impl<const N: usize> PartialEq<&str> for TinyAsciiStr<N> { + fn eq(&self, other: &&str) -> bool { + self.deref() == *other + } +} + +#[cfg(feature = "alloc")] +impl<const N: usize> PartialEq<alloc::string::String> for TinyAsciiStr<N> { + fn eq(&self, other: &alloc::string::String) -> bool { + self.deref() == other.deref() + } +} + +#[cfg(feature = "alloc")] +impl<const N: usize> PartialEq<TinyAsciiStr<N>> for alloc::string::String { + fn eq(&self, other: &TinyAsciiStr<N>) -> bool { + self.deref() == other.deref() + } +} + +#[cfg(test)] +mod test { + use super::*; + use rand::distributions::Distribution; + use rand::distributions::Standard; + use rand::rngs::SmallRng; + use rand::seq::SliceRandom; + use rand::SeedableRng; + + const STRINGS: [&str; 26] = [ + "Latn", + "laTn", + "windows", + "AR", + "Hans", + "macos", + "AT", + "infiniband", + "FR", + "en", + "Cyrl", + "FromIntegral", + "NO", + "419", + "MacintoshOSX2019", + "a3z", + "A3z", + "A3Z", + "a3Z", + "3A", + "3Z", + "3a", + "3z", + "@@[`{", + "UK", + "E12", + ]; + + fn gen_strings(num_strings: usize, allowed_lengths: &[usize]) -> Vec<String> { + let mut rng = SmallRng::seed_from_u64(2022); + // Need to do this in 2 steps since the RNG is needed twice + let string_lengths = core::iter::repeat_with(|| *allowed_lengths.choose(&mut rng).unwrap()) + .take(num_strings) + .collect::<Vec<usize>>(); + string_lengths + .iter() + .map(|len| { + Standard + .sample_iter(&mut rng) + .filter(|b: &u8| *b > 0 && *b < 0x80) + .take(*len) + .collect::<Vec<u8>>() + }) + .map(|byte_vec| String::from_utf8(byte_vec).expect("All ASCII")) + .collect() + } + + fn check_operation<T, F1, F2, const N: usize>(reference_f: F1, tinystr_f: F2) + where + F1: Fn(&str) -> T, + F2: Fn(TinyAsciiStr<N>) -> T, + T: core::fmt::Debug + core::cmp::PartialEq, + { + for s in STRINGS + .into_iter() + .map(str::to_owned) + .chain(gen_strings(100, &[3, 4, 5, 8, 12])) + { + let t = match TinyAsciiStr::<N>::from_str(&s) { + Ok(t) => t, + Err(ParseError::TooLong { .. }) => continue, + Err(e) => panic!("{}", e), + }; + let expected = reference_f(&s); + let actual = tinystr_f(t); + assert_eq!(expected, actual, "TinyAsciiStr<{N}>: {s:?}"); + + let s_utf16: Vec<u16> = s.encode_utf16().collect(); + let t = match TinyAsciiStr::<N>::try_from_utf16(&s_utf16) { + Ok(t) => t, + Err(ParseError::TooLong { .. }) => continue, + Err(e) => panic!("{}", e), + }; + let expected = reference_f(&s); + let actual = tinystr_f(t); + assert_eq!(expected, actual, "TinyAsciiStr<{N}>: {s:?}"); + } + } + + #[test] + fn test_is_ascii_alphabetic() { + fn check<const N: usize>() { + check_operation( + |s| s.chars().all(|c| c.is_ascii_alphabetic()), + |t: TinyAsciiStr<N>| TinyAsciiStr::is_ascii_alphabetic(&t), + ) + } + check::<2>(); + check::<3>(); + check::<4>(); + check::<5>(); + check::<8>(); + check::<16>(); + } + + #[test] + fn test_is_ascii_alphanumeric() { + fn check<const N: usize>() { + check_operation( + |s| s.chars().all(|c| c.is_ascii_alphanumeric()), + |t: TinyAsciiStr<N>| TinyAsciiStr::is_ascii_alphanumeric(&t), + ) + } + check::<2>(); + check::<3>(); + check::<4>(); + check::<5>(); + check::<8>(); + check::<16>(); + } + + #[test] + fn test_is_ascii_numeric() { + fn check<const N: usize>() { + check_operation( + |s| s.chars().all(|c| c.is_ascii_digit()), + |t: TinyAsciiStr<N>| TinyAsciiStr::is_ascii_numeric(&t), + ) + } + check::<2>(); + check::<3>(); + check::<4>(); + check::<5>(); + check::<8>(); + check::<16>(); + } + + #[test] + fn test_is_ascii_lowercase() { + fn check<const N: usize>() { + check_operation( + |s| { + s == TinyAsciiStr::<16>::try_from_str(s) + .unwrap() + .to_ascii_lowercase() + .as_str() + }, + |t: TinyAsciiStr<N>| TinyAsciiStr::is_ascii_lowercase(&t), + ) + } + check::<2>(); + check::<3>(); + check::<4>(); + check::<5>(); + check::<8>(); + check::<16>(); + } + + #[test] + fn test_is_ascii_titlecase() { + fn check<const N: usize>() { + check_operation( + |s| { + s == TinyAsciiStr::<16>::try_from_str(s) + .unwrap() + .to_ascii_titlecase() + .as_str() + }, + |t: TinyAsciiStr<N>| TinyAsciiStr::is_ascii_titlecase(&t), + ) + } + check::<2>(); + check::<3>(); + check::<4>(); + check::<5>(); + check::<8>(); + check::<16>(); + } + + #[test] + fn test_is_ascii_uppercase() { + fn check<const N: usize>() { + check_operation( + |s| { + s == TinyAsciiStr::<16>::try_from_str(s) + .unwrap() + .to_ascii_uppercase() + .as_str() + }, + |t: TinyAsciiStr<N>| TinyAsciiStr::is_ascii_uppercase(&t), + ) + } + check::<2>(); + check::<3>(); + check::<4>(); + check::<5>(); + check::<8>(); + check::<16>(); + } + + #[test] + fn test_is_ascii_alphabetic_lowercase() { + fn check<const N: usize>() { + check_operation( + |s| { + // Check alphabetic + s.chars().all(|c| c.is_ascii_alphabetic()) && + // Check lowercase + s == TinyAsciiStr::<16>::try_from_str(s) + .unwrap() + .to_ascii_lowercase() + .as_str() + }, + |t: TinyAsciiStr<N>| TinyAsciiStr::is_ascii_alphabetic_lowercase(&t), + ) + } + check::<2>(); + check::<3>(); + check::<4>(); + check::<5>(); + check::<8>(); + check::<16>(); + } + + #[test] + fn test_is_ascii_alphabetic_titlecase() { + fn check<const N: usize>() { + check_operation( + |s| { + // Check alphabetic + s.chars().all(|c| c.is_ascii_alphabetic()) && + // Check titlecase + s == TinyAsciiStr::<16>::try_from_str(s) + .unwrap() + .to_ascii_titlecase() + .as_str() + }, + |t: TinyAsciiStr<N>| TinyAsciiStr::is_ascii_alphabetic_titlecase(&t), + ) + } + check::<2>(); + check::<3>(); + check::<4>(); + check::<5>(); + check::<8>(); + check::<16>(); + } + + #[test] + fn test_is_ascii_alphabetic_uppercase() { + fn check<const N: usize>() { + check_operation( + |s| { + // Check alphabetic + s.chars().all(|c| c.is_ascii_alphabetic()) && + // Check uppercase + s == TinyAsciiStr::<16>::try_from_str(s) + .unwrap() + .to_ascii_uppercase() + .as_str() + }, + |t: TinyAsciiStr<N>| TinyAsciiStr::is_ascii_alphabetic_uppercase(&t), + ) + } + check::<2>(); + check::<3>(); + check::<4>(); + check::<5>(); + check::<8>(); + check::<16>(); + } + + #[test] + fn test_to_ascii_lowercase() { + fn check<const N: usize>() { + check_operation( + |s| { + s.chars() + .map(|c| c.to_ascii_lowercase()) + .collect::<String>() + }, + |t: TinyAsciiStr<N>| TinyAsciiStr::to_ascii_lowercase(t).as_str().to_owned(), + ) + } + check::<2>(); + check::<3>(); + check::<4>(); + check::<5>(); + check::<8>(); + check::<16>(); + } + + #[test] + fn test_to_ascii_titlecase() { + fn check<const N: usize>() { + check_operation( + |s| { + let mut r = s + .chars() + .map(|c| c.to_ascii_lowercase()) + .collect::<String>(); + // Safe because the string is nonempty and an ASCII string + unsafe { r.as_bytes_mut()[0].make_ascii_uppercase() }; + r + }, + |t: TinyAsciiStr<N>| TinyAsciiStr::to_ascii_titlecase(t).as_str().to_owned(), + ) + } + check::<2>(); + check::<3>(); + check::<4>(); + check::<5>(); + check::<8>(); + check::<16>(); + } + + #[test] + fn test_to_ascii_uppercase() { + fn check<const N: usize>() { + check_operation( + |s| { + s.chars() + .map(|c| c.to_ascii_uppercase()) + .collect::<String>() + }, + |t: TinyAsciiStr<N>| TinyAsciiStr::to_ascii_uppercase(t).as_str().to_owned(), + ) + } + check::<2>(); + check::<3>(); + check::<4>(); + check::<5>(); + check::<8>(); + check::<16>(); + } + + #[test] + fn lossy_constructor() { + assert_eq!(TinyAsciiStr::<4>::from_utf8_lossy(b"", b'?').as_str(), ""); + assert_eq!( + TinyAsciiStr::<4>::from_utf8_lossy(b"oh\0o", b'?').as_str(), + "oh?o" + ); + assert_eq!( + TinyAsciiStr::<4>::from_utf8_lossy(b"\0", b'?').as_str(), + "?" + ); + assert_eq!( + TinyAsciiStr::<4>::from_utf8_lossy(b"toolong", b'?').as_str(), + "tool" + ); + assert_eq!( + TinyAsciiStr::<4>::from_utf8_lossy(&[b'a', 0x80, 0xFF, b'1'], b'?').as_str(), + "a??1" + ); + } +} diff --git a/vendor/tinystr/src/asciibyte.rs b/vendor/tinystr/src/asciibyte.rs new file mode 100644 index 00000000..f41a0334 --- /dev/null +++ b/vendor/tinystr/src/asciibyte.rs @@ -0,0 +1,145 @@ +// This file is part of ICU4X. For terms of use, please see the file +// called LICENSE at the top level of the ICU4X source tree +// (online at: https://github.com/unicode-org/icu4x/blob/main/LICENSE ). + +#[repr(u8)] +#[allow(dead_code)] +#[derive(PartialEq, Eq, Ord, PartialOrd, Copy, Clone, Hash)] +pub enum AsciiByte { + B0 = 0, + B1 = 1, + B2 = 2, + B3 = 3, + B4 = 4, + B5 = 5, + B6 = 6, + B7 = 7, + B8 = 8, + B9 = 9, + B10 = 10, + B11 = 11, + B12 = 12, + B13 = 13, + B14 = 14, + B15 = 15, + B16 = 16, + B17 = 17, + B18 = 18, + B19 = 19, + B20 = 20, + B21 = 21, + B22 = 22, + B23 = 23, + B24 = 24, + B25 = 25, + B26 = 26, + B27 = 27, + B28 = 28, + B29 = 29, + B30 = 30, + B31 = 31, + B32 = 32, + B33 = 33, + B34 = 34, + B35 = 35, + B36 = 36, + B37 = 37, + B38 = 38, + B39 = 39, + B40 = 40, + B41 = 41, + B42 = 42, + B43 = 43, + B44 = 44, + B45 = 45, + B46 = 46, + B47 = 47, + B48 = 48, + B49 = 49, + B50 = 50, + B51 = 51, + B52 = 52, + B53 = 53, + B54 = 54, + B55 = 55, + B56 = 56, + B57 = 57, + B58 = 58, + B59 = 59, + B60 = 60, + B61 = 61, + B62 = 62, + B63 = 63, + B64 = 64, + B65 = 65, + B66 = 66, + B67 = 67, + B68 = 68, + B69 = 69, + B70 = 70, + B71 = 71, + B72 = 72, + B73 = 73, + B74 = 74, + B75 = 75, + B76 = 76, + B77 = 77, + B78 = 78, + B79 = 79, + B80 = 80, + B81 = 81, + B82 = 82, + B83 = 83, + B84 = 84, + B85 = 85, + B86 = 86, + B87 = 87, + B88 = 88, + B89 = 89, + B90 = 90, + B91 = 91, + B92 = 92, + B93 = 93, + B94 = 94, + B95 = 95, + B96 = 96, + B97 = 97, + B98 = 98, + B99 = 99, + B100 = 100, + B101 = 101, + B102 = 102, + B103 = 103, + B104 = 104, + B105 = 105, + B106 = 106, + B107 = 107, + B108 = 108, + B109 = 109, + B110 = 110, + B111 = 111, + B112 = 112, + B113 = 113, + B114 = 114, + B115 = 115, + B116 = 116, + B117 = 117, + B118 = 118, + B119 = 119, + B120 = 120, + B121 = 121, + B122 = 122, + B123 = 123, + B124 = 124, + B125 = 125, + B126 = 126, + B127 = 127, +} + +impl AsciiByte { + // Convert [u8; N] to [AsciiByte; N] + #[inline] + pub const unsafe fn to_ascii_byte_array<const N: usize>(bytes: &[u8; N]) -> [AsciiByte; N] { + *(bytes as *const [u8; N] as *const [AsciiByte; N]) + } +} diff --git a/vendor/tinystr/src/databake.rs b/vendor/tinystr/src/databake.rs new file mode 100644 index 00000000..748d23ac --- /dev/null +++ b/vendor/tinystr/src/databake.rs @@ -0,0 +1,75 @@ +// This file is part of ICU4X. For terms of use, please see the file +// called LICENSE at the top level of the ICU4X source tree +// (online at: https://github.com/unicode-org/icu4x/blob/main/LICENSE ). + +use crate::TinyAsciiStr; +use crate::UnvalidatedTinyAsciiStr; +use databake::*; + +impl<const N: usize> Bake for TinyAsciiStr<N> { + fn bake(&self, env: &CrateEnv) -> TokenStream { + env.insert("tinystr"); + let string = self.as_str(); + quote! { + tinystr::tinystr!(#N, #string) + } + } +} + +impl<const N: usize> BakeSize for TinyAsciiStr<N> { + fn borrows_size(&self) -> usize { + 0 + } +} + +impl<const N: usize> databake::Bake for UnvalidatedTinyAsciiStr<N> { + fn bake(&self, env: &databake::CrateEnv) -> databake::TokenStream { + match self.try_into_tinystr() { + Ok(tiny) => { + let tiny = tiny.bake(env); + databake::quote! { + #tiny.to_unvalidated() + } + } + Err(_) => { + let bytes = self.0.bake(env); + env.insert("tinystr"); + databake::quote! { + tinystr::UnvalidatedTinyAsciiStr::from_utf8_unchecked(#bytes) + } + } + } + } +} + +impl<const N: usize> databake::BakeSize for UnvalidatedTinyAsciiStr<N> { + fn borrows_size(&self) -> usize { + 0 + } +} + +#[test] +fn test() { + test_bake!( + TinyAsciiStr<10>, + const, + crate::tinystr!(10usize, "foo"), + tinystr + ); +} + +#[test] +fn test_unvalidated() { + test_bake!( + UnvalidatedTinyAsciiStr<10>, + const, + crate::tinystr!(10usize, "foo").to_unvalidated(), + tinystr + ); + test_bake!( + UnvalidatedTinyAsciiStr<3>, + const, + crate::UnvalidatedTinyAsciiStr::from_utf8_unchecked(*b"AB\xCD"), + tinystr + ); +} diff --git a/vendor/tinystr/src/error.rs b/vendor/tinystr/src/error.rs new file mode 100644 index 00000000..ea1ab212 --- /dev/null +++ b/vendor/tinystr/src/error.rs @@ -0,0 +1,18 @@ +// This file is part of ICU4X. For terms of use, please see the file +// called LICENSE at the top level of the ICU4X source tree +// (online at: https://github.com/unicode-org/icu4x/blob/main/LICENSE ). + +use displaydoc::Display; + +impl core::error::Error for ParseError {} + +#[derive(Display, Debug, PartialEq, Eq)] +#[non_exhaustive] +pub enum ParseError { + #[displaydoc("found string of larger length {len} when constructing string of length {max}")] + TooLong { max: usize, len: usize }, + #[displaydoc("tinystr types do not support strings with null bytes")] + ContainsNull, + #[displaydoc("attempted to construct TinyAsciiStr from a non-ASCII string")] + NonAscii, +} diff --git a/vendor/tinystr/src/int_ops.rs b/vendor/tinystr/src/int_ops.rs new file mode 100644 index 00000000..7bdb3633 --- /dev/null +++ b/vendor/tinystr/src/int_ops.rs @@ -0,0 +1,315 @@ +// This file is part of ICU4X. For terms of use, please see the file +// called LICENSE at the top level of the ICU4X source tree +// (online at: https://github.com/unicode-org/icu4x/blob/main/LICENSE ). + +use crate::asciibyte::AsciiByte; + +/// Internal helper struct that performs operations on aligned integers. +/// Supports strings up to 4 bytes long. +#[repr(transparent)] +pub struct Aligned4(u32); + +impl Aligned4 { + /// # Panics + /// Panics if N is greater than 4 + #[inline] + pub const fn from_utf8<const N: usize>(src: &[u8; N]) -> Self { + let mut bytes = [0; 4]; + let mut i = 0; + // The function documentation defines when panics may occur + #[allow(clippy::indexing_slicing)] + while i < N { + bytes[i] = src[i]; + i += 1; + } + Self(u32::from_ne_bytes(bytes)) + } + + #[inline] + pub const fn from_ascii_bytes<const N: usize>(src: &[AsciiByte; N]) -> Self { + Self::from_utf8::<N>(unsafe { core::mem::transmute::<&[AsciiByte; N], &[u8; N]>(src) }) + } + + #[inline] + pub const fn to_bytes(&self) -> [u8; 4] { + self.0.to_ne_bytes() + } + + #[inline] + pub const fn to_ascii_bytes(&self) -> [AsciiByte; 4] { + unsafe { core::mem::transmute(self.to_bytes()) } + } + + pub const fn len(&self) -> usize { + let word = self.0; + #[cfg(target_endian = "little")] + let len = (4 - word.leading_zeros() / 8) as usize; + #[cfg(target_endian = "big")] + let len = (4 - word.trailing_zeros() / 8) as usize; + len + } + + pub const fn is_ascii_alphabetic(&self) -> bool { + let word = self.0; + // Each of the following bitmasks set *the high bit* (0x8) to 0 for valid and 1 for invalid. + // `mask` sets all NUL bytes to 0. + let mask = (word + 0x7f7f_7f7f) & 0x8080_8080; + // `lower` converts the string to lowercase. It may also change the value of non-alpha + // characters, but this does not matter for the alphabetic test that follows. + let lower = word | 0x2020_2020; + // `alpha` sets all alphabetic bytes to 0. We only need check for lowercase characters. + let alpha = !