feat: add display UI with splash screen, dashboard, and RSSI sparkline

Implement full SSD1306 OLED display support:
- Splash screen on boot (1.5s) with project name, version, board
- Dashboard showing radio state, config, counters, RSSI/SNR
- RSSI history sparkline (64-sample ring buffer, 2px-wide bars)
- DisplayOn/DisplayOff host commands routed via dedicated channel
- Pure render functions against generic DrawTarget (no hardware coupling)
- Display on/off control bypasses radio task entirely

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
This commit is contained in:
swaitsandClaude Opus 4.6 committed 2026-03-29 20:42:54 -06:00
1 parent 7064110293
commit 3102fc747a
17624 files changed
+4564928 -38

No files matched your search

+134
View File
@@ -0,0 +1,134 @@
use winnow::combinator::cut_err;
use winnow::combinator::delimited;
use winnow::combinator::opt;
use winnow::combinator::peek;
use winnow::combinator::separated;
use winnow::combinator::trace;
use crate::parser::trivia::ws_comment_newline;
use crate::parser::value::value;
use crate::{Array, Item, RawString};
use crate::parser::prelude::*;
// ;; Array
// array = array-open array-values array-close
pub(crate) fn array<'i>(input: &mut Input<'i>) -> ModalResult<Array> {
trace("array", move |input: &mut Input<'i>| {
delimited(
ARRAY_OPEN,
cut_err(array_values),
cut_err(ARRAY_CLOSE)
.context(StrContext::Label("array"))
.context(StrContext::Expected(StrContextValue::CharLiteral(']'))),
)
.parse_next(input)
})
.parse_next(input)
}
// note: we're omitting ws and newlines here, because
// they should be part of the formatted values
// array-open = %x5B ws-newline ; [
pub(crate) const ARRAY_OPEN: u8 = b'[';
// array-close = ws-newline %x5D ; ]
const ARRAY_CLOSE: u8 = b']';
// array-sep = ws %x2C ws ; , Comma
const ARRAY_SEP: u8 = b',';
// array-values = ws-comment-newline val ws-comment-newline array-sep array-values
// array-values =/ ws-comment-newline val ws-comment-newline [ array-sep ]
fn array_values(input: &mut Input<'_>) -> ModalResult<Array> {
if peek(opt(ARRAY_CLOSE)).parse_next(input)?.is_some() {
// Optimize for empty arrays, avoiding `value` from being expected to fail
return Ok(Array::new());
}
let array = separated(0.., array_value, ARRAY_SEP).parse_next(input)?;
let mut array = Array::with_vec(array);
if !array.is_empty() {
let comma = opt(ARRAY_SEP).parse_next(input)?.is_some();
array.set_trailing_comma(comma);
}
let trailing = ws_comment_newline.span().parse_next(input)?;
array.set_trailing(RawString::with_span(trailing));
Ok(array)
}
fn array_value(input: &mut Input<'_>) -> ModalResult<Item> {
let prefix = ws_comment_newline.span().parse_next(input)?;
let value = value.parse_next(input)?;
let suffix = ws_comment_newline.span().parse_next(input)?;
let value = value.decorated(RawString::with_span(prefix), RawString::with_span(suffix));
let value = Item::Value(value);
Ok(value)
}
#[cfg(test)]
#[cfg(feature = "parse")]
#[cfg(feature = "display")]
mod test {
use super::*;
#[test]
fn arrays() {
let inputs = [
r#"[]"#,
r#"[ ]"#,
r#"[
1, 2, 3
]"#,
r#"[
1,
2, # this is ok
]"#,
r#"[# comment
# comment2
]"#,
r#"[# comment
# comment2
1
#sd
,
# comment3
]"#,
r#"[1]"#,
r#"[1,]"#,
r#"[ "all", 'strings', """are the same""", '''type''']"#,
r#"[ 100, -2,]"#,
r#"[1, 2, 3]"#,
r#"[1.1, 2.1, 3.1]"#,
r#"["a", "b", "c"]"#,
r#"[ [ 1, 2 ], [3, 4, 5] ]"#,
r#"[ [ 1, 2 ], ["a", "b", "c"] ]"#,
r#"[ { x = 1, a = "2" }, {a = "a",b = "b", c = "c"} ]"#,
];
for input in inputs {
dbg!(input);
let mut parsed = array.parse(new_input(input));
if let Ok(parsed) = &mut parsed {
parsed.despan(input);
}
assert_eq!(parsed.map(|a| a.to_string()), Ok(input.to_owned()));
}
}
#[test]
fn invalid_arrays() {
let invalid_inputs = [r#"["#, r#"[,]"#, r#"[,2]"#, r#"[1e165,,]"#];
for input in invalid_inputs {
dbg!(input);
let mut parsed = array.parse(new_input(input));
if let Ok(parsed) = &mut parsed {
parsed.despan(input);
}
assert!(parsed.is_err());
}
}
}
+469
View File
@@ -0,0 +1,469 @@
use std::ops::RangeInclusive;
use crate::parser::error::CustomError;
use crate::parser::prelude::*;
use crate::parser::trivia::from_utf8_unchecked;
use toml_datetime::{Date, Datetime, Offset, Time};
use winnow::combinator::alt;
use winnow::combinator::cut_err;
use winnow::combinator::opt;
use winnow::combinator::preceded;
use winnow::combinator::trace;
use winnow::stream::Stream as _;
use winnow::token::one_of;
use winnow::token::take_while;
// ;; Date and Time (as defined in RFC 3339)
// date-time = offset-date-time / local-date-time / local-date / local-time
// offset-date-time = full-date time-delim full-time
// local-date-time = full-date time-delim partial-time
// local-date = full-date
// local-time = partial-time
// full-time = partial-time time-offset
pub(crate) fn date_time(input: &mut Input<'_>) -> ModalResult<Datetime> {
trace(
"date-time",
alt((
(full_date, opt((time_delim, partial_time, opt(time_offset))))
.map(|(date, opt)| {
match opt {
// Offset Date-Time
Some((_, time, offset)) => Datetime {
date: Some(date),
time: Some(time),
offset,
},
// Local Date
None => Datetime {
date: Some(date),
time: None,
offset: None,
},
}
})
.context(StrContext::Label("date-time")),
partial_time
.map(|t| t.into())
.context(StrContext::Label("time")),
)),
)
.parse_next(input)
}
// full-date = date-fullyear "-" date-month "-" date-mday
fn full_date(input: &mut Input<'_>) -> ModalResult<Date> {
trace("full-date", full_date_).parse_next(input)
}
fn full_date_(input: &mut Input<'_>) -> ModalResult<Date> {
let year = date_fullyear.parse_next(input)?;
let _ = b'-'.parse_next(input)?;
let month = cut_err(date_month).parse_next(input)?;
let _ = cut_err(b'-').parse_next(input)?;
let day_start = input.checkpoint();
let day = cut_err(date_mday).parse_next(input)?;
let is_leap_year = (year % 4 == 0) && ((year % 100 != 0) || (year % 400 == 0));
let max_days_in_month = match month {
2 if is_leap_year => 29,
2 => 28,
4 | 6 | 9 | 11 => 30,
_ => 31,
};
if max_days_in_month < day {
input.reset(&day_start);
return Err(
winnow::error::ErrMode::from_external_error(input, CustomError::OutOfRange).cut(),
);
}
Ok(Date { year, month, day })
}
// partial-time = time-hour ":" time-minute ":" time-second [time-secfrac]
fn partial_time(input: &mut Input<'_>) -> ModalResult<Time> {
trace(
"partial-time",
(
time_hour,
b':',
cut_err((time_minute, b':', time_second, opt(time_secfrac))),
)
.map(|(hour, _, (minute, _, second, nanosecond))| Time {
hour,
minute,
second,
nanosecond: nanosecond.unwrap_or_default(),
}),
)
.parse_next(input)
}
// time-offset = "Z" / time-numoffset
// time-numoffset = ( "+" / "-" ) time-hour ":" time-minute
fn time_offset(input: &mut Input<'_>) -> ModalResult<Offset> {
trace(
"time-offset",
alt((
one_of((b'Z', b'z')).value(Offset::Z),
(
one_of((b'+', b'-')),
cut_err((time_hour, b':', time_minute)),
)
.map(|(sign, (hours, _, minutes))| {
let sign = match sign {
b'+' => 1,
b'-' => -1,
_ => unreachable!("Parser prevents this"),
};
sign * (hours as i16 * 60 + minutes as i16)
})
.verify(|minutes| ((-24 * 60)..=(24 * 60)).contains(minutes))
.map(|minutes| Offset::Custom { minutes }),
))
.context(StrContext::Label("time offset")),
)
.parse_next(input)
}
// date-fullyear = 4DIGIT
fn date_fullyear(input: &mut Input<'_>) -> ModalResult<u16> {
unsigned_digits::<4, 4>
.map(|s: &str| s.parse::<u16>().expect("4DIGIT should match u8"))
.parse_next(input)
}
// date-month = 2DIGIT ; 01-12
fn date_month(input: &mut Input<'_>) -> ModalResult<u8> {
unsigned_digits::<2, 2>
.try_map(|s: &str| {
let d = s.parse::<u8>().expect("2DIGIT should match u8");
if (1..=12).contains(&d) {
Ok(d)
} else {
Err(CustomError::OutOfRange)
}
})
.parse_next(input)
}
// date-mday = 2DIGIT ; 01-28, 01-29, 01-30, 01-31 based on month/year
fn date_mday(input: &mut Input<'_>) -> ModalResult<u8> {
