diff options
author | Daniel Baumann <daniel.baumann@progress-linux.org> | 2024-04-19 00:47:55 +0000 |
---|---|---|
committer | Daniel Baumann <daniel.baumann@progress-linux.org> | 2024-04-19 00:47:55 +0000 |
commit | 26a029d407be480d791972afb5975cf62c9360a6 (patch) | |
tree | f435a8308119effd964b339f76abb83a57c29483 /third_party/rust/time-macros/src/format_description | |
parent | Initial commit. (diff) | |
download | firefox-26a029d407be480d791972afb5975cf62c9360a6.tar.xz firefox-26a029d407be480d791972afb5975cf62c9360a6.zip |
Adding upstream version 124.0.1.upstream/124.0.1
Signed-off-by: Daniel Baumann <daniel.baumann@progress-linux.org>
Diffstat (limited to 'third_party/rust/time-macros/src/format_description')
7 files changed, 1464 insertions, 0 deletions
diff --git a/third_party/rust/time-macros/src/format_description/ast.rs b/third_party/rust/time-macros/src/format_description/ast.rs new file mode 100644 index 0000000000..b75056bc2f --- /dev/null +++ b/third_party/rust/time-macros/src/format_description/ast.rs @@ -0,0 +1,253 @@ +use std::boxed::Box; +use std::iter; + +use super::{lexer, unused, Error, Location, Spanned, SpannedValue, Unused}; + +pub(super) enum Item<'a> { + Literal(Spanned<&'a [u8]>), + EscapedBracket { + _first: Unused<Location>, + _second: Unused<Location>, + }, + Component { + _opening_bracket: Unused<Location>, + _leading_whitespace: Unused<Option<Spanned<&'a [u8]>>>, + name: Spanned<&'a [u8]>, + modifiers: Box<[Modifier<'a>]>, + _trailing_whitespace: Unused<Option<Spanned<&'a [u8]>>>, + _closing_bracket: Unused<Location>, + }, + Optional { + opening_bracket: Location, + _leading_whitespace: Unused<Option<Spanned<&'a [u8]>>>, + _optional_kw: Unused<Spanned<&'a [u8]>>, + _whitespace: Unused<Spanned<&'a [u8]>>, + nested_format_description: NestedFormatDescription<'a>, + closing_bracket: Location, + }, + First { + opening_bracket: Location, + _leading_whitespace: Unused<Option<Spanned<&'a [u8]>>>, + _first_kw: Unused<Spanned<&'a [u8]>>, + _whitespace: Unused<Spanned<&'a [u8]>>, + nested_format_descriptions: Box<[NestedFormatDescription<'a>]>, + closing_bracket: Location, + }, +} + +pub(super) struct NestedFormatDescription<'a> { + pub(super) _opening_bracket: Unused<Location>, + pub(super) items: Box<[Item<'a>]>, + pub(super) _closing_bracket: Unused<Location>, + pub(super) _trailing_whitespace: Unused<Option<Spanned<&'a [u8]>>>, +} + +pub(super) struct Modifier<'a> { + pub(super) _leading_whitespace: Unused<Spanned<&'a [u8]>>, + pub(super) key: Spanned<&'a [u8]>, + pub(super) _colon: Unused<Location>, + pub(super) value: Spanned<&'a [u8]>, +} + +pub(super) fn parse< + 'item: 'iter, + 'iter, + I: Iterator<Item = Result<lexer::Token<'item>, Error>>, + const VERSION: u8, +>( + tokens: &'iter mut lexer::Lexed<I>, +) -> impl Iterator<Item = Result<Item<'item>, Error>> + 'iter { + assert!(version!(1..=2)); + parse_inner::<_, false, VERSION>(tokens) +} + +fn parse_inner< + 'item, + I: Iterator<Item = Result<lexer::Token<'item>, Error>>, + const NESTED: bool, + const VERSION: u8, +>( + tokens: &mut lexer::Lexed<I>, +) -> impl Iterator<Item = Result<Item<'item>, Error>> + '_ { + iter::from_fn(move || { + if NESTED && tokens.peek_closing_bracket().is_some() { + return None; + } + + let next = match tokens.next()? { + Ok(token) => token, + Err(err) => return Some(Err(err)), + }; + + Some(match next { + lexer::Token::Literal(Spanned { value: _, span: _ }) if NESTED => { + bug!("literal should not be present in nested description") + } + lexer::Token::Literal(value) => Ok(Item::Literal(value)), + lexer::Token::Bracket { + kind: lexer::BracketKind::Opening, + location, + } => { + if version!(..=1) { + if let Some(second_location) = tokens.next_if_opening_bracket() { + Ok(Item::EscapedBracket { + _first: unused(location), + _second: unused(second_location), + }) + } else { + parse_component::<_, VERSION>(location, tokens) + } + } else { + parse_component::<_, VERSION>(location, tokens) + } + } + lexer::Token::Bracket { + kind: lexer::BracketKind::Closing, + location: _, + } if NESTED => { + bug!("closing bracket should be caught by the `if` statement") + } + lexer::Token::Bracket { + kind: lexer::BracketKind::Closing, + location: _, + } => { + bug!("closing bracket should have been consumed by `parse_component`") + } + lexer::Token::ComponentPart { kind: _, value } if NESTED => Ok(Item::Literal(value)), + lexer::Token::ComponentPart { kind: _, value: _ } => { + bug!