(lower + 0x1f1f_1f1f) | (lower + 0x0505_0505); + // The overall string is valid if every character passes at least one test. + // We performed two tests here: non-NUL (`mask`) and alphabetic (`alpha`). + (alpha & mask) == 0 + } + + pub const fn is_ascii_alphanumeric(&self) -> bool { + let word = self.0; + // See explanatory comments in is_ascii_alphabetic + let mask = (word + 0x7f7f_7f7f) & 0x8080_8080; + let numeric = !(word + 0x5050_5050) | (word + 0x4646_4646); + let lower = word | 0x2020_2020; + let alpha = !(lower + 0x1f1f_1f1f) | (lower + 0x0505_0505); + (alpha & numeric & mask) == 0 + } + + pub const fn is_ascii_numeric(&self) -> bool { + let word = self.0; + // See explanatory comments in is_ascii_alphabetic + let mask = (word + 0x7f7f_7f7f) & 0x8080_8080; + let numeric = !(word + 0x5050_5050) | (word + 0x4646_4646); + (numeric & mask) == 0 + } + + pub const fn is_ascii_lowercase(&self) -> bool { + let word = self.0; + // For efficiency, this function tests for an invalid string rather than a valid string. + // A string is ASCII lowercase iff it contains no uppercase ASCII characters. + // `invalid_case` sets all uppercase ASCII characters to 0 and all others to 1. + let invalid_case = !(word + 0x3f3f_3f3f) | (word + 0x2525_2525); + // The string is valid if it contains no invalid characters (if all high bits are 1). + (invalid_case & 0x8080_8080) == 0x8080_8080 + } + + pub const fn is_ascii_titlecase(&self) -> bool { + let word = self.0; + // See explanatory comments in is_ascii_lowercase + let invalid_case = if cfg!(target_endian = "little") { + !(word + 0x3f3f_3f1f) | (word + 0x2525_2505) + } else { + !(word + 0x1f3f_3f3f) | (word + 0x0525_2525) + }; + (invalid_case & 0x8080_8080) == 0x8080_8080 + } + + pub const fn is_ascii_uppercase(&self) -> bool { + let word = self.0; + // See explanatory comments in is_ascii_lowercase + let invalid_case = !(word + 0x1f1f_1f1f) | (word + 0x0505_0505); + (invalid_case & 0x8080_8080) == 0x8080_8080 + } + + pub const fn is_ascii_alphabetic_lowercase(&self) -> bool { + let word = self.0; + // `mask` sets all NUL bytes to 0. + let mask = (word + 0x7f7f_7f7f) & 0x8080_8080; + // `lower_alpha` sets all lowercase ASCII characters to 0 and all others to 1. + let lower_alpha = !(word + 0x1f1f_1f1f) | (word + 0x0505_0505); + // The overall string is valid if every character passes at least one test. + // We performed two tests here: non-NUL (`mask`) and lowercase ASCII character (`alpha`). + (lower_alpha & mask) == 0 + } + + pub const fn is_ascii_alphabetic_titlecase(&self) -> bool { + let word = self.0; + // See explanatory comments in is_ascii_alphabetic_lowercase + let mask = (word + 0x7f7f_7f7f) & 0x8080_8080; + let title_case = if cfg!(target_endian = "little") { + !(word + 0x1f1f_1f3f) | (word + 0x0505_0525) + } else { + !(word + 0x3f1f_1f1f) | (word + 0x2505_0505) + }; + (title_case & mask) == 0 + } + + pub const fn is_ascii_alphabetic_uppercase(&self) -> bool { + let word = self.0; + // See explanatory comments in is_ascii_alphabetic_lowercase + let mask = (word + 0x7f7f_7f7f) & 0x8080_8080; + let upper_alpha = !(word + 0x3f3f_3f3f) | (word + 0x2525_2525); + (upper_alpha & mask) == 0 + } + + pub const fn to_ascii_lowercase(&self) -> Self { + let word = self.0; + let result = word | (((word + 0x3f3f_3f3f) & !(word + 0x2525_2525) & 0x8080_8080) >> 2); + Self(result) + } + + pub const fn to_ascii_titlecase(&self) -> Self { + let word = self.0.to_le(); + let mask = ((word + 0x3f3f_3f1f) & !(word + 0x2525_2505) & 0x8080_8080) >> 2; + let result = (word | mask) & !(0x20 & mask); + Self(u32::from_le(result)) + } + + pub const fn to_ascii_uppercase(&self) -> Self { + let word = self.0; + let result = word & !(((word + 0x1f1f_1f1f) & !(word + 0x0505_0505) & 0x8080_8080) >> 2); + Self(result) + } +} + +/// Internal helper struct that performs operations on aligned integers. +/// Supports strings up to 8 bytes long. +#[repr(transparent)] +pub struct Aligned8(u64); + +impl Aligned8 { + /// # Panics + /// Panics if N is greater than 8 + #[inline] + pub const fn from_utf8<const N: usize>(src: &[u8; N]) -> Self { + let mut bytes = [0; 8]; + let mut i = 0; + // The function documentation defines when panics may occur + #[allow(clippy::indexing_slicing)] + while i < N { + bytes[i] = src[i]; + i += 1; + } + Self(u64::from_ne_bytes(bytes)) + } + + #[inline] + pub const fn from_ascii_bytes<const N: usize>(src: &[AsciiByte; N]) -> Self { + Self::from_utf8::<N>(unsafe { core::mem::transmute::<&[AsciiByte; N], &[u8; N]>(src) }) + } + + #[inline] + pub const fn to_bytes(&self) -> [u8; 8] { + self.0.to_ne_bytes() + } + + #[inline] + pub const fn to_ascii_bytes(&self) -> [AsciiByte; 8] { + unsafe { core::mem::transmute(self.to_bytes()) } + } + + pub const fn len(&self) -> usize { + let word = self.0; + #[cfg(target_endian = "little")] + let len = (8 - word.leading_zeros() / 8) as usize; + #[cfg(target_endian = "big")] + let len = (8 - word.trailing_zeros() / 8) as usize; + len + } + + pub const fn is_ascii_alphabetic(&self) -> bool { + let word = self.0; + let mask = (word + 0x7f7f_7f7f_7f7f_7f7f) & 0x8080_8080_8080_8080; + let lower = word | 0x2020_2020_2020_2020; + let alpha = !