unsigned_digits::<2, 2>
.try_map(|s: &str| {
let d = s.parse::<u8>().expect("2DIGIT should match u8");
if (1..=31).contains(&d) {
Ok(d)
} else {
Err(CustomError::OutOfRange)
}
})
.parse_next(input)
}
// time-delim = "T" / %x20 ; T, t, or space
fn time_delim(input: &mut Input<'_>) -> ModalResult<u8> {
one_of(TIME_DELIM).parse_next(input)
}
const TIME_DELIM: (u8, u8, u8) = (b'T', b't', b' ');
// time-hour = 2DIGIT ; 00-23
fn time_hour(input: &mut Input<'_>) -> ModalResult<u8> {
unsigned_digits::<2, 2>
.try_map(|s: &str| {
let d = s.parse::<u8>().expect("2DIGIT should match u8");
if (0..=23).contains(&d) {
Ok(d)
} else {
Err(CustomError::OutOfRange)
}
})
.parse_next(input)
}
// time-minute = 2DIGIT ; 00-59
fn time_minute(input: &mut Input<'_>) -> ModalResult<u8> {
unsigned_digits::<2, 2>
.try_map(|s: &str| {
let d = s.parse::<u8>().expect("2DIGIT should match u8");
if (0..=59).contains(&d) {
Ok(d)
} else {
Err(CustomError::OutOfRange)
}
})
.parse_next(input)
}
// time-second = 2DIGIT ; 00-58, 00-59, 00-60 based on leap second rules
fn time_second(input: &mut Input<'_>) -> ModalResult<u8> {
unsigned_digits::<2, 2>
.try_map(|s: &str| {
let d = s.parse::<u8>().expect("2DIGIT should match u8");
if (0..=60).contains(&d) {
Ok(d)
} else {
Err(CustomError::OutOfRange)
}
})
.parse_next(input)
}
// time-secfrac = "." 1*DIGIT
fn time_secfrac(input: &mut Input<'_>) -> ModalResult<u32> {
static SCALE: [u32; 10] = [
0,
100_000_000,
10_000_000,
1_000_000,
100_000,
10_000,
1_000,
100,
10,
1,
];
const INF: usize = usize::MAX;
preceded(b'.', unsigned_digits::<1, INF>)
.try_map(|mut repr: &str| -> Result<u32, CustomError> {
let max_digits = SCALE.len() - 1;
if max_digits < repr.len() {
// Millisecond precision is required. Further precision of fractional seconds is
// implementation-specific. If the value contains greater precision than the
// implementation can support, the additional precision must be truncated, not rounded.
repr = &repr[0..max_digits];
}
let v = repr.parse::<u32>().map_err(|_| CustomError::OutOfRange)?;
let num_digits = repr.len();
// scale the number accordingly.
let scale = SCALE.get(num_digits).ok_or(CustomError::OutOfRange)?;
let v = v.checked_mul(*scale).ok_or(CustomError::OutOfRange)?;
Ok(v)
})
.parse_next(input)
}
fn unsigned_digits<'i, const MIN: usize, const MAX: usize>(
input: &mut Input<'i>,
) -> ModalResult<&'i str> {
take_while(MIN..=MAX, DIGIT)
.map(|b: &[u8]| unsafe { from_utf8_unchecked(b, "`is_ascii_digit` filters out on-ASCII") })
.parse_next(input)
}
// DIGIT = %x30-39 ; 0-9
const DIGIT: RangeInclusive<u8> = b'0'..=b'9';
#[cfg(test)]
#[cfg(feature = "parse")]
#[cfg(feature = "display")]
mod test {
use super::*;
#[test]
fn offset_date_time() {
let inputs = [
(
"1979-05-27T07:32:00Z",
Datetime {
date: Some(Date {
year: 1979,
month: 5,
day: 27,
}),
time: Some(Time {
hour: 7,
minute: 32,
second: 0,
nanosecond: 0,
}),
offset: Some(Offset::Z),
},
),
(
"1979-05-27T00:32:00-07:00",
Datetime {
date: Some(Date {
year: 1979,
month: 5,
day: 27,
}),
time: Some(Time {
hour: 0,
minute: 32,
second: 0,
nanosecond: 0,
}),
offset: Some(Offset::Custom { minutes: -7 * 60 }),
},
),
(
"1979-05-27T00:32:00-00:36",
Datetime {
date: Some(Date {
year: 1979,
month: 5,
day: 27,
}),
time: Some(Time {
hour: 0,
minute: 32,
second: 0,
nanosecond: 0,
}),
offset: Some(Offset::Custom { minutes: -36 }),
},
),
(
"1979-05-27T00:32:00.999999",
Datetime {
date: Some(Date {
year: 1979,
month: 5,
day: 27,
}),
time: Some(Time {
hour: 0,
minute: 32,
second: 0,
nanosecond: 999999000,
}),
offset: None,
},
),
];
for (input, expected) in inputs {
dbg!(input);
let actual = date_time.parse(new_input(input)).unwrap();
assert_eq!(expected, actual);
}
}
#[test]
fn local_date_time() {
let inputs = [
(
"1979-05-27T07:32:00",
Datetime {
date: Some(Date {
year: 1979,
month: 5,
day: 27,
}),
time: Some(Time {
hour: 7,
minute: 32,
second: 0,
nanosecond: 0,
}),
offset: None,
},
),
(
"1979-05-27T00:32:00.999999",
Datetime {
date: Some(Date {
year: 1979,
month: 5,
day: 27,
}),
time: Some(Time {
hour: 0,
minute: 32,
second: 0,
nanosecond: 999999000,
}),
offset: None,
},
),
];
for (input, expected) in inputs {
dbg!(input);
let actual = date_time.parse(new_input(input)).unwrap();
assert_eq!(expected, actual);
}
}
#[test]
fn local_date() {
let inputs = [
(
"1979-05-27",
Datetime {
date: Some(Date {
year: 1979,
month: 5,
day: 27,
}),
time: None,
offset: None,
},
),
(
"2017-07-20",
Datetime {
date: Some(Date {
year: 2017,
month: 7,
day: 20,
}),
time: None,
offset: None,
},
),
];
for (input, expected) in inputs {
dbg!(input);
let actual = date_time.parse(new_input(input)).unwrap();
assert_eq!(expected, actual);
}
}
#[test]
fn local_time() {
let inputs = [
(
"07:32:00",
Datetime {
date: None,
time: Some(Time {
hour: 7,
minute: 32,
second: 0,
nanosecond: 0,
}),
offset: None,
},
),
(
"00:32:00.999999",
Datetime {
date: None,
time: Some(Time {
hour: 0,
minute: 32,
second: 0,
nanosecond: 999999000,
}),
offset: None,
},
),
];
for (input, expected) in inputs {
dbg!(input);
let actual = date_time.parse(new_input(input)).unwrap();
assert_eq!(expected, actual);
}
}
#[test]
fn time_fraction_truncated() {
let input = "1987-07-05T17:45:00.123456789012345Z";
date_time.parse(new_input(input)).unwrap();
}
}
+130
View File
@@ -0,0 +1,130 @@
use std::cell::RefCell;
use winnow::combinator::cut_err;
use winnow::combinator::eof;
use winnow::combinator::opt;
use winnow::combinator::peek;
use winnow::combinator::repeat;
use winnow::combinator::trace;
use winnow::token::any;
use winnow::token::one_of;
use crate::key::Key;
use crate::parser::inline_table::KEYVAL_SEP;
use crate::parser::key::key;
use crate::parser::prelude::*;
use crate::parser::state::ParseState;
use crate::parser::table::table;
use crate::parser::trivia::{comment, line_ending, line_trailing, newline, ws};
use crate::parser::value::value;
use crate::Item;
use crate::RawString;
// ;; TOML
// toml = expression *( newline expression )
// expression = ( ( ws comment ) /
// ( ws keyval ws [ comment ] ) /
// ( ws table ws [ comment ] ) /
// ws )
pub(crate) fn document<'s, 'i>(
state_ref: &'s RefCell<ParseState>,
) -> impl ModalParser<Input<'i>, (), ContextError> + 's {
move |i: &mut Input<'i>| {
(
// Remove BOM if present
opt(b"\xEF\xBB\xBF"),
parse_ws(state_ref),
repeat(0.., (
dispatch! {peek(any);
crate::parser::trivia::COMMENT_START_SYMBOL => cut_err(parse_comment(state_ref)),
crate::parser::table::STD_TABLE_OPEN => cut_err(table(state_ref)),
crate::parser::trivia::LF |
crate::parser::trivia::CR => parse_newline(state_ref),
_ => cut_err(keyval(state_ref)),
},
parse_ws(state_ref),
))
.map(|()| ()),
eof,
).void().parse_next(i)
}
}
fn parse_comment<'s, 'i>(
state: &'s RefCell<ParseState>,
) -> impl ModalParser<Input<'i>, (), ContextError> + 's {
move |i: &mut Input<'i>| {
(comment, line_ending)
.span()
.map(|span| {
state.borrow_mut().on_comment(span);
})
.parse_next(i)
}
}
fn parse_ws<'s, 'i>(
state: &'s RefCell<ParseState>,
) -> impl ModalParser<Input<'i>, (), ContextError> + 's {
move |i: &mut Input<'i>| {
ws.span()
.map(|span| state.borrow_mut().on_ws(span))
.parse_next(i)
}
}
fn parse_newline<'s, 'i>(
state: &'s RefCell<ParseState>,
) -> impl ModalParser<Input<'i>, (), ContextError> + 's {
move |i: &mut Input<'i>| {
newline
.span()
.map(|span| state.borrow_mut().on_ws(span))
.parse_next(i)
}
}
fn keyval<'s, 'i>(
state: &'s RefCell<ParseState>,
) -> impl ModalParser<Input<'i>, (), ContextError> + 's {
move |i: &mut Input<'i>| {
parse_keyval
.try_map(|(p, kv)| state.borrow_mut().on_keyval(p, kv))
.parse_next(i)
}
}
// keyval = key keyval-sep val
fn parse_keyval(input: &mut Input<'_>) -> ModalResult<(Vec<Key>, (Key, Item))> {
trace(
"keyval",
(
key,
cut_err((
one_of(KEYVAL_SEP)
.context(StrContext::Expected(StrContextValue::CharLiteral('.')))