("component part should have been consumed by `parse_component`") + } + }) + }) +} + +fn parse_component<'a, I: Iterator<Item = Result<lexer::Token<'a>, Error>>, const VERSION: u8>( + opening_bracket: Location, + tokens: &mut lexer::Lexed<I>, +) -> Result<Item<'a>, Error> { + let leading_whitespace = tokens.next_if_whitespace(); + + let Some(name) = tokens.next_if_not_whitespace() else { + let span = match leading_whitespace { + Some(Spanned { value: _, span }) => span, + None => opening_bracket.to(opening_bracket), + }; + return Err(span.error("expected component name")); + }; + + if *name == b"optional" { + let Some(whitespace) = tokens.next_if_whitespace() else { + return Err(name.span.error("expected whitespace after `optional`")); + }; + + let nested = parse_nested::<_, VERSION>(whitespace.span.end, tokens)?; + + let Some(closing_bracket) = tokens.next_if_closing_bracket() else { + return Err(opening_bracket.error("unclosed bracket")); + }; + + return Ok(Item::Optional { + opening_bracket, + _leading_whitespace: unused(leading_whitespace), + _optional_kw: unused(name), + _whitespace: unused(whitespace), + nested_format_description: nested, + closing_bracket, + }); + } + + if *name == b"first" { + let Some(whitespace) = tokens.next_if_whitespace() else { + return Err(name.span.error("expected whitespace after `first`")); + }; + + let mut nested_format_descriptions = Vec::new(); + while let Ok(description) = parse_nested::<_, VERSION>(whitespace.span.end, tokens) { + nested_format_descriptions.push(description); + } + + let Some(closing_bracket) = tokens.next_if_closing_bracket() else { + return Err(opening_bracket.error("unclosed bracket")); + }; + + return Ok(Item::First { + opening_bracket, + _leading_whitespace: unused(leading_whitespace), + _first_kw: unused(name), + _whitespace: unused(whitespace), + nested_format_descriptions: nested_format_descriptions.into_boxed_slice(), + closing_bracket, + }); + } + + let mut modifiers = Vec::new(); + let trailing_whitespace = loop { + let Some(whitespace) = tokens.next_if_whitespace() else { + break None; + }; + + if let Some(location) = tokens.next_if_opening_bracket() { + return Err(location + .to(location) + .error("modifier must be of the form `key:value`")); + } + + let Some(Spanned { value, span }) = tokens.next_if_not_whitespace() else { + break Some(whitespace); + }; + + let Some(colon_index) = value.iter().position(|&b| b == b':') else { + return Err(span.error("modifier must be of the form `key:value`")); + }; + let key = &value[..colon_index]; + let value = &value[colon_index + 1..]; + + if key.is_empty() { + return Err(span.shrink_to_start().error("expected modifier key")); + } + if value.is_empty() { + return Err(span.shrink_to_end().error("expected modifier value")); + } + + modifiers.push(Modifier { + _leading_whitespace: unused(whitespace), + key: key.spanned(span.shrink_to_before(colon_index as _)), + _colon: unused(span.start.offset(colon_index as _)), + value: value.spanned(span.shrink_to_after(colon_index as _)), + }); + }; + + let Some(closing_bracket) = tokens.next_if_closing_bracket() else { + return Err(opening_bracket.error("unclosed bracket")); + }; + + Ok(Item::Component { + _opening_bracket: unused(opening_bracket), + _leading_whitespace: unused(leading_whitespace), + name, + modifiers: modifiers.into_boxed_slice(), + _trailing_whitespace: unused(trailing_whitespace), + _closing_bracket: unused(closing_bracket), + }) +} + +fn parse_nested<'a, I: Iterator<Item = Result<lexer::Token<'a>, Error>>, const VERSION: u8>( + last_location: Location, + tokens: &mut lexer::Lexed<I>, +) -> Result<NestedFormatDescription<'a>, Error> { + let Some(opening_bracket) = tokens.next_if_opening_bracket() else { + return Err(last_location.error("expected opening bracket")); + }; + let items = parse_inner::<_, true, VERSION>(tokens).collect::<Result<_, _>>()?; + let Some(closing_bracket) = tokens.next_if_closing_bracket() else { + return Err(opening_bracket.error("unclosed bracket")); + }; + let trailing_whitespace = tokens.next_if_whitespace(); + + Ok(NestedFormatDescription { + _opening_bracket: unused(opening_bracket), + items, + _closing_bracket: unused(closing_bracket), + _trailing_whitespace: unused(trailing_whitespace), + }) +} diff --git a/third_party/rust/time-macros/src/format_description/format_item.rs b/third_party/rust/time-macros/src/format_description/format_item.rs new file mode 100644 index 0000000000..6a8cf555ee --- /dev/null +++ b/third_party/rust/time-macros/src/format_description/format_item.rs @@ -0,0 +1,442 @@ +use std::boxed::Box; +use std::num::NonZeroU16; +use