(lower + 0x1f1f_1f1f_1f1f_1f1f) | (lower + 0x0505_0505_0505_0505); + (alpha & mask) == 0 + } + + pub const fn is_ascii_alphanumeric(&self) -> bool { + let word = self.0; + let mask = (word + 0x7f7f_7f7f_7f7f_7f7f) & 0x8080_8080_8080_8080; + let numeric = !(word + 0x5050_5050_5050_5050) | (word + 0x4646_4646_4646_4646); + let lower = word | 0x2020_2020_2020_2020; + let alpha = !(lower + 0x1f1f_1f1f_1f1f_1f1f) | (lower + 0x0505_0505_0505_0505); + (alpha & numeric & mask) == 0 + } + + pub const fn is_ascii_numeric(&self) -> bool { + let word = self.0; + let mask = (word + 0x7f7f_7f7f_7f7f_7f7f) & 0x8080_8080_8080_8080; + let numeric = !(word + 0x5050_5050_5050_5050) | (word + 0x4646_4646_4646_4646); + (numeric & mask) == 0 + } + + pub const fn is_ascii_lowercase(&self) -> bool { + let word = self.0; + let invalid_case = !(word + 0x3f3f_3f3f_3f3f_3f3f) | (word + 0x2525_2525_2525_2525); + (invalid_case & 0x8080_8080_8080_8080) == 0x8080_8080_8080_8080 + } + + pub const fn is_ascii_titlecase(&self) -> bool { + let word = self.0; + let invalid_case = if cfg!(target_endian = "little") { + !(word + 0x3f3f_3f3f_3f3f_3f1f) | (word + 0x2525_2525_2525_2505) + } else { + !(word + 0x1f3f_3f3f_3f3f_3f3f) | (word + 0x0525_2525_2525_2525) + }; + (invalid_case & 0x8080_8080_8080_8080) == 0x8080_8080_8080_8080 + } + + pub const fn is_ascii_uppercase(&self) -> bool { + let word = self.0; + let invalid_case = !(word + 0x1f1f_1f1f_1f1f_1f1f) | (word + 0x0505_0505_0505_0505); + (invalid_case & 0x8080_8080_8080_8080) == 0x8080_8080_8080_8080 + } + + pub const fn is_ascii_alphabetic_lowercase(&self) -> bool { + let word = self.0; + // `mask` sets all NUL bytes to 0. + let mask = (word + 0x7f7f_7f7f_7f7f_7f7f) & 0x8080_8080_8080_8080; + // `lower_alpha` sets all lowercase ASCII characters to 0 and all others to 1. + let lower_alpha = !(word + 0x1f1f_1f1f_1f1f_1f1f) | (word + 0x0505_0505_0505_0505); + // The overall string is valid if every character passes at least one test. + // We performed two tests here: non-NUL (`mask`) and lowercase ASCII character (`alpha`). + (lower_alpha & mask) == 0 + } + + pub const fn is_ascii_alphabetic_titlecase(&self) -> bool { + let word = self.0; + // See explanatory comments in is_ascii_alphabetic_lowercase + let mask = (word + 0x7f7f_7f7f_7f7f_7f7f) & 0x8080_8080_8080_8080; + let title_case = if cfg!(target_endian = "little") { + !(word + 0x1f1f_1f1f_1f1f_1f3f) | (word + 0x0505_0505_0505_0525) + } else { + !(word + 0x3f1f_1f1f_1f1f_1f1f) | (word + 0x2505_0505_0505_0505) + }; + (title_case & mask) == 0 + } + + pub const fn is_ascii_alphabetic_uppercase(&self) -> bool { + let word = self.0; + // See explanatory comments in is_ascii_alphabetic_lowercase + let mask = (word + 0x7f7f_7f7f_7f7f_7f7f) & 0x8080_8080_8080_8080; + let upper_alpha = !(word + 0x3f3f_3f3f_3f3f_3f3f) | (word + 0x2525_2525_2525_2525); + (upper_alpha & mask) == 0 + } + + pub const fn to_ascii_lowercase(&self) -> Self { + let word = self.0; + let result = word + | (((word + 0x3f3f_3f3f_3f3f_3f3f) + & !(word + 0x2525_2525_2525_2525) + & 0x8080_8080_8080_8080) + >> 2); + Self(result) + } + + pub const fn to_ascii_titlecase(&self) -> Self { + let word = self.0.to_le(); + let mask = ((word + 0x3f3f_3f3f_3f3f_3f1f) + & !(word + 0x2525_2525_2525_2505) + & 0x8080_8080_8080_8080) + >> 2; + let result = (word | mask) & !(0x20 & mask); + Self(u64::from_le(result)) + } + + pub const fn to_ascii_uppercase(&self) -> Self { + let word = self.0; + let result = word + & !(((word + 0x1f1f_1f1f_1f1f_1f1f) + & !(word + 0x0505_0505_0505_0505) + & 0x8080_8080_8080_8080) + >> 2); + Self(result) + } +} diff --git a/vendor/tinystr/src/lib.rs b/vendor/tinystr/src/lib.rs new file mode 100644 index 00000000..19440720 --- /dev/null +++ b/vendor/tinystr/src/lib.rs @@ -0,0 +1,114 @@ +// This file is part of ICU4X. For terms of use, please see the file +// called LICENSE at the top level of the ICU4X source tree +// (online at: https://github.com/unicode-org/icu4x/blob/main/LICENSE ). + +//! `tinystr` is a utility crate of the [`ICU4X`] project. +//! +//! It includes [`TinyAsciiStr`], a core API for representing small ASCII-only bounded length strings. +//! +//! It is optimized for operations on strings of size 8 or smaller. When use cases involve comparison +//! and conversion of strings for lowercase/uppercase/titlecase, or checking +//! numeric/alphabetic/alphanumeric, `TinyAsciiStr` is the edge performance library. +//! +//! # Examples +//! +//! ```rust +//! use tinystr::TinyAsciiStr; +//! +//! let s1: TinyAsciiStr<4> = "tEsT".parse().expect("Failed to parse."); +//! +//! assert_eq!