.context(StrContext::Expected(StrContextValue::CharLiteral('='))),
(
ws.span(),
value,
line_trailing
.context(StrContext::Expected(StrContextValue::CharLiteral('\n')))
.context(StrContext::Expected(StrContextValue::CharLiteral('#'))),
),
)),
)
.try_map::<_, _, std::str::Utf8Error>(|(key, (_, v))| {
let mut path = key;
let key = path.pop().expect("grammar ensures at least 1");
let (pre, v, suf) = v;
let pre = RawString::with_span(pre);
let suf = RawString::with_span(suf);
let v = v.decorated(pre, suf);
Ok((path, (key, Item::Value(v))))
}),
)
.parse_next(input)
}
+86
View File
@@ -0,0 +1,86 @@
use std::error::Error as StdError;
use std::fmt::{Display, Formatter, Result};
use crate::Key;
#[derive(Debug, Clone)]
pub(crate) enum CustomError {
DuplicateKey {
key: String,
table: Option<Vec<Key>>,
},
DottedKeyExtendWrongType {
key: Vec<Key>,
actual: &'static str,
},
OutOfRange,
#[cfg_attr(feature = "unbounded", allow(dead_code))]
RecursionLimitExceeded,
}
impl CustomError {
pub(crate) fn duplicate_key(path: &[Key], i: usize) -> Self {
assert!(i < path.len());
let key = &path[i];
let repr = key
.as_repr()
.and_then(|key| key.as_raw().as_str())
.map(|s| s.to_owned())
.unwrap_or_else(|| {
#[cfg(feature = "display")]
{
key.default_repr().as_raw().as_str().unwrap().to_owned()
}
#[cfg(not(feature = "display"))]
{
format!("{:?}", key.get())
}
});
Self::DuplicateKey {
key: repr,
table: Some(path[..i].to_vec()),
}
}
pub(crate) fn extend_wrong_type(path: &[Key], i: usize, actual: &'static str) -> Self {
assert!(i < path.len());
Self::DottedKeyExtendWrongType {
key: path[..=i].to_vec(),
actual,
}
}
}
impl StdError for CustomError {
fn description(&self) -> &'static str {
"TOML parse error"
}
}
impl Display for CustomError {
fn fmt(&self, f: &mut Formatter<'_>) -> Result {
match self {
CustomError::DuplicateKey { key, table } => {
if let Some(table) = table {
if table.is_empty() {
write!(f, "duplicate key `{key}` in document root")
} else {
let path = table.iter().map(|k| k.get()).collect::<Vec<_>>().join(".");
write!(f, "duplicate key `{key}` in table `{path}`")
}
} else {
write!(f, "duplicate key `{key}`")
}
}
CustomError::DottedKeyExtendWrongType { key, actual } => {
let path = key.iter().map(|k| k.get()).collect::<Vec<_>>().join(".");
write!(
f,
"dotted key `{path}` attempted to extend non-table type ({actual})"
)
}
CustomError::OutOfRange => write!(f, "value is out of range"),
CustomError::RecursionLimitExceeded => write!(f, "recursion limit exceeded"),
}
}
}
+195
View File
@@ -0,0 +1,195 @@
use winnow::combinator::cut_err;
use winnow::combinator::delimited;
use winnow::combinator::separated;
use winnow::combinator::trace;
use winnow::token::one_of;
use crate::key::Key;
use crate::parser::error::CustomError;
use crate::parser::key::key;
use crate::parser::prelude::*;
use crate::parser::trivia::ws;
use crate::parser::value::value;
use crate::{InlineTable, Item, RawString, Value};
use indexmap::map::Entry;
// ;; Inline Table
// inline-table = inline-table-open inline-table-keyvals inline-table-close
pub(crate) fn inline_table<'i>(input: &mut Input<'i>) -> ModalResult<InlineTable> {
trace("inline-table", move |input: &mut Input<'i>| {
delimited(
INLINE_TABLE_OPEN,
cut_err(inline_table_keyvals.try_map(|(kv, p)| table_from_pairs(kv, p))),
cut_err(INLINE_TABLE_CLOSE)
.context(StrContext::Label("inline table"))
.context(StrContext::Expected(StrContextValue::CharLiteral('}'))),
)
.parse_next(input)
})
.parse_next(input)
}
fn table_from_pairs(
v: Vec<(Vec<Key>, (Key, Item))>,
preamble: RawString,
) -> Result<InlineTable, CustomError> {
let mut root = InlineTable::new();
root.set_preamble(preamble);
// Assuming almost all pairs will be directly in `root`
root.items.reserve(v.len());
for (path, (key, value)) in v {
let table = descend_path(&mut root, &path, true)?;
// "Likewise, using dotted keys to redefine tables already defined in [table] form is not allowed"
let mixed_table_types = table.is_dotted() == path.is_empty();
if mixed_table_types {
return Err(CustomError::DuplicateKey {
key: key.get().into(),
table: None,
});
}
match table.items.entry(key) {
Entry::Vacant(o) => {
o.insert(value);
}
Entry::Occupied(o) => {
return Err(CustomError::DuplicateKey {
key: o.key().get().into(),
table: None,
});
}
}
}
Ok(root)
}
fn descend_path<'a>(
mut table: &'a mut InlineTable,
path: &'a [Key],
dotted: bool,
) -> Result<&'a mut InlineTable, CustomError> {
for (i, key) in path.iter().enumerate() {
table = match table.entry_format(key) {
crate::InlineEntry::Vacant(entry) => {
let mut new_table = InlineTable::new();
new_table.set_implicit(true);
new_table.set_dotted(dotted);
entry
.insert(Value::InlineTable(new_table))
.as_inline_table_mut()
.unwrap()
}
crate::InlineEntry::Occupied(entry) => {
match entry.into_mut() {
Value::InlineTable(ref mut sweet_child_of_mine) => {
// Since tables cannot be defined more than once, redefining such tables using a
// [table] header is not allowed. Likewise, using dotted keys to redefine tables
// already defined in [table] form is not allowed.
if dotted && !sweet_child_of_mine.is_implicit() {
return Err(CustomError::DuplicateKey {
key: key.get().into(),
table: None,
});
}
sweet_child_of_mine
}
ref v => {
return Err(CustomError::extend_wrong_type(path, i, v.type_name()));
}
}
}
};
}
Ok(table)
}
// inline-table-open = %x7B ws ; {
pub(crate) const INLINE_TABLE_OPEN: u8 = b'{';
// inline-table-close = ws %x7D ; }
const INLINE_TABLE_CLOSE: u8 = b'}';
// inline-table-sep = ws %x2C ws ; , Comma
const INLINE_TABLE_SEP: u8 = b',';
// keyval-sep = ws %x3D ws ; =
pub(crate) const KEYVAL_SEP: u8 = b'=';
// inline-table-keyvals = [ inline-table-keyvals-non-empty ]
// inline-table-keyvals-non-empty =
// ( key keyval-sep val inline-table-sep inline-table-keyvals-non-empty ) /
// ( key keyval-sep val )
fn inline_table_keyvals(
input: &mut Input<'_>,
) -> ModalResult<(Vec<(Vec<Key>, (Key, Item))>, RawString)> {
(
separated(0.., keyval, INLINE_TABLE_SEP),
ws.span().map(RawString::with_span),
)
.parse_next(input)
}
fn keyval(input: &mut Input<'_>) -> ModalResult<(Vec<Key>, (Key, Item))> {
(
key,
cut_err((
one_of(KEYVAL_SEP)
.context(StrContext::Expected(StrContextValue::CharLiteral('.')))