std::str::{self, FromStr}; + +use super::{ast, unused, Error, Span, Spanned, Unused}; + +pub(super) fn parse<'a>( + ast_items: impl Iterator<Item = Result<ast::Item<'a>, Error>>, +) -> impl Iterator<Item = Result<Item<'a>, Error>> { + ast_items.map(|ast_item| ast_item.and_then(Item::from_ast)) +} + +pub(super) enum Item<'a> { + Literal(&'a [u8]), + Component(Component), + Optional { + value: Box<[Self]>, + _span: Unused<Span>, + }, + First { + value: Box<[Box<[Self]>]>, + _span: Unused<Span>, + }, +} + +impl Item<'_> { + pub(super) fn from_ast(ast_item: ast::Item<'_>) -> Result<Item<'_>, Error> { + Ok(match ast_item { + ast::Item::Component { + _opening_bracket: _, + _leading_whitespace: _, + name, + modifiers, + _trailing_whitespace: _, + _closing_bracket: _, + } => Item::Component(component_from_ast(&name, &modifiers)?), + ast::Item::Literal(Spanned { value, span: _ }) => Item::Literal(value), + ast::Item::EscapedBracket { + _first: _, + _second: _, + } => Item::Literal(b"["), + ast::Item::Optional { + opening_bracket, + _leading_whitespace: _, + _optional_kw: _, + _whitespace: _, + nested_format_description, + closing_bracket, + } => { + let items = nested_format_description + .items + .into_vec() + .into_iter() + .map(Item::from_ast) + .collect::<Result<_, _>>()?; + Item::Optional { + value: items, + _span: unused(opening_bracket.to(closing_bracket)), + } + } + ast::Item::First { + opening_bracket, + _leading_whitespace: _, + _first_kw: _, + _whitespace: _, + nested_format_descriptions, + closing_bracket, + } => { + let items = nested_format_descriptions + .into_vec() + .into_iter() + .map(|nested_format_description| { + nested_format_description + .items + .into_vec() + .into_iter() + .map(Item::from_ast) + .collect() + }) + .collect::<Result<_, _>>()?; + Item::First { + value: items, + _span: unused(opening_bracket.to(closing_bracket)), + } + } + }) + } +} + +impl From<Item<'_>> for crate::format_description::public::OwnedFormatItem { + fn from(item: Item<'_>) -> Self { + match item { + Item::Literal(literal) => Self::Literal(literal.to_vec().into_boxed_slice()), + Item::Component(component) => Self::Component(component.into()), + Item::Optional { value, _span: _ } => Self::Optional(Box::new(value.into())), + Item::First { value, _span: _ } => { + Self::First(value.into_vec().into_iter().map(Into::into).collect()) + } + } + } +} + +impl<'a> From<Box<[Item<'a>]>> for crate::format_description::public::OwnedFormatItem { + fn from(items: Box<[Item<'a>]>) -> Self { + let items = items.into_vec(); + if items.len() == 1 { + if let Ok([item]) = <[_; 1]>::try_from(items) { + item.into() + } else { + bug!("the length was just checked to be 1") + } + } else { + Self::Compound(items.into_iter().map(Self::from).collect()) + } + } +} + +macro_rules! component_definition { + (@if_required required then { $($then:tt)* } $(else { $($else:tt)* })?) => { $($then)* }; + (@if_required then { $($then:tt)* } $(else { $($else:tt)* })?) => { $($($else)*)? }; + (@if_from_str from_str then { $($then:tt)* } $(else { $($else:tt)* })?) => { $($then)* }; + (@if_from_str then { $($then:tt)* } $(else { $($else:tt)* })?) => { $($($else)*)? }; + + ($vis:vis enum $name:ident { + $($variant:ident = $parse_variant:literal {$( + $(#[$required:tt])? + $field:ident = $parse_field:literal: + Option<$(#[$from_str:tt])? $field_type:ty> + => $target_field:ident + ),* $(,)?}),* $(,)? + }) => { + $vis enum $name { + $($variant($variant),)* + } + + $($vis struct $variant { + $($field: Option<$field_type>),* + })* + + $(impl $variant { + fn with_modifiers( + modifiers: &[ast::Modifier<'_>], + _component_span: Span, + ) -> Result<Self, Error> + { + let mut this = Self { + $($field: None),* + }; + + for modifier in modifiers { + $(#[allow(clippy::string_lit_as_bytes)] + if modifier.key.eq_ignore_ascii_case($parse_field.as_bytes()) { + this.$field = component_definition!(@if_from_str $($from_str)? + then { + parse_from_modifier_value::<$field_type>(&modifier.value)? + } else { + <$field_type>::from_modifier_value(&modifier.value)? + }); + continue; + })* + return Err(modifier.key.span.error("invalid modifier key")); + } + + $(component_definition! { @if_required $($required)? then { + if this.$field.is_none() { + return Err(_component_span.error("missing required modifier")); + } + }})* + + Ok(this) + } + })* + + impl From<$name> for crate::format_description::public::Component { + fn from(component: $name) -> Self { + match component {$( + $name::$variant($variant { $($field),* }) => { + $crate::format_description::public::Component::$variant( + super::public::modifier::$variant {$( + $target_field: component_definition! { @if_required $($required)? + then { + match $field { + Some(value) => value.into(), + None => bug!