(s1, "tEsT"); +//! assert_eq!(s1.to_ascii_uppercase(), "TEST"); +//! assert_eq!(s1.to_ascii_lowercase(), "test"); +//! assert_eq!(s1.to_ascii_titlecase(), "Test"); +//! assert!(s1.is_ascii_alphanumeric()); +//! assert!(!s1.is_ascii_numeric()); +//! +//! let s2 = TinyAsciiStr::<8>::try_from_raw(*b"New York") +//! .expect("Failed to parse."); +//! +//! assert_eq!(s2, "New York"); +//! assert_eq!(s2.to_ascii_uppercase(), "NEW YORK"); +//! assert_eq!(s2.to_ascii_lowercase(), "new york"); +//! assert_eq!(s2.to_ascii_titlecase(), "New york"); +//! assert!(!s2.is_ascii_alphanumeric()); +//! ``` +//! +//! # Details +//! +//! When strings are of size 8 or smaller, the struct transforms the strings as `u32`/`u64` and uses +//! bitmasking to provide basic string manipulation operations: +//! * `is_ascii_numeric` +//! * `is_ascii_alphabetic` +//! * `is_ascii_alphanumeric` +//! * `to_ascii_lowercase` +//! * `to_ascii_uppercase` +//! * `to_ascii_titlecase` +//! * `PartialEq` +//! +//! `TinyAsciiStr` will fall back to `u8` character manipulation for strings of length greater than 8. + +//! +//! [`ICU4X`]: ../icu/index.html + +// https://github.com/unicode-org/icu4x/blob/main/documents/process/boilerplate.md#library-annotations +#![cfg_attr(not(any(test, doc)), no_std)] +#![cfg_attr( + not(test), + deny( + clippy::indexing_slicing, + clippy::unwrap_used, + clippy::expect_used, + clippy::panic, + clippy::exhaustive_structs, + clippy::exhaustive_enums, + clippy::trivially_copy_pass_by_ref, + missing_debug_implementations, + ) +)] + +mod macros; + +mod ascii; +mod asciibyte; +mod error; +mod int_ops; +mod unvalidated; + +#[cfg(feature = "serde")] +mod serde; + +#[cfg(feature = "databake")] +mod databake; + +#[cfg(feature = "zerovec")] +mod ule; + +#[cfg(any(feature = "serde", feature = "alloc"))] +extern crate alloc; + +pub use ascii::TinyAsciiStr; +pub use error::ParseError; +pub use unvalidated::UnvalidatedTinyAsciiStr; + +/// These are temporary compatability reexports that will be removed +/// in a future version. +pub type TinyStr4 = TinyAsciiStr<4>; +/// These are temporary compatability reexports that will be removed +/// in a future version. +pub type TinyStr8 = TinyAsciiStr<8>; +/// These are temporary compatability reexports that will be removed +/// in a future version. +pub type TinyStr16 = TinyAsciiStr<16>; + +#[test] +fn test_size() { + assert_eq!( + core::mem::size_of::<TinyStr4>(), + core::mem::size_of::<Option<TinyStr4>>() + ); + assert_eq!( + core::mem::size_of::<TinyStr8>(), + core::mem::size_of::<Option<TinyStr8>>() + ); +} diff --git a/vendor/tinystr/src/macros.rs b/vendor/tinystr/src/macros.rs new file mode 100644 index 00000000..30dc445e --- /dev/null +++ b/vendor/tinystr/src/macros.rs @@ -0,0 +1,32 @@ +// This file is part of ICU4X. For terms of use, please see the file +// called LICENSE at the top level of the ICU4X source tree +// (online at: https://github.com/unicode-org/icu4x/blob/main/LICENSE ). + +#[macro_export] +macro_rules! tinystr { + ($n:literal, $s:literal) => {{ + // Force it into a const context; otherwise it may get evaluated at runtime instead. + const TINYSTR_MACRO_CONST: $crate::TinyAsciiStr<$n> = { + match $crate::TinyAsciiStr::try_from_utf8($s.as_bytes()) { + Ok(s) => s, + // We are okay with panicking here because this is in a const context + #[allow(clippy::panic)] + // Cannot format the error since formatting isn't const yet + Err(_) => panic!(concat!("Failed to construct tinystr from ", $s)), + } + }; + TINYSTR_MACRO_CONST + }}; +} + +#[cfg(test)] +mod tests { + #[test] + fn test_macro_construction() { + let s1 = tinystr!(8, "foobar"); + assert_eq!(&*s1, "foobar"); + + let s1 = tinystr!(12, "foobarbaz"); + assert_eq!