.context(StrContext::Expected(StrContextValue::CharLiteral('='))),
(ws.span(), value, ws.span()),
)),
)
.map(|(key, (_, v))| {
let mut path = key;
let key = path.pop().expect("grammar ensures at least 1");
let (pre, v, suf) = v;
let pre = RawString::with_span(pre);
let suf = RawString::with_span(suf);
let v = v.decorated(pre, suf);
(path, (key, Item::Value(v)))
})
.parse_next(input)
}
#[cfg(test)]
#[cfg(feature = "parse")]
#[cfg(feature = "display")]
mod test {
use super::*;
#[test]
fn inline_tables() {
let inputs = [
r#"{}"#,
r#"{ }"#,
r#"{a = 1e165}"#,
r#"{ hello = "world", a = 1}"#,
r#"{ hello.world = "a" }"#,
];
for input in inputs {
dbg!(input);
let mut parsed = inline_table.parse(new_input(input));
if let Ok(parsed) = &mut parsed {
parsed.despan(input);
}
assert_eq!(parsed.map(|a| a.to_string()), Ok(input.to_owned()));
}
}
#[test]
fn invalid_inline_tables() {
let invalid_inputs = [r#"{a = 1e165"#, r#"{ hello = "world", a = 2, hello = 1}"#];
for input in invalid_inputs {
dbg!(input);
let mut parsed = inline_table.parse(new_input(input));
if let Ok(parsed) = &mut parsed {
parsed.despan(input);
}
assert!(parsed.is_err());
}
}
}
+134
View File
@@ -0,0 +1,134 @@
use std::ops::RangeInclusive;
use winnow::combinator::peek;
use winnow::combinator::separated;
use winnow::combinator::trace;
use winnow::token::any;
use winnow::token::take_while;
use crate::key::Key;
use crate::parser::error::CustomError;
use crate::parser::prelude::*;
use crate::parser::strings::{basic_string, literal_string};
use crate::parser::trivia::{from_utf8_unchecked, ws};
use crate::repr::{Decor, Repr};
use crate::InternalString;
use crate::RawString;
// key = simple-key / dotted-key
// dotted-key = simple-key 1*( dot-sep simple-key )
pub(crate) fn key(input: &mut Input<'_>) -> ModalResult<Vec<Key>> {
let mut key_path = trace(
"dotted-key",
separated(
1..,
(ws.span(), simple_key, ws.span()).map(|(pre, (raw, key), suffix)| {
Key::new(key)
.with_repr_unchecked(Repr::new_unchecked(raw))
.with_dotted_decor(Decor::new(
RawString::with_span(pre),
RawString::with_span(suffix),
))
}),
DOT_SEP,
)
.context(StrContext::Label("key"))
.try_map(|k: Vec<_>| {
// Inserting the key will require recursion down the line
RecursionCheck::check_depth(k.len())?;
Ok::<_, CustomError>(k)
}),
)
.parse_next(input)?;
let mut leaf_decor = Decor::new("", "");
{
let first_dotted_decor = key_path
.first_mut()
.expect("always at least one key")
.dotted_decor_mut();
if let Some(prefix) = first_dotted_decor.prefix().cloned() {
leaf_decor.set_prefix(prefix);
first_dotted_decor.set_prefix("");
}
}
let last_key = &mut key_path.last_mut().expect("always at least one key");
{
let last_dotted_decor = last_key.dotted_decor_mut();
if let Some(suffix) = last_dotted_decor.suffix().cloned() {
leaf_decor.set_suffix(suffix);
last_dotted_decor.set_suffix("");
}
}
*last_key.leaf_decor_mut() = leaf_decor;
Ok(key_path)
}
// simple-key = quoted-key / unquoted-key
// quoted-key = basic-string / literal-string
pub(crate) fn simple_key(input: &mut Input<'_>) -> ModalResult<(RawString, InternalString)> {
trace(
"simple-key",
dispatch! {peek(any);
crate::parser::strings::QUOTATION_MARK => basic_string
.map(|s: std::borrow::Cow<'_, str>| s.as_ref().into()),
crate::parser::strings::APOSTROPHE => literal_string.map(|s: &str| s.into()),
_ => unquoted_key.map(|s: &str| s.into()),
}
.with_span()
.map(|(k, span)| {
let raw = RawString::with_span(span);
(raw, k)
}),
)
.parse_next(input)
}
// unquoted-key = 1*( ALPHA / DIGIT / %x2D / %x5F ) ; A-Z / a-z / 0-9 / - / _
fn unquoted_key<'i>(input: &mut Input<'i>) -> ModalResult<&'i str> {
trace(
"unquoted-key",
take_while(1.., UNQUOTED_CHAR)
.map(|b| unsafe { from_utf8_unchecked(b, "`is_unquoted_char` filters out on-ASCII") }),
)
.parse_next(input)
}
const UNQUOTED_CHAR: (
RangeInclusive<u8>,
RangeInclusive<u8>,
RangeInclusive<u8>,
u8,
u8,
) = (b'A'..=b'Z', b'a'..=b'z', b'0'..=b'9', b'-', b'_');
// dot-sep = ws %x2E ws ; . Period
const DOT_SEP: u8 = b'.';
#[cfg(test)]
#[cfg(feature = "parse")]
#[cfg(feature = "display")]
mod test {
use super::*;
#[test]
fn keys() {
let cases = [
("a", "a"),
(r#""hello\n ""#, "hello\n "),
(r"'hello\n '", "hello\\n "),
];
for (input, expected) in cases {
dbg!(input);
let parsed = simple_key.parse(new_input(input));
assert_eq!(
parsed,
Ok((RawString::with_span(0..(input.len())), expected.into())),
"Parsing {input:?}"
);
}
}
}
+255
View File
@@ -0,0 +1,255 @@
#![allow(clippy::type_complexity)]
use std::cell::RefCell;
pub(crate) mod array;
pub(crate) mod datetime;
pub(crate) mod document;
pub(crate) mod error;
pub(crate) mod inline_table;
pub(crate) mod key;
pub(crate) mod numbers;
pub(crate) mod state;
pub(crate) mod strings;
pub(crate) mod table;
pub(crate) mod trivia;
pub(crate) mod value;
pub(crate) use crate::error::TomlError;
pub(crate) fn parse_document<S: AsRef<str>>(raw: S) -> Result<crate::ImDocument<S>, TomlError> {
use prelude::*;
let b = new_input(raw.as_ref());
let state = RefCell::new(state::ParseState::new());
let state_ref = &state;
document::document(state_ref)
.parse(b.clone())
.map_err(|e| TomlError::new(e, b))?;
let doc = state
.into_inner()
.into_document(raw)
.map_err(|e| TomlError::custom(e.to_string(), None))?;
Ok(doc)
}
pub(crate) fn parse_key(raw: &str) -> Result<crate::Key, TomlError> {
use prelude::*;
let b = new_input(raw);
let result = key::simple_key.parse(b.clone());
match result {
Ok((raw, key)) => {
Ok(crate::Key::new(key).with_repr_unchecked(crate::Repr::new_unchecked(raw)))
}
Err(e) => Err(TomlError::new(e, b)),
}
}
pub(crate) fn parse_key_path(raw: &str) -> Result<Vec<crate::Key>, TomlError> {
use prelude::*;
let b = new_input(raw);
let result = key::key.parse(b.clone());
match result {
Ok(keys) => Ok(keys),
Err(e) => Err(TomlError::new(e, b)),
}
}
pub(crate) fn parse_value(raw: &str) -> Result<crate::Value, TomlError> {
use prelude::*;
let b = new_input(raw);
let parsed = value::value.parse(b.clone());
match parsed {
Ok(value) => Ok(value),
Err(e) => Err(TomlError::new(e, b)),
}
}
pub(crate) mod prelude {
pub(crate) use winnow::combinator::dispatch;
pub(crate) use winnow::error::ContextError;
pub(crate) use winnow::error::FromExternalError;
pub(crate) use winnow::error::StrContext;
pub(crate) use winnow::error::StrContextValue;
pub(crate) use winnow::ModalParser;
pub(crate) use winnow::ModalResult;
pub(crate) use winnow::Parser as _;
pub(crate) type Input<'b> =
winnow::Stateful<winnow::LocatingSlice<&'b winnow::BStr>, RecursionCheck>;
pub(crate) fn new_input(s: &str) -> Input<'_> {
winnow::Stateful {
input: winnow::LocatingSlice::new(winnow::BStr::new(s)),
state: Default::default(),
}
}
#[derive(Clone, Debug, Default, PartialEq, Eq)]
pub(crate) struct RecursionCheck {
#[cfg(not(feature = "unbounded"))]
current: usize,
}
#[cfg(not(feature = "unbounded"))]
const LIMIT: usize = 80;
impl RecursionCheck {
pub(crate) fn check_depth(_depth: usize) -> Result<(), super::error::CustomError> {
#[cfg(not(feature = "unbounded"))]
if LIMIT <= _depth {
return Err(super::error::CustomError::RecursionLimitExceeded);
}
Ok(())
}
fn enter(&mut self) -> Result<(), super::error::CustomError> {
#[cfg(not(feature = "unbounded"))]
{
self.current += 1;
if LIMIT <= self.current {
return Err(super::error::CustomError::RecursionLimitExceeded);
}
}
Ok(())
}
fn exit(&mut self) {
#[cfg(not(feature = "unbounded"))]
{
self.current -= 1;
}
}
}
pub(crate) fn check_recursion<'b, O>(
mut parser: impl ModalParser<Input<'b>, O, ContextError>,
) -> impl ModalParser<Input<'b>, O, ContextError> {
move |input: &mut Input<'b>| {
input
.state
.enter()
.map_err(|err| winnow::error::ErrMode::from_external_error(input, err).cut())?;
let result = parser.parse_next(input);
input.state.exit();
result
}
}
}
#[cfg(test)]
#[cfg(feature = "parse")]
#[cfg(feature = "display")]
mod test {
use super::*;
use snapbox::assert_data_eq;
use snapbox::prelude::*;
#[test]
fn documents() {
let documents = [
"",
r#"