("required modifier was not set"), + } + } else { + $field.unwrap_or_default().into() + } + } + ),*} + ) + } + )*} + } + } + + fn component_from_ast( + name: &Spanned<&[u8]>, + modifiers: &[ast::Modifier<'_>], + ) -> Result<Component, Error> { + $(#[allow(clippy::string_lit_as_bytes)] + if name.eq_ignore_ascii_case($parse_variant.as_bytes()) { + return Ok(Component::$variant($variant::with_modifiers(&modifiers, name.span)?)); + })* + Err(name.span.error("invalid component")) + } + } +} + +component_definition! { + pub(super) enum Component { + Day = "day" { + padding = "padding": Option<Padding> => padding, + }, + Hour = "hour" { + padding = "padding": Option<Padding> => padding, + base = "repr": Option<HourBase> => is_12_hour_clock, + }, + Ignore = "ignore" { + #[required] + count = "count": Option<#[from_str] NonZeroU16> => count, + }, + Minute = "minute" { + padding = "padding": Option<Padding> => padding, + }, + Month = "month" { + padding = "padding": Option<Padding> => padding, + repr = "repr": Option<MonthRepr> => repr, + case_sensitive = "case_sensitive": Option<MonthCaseSensitive> => case_sensitive, + }, + OffsetHour = "offset_hour" { + sign_behavior = "sign": Option<SignBehavior> => sign_is_mandatory, + padding = "padding": Option<Padding> => padding, + }, + OffsetMinute = "offset_minute" { + padding = "padding": Option<Padding> => padding, + }, + OffsetSecond = "offset_second" { + padding = "padding": Option<Padding> => padding, + }, + Ordinal = "ordinal" { + padding = "padding": Option<Padding> => padding, + }, + Period = "period" { + case = "case": Option<PeriodCase> => is_uppercase, + case_sensitive = "case_sensitive": Option<PeriodCaseSensitive> => case_sensitive, + }, + Second = "second" { + padding = "padding": Option<Padding> => padding, + }, + Subsecond = "subsecond" { + digits = "digits": Option<SubsecondDigits> => digits, + }, + UnixTimestamp = "unix_timestamp" { + precision = "precision": Option<UnixTimestampPrecision> => precision, + sign_behavior = "sign": Option<SignBehavior> => sign_is_mandatory, + }, + Weekday = "weekday" { + repr = "repr": Option<WeekdayRepr> => repr, + one_indexed = "one_indexed": Option<WeekdayOneIndexed> => one_indexed, + case_sensitive = "case_sensitive": Option<WeekdayCaseSensitive> => case_sensitive, + }, + WeekNumber = "week_number" { + padding = "padding": Option<Padding> => padding, + repr = "repr": Option<WeekNumberRepr> => repr, + }, + Year = "year" { + padding = "padding": Option<Padding> => padding, + repr = "repr": Option<YearRepr> => repr, + base = "base": Option<YearBase> => iso_week_based, + sign_behavior = "sign": Option<SignBehavior> => sign_is_mandatory, + }, + } +} + +macro_rules! target_ty { + ($name:ident $type:ty) => { + $type + }; + ($name:ident) => { + super::public::modifier::$name + }; +} + +/// Get the target value for a given enum. +macro_rules! target_value { + ($name:ident $variant:ident $value:expr) => { + $value + }; + ($name:ident $variant:ident) => { + super::public::modifier::$name::$variant + }; +} + +macro_rules! modifier { + ($( + enum $name:ident $(($target_ty:ty))? { + $( + $(#[$attr:meta])? + $variant:ident $(($target_value:expr))? = $parse_variant:literal + ),* $(,)? + } + )+) => {$( + #[derive(Default)] + enum $name { + $($(#[$attr])? $variant),* + } + + impl $name { + /// Parse the modifier from its string representation. + fn from_modifier_value(value: &Spanned<&[u8]>) -> Result<Option<Self>, Error> { + $(if value.eq_ignore_ascii_case($parse_variant) { + return Ok(Some(Self::$variant)); + })* + Err(value.span.error("invalid modifier value")) + } + } + + impl From<$name> for target_ty!($name $($target_ty)?) { + fn from(modifier: $name) -> Self { + match modifier { + $($name::$variant => target_value!