(&*s1, "foobarbaz"); + } +} diff --git a/vendor/tinystr/src/serde.rs b/vendor/tinystr/src/serde.rs new file mode 100644 index 00000000..529902b5 --- /dev/null +++ b/vendor/tinystr/src/serde.rs @@ -0,0 +1,91 @@ +// This file is part of ICU4X. For terms of use, please see the file +// called LICENSE at the top level of the ICU4X source tree +// (online at: https://github.com/unicode-org/icu4x/blob/main/LICENSE ). + +use crate::TinyAsciiStr; +use alloc::borrow::Cow; +use alloc::string::ToString; +use core::fmt; +use core::marker::PhantomData; +use core::ops::Deref; +use serde::de::{Error, SeqAccess, Visitor}; +use serde::ser::SerializeTuple; +use serde::{Deserialize, Deserializer, Serialize, Serializer}; + +impl<const N: usize> Serialize for TinyAsciiStr<N> { + #[inline] + fn serialize<S>(&self, serializer: S) -> Result<S::Ok, S::Error> + where + S: Serializer, + { + if serializer.is_human_readable() { + self.deref().serialize(serializer) + } else { + let mut seq = serializer.serialize_tuple(N)?; + for byte in self.all_bytes() { + seq.serialize_element(byte)?; + } + seq.end() + } + } +} + +struct TinyAsciiStrVisitor<const N: usize> { + marker: PhantomData<TinyAsciiStr<N>>, +} + +impl<const N: usize> TinyAsciiStrVisitor<N> { + fn new() -> Self { + TinyAsciiStrVisitor { + marker: PhantomData, + } + } +} + +impl<'de, const N: usize> Visitor<'de> for TinyAsciiStrVisitor<N> { + type Value = TinyAsciiStr<N>; + + fn expecting(&self, formatter: &mut fmt::Formatter) -> fmt::Result { + write!(formatter, "a TinyAsciiStr<{N}>") + } + + #[inline] + fn visit_seq<A>(self, mut seq: A) -> Result<Self::Value, A::Error> + where + A: SeqAccess<'de>, + { + let mut bytes = [0u8; N]; + let mut zeroes = false; + for out in &mut bytes.iter_mut().take(N) { + let byte = seq + .next_element()? + .ok_or_else(|| Error::invalid_length(N, &self))?; + if byte == 0 { + zeroes = true; + } else if zeroes { + return Err(Error::custom("TinyAsciiStr cannot contain null bytes")); + } + + if byte >= 0x80 { + return Err(Error::custom("TinyAsciiStr cannot contain non-ascii bytes")); + } + *out = byte; + } + + Ok(unsafe { TinyAsciiStr::from_utf8_unchecked(bytes) }) + } +} + +impl<'de, const N: usize> Deserialize<'de> for TinyAsciiStr<N> { + fn deserialize<D>(deserializer: D) -> Result<Self, D::Error> + where + D: Deserializer<'de>, + { + if deserializer.is_human_readable() { + let x: Cow<'de, str> = Deserialize::deserialize(deserializer)?; + TinyAsciiStr::try_from_str(&x).map_err(|e| Error::custom(e.to_string())) + } else { + deserializer.deserialize_tuple(N, TinyAsciiStrVisitor::<N>::new()) + } + } +} diff --git a/vendor/tinystr/src/ule.rs b/vendor/tinystr/src/ule.rs new file mode 100644 index 00000000..0dd10ff7 --- /dev/null +++ b/vendor/tinystr/src/ule.rs @@ -0,0 +1,125 @@ +// This file is part of ICU4X. For terms of use, please see the file +// called LICENSE at the top level of the ICU4X source tree +// (online at: https://github.com/unicode-org/icu4x/blob/main/LICENSE ). + +use crate::{TinyAsciiStr, UnvalidatedTinyAsciiStr}; +#[cfg(feature = "alloc")] +use zerovec::maps::ZeroMapKV; +use zerovec::ule::*; +#[cfg(feature = "alloc")] +use zerovec::{ZeroSlice, ZeroVec}; + +// Safety (based on the safety checklist on the ULE trait): +// 1. TinyAsciiStr does not include any uninitialized or padding bytes. +// (achieved by `#[repr(transparent)]` on a type that satisfies this invariant) +// 2. TinyAsciiStr is aligned to 1 byte. +// (achieved by `#[repr(transparent)]` on a type that satisfies this invariant) +// 3. The impl of validate_bytes() returns an error if any byte is not valid. +// 4. The impl of validate_bytes() returns an error if there are extra bytes. +// 5. The other ULE methods use the default impl. +// 6. TinyAsciiStr byte equality is semantic equality +unsafe impl<const N: usize> ULE for TinyAsciiStr<N> { + #[inline] + fn validate_bytes(bytes: &[u8]) -> Result<(), UleError> { + if bytes.len() % N != 0 { + return Err(UleError::length::<Self>(bytes.len())); + } + // Validate the bytes + for chunk in bytes.chunks_exact(N) { + let _ = TinyAsciiStr::<N>::try_from_utf8_inner(chunk, true) + .map_err(|_| UleError::parse::<Self>())?; + } + Ok(()) + } +} + +impl<const N: usize> NicheBytes<N> for TinyAsciiStr<N> { + // AsciiByte is 0..128 + const NICHE_BIT_PATTERN: [u8; N] = [255; N]; +} + +impl<const N: usize> AsULE for TinyAsciiStr<N> { + type ULE = Self; + + #[inline] + fn to_unaligned(self) -> Self::ULE { + self + } + + #[inline] + fn from_unaligned(unaligned: Self::ULE) -> Self { + unaligned + } +} + +#[cfg(feature = "alloc")] +impl<'a, const N: usize> ZeroMapKV<'a> for TinyAsciiStr<N> { + type Container = ZeroVec<'a, TinyAsciiStr<N>>; + type Slice = ZeroSlice<TinyAsciiStr<N>>; + type GetType = TinyAsciiStr<N>; + type OwnedType = TinyAsciiStr<N>; +} + +// Safety (based on the safety checklist on the ULE trait): +// 1. UnvalidatedTinyAsciiStr does not include any uninitialized or padding bytes. +// (achieved by `#[repr(transparent)]` on a type that satisfies this invariant) +// 2. UnvalidatedTinyAsciiStr is aligned to 1 byte. +// (achieved by `#[repr(transparent)]` on a type that satisfies this invariant) +// 3. The impl of validate_bytes() returns an error if any byte is not valid. +// 4. The impl of validate_bytes() returns an error if there are extra bytes. +// 5. The other ULE methods use the default impl. +// 6. UnvalidatedTinyAsciiStr byte equality is semantic equality +unsafe impl<const N: usize> ULE for