# This is a TOML document.
title = "TOML Example"
[owner]
name = "Tom Preston-Werner"
dob = 1979-05-27T07:32:00-08:00 # First class dates
[database]
server = "192.168.1.1"
ports = [ 8001, 8001, 8002 ]
connection_max = 5000
enabled = true
[servers]
# Indentation (tabs and/or spaces) is allowed but not required
[servers.alpha]
ip = "10.0.0.1"
dc = "eqdc10"
[servers.beta]
ip = "10.0.0.2"
dc = "eqdc10"
[clients]
data = [ ["gamma", "delta"], [1, 2] ]
# Line breaks are OK when inside arrays
hosts = [
"alpha",
"omega"
]
'some.weird .stuff' = """
like
that
# """ # this broke my syntax highlighting
" also. like " = '''
that
'''
double = 2e39 # this number looks familiar
# trailing comment"#,
r#""#,
r#" "#,
r#" hello = 'darkness' # my old friend
"#,
r#"[parent . child]
key = "value"
"#,
r#"hello.world = "a"
"#,
r#"foo = 1979-05-27 # Comment
"#,
];
for input in documents {
dbg!(input);
let parsed = parse_document(input).map(|d| d.into_mut());
let doc = match parsed {
Ok(doc) => doc,
Err(err) => {
panic!("Parse error: {err:?}\nFailed to parse:\n```\n{input}\n```")
}
};
assert_data_eq!(doc.to_string(), input.raw());
}
}
#[test]
fn documents_parse_only() {
let parse_only = ["\u{FEFF}
[package]
name = \"foo\"
version = \"0.0.1\"
authors = []
"];
for input in parse_only {
dbg!(input);
let parsed = parse_document(input).map(|d| d.into_mut());
match parsed {
Ok(_) => (),
Err(err) => {
panic!("Parse error: {err:?}\nFailed to parse:\n```\n{input}\n```")
}
}
}
}
#[test]
fn invalid_documents() {
let invalid_inputs = [r#" hello = 'darkness' # my old friend
$"#];
for input in invalid_inputs {
dbg!(input);
let parsed = parse_document(input).map(|d| d.into_mut());
assert!(parsed.is_err(), "Input: {input:?}");
}
}
}
+399
View File
@@ -0,0 +1,399 @@
use std::ops::RangeInclusive;
use winnow::combinator::alt;
use winnow::combinator::cut_err;
use winnow::combinator::opt;
use winnow::combinator::peek;
use winnow::combinator::preceded;
use winnow::combinator::repeat;
use winnow::combinator::trace;
use winnow::token::one_of;
use winnow::token::rest;
use winnow::token::take;
use crate::parser::prelude::*;
use crate::parser::trivia::from_utf8_unchecked;
// ;; Boolean
// boolean = true / false
#[allow(dead_code)] // directly define in `fn value`
pub(crate) fn boolean(input: &mut Input<'_>) -> ModalResult<bool> {
trace("boolean", alt((true_, false_))).parse_next(input)
}
pub(crate) fn true_(input: &mut Input<'_>) -> ModalResult<bool> {
(peek(TRUE[0]), cut_err(TRUE)).value(true).parse_next(input)
}
const TRUE: &[u8] = b"true";
pub(crate) fn false_(input: &mut Input<'_>) -> ModalResult<bool> {
(peek(FALSE[0]), cut_err(FALSE))
.value(false)
.parse_next(input)
}
const FALSE: &[u8] = b"false";
// ;; Integer
// integer = dec-int / hex-int / oct-int / bin-int
pub(crate) fn integer(input: &mut Input<'_>) -> ModalResult<i64> {
trace("integer",
dispatch! {peek(opt::<_, &[u8], _, _>(take(2usize)));
Some(b"0x") => cut_err(hex_int.try_map(|s| i64::from_str_radix(&s.replace('_', ""), 16))),
Some(b"0o") => cut_err(oct_int.try_map(|s| i64::from_str_radix(&s.replace('_', ""), 8))),
Some(b"0b") => cut_err(bin_int.try_map(|s| i64::from_str_radix(&s.replace('_', ""), 2))),
_ => dec_int.and_then(cut_err(rest
.try_map(|s: &str| s.replace('_', "").parse())))
})
.parse_next(input)
}
// dec-int = [ minus / plus ] unsigned-dec-int
// unsigned-dec-int = DIGIT / digit1-9 1*( DIGIT / underscore DIGIT )
pub(crate) fn dec_int<'i>(input: &mut Input<'i>) -> ModalResult<&'i str> {
trace(
"dec-int",
(
opt(one_of((b'+', b'-'))),
alt((
(
one_of(DIGIT1_9),
repeat(
0..,
alt((
digit.void(),
(
one_of(b'_'),
cut_err(digit).context(StrContext::Expected(
StrContextValue::Description("digit"),
)),
)
.void(),
)),
)
.map(|()| ()),
)
.void(),
digit.void(),
)),
)
.take()
.map(|b: &[u8]| unsafe {
from_utf8_unchecked(b, "`digit` and `_` filter out non-ASCII")
})
.context(StrContext::Label("integer")),
)
.parse_next(input)
}
const DIGIT1_9: RangeInclusive<u8> = b'1'..=b'9';
// hex-prefix = %x30.78 ; 0x
// hex-int = hex-prefix HEXDIG *( HEXDIG / underscore HEXDIG )
pub(crate) fn hex_int<'i>(input: &mut Input<'i>) -> ModalResult<&'i str> {
trace(
"hex-int",
preceded(
HEX_PREFIX,
cut_err((
hexdig,
repeat(
0..,
alt((
hexdig.void(),
(
one_of(b'_'),
cut_err(hexdig).context(StrContext::Expected(
StrContextValue::Description("digit"),
)),
)
.void(),
)),
)
.map(|()| ()),
))
.take(),
)
.map(|b| unsafe { from_utf8_unchecked(b, "`hexdig` and `_` filter out non-ASCII") })
.context(StrContext::Label("hexadecimal integer")),
)
.parse_next(input)
}
const HEX_PREFIX: &[u8] = b"0x";
// oct-prefix = %x30.6F ; 0o
// oct-int = oct-prefix digit0-7 *( digit0-7 / underscore digit0-7 )
pub(crate) fn oct_int<'i>(input: &mut Input<'i>) -> ModalResult<&'i str> {
trace(
"oct-int",
preceded(
OCT_PREFIX,
cut_err((
one_of(DIGIT0_7),
repeat(
0..,
alt((
one_of(DIGIT0_7).void(),
(
one_of(b'_'),
cut_err(one_of(DIGIT0_7)).context(StrContext::Expected(
StrContextValue::Description("digit"),
)),
)
.void(),
)),
)
.map(|()| ()),
))
.take(),
)
.map(|b| unsafe { from_utf8_unchecked(b, "`DIGIT0_7` and `_` filter out non-ASCII") })
.context(StrContext::Label("octal integer")),
)
.parse_next(input)
}
const OCT_PREFIX: &[u8] = b"0o";
const DIGIT0_7: RangeInclusive<u8> = b'0'..=b'7';
// bin-prefix = %x30.62 ; 0b
// bin-int = bin-prefix digit0-1 *( digit0-1 / underscore digit0-1 )
pub(crate) fn bin_int<'i>(input: &mut Input<'i>) -> ModalResult<&'i str> {
trace(
"bin-int",
preceded(
BIN_PREFIX,
cut_err((
one_of(DIGIT0_1),
repeat(
0..,
alt((
one_of(DIGIT0_1).void(),
(
one_of(b'_'),
cut_err(one_of(DIGIT0_1)).context(StrContext::Expected(
StrContextValue::Description("digit"),
)),
)
.void(),
)),
)
.map(|()| ()),
))
.take(),
)
.map(|b| unsafe { from_utf8_unchecked(b, "`DIGIT0_1` and `_` filter out non-ASCII") })
.context(StrContext::Label("binary integer")),
)
.parse_next(input)
}
const BIN_PREFIX: &[u8] = b"0b";
const DIGIT0_1: RangeInclusive<u8> = b'0'..=b'1';
// ;; Float
// float = float-int-part ( exp / frac [ exp ] )
// float =/ special-float
// float-int-part = dec-int
pub(crate) fn float(input: &mut Input<'_>) -> ModalResult<f64> {
trace(
"float",
alt((
float_.and_then(cut_err(
rest.try_map(|s: &str| s.replace('_', "").parse())
.verify(|f: &f64| *f != f64::INFINITY),
)),
special_float,
))
.context(StrContext::Label("floating-point number")),
)
.parse_next(input)
}
fn float_<'i>(input: &mut Input<'i>) -> ModalResult<&'i str> {
(
dec_int,
alt((exp.void(), (frac.void(), opt(exp.void())).void())),
)
.take()
.map(|b: &[u8]| unsafe {
from_utf8_unchecked(
b,
"`dec_int`, `one_of`, `exp`, and `frac` filter out non-ASCII",
)
})
.parse_next(input)
}
// frac = decimal-point zero-prefixable-int
// decimal-point = %x2E ; .