($name $variant $($target_value)?)),* + } + } + } + )+}; +} + +modifier! { + enum HourBase(bool) { + Twelve(true) = b"12", + #[default] + TwentyFour(false) = b"24", + } + + enum MonthCaseSensitive(bool) { + False(false) = b"false", + #[default] + True(true) = b"true", + } + + enum MonthRepr { + #[default] + Numerical = b"numerical", + Long = b"long", + Short = b"short", + } + + enum Padding { + Space = b"space", + #[default] + Zero = b"zero", + None = b"none", + } + + enum PeriodCase(bool) { + Lower(false) = b"lower", + #[default] + Upper(true) = b"upper", + } + + enum PeriodCaseSensitive(bool) { + False(false) = b"false", + #[default] + True(true) = b"true", + } + + enum SignBehavior(bool) { + #[default] + Automatic(false) = b"automatic", + Mandatory(true) = b"mandatory", + } + + enum SubsecondDigits { + One = b"1", + Two = b"2", + Three = b"3", + Four = b"4", + Five = b"5", + Six = b"6", + Seven = b"7", + Eight = b"8", + Nine = b"9", + #[default] + OneOrMore = b"1+", + } + + enum UnixTimestampPrecision { + #[default] + Second = b"second", + Millisecond = b"millisecond", + Microsecond = b"microsecond", + Nanosecond = b"nanosecond", + } + + enum WeekNumberRepr { + #[default] + Iso = b"iso", + Sunday = b"sunday", + Monday = b"monday", + } + + enum WeekdayCaseSensitive(bool) { + False(false) = b"false", + #[default] + True(true) = b"true", + } + + enum WeekdayOneIndexed(bool) { + False(false) = b"false", + #[default] + True(true) = b"true", + } + + enum WeekdayRepr { + Short = b"short", + #[default] + Long = b"long", + Sunday = b"sunday", + Monday = b"monday", + } + + enum YearBase(bool) { + #[default] + Calendar(false) = b"calendar", + IsoWeek(true) = b"iso_week", + } + + enum YearRepr { + #[default] + Full = b"full", + LastTwo = b"last_two", + } +} + +fn parse_from_modifier_value<T: FromStr>(value: &Spanned<&[u8]>) -> Result<Option<T>, Error> { + str::from_utf8(value) + .ok() + .and_then(|val| val.parse::<T>().ok()) + .map(|val| Some(val)) + .ok_or_else(|| value.span.error("invalid modifier value")) +} diff --git a/third_party/rust/time-macros/src/format_description/lexer.rs b/third_party/rust/time-macros/src/format_description/lexer.rs new file mode 100644 index 0000000000..2c927cb94d --- /dev/null +++ b/third_party/rust/time-macros/src/format_description/lexer.rs @@ -0,0 +1,248 @@ +use core::iter; + +use super::{Error, Location, Spanned, SpannedValue}; + +pub(super) struct Lexed<I: Iterator> { + iter: core::iter::Peekable<I>, +} + +impl<I: Iterator> Iterator for Lexed<I> { + type Item = I::Item; + + fn next(&mut self) -> Option<Self::Item> { + self.iter.next() + } +} + +impl<'iter, 'token: 'iter, I: Iterator<Item = Result<Token<'token>, Error>> + 'iter> Lexed<I> { + pub(super) fn peek(&mut self) -> Option<&I::Item> { + self.iter.peek() + } + + pub(super) fn next_if_whitespace(&mut self) -> Option<Spanned<&'token [u8]>> { + if let Some(&Ok(Token::ComponentPart { + kind: ComponentKind::Whitespace, + value, + })) = self.peek() + { + self.next(); // consume + Some(value) + } else { + None + } + } + + pub(super) fn next_if_not_whitespace(&mut self) -> Option<Spanned<&'token [u8]>> { + if let Some(&Ok(Token::ComponentPart { + kind: ComponentKind::NotWhitespace, + value, + })) = self.peek() + { + self.next(); + Some(value) + } else { + None + } + } + + pub(super) fn next_if_opening_bracket(&mut self) -> Option<Location> { + if let Some(&Ok(Token::Bracket { + kind: BracketKind::Opening, + location, + })) = self.peek() + { + self.next(); + Some(location) + } else { + None + } + } + + pub(super) fn peek_closing_bracket(&'iter mut self) -> Option<&'iter Location> { + if let Some(Ok(Token::Bracket { + kind: BracketKind::Closing, + location, + })) = self.peek() + { + Some(location) + } else { + None + } + } + + pub(super) fn next_if_closing_bracket(&mut self) -> Option<Location> { + if let Some(&Ok(Token::Bracket { + kind: BracketKind::Closing, + location, + })) = self.peek() + { + self.next(); + Some(location) + } else { + None + } + } +} + +pub(super) enum Token<'a> { + Literal(Spanned<&'a [u8]>), + Bracket { + kind: BracketKind, + location: Location, + }, + ComponentPart { + kind: ComponentKind, + value: Spanned<&'a [u8]>, + }, +} + +pub(super) enum BracketKind { + Opening, + Closing, +} + +pub(super) enum ComponentKind { + #[allow(clippy::missing_docs_in_private_items)] + Whitespace, + #[allow(clippy::missing_docs_in_private_items)] + NotWhitespace, +} + +fn attach_location<'item>( + iter: impl Iterator<Item = &'item u8>, + proc_span: proc_macro::Span, +) -> impl Iterator<Item = (&'item u8, Location)> { + let mut byte_pos = 0; + + iter.map(move |byte| { + let location = Location { + byte: byte_pos, + proc_span, + }; + byte_pos += 1; + (byte, location) + }) +} + +#[allow(clippy::unused_peekable)] // false positive +pub(super) fn lex<const VERSION: u8>( + mut input: &[u8], + proc_span: proc_macro::Span, +) -> Lexed<impl Iterator<Item = Result<Token<'_>, Error>>> { + assert!(version!(1..=2)); + + let mut depth: u8 = 0; + let mut iter = attach_location(input.iter(), proc_span).peekable(); + let mut second_bracket_location = None; + + let iter = iter::from_fn(move || { + if version!