UnvalidatedTinyAsciiStr<N> { + #[inline] + fn validate_bytes(bytes: &[u8]) -> Result<(), UleError> { + if bytes.len() % N != 0 { + return Err(UleError::length::<Self>(bytes.len())); + } + Ok(()) + } +} + +impl<const N: usize> AsULE for UnvalidatedTinyAsciiStr<N> { + type ULE = Self; + + #[inline] + fn to_unaligned(self) -> Self::ULE { + self + } + + #[inline] + fn from_unaligned(unaligned: Self::ULE) -> Self { + unaligned + } +} + +#[cfg(feature = "alloc")] +impl<'a, const N: usize> ZeroMapKV<'a> for UnvalidatedTinyAsciiStr<N> { + type Container = ZeroVec<'a, UnvalidatedTinyAsciiStr<N>>; + type Slice = ZeroSlice<UnvalidatedTinyAsciiStr<N>>; + type GetType = UnvalidatedTinyAsciiStr<N>; + type OwnedType = UnvalidatedTinyAsciiStr<N>; +} + +#[cfg(test)] +mod test { + use crate::*; + use zerovec::*; + + #[test] + fn test_zerovec() { + let mut vec = ZeroVec::<TinyAsciiStr<7>>::new(); + + vec.with_mut(|v| v.push("foobar".parse().unwrap())); + vec.with_mut(|v| v.push("baz".parse().unwrap())); + vec.with_mut(|v| v.push("quux".parse().unwrap())); + + let bytes = vec.as_bytes(); + + let vec: ZeroVec<TinyAsciiStr<7>> = ZeroVec::parse_bytes(bytes).unwrap(); + + assert_eq!(&*vec.get(0).unwrap(), "foobar"); + assert_eq!(&*vec.get(1).unwrap(), "baz"); + assert_eq!(&*vec.get(2).unwrap(), "quux"); + } +} diff --git a/vendor/tinystr/src/unvalidated.rs b/vendor/tinystr/src/unvalidated.rs new file mode 100644 index 00000000..3758b64f --- /dev/null +++ b/vendor/tinystr/src/unvalidated.rs @@ -0,0 +1,122 @@ +// This file is part of ICU4X. For terms of use, please see the file +// called LICENSE at the top level of the ICU4X source tree +// (online at: https://github.com/unicode-org/icu4x/blob/main/LICENSE ). + +use crate::ParseError; +use crate::TinyAsciiStr; +use core::fmt; + +/// A fixed-length bytes array that is expected to be an ASCII string but does not enforce that invariant. +/// +/// Use this type instead of `TinyAsciiStr` if you don't need to enforce ASCII during deserialization. For +/// example, strings that are keys of a map don't need to ever be reified as `TinyAsciiStr`s. +/// +/// The main advantage of this type over `[u8; N]` is that it serializes as a string in +/// human-readable formats like JSON. +#[derive(PartialEq, PartialOrd, Eq, Ord, Clone, Copy)] +pub struct UnvalidatedTinyAsciiStr<const N: usize>(pub(crate) [u8; N]); + +impl<const N: usize> fmt::Debug for UnvalidatedTinyAsciiStr<N> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + // Debug as a string if possible + match self.try_into_tinystr() { + Ok(s) => fmt::Debug::fmt(&s, f), + Err(_) => fmt::Debug::fmt(&self.0, f), + } + } +} + +impl<const N: usize> UnvalidatedTinyAsciiStr<N> { + #[inline] + /// Converts into a [`TinyAsciiStr`]. Fails if the bytes are not valid ASCII. + pub fn try_into_tinystr(self) -> Result<TinyAsciiStr<N>, ParseError> { + TinyAsciiStr::try_from_raw(self.0) + } + + #[inline] + /// Unsafely converts into a [`TinyAsciiStr`]. + pub const fn from_utf8_unchecked(bytes: [u8; N]) -> Self { + Self(bytes) + } +} + +impl<const N: usize> TinyAsciiStr<N> { + #[inline] + // Converts into a [`UnvalidatedTinyAsciiStr`] + pub const fn to_unvalidated(self) -> UnvalidatedTinyAsciiStr<N> { + UnvalidatedTinyAsciiStr(*self.all_bytes()) + } +} + +impl<const N: usize> From<TinyAsciiStr<N>> for UnvalidatedTinyAsciiStr<N> { + fn from(other: TinyAsciiStr<N>) -> Self { + other.to_unvalidated() + } +} + +#[cfg(feature = "serde")] +impl<const N: usize> serde::Serialize for UnvalidatedTinyAsciiStr<N> { + fn serialize<S>(&self, serializer: S) -> Result<S::Ok, S::Error> + where + S: serde::Serializer, + { + use serde::ser::Error; + self.try_into_tinystr() + .map_err(|_| S::Error::custom("invalid ascii in UnvalidatedTinyAsciiStr"))? + .serialize(serializer) + } +} + +macro_rules! deserialize { + ($size:literal) => { + #[cfg(feature = "serde")] + impl<'de, 'a> serde::Deserialize<'de> for UnvalidatedTinyAsciiStr<$size> + where + 'de: 'a, + { + fn deserialize<D>(deserializer: D) -> Result<Self, D::Error> + where + D: serde::Deserializer<'de>, + { + if deserializer.is_human_readable() { + Ok(TinyAsciiStr::deserialize(deserializer)?.to_unvalidated()) + } else { + Ok(Self(<[u8; $size]>::deserialize(deserializer)?)) + } + } + } + }; +} + +deserialize!(1); +deserialize!(2); +deserialize!(3); +deserialize!(4); +deserialize!(5); +deserialize!(6); +deserialize!(7); +deserialize!(8); +deserialize!(9); +deserialize!(10); +deserialize!(11); +deserialize!(12); +deserialize!(13); +deserialize!(14); +deserialize!(15); +deserialize!(16); +deserialize!(17); +deserialize!(18); +deserialize!(19); +deserialize!(20); +deserialize!(21); +deserialize!(22); +deserialize!(23); +deserialize!(24); +deserialize!(25); +deserialize!(26); +deserialize!(27); +deserialize!(28); +deserialize!(29); +deserialize!(30); +deserialize!(31); +deserialize!(32); |