fn frac<'i>(input: &mut Input<'i>) -> ModalResult<&'i str> {
(
b'.',
cut_err(zero_prefixable_int)
.context(StrContext::Expected(StrContextValue::Description("digit"))),
)
.take()
.map(|b: &[u8]| unsafe {
from_utf8_unchecked(
b,
"`.` and `parse_zero_prefixable_int` filter out non-ASCII",
)
})
.parse_next(input)
}
// zero-prefixable-int = DIGIT *( DIGIT / underscore DIGIT )
fn zero_prefixable_int<'i>(input: &mut Input<'i>) -> ModalResult<&'i str> {
(
digit,
repeat(
0..,
alt((
digit.void(),
(
one_of(b'_'),
cut_err(digit)
.context(StrContext::Expected(StrContextValue::Description("digit"))),
)
.void(),
)),
)
.map(|()| ()),
)
.take()
.map(|b: &[u8]| unsafe { from_utf8_unchecked(b, "`digit` and `_` filter out non-ASCII") })
.parse_next(input)
}
// exp = "e" float-exp-part
// float-exp-part = [ minus / plus ] zero-prefixable-int
fn exp<'i>(input: &mut Input<'i>) -> ModalResult<&'i str> {
(
one_of((b'e', b'E')),
opt(one_of([b'+', b'-'])),
cut_err(zero_prefixable_int),
)
.take()
.map(|b: &[u8]| unsafe {
from_utf8_unchecked(
b,
"`one_of` and `parse_zero_prefixable_int` filter out non-ASCII",
)
})
.parse_next(input)
}
// special-float = [ minus / plus ] ( inf / nan )
fn special_float(input: &mut Input<'_>) -> ModalResult<f64> {
(opt(one_of((b'+', b'-'))), alt((inf, nan)))
.map(|(s, f)| match s {
Some(b'+') | None => f,
Some(b'-') => -f,
_ => unreachable!("one_of should prevent this"),
})
.parse_next(input)
}
// inf = %x69.6e.66 ; inf
pub(crate) fn inf(input: &mut Input<'_>) -> ModalResult<f64> {
INF.value(f64::INFINITY).parse_next(input)
}
const INF: &[u8] = b"inf";
// nan = %x6e.61.6e ; nan
pub(crate) fn nan(input: &mut Input<'_>) -> ModalResult<f64> {
NAN.value(f64::NAN.copysign(1.0)).parse_next(input)
}
const NAN: &[u8] = b"nan";
// DIGIT = %x30-39 ; 0-9
fn digit(input: &mut Input<'_>) -> ModalResult<u8> {
one_of(DIGIT).parse_next(input)
}
const DIGIT: RangeInclusive<u8> = b'0'..=b'9';
// HEXDIG = DIGIT / "A" / "B" / "C" / "D" / "E" / "F"
fn hexdig(input: &mut Input<'_>) -> ModalResult<u8> {
one_of(HEXDIG).parse_next(input)
}
pub(crate) const HEXDIG: (RangeInclusive<u8>, RangeInclusive<u8>, RangeInclusive<u8>) =
(DIGIT, b'A'..=b'F', b'a'..=b'f');
#[cfg(test)]
#[cfg(feature = "parse")]
#[cfg(feature = "display")]
mod test {
use super::*;
#[test]
fn integers() {
let cases = [
("+99", 99),
("42", 42),
("0", 0),
("-17", -17),
("1_000", 1_000),
("5_349_221", 5_349_221),
("1_2_3_4_5", 1_2_3_4_5),
("0xF", 15),
("0o0_755", 493),
("0b1_0_1", 5),
(&i64::MIN.to_string()[..], i64::MIN),
(&i64::MAX.to_string()[..], i64::MAX),
];
for &(input, expected) in &cases {
dbg!(input);
let parsed = integer.parse(new_input(input));
assert_eq!(parsed, Ok(expected), "Parsing {input:?}");
}
let overflow = "1000000000000000000000000000000000";
let parsed = integer.parse(new_input(overflow));
assert!(parsed.is_err());
}
#[track_caller]
fn assert_float_eq(actual: f64, expected: f64) {
if expected.is_nan() {
assert!(actual.is_nan());
assert_eq!(expected.is_sign_positive(), actual.is_sign_positive());
} else if expected.is_infinite() {
assert!(actual.is_infinite());
assert_eq!(expected.is_sign_positive(), actual.is_sign_positive());
} else {
dbg!(expected);
dbg!(actual);
assert!((expected - actual).abs() < f64::EPSILON);
}
}
#[test]
fn floats() {
let cases = [
("+1.0", 1.0),
("3.1419", 3.1419),
("-0.01", -0.01),
("5e+22", 5e+22),
("1e6", 1e6),
("-2E-2", -2E-2),
("6.626e-34", 6.626e-34),
("9_224_617.445_991_228_313", 9_224_617.445_991_227),
("-1.7976931348623157e+308", f64::MIN),
("1.7976931348623157e+308", f64::MAX),
("nan", f64::NAN.copysign(1.0)),
("+nan", f64::NAN.copysign(1.0)),
("-nan", f64::NAN.copysign(-1.0)),
("inf", f64::INFINITY),
("+inf", f64::INFINITY),
("-inf", f64::NEG_INFINITY),
// ("1e+400", f64::INFINITY),
];
for &(input, expected) in &cases {
dbg!(input);
let parsed = float.parse(new_input(input)).unwrap();
assert_float_eq(parsed, expected);
let overflow = "9e99999";
let parsed = float.parse(new_input(overflow));
assert!(parsed.is_err(), "{parsed:?}");
}
}
}
+320
View File
@@ -0,0 +1,320 @@
use crate::key::Key;
use crate::parser::error::CustomError;
use crate::repr::Decor;
use crate::{ArrayOfTables, ImDocument, Item, RawString, Table};
pub(crate) struct ParseState {
root: Table,
trailing: Option<std::ops::Range<usize>>,
current_table_position: usize,
current_table: Table,
current_is_array: bool,
current_table_path: Vec<Key>,
}
impl ParseState {
pub(crate) fn new() -> Self {
let mut root = Table::new();
root.span = Some(0..0);
Self {
root: Table::new(),
trailing: None,
current_table_position: 0,
current_table: root,
current_is_array: false,
current_table_path: Vec::new(),
}
}
pub(crate) fn into_document<S>(mut self, raw: S) -> Result<ImDocument<S>, CustomError> {
self.finalize_table()?;
let trailing = self.trailing.map(RawString::with_span).unwrap_or_default();
Ok(ImDocument {
root: Item::Table(self.root),
trailing,
raw,
})
}
pub(crate) fn on_ws(&mut self, span: std::ops::Range<usize>) {
if let Some(old) = self.trailing.take() {
self.trailing = Some(old.start..span.end);
} else {
self.trailing = Some(span);
}
}
pub(crate) fn on_comment(&mut self, span: std::ops::Range<usize>) {
if let Some(old) = self.trailing.take() {
self.trailing = Some(old.start..span.end);
} else {
self.trailing = Some(span);
}
}
pub(crate) fn on_keyval(
&mut self,
path: Vec<Key>,
(mut key, value): (Key, Item),
) -> Result<(), CustomError> {
{
let mut prefix = self.trailing.take();
let prefix = match (
prefix.take(),
key.leaf_decor.prefix().and_then(|d| d.span()),
) {
(Some(p), Some(k)) => Some(p.start..k.end),
(Some(p), None) | (None, Some(p)) => Some(p),
(None, None) => None,
};
key.leaf_decor
.set_prefix(prefix.map(RawString::with_span).unwrap_or_default());
}
if let (Some(existing), Some(value)) = (self.current_table.span(), value.span()) {
self.current_table.span = Some((existing.start)..(value.end));
}
let table = &mut self.current_table;
let table = Self::descend_path(table, &path, true)?;
// "Likewise, using dotted keys to redefine tables already defined in [table] form is not allowed"
let mixed_table_types = table.is_dotted() == path.is_empty();
if mixed_table_types {
return Err(CustomError::DuplicateKey {
key: key.get().into(),
table: None,
});
}
match table.items.entry(key) {
indexmap::map::Entry::Vacant(o) => {
o.insert(value);
}
indexmap::map::Entry::Occupied(o) => {
// "Since tables cannot be defined more than once, redefining such tables using a [table] header is not allowed"
return Err(CustomError::DuplicateKey {
key: o.key().get().into(),
table: Some(self.current_table_path.clone()),
});
}
}
Ok(())
}
pub(crate) fn start_array_table(
&mut self,
path: Vec<Key>,
decor: Decor,
span: std::ops::Range<usize>,
) -> Result<(), CustomError> {
debug_assert!(!path.is_empty());
debug_assert!(self.current_table.is_empty());
debug_assert!(self.current_table_path.is_empty());
// Look up the table on start to ensure the duplicate_key error points to the right line
let root = &mut self.root;
let parent_table = Self::descend_path(root, &path[..path.len() - 1], false)?;
let key = &path[path.len() - 1];
let entry = parent_table
.entry_format(key)
.or_insert(Item::ArrayOfTables(ArrayOfTables::new()));
entry
.as_array_of_tables()
.ok_or_else(|| CustomError::duplicate_key(&path, path.len() - 1))?;
self.current_table_position += 1;
self.current_table.decor = decor;
self.current_table.set_implicit(false);
self.current_table.set_dotted(false);
self.current_table.set_position(self.current_table_position);
self.current_table.span = Some(span);
self.current_is_array = true;
self.current_table_path = path;
Ok(())
}
pub(crate) fn start_table(
&mut self,
path: Vec<Key>,
decor: Decor,
span: std::ops::Range<usize>,
) -> Result<(), CustomError> {
debug_assert!(!path.is_empty());
debug_assert!(self.current_table.is_empty());
debug_assert!(self.current_table_path.is_empty());
// 1. Look up the table on start to ensure the duplicate_key error points to the right line
// 2. Ensure any child tables from an implicit table are preserved
let root = &mut self.root;
let parent_table = Self::descend_path(root, &path[..path.len() - 1], false)?;
let key = &path[path.len() - 1];
if let Some(entry) = parent_table.remove(key.get()) {
match entry {
Item::Table(t) if t.implicit && !t.is_dotted() => {
self.current_table = t;
}
// Since tables cannot be defined more than once, redefining such tables using a [table] header is not allowed. Likewise, using dotted keys to redefine tables already defined in [table] form is not allowed.
_ => return Err(CustomError::duplicate_key(&path, path.len() - 1)),
}
}
self.current_table_position += 1;
self.current_table.decor = decor;
self.current_table.set_implicit(false);
self.current_table.set_dotted(false);
self.current_table.set_position(self.current_table_position);
self.current_table.span = Some(span);
self.current_is_array = false;
self.current_table_path = path;
Ok(())
}
pub(crate) fn finalize_table(&mut self) -> Result<(), CustomError> {
let mut table = std::mem::take(&mut self.current_table);
let path = std::mem::take(&mut self.current_table_path);
let root = &mut self.root;
if path.is_empty() {
assert!(root.is_empty());
std::mem::swap(&mut table, root);
} else if self.current_is_array {
let parent_table = Self::descend_path(root, &path[..path.len() - 1], false)?;
let key = &path[path.len() - 1];
let entry = parent_table
.entry_format(key)
.or_insert(Item::ArrayOfTables(ArrayOfTables::new()));
let array = entry
.as_array_of_tables_mut()
.ok_or_else(|| CustomError::duplicate_key(&path, path.len() - 1))?;
array.push(table);
let span = if let (Some(first), Some(last)) = (
array.values.first().and_then(|t| t.span()),
array.values.last().and_then(|t| t.span()),
) {
Some((first.start)..(last.end))
} else {
None
};
array.span = span;
} else {
let parent_table = Self::descend_path(root, &path[..path.len() - 1], false)?;
let key = &path[path.len() - 1];
let entry = parent_table.entry_format(key);
match entry {
crate::Entry::Occupied(entry) => {
match entry.into_mut() {
// if [a.b.c] header preceded [a.b]
Item::Table(ref mut t) if t.implicit => {
std::mem::swap(t, &mut table);
}
_ => return Err(CustomError::duplicate_key(&path, path.len() - 1)),
}
}
crate::Entry::Vacant(entry) => {
let item = Item::Table(table);
entry.insert(item);
}
}
}
Ok(())
}
pub(crate) fn descend_path<'t>(
mut table: &'t mut Table,
path: &[Key],
dotted: bool,
) -> Result<&'t mut Table, CustomError> {
for (i, key) in path.iter().enumerate() {
table = match table.entry_format(key) {
crate::Entry::Vacant(entry) => {
let mut new_table = Table::new();
new_table.set_implicit(true);
new_table.set_dotted(dotted);
entry.insert(Item::Table(new_table)).as_table_mut().unwrap()
}
crate::Entry::Occupied(entry) => {
match entry.into_mut() {
Item::Value(ref v) => {
return Err(CustomError::extend_wrong_type(path, i, v.type_name()));
}
Item::ArrayOfTables(ref mut array) => {
debug_assert!(!array.is_empty());
let index = array.len() - 1;
let last_child = array.get_mut(index).unwrap();
last_child
}
Item::Table(ref mut sweet_child_of_mine) => {
// Since tables cannot be defined more than once, redefining such tables using a
// [table] header is not allowed. Likewise, using dotted keys to redefine tables
// already defined in [table] form is not allowed.
if dotted && !sweet_child_of_mine.is_implicit() {
return Err(CustomError::DuplicateKey {
key: key.get().into(),
table: None,
});
}
sweet_child_of_mine
}
Item::None => unreachable!(),
}
}
};
}
Ok(table)
}
pub(crate) fn on_std_header(
&mut self,
path: Vec<Key>,
trailing: std::ops::Range<usize>,
span: std::ops::Range<usize>,
) -> Result<(), CustomError> {
debug_assert!(!path.is_empty());
self.finalize_table()?;
let leading = self
.trailing
.take()
.map(RawString::with_span)
.unwrap_or_default();
self.start_table(
path,
Decor::new(leading, RawString::with_span(trailing)),
span,
)?;
Ok(())
}
pub(crate) fn on_array_header(
&mut self,
path: Vec<Key>,
trailing: std::ops::Range<usize>,
span: std::ops::Range<usize>,
) -> Result<(), CustomError> {
debug_assert!(!path.is_empty());
self.finalize_table()?;
let leading = self
.trailing
.take()
.map(RawString::with_span)
.unwrap_or_default();
self.start_array_table(
path,
Decor::new(leading, RawString::with_span(trailing)),
span,
)?;
Ok(())
}
}
+479
View File
@@ -0,0 +1,479 @@
use std::borrow::Cow;
use std::char;
use std::ops::RangeInclusive;
use winnow::combinator::alt;
use winnow::combinator::cut_err;
use winnow::combinator::delimited;
use winnow::combinator::empty;
use winnow::combinator::fail;
use winnow::combinator::opt;
use winnow::combinator::peek;
use winnow::combinator::preceded;
use winnow::combinator::repeat;
use winnow::combinator::terminated;
use winnow::combinator::trace;
use winnow::prelude::*;
use winnow::stream::Stream;
use winnow::token::any;
use winnow::token::none_of;
use winnow::token::one_of;
use winnow::token::take_while;
use crate::parser::error::CustomError;
use crate::parser::numbers::HEXDIG;
use crate::parser::prelude::*;
use crate::parser::trivia::{from_utf8_unchecked, newline, ws, ws_newlines, NON_ASCII, WSCHAR};
// ;; String
// string = ml-basic-string / basic-string / ml-literal-string / literal-string
pub(crate) fn string<'i>(input: &mut Input<'i>) -> ModalResult<Cow<'i, str>> {
trace(
"string",
alt((
ml_basic_string,
basic_string,
ml_literal_string,
literal_string.map(Cow::Borrowed),
)),
)
.parse_next(input)
}
// ;; Basic String
// basic-string = quotation-mark *basic-char quotation-mark
pub(crate) fn basic_string<'i>(input: &mut Input<'i>) -> ModalResult<Cow<'i, str>> {
trace("basic-string", |input: &mut Input<'i>| {
let _ = one_of(QUOTATION_MARK).parse_next(input)?;
let mut c = Cow::Borrowed("");
if let Some(ci) = opt(basic_chars).parse_next(input)? {
c = ci;
}
while let Some(ci) = opt(basic_chars).parse_next(input)? {
c.to_mut().push_str(&ci);
}
let _ = cut_err(one_of(QUOTATION_MARK))
.context(StrContext::Label("basic string"))
.parse_next(input)?;
Ok(c)
})
.parse_next(input)
}
// quotation-mark = %x22 ; "
pub(crate) const QUOTATION_MARK: u8 = b'"';
// basic-char = basic-unescaped / escaped
fn basic_chars<'i>(input: &mut Input<'i>) -> ModalResult<Cow<'i, str>> {
alt((
// Deviate from the official grammar by batching the unescaped chars so we build a string a
// chunk at a time, rather than a `char` at a time.
take_while(1.., BASIC_UNESCAPED)
.try_map(std::str::from_utf8)
.map(Cow::Borrowed),
escaped.map(|c| Cow::Owned(String::from(c))),
))
.parse_next(input)
}
// basic-unescaped = wschar / %x21 / %x23-5B / %x5D-7E / non-ascii
const BASIC_UNESCAPED: (
(u8, u8),
u8,
RangeInclusive<u8>,
RangeInclusive<u8>,
RangeInclusive<u8>,
) = (WSCHAR, 0x21, 0x23..=0x5B, 0x5D..=0x7E, NON_ASCII);
// escaped = escape escape-seq-char
fn escaped(input: &mut Input<'_>) -> ModalResult<char> {
preceded(ESCAPE, escape_seq_char).parse_next(input)
}
// escape = %x5C ; \
const ESCAPE: u8 = b'\\';
// escape-seq-char = %x22 ; " quotation mark U+0022
// escape-seq-char =/ %x5C ; \ reverse solidus U+005C
// escape-seq-char =/ %x62 ; b backspace U+0008
// escape-seq-char =/ %x66 ; f form feed U+000C
// escape-seq-char =/ %x6E ; n line feed U+000A
// escape-seq-char =/ %x72 ; r carriage return U+000D
// escape-seq-char =/ %x74 ; t tab U+0009
// escape-seq-char =/ %x75 4HEXDIG ; uXXXX U+XXXX
// escape-seq-char =/ %x55 8HEXDIG ; UXXXXXXXX U+XXXXXXXX
fn escape_seq_char(input: &mut Input<'_>) -> ModalResult<char> {
dispatch! {any;
b'b' => empty.value('\u{8}'),
b'f' => empty.value('\u{c}'),
b'n' => empty.value('\n'),
b'r' => empty.value('\r'),
b't' => empty.value('\t'),
b'u' => cut_err(hexescape::<4>).context(StrContext::Label("unicode 4-digit hex code")),
b'U' => cut_err(hexescape::<8>).context(StrContext::Label("unicode 8-digit hex code")),
b'\\' => empty.value('\\'),
b'"' => empty.value('"'),
_ => {
cut_err(fail::<_, char, _>)
.context(StrContext::Label("escape sequence"))
.context(StrContext::Expected(StrContextValue::CharLiteral('b')))
.context(StrContext::Expected(StrContextValue::CharLiteral('f')))
.context(StrContext::Expected(StrContextValue::CharLiteral('n')))
.context(StrContext::Expected(StrContextValue::CharLiteral('r')))
.context(StrContext::Expected(StrContextValue::CharLiteral('t')))
.context(StrContext::Expected(StrContextValue::CharLiteral('u')))
.context(StrContext::Expected(StrContextValue::CharLiteral('U')))
.context(StrContext::Expected(StrContextValue::CharLiteral('\\')))
.context(StrContext::Expected(StrContextValue::CharLiteral('"')))
}
}
.parse_next(input)
}
fn hexescape<const N: usize>(input: &mut Input<'_>) -> ModalResult<char> {
take_while(0..=N, HEXDIG)
.verify(|b: &[u8]| b.len() == N)
.map(|b: &[u8]| unsafe { from_utf8_unchecked(b, "`is_ascii_digit` filters out on-ASCII") })
.verify_map(|s| u32::from_str_radix(s, 16).ok())
.try_map(|h| char::from_u32(h).ok_or(CustomError::OutOfRange))
.parse_next(input)
}
// ;; Multiline Basic String
// ml-basic-string = ml-basic-string-delim [ newline ] ml-basic-body
// ml-basic-string-delim
fn ml_basic_string<'i>(input: &mut Input<'i>) -> ModalResult<Cow<'i, str>> {
trace(
"ml-basic-string",
delimited(
ML_BASIC_STRING_DELIM,
preceded(opt(newline), cut_err(ml_basic_body))
.context(StrContext::Label("multiline basic string")),
cut_err(ML_BASIC_STRING_DELIM).context(StrContext::Label("multiline basic string")),
),
)
.parse_next(input)
}
// ml-basic-string-delim = 3quotation-mark
const ML_BASIC_STRING_DELIM: &[u8] = b"\"\"\"";
// ml-basic-body = *mlb-content *( mlb-quotes 1*mlb-content ) [ mlb-quotes ]
fn ml_basic_body<'i>(input: &mut Input<'i>) -> ModalResult<Cow<'i, str>> {
let mut c = Cow::Borrowed("");
if let Some(ci) = opt(mlb_content).parse_next(input)? {
c = ci;
}
while let Some(ci) = opt(mlb_content).parse_next(input)? {
c.to_mut().push_str(&ci);
}
while let Some(qi) = opt(mlb_quotes(none_of(b'\"').value(()))).parse_next(input)? {
if let Some(ci) = opt(mlb_content).parse_next(input)? {
c.to_mut().push_str(qi);
c.to_mut().push_str(&ci);
while let Some(ci) = opt(mlb_content).parse_next(input)? {
c.to_mut().push_str(&ci);
}
} else {
break;
}
}
if let Some(qi) = opt(mlb_quotes(ML_BASIC_STRING_DELIM.void())).parse_next(input)? {
c.to_mut().push_str(qi);
}
Ok(c)
}
// mlb-content = mlb-char / newline / mlb-escaped-nl
// mlb-char = mlb-unescaped / escaped
fn mlb_content<'i>(input: &mut Input<'i>) -> ModalResult<Cow<'i, str>> {
alt((
// Deviate from the official grammar by batching the unescaped chars so we build a string a
// chunk at a time, rather than a `char` at a time.
take_while(1.., MLB_UNESCAPED)
.try_map(std::str::from_utf8)
.map(Cow::Borrowed),
// Order changed fromg grammar so `escaped` can more easily `cut_err` on bad escape sequences
mlb_escaped_nl.map(|_| Cow::Borrowed("")),
escaped.map(|c| Cow::Owned(String::from(c))),
newline.map(|_| Cow::Borrowed("\n")),
))
.parse_next(input)
}
// mlb-quotes = 1*2quotation-mark
fn mlb_quotes<'i>(
mut term: impl ModalParser<Input<'i>, (), ContextError>,
) -> impl ModalParser<Input<'i>, &'i str, ContextError> {
move |input: &mut Input<'i>| {
let start = input.checkpoint();
let res = terminated(b"\"\"", peek(term.by_ref()))
.map(|b| unsafe { from_utf8_unchecked(b, "`bytes` out non-ASCII") })
.parse_next(input);
match res {
Err(winnow::error::ErrMode::Backtrack(_)) => {
input.reset(&start);
terminated(b"\"", peek(term.by_ref()))
.map(|b| unsafe { from_utf8_unchecked(b, "`bytes` out non-ASCII") })
.parse_next(input)
}
res => res,
}
}
}
// mlb-unescaped = wschar / %x21 / %x23-5B / %x5D-7E / non-ascii
const MLB_UNESCAPED: (
(u8, u8),
u8,
RangeInclusive<u8>,
RangeInclusive<u8>,
RangeInclusive<u8>,
) = (WSCHAR, 0x21, 0x23..=0x5B, 0x5D..=0x7E, NON_ASCII);
// mlb-escaped-nl = escape ws newline *( wschar / newline
// When the last non-whitespace character on a line is a \,
// it will be trimmed along with all whitespace
// (including newlines) up to the next non-whitespace
// character or closing delimiter.
fn mlb_escaped_nl(input: &mut Input<'_>) -> ModalResult<()> {
repeat(1.., (ESCAPE, ws, ws_newlines))
.map(|()| ())
.value(())
.parse_next(input)
}
// ;; Literal String
// literal-string = apostrophe *literal-char apostrophe
pub(crate) fn literal_string<'i>(input: &mut Input<'i>) -> ModalResult<&'i str> {
trace(
"literal-string",
delimited(
APOSTROPHE,
cut_err(take_while(0.., LITERAL_CHAR)),
cut_err(APOSTROPHE),
)
.try_map(std::str::from_utf8)
.context(StrContext::Label("literal string")),
)
.parse_next(input)
}
// apostrophe = %x27 ; ' apostrophe
pub(crate) const APOSTROPHE: u8 = b'\'';
// literal-char = %x09 / %x20-26 / %x28-7E / non-ascii
const LITERAL_CHAR: (
u8,
RangeInclusive<u8>,
RangeInclusive<u8>,
RangeInclusive<u8>,