(..=1) { + if let Some(location) = second_bracket_location.take() { + return Some(Ok(Token::Bracket { + kind: BracketKind::Opening, + location, + })); + } + } + + Some(Ok(match iter.next()? { + (b'\\', backslash_loc) if version!(2..) => match iter.next() { + Some((b'\\' | b'[' | b']', char_loc)) => { + let char = &input[1..2]; + input = &input[2..]; + if depth == 0 { + Token::Literal(char.spanned(backslash_loc.to(char_loc))) + } else { + Token::ComponentPart { + kind: ComponentKind::NotWhitespace, + value: char.spanned(backslash_loc.to(char_loc)), + } + } + } + Some((_, loc)) => { + return Some(Err(loc.error("invalid escape sequence"))); + } + None => { + return Some(Err(backslash_loc.error("unexpected end of input"))); + } + }, + (b'[', location) if version!(..=1) => { + if let Some((_, second_location)) = iter.next_if(|&(&byte, _)| byte == b'[') { + second_bracket_location = Some(second_location); + input = &input[2..]; + } else { + depth += 1; + input = &input[1..]; + } + + Token::Bracket { + kind: BracketKind::Opening, + location, + } + } + (b'[', location) => { + depth += 1; + input = &input[1..]; + + Token::Bracket { + kind: BracketKind::Opening, + location, + } + } + (b']', location) if depth > 0 => { + depth -= 1; + input = &input[1..]; + + Token::Bracket { + kind: BracketKind::Closing, + location, + } + } + (_, start_location) if depth == 0 => { + let mut bytes = 1; + let mut end_location = start_location; + + while let Some((_, location)) = + iter.next_if(|&(&byte, _)| !((version!(2..) && byte == b'\\') || byte == b'[')) + { + end_location = location; + bytes += 1; + } + + let value = &input[..bytes]; + input = &input[bytes..]; + + Token::Literal(value.spanned(start_location.to(end_location))) + } + (byte, start_location) => { + let mut bytes = 1; + let mut end_location = start_location; + let is_whitespace = byte.is_ascii_whitespace(); + + while let Some((_, location)) = iter.next_if(|&(byte, _)| { + !matches!(byte, b'\\' | b'[' | b']') + && is_whitespace == byte.is_ascii_whitespace() + }) { + end_location = location; + bytes += 1; + } + + let value = &input[..bytes]; + input = &input[bytes..]; + + Token::ComponentPart { + kind: if is_whitespace { + ComponentKind::Whitespace + } else { + ComponentKind::NotWhitespace + }, + value: value.spanned(start_location.to(end_location)), + } + } + })) + }); + + Lexed { + iter: iter.peekable(), + } +} diff --git a/third_party/rust/time-macros/src/format_description/mod.rs b/third_party/rust/time-macros/src/format_description/mod.rs new file mode 100644 index 0000000000..fde1272f6a --- /dev/null +++ b/third_party/rust/time-macros/src/format_description/mod.rs @@ -0,0 +1,171 @@ +//! Parser for format descriptions. + +use std::vec::Vec; + +macro_rules! version { + ($range:expr) => { + $range.contains(&VERSION) + }; +} + +mod ast; +mod format_item; +mod lexer; +mod public; + +pub(crate) fn parse_with_version( + version: Option<crate::FormatDescriptionVersion>, + s: &[u8], + proc_span: proc_macro::Span, +) -> Result<Vec<crate::format_description::public::OwnedFormatItem>, crate::Error> { + match version { + Some(crate::FormatDescriptionVersion::V1) | None => parse::<1>(s, proc_span), + Some(crate::FormatDescriptionVersion::V2) => parse::<2>(s, proc_span), + } +} + +fn parse<const VERSION: u8>( + s: &[u8], + proc_span: proc_macro::Span, +) -> Result<Vec<crate::format_description::public::OwnedFormatItem>, crate::Error> { + let mut lexed = lexer::lex::<VERSION>(s, proc_span); + let ast = ast::parse::<_, VERSION>(&mut lexed); + let format_items = format_item::parse(ast); + Ok(format_items + .map(|res| res.map(Into::into)) + .collect::<Result<_, _>>()?) +} + +#[derive(Clone, Copy)] +struct Location { + byte: u32, + proc_span: proc_macro::Span, +} + +impl Location { + fn to(self, end: Self) -> Span { + Span { start: self, end } + } + + #[must_use = "this does not modify the original value"] + fn offset(&self, offset: u32) -> Self { + Self { + byte: self.byte + offset, + proc_span: self.proc_span, + } + } + + fn error(self, message: &'static str) -> Error { + Error { + message, + _span: unused(Span { + start: self, + end: self, + }), + proc_span: self.proc_span, + } + } +} + +#[derive(Clone, Copy)] +struct Span { + #[allow(clippy::missing_docs_in_private_items)] + start: Location, + #[allow(clippy::missing_docs_in_private_items)] + end: Location, +} + +impl Span { + #[must_use = "this does not modify the original value"] + const fn shrink_to_start(&self) -> Self { + Self { + start: self.start, + end: self.start, + } + } + + #[must_use = "this does not modify the original value"] + const fn shrink_to_end(&self) -> Self { + Self { + start: self.end, + end: self.end, + } + } + + #[must_use = "this does not modify the original value"] + const fn shrink_to_before(&self, pos: u32) -> Self { + Self { + start: self.start, + end: Location { + byte: self.start.byte + pos - 1, + proc_span: self.start.proc_span, + }, + } + } + + #[must_use = "this does not modify the original value"] + fn shrink_to_after(&self, pos: u32) -> Self { + Self { + start: Location { + byte: self.start.byte + pos + 1, + proc_span: self.start.proc_span, + }, + end: self.end, + } + } + + fn error(self, message: &'static str) -> Error { + Error { + message, + _span: unused(self), + proc_span: self.start.proc_span, + } + } +} + +#[derive(Clone, Copy)] +struct Spanned<T> { + value: T, + span: Span, +} + +impl<T> core::ops::Deref for Spanned<T> { + type Target = T; + + fn deref(&self) -> &Self::Target { + &self.value + } +} + +trait SpannedValue: Sized { + fn spanned(self, span: Span) -> Spanned<Self>; +} + +impl<T> SpannedValue for T { + fn spanned(self, span: Span) -> Spanned<Self> { + Spanned { value: self, span } + } +} + +struct Error { + message: &'static str, + _span: Unused<Span>, + proc_span: proc_macro::Span, +} + +impl From<Error> for crate::Error { + fn from(error: Error) -> Self { + Self::Custom { + message: error.message.into(), + span_start: Some(error.proc_span), + span_end: Some(error.proc_span), + } + } +} + +struct Unused<T>(core::marker::PhantomData<T>); + +#[allow(clippy::missing_const_for_fn)] // false positive +fn unused<T>(_: T) -> Unused<T> { + Unused(core::marker::PhantomData) +} diff --git a/third_party/rust/time-macros/src/format_description/public/component.rs b/third_party/rust/time-macros/src/format_description/public/component.rs new file mode 100644 index 0000000000..4737c6ce5c --- /dev/null +++ b/third_party/rust/time-macros/src/format_description/public/component.rs @@ -0,0 +1,49 @@ +use proc_macro::{Ident, Span, TokenStream}; + +use super::modifier; +use crate::to_tokens::ToTokenStream; + +macro_rules! declare_component { + ($($name:ident)*) => { + pub(crate) enum Component {$( + $name(modifier::$name), + )*} + + impl ToTokenStream for Component { + fn append_to(self, ts: &mut TokenStream) { + let mut mts = TokenStream::new(); + + let component = match self {$( + Self::$name(modifier) => { + modifier.append_to(&mut mts); + stringify!($name) + } + )*}; + let component = Ident::new(component, Span::mixed_site()); + + quote_append! { ts + ::time::format_description::Component::#(component)(#S(mts)) + } + } + } + }; +} + +declare_component! { + Day + Month + Ordinal + Weekday + WeekNumber + Year + Hour + Minute + Period + Second + Subsecond + OffsetHour + OffsetMinute + OffsetSecond + Ignore + UnixTimestamp +} diff --git a/third_party/rust/time-macros/src/format_description/public/mod.rs b/third_party/rust/time-macros/src/format_description/public/mod.rs new file mode 100644 index 0000000000..ccb0b6e2a3 --- /dev/null +++ b/third_party/rust/time-macros/src/format_description/public/mod.rs @@ -0,0 +1,54 @@ +mod component; +pub(super) mod modifier; + +use proc_macro::{Literal, TokenStream}; + +pub(crate) use self::component::Component; +use crate::to_tokens::ToTokenStream; + +#[allow(variant_size_differences)] +pub(crate) enum OwnedFormatItem { + Literal(Box<[u8]>), + Component(Component), + Compound(Box<[Self]>), + Optional(Box<Self>), + First(Box<[Self]>), +} + +impl ToTokenStream for OwnedFormatItem { + fn append_to(self, ts: &mut TokenStream) { + match self { + Self::Literal(bytes) => quote_append! { ts + ::time::format_description::FormatItem::Literal { + 0: #(Literal::byte_string(bytes.as_ref())) + } + }, + Self::Component(component) => quote_append! { ts + ::time::format_description::FormatItem::Component { 0: #S(component) } + }, + Self::Compound(items) => { + let items = items + .into_vec() + .into_iter() + .map(|item| quote! { #S(item), }) + .collect::<TokenStream>(); + quote_append! { ts + ::time::format_description::FormatItem::Compound { 0: &[#S(items)] } + } + } + Self::Optional(item) => quote_append! {ts + ::time::format_description::FormatItem::Optional { 0: &#S(*item) } + }, + Self::First(items) => { + let items = items + .into_vec() + .into_iter() + .map(|item| quote! { #S(item), }) + .collect::<TokenStream>(); + quote_append! { ts + ::time::format_description::FormatItem::First { 0: &[#S(items)] } + } + } + } + } +} diff --git a/third_party/rust/time-macros/src/format_description/public/modifier.rs b/third_party/rust/time-macros/src/format_description/public/modifier.rs new file mode 100644 index 0000000000..e39c6bf552 --- /dev/null +++ b/third_party/rust/time-macros/src/format_description/public/modifier.rs @@ -0,0 +1,247 @@ +use std::num::NonZeroU16; + +use proc_macro::{Ident, Span, TokenStream, TokenTree}; + +use crate::to_tokens::{ToTokenStream, ToTokenTree}; + +macro_rules! to_tokens { + ( + $(#[$struct_attr:meta])* + $struct_vis:vis struct $struct_name:ident {$( + $(#[$field_attr:meta])* + $field_vis:vis $field_name:ident : $field_ty:ty + ),+ $(,)?} + ) => { + $(#[$struct_attr])* + $struct_vis struct $struct_name {$( + $(#[$field_attr])* + $field_vis $field_name: $field_ty + ),+} + + impl ToTokenTree for $struct_name { + fn into_token_tree(self) -> TokenTree { + let mut tokens = TokenStream::new(); + let Self {$($field_name),+} = self; + + quote_append! { tokens + let mut value = ::time::format_description::modifier::$struct_name::default(); + }; + $( + quote_append!(tokens value.$field_name =); + $field_name.append_to(&mut tokens); + quote_append!(tokens ;); + )+ + quote_append!(tokens value); + + proc_macro::TokenTree::Group(proc_macro::Group::new( + proc_macro::Delimiter::Brace, + tokens, + )) + } + } + }; + + ( + $(#[$enum_attr:meta])* + $enum_vis:vis enum $enum_name:ident {$( + $(#[$variant_attr:meta])* + $variant_name:ident + ),+ $(,)?} + ) => { + $(#[$enum_attr])* + $enum_vis enum $enum_name {$( + $(#[$variant_attr])* + $variant_name + ),+} + + impl ToTokenStream for $enum_name { + fn append_to(self, ts: &mut TokenStream) { + quote_append! { ts + ::time::format_description::modifier::$enum_name:: + }; + let name = match self { + $(Self::$variant_name => stringify!($variant_name)),+ + }; + ts.extend([TokenTree::Ident(Ident::new(name, Span::mixed_site()))]); + } + } + } +} + +to_tokens! { + pub(crate) struct Day { + pub(crate) padding: Padding, + } +} + +to_tokens! { + pub(crate) enum MonthRepr { + Numerical, + Long, + Short, + } +} + +to_tokens! { + pub(crate) struct Month { + pub(crate) padding: Padding, + pub(crate) repr: MonthRepr, + pub(crate) case_sensitive: bool, + } +} + +to_tokens! { + pub(crate) struct Ordinal { + pub(crate) padding: Padding, + } +} + +to_tokens! { + pub(crate) enum WeekdayRepr { + Short, + Long, + Sunday, + Monday, + } +} + +to_tokens! { + pub(crate) struct Weekday { + pub(crate) repr: WeekdayRepr, + pub(crate) one_indexed: bool, + pub(crate) case_sensitive: bool, + } +} + +to_tokens! { + pub(crate) enum WeekNumberRepr { + Iso, + Sunday, + Monday, + } +} + +to_tokens! { + pub(crate) struct WeekNumber { + pub(crate) padding: Padding, + pub(crate) repr: WeekNumberRepr, + } +} + +to_tokens! { + pub(crate) enum YearRepr { + Full, + LastTwo, + } +} + +to_tokens! { + pub(crate) struct Year { + pub(crate) padding: Padding, + pub(crate) repr: YearRepr, + pub(crate) iso_week_based: bool, + pub(crate) sign_is_mandatory: bool, + } +} + +to_tokens! { + pub(crate) struct Hour { + pub(crate) padding: Padding, + pub(crate) is_12_hour_clock: bool, + } +} + +to_tokens! { + pub(crate) struct Minute { + pub(crate) padding: Padding, + } +} + +to_tokens! { + pub(crate) struct Period { + pub(crate) is_uppercase: bool, + pub(crate) case_sensitive: bool, + } +} + +to_tokens! { + pub(crate) struct Second { + pub(crate) padding: Padding, + } +} + +to_tokens! { + pub(crate) enum SubsecondDigits { + One, + Two, + Three, + Four, + Five, + Six, + Seven, + Eight, + Nine, + OneOrMore, + } +} + +to_tokens! { + pub(crate) struct Subsecond { + pub(crate) digits: SubsecondDigits, + } +} + +to_tokens! { + pub(crate) struct OffsetHour { + pub(crate) sign_is_mandatory: bool, + pub(crate) padding: Padding, + } +} + +to_tokens! { + pub(crate) struct OffsetMinute { + pub(crate) padding: Padding, + } +} + +to_tokens! { + pub(crate) struct OffsetSecond { + pub(crate) padding: Padding, + } +} + +to_tokens! { + pub(crate) enum Padding { + Space, + Zero, + None, + } +} + +pub(crate) struct Ignore { + pub(crate) count: NonZeroU16, +} + +impl ToTokenTree for Ignore { + fn into_token_tree(self) -> TokenTree { + quote_group! {{ + ::time::format_description::modifier::Ignore::count(#(self.count)) + }} + } +} + +to_tokens! { + pub(crate) enum UnixTimestampPrecision { + Second, + Millisecond, + Microsecond, + Nanosecond, + } +} + +to_tokens! { + pub(crate) struct UnixTimestamp { + pub(crate) precision: UnixTimestampPrecision, + pub(crate) sign_is_mandatory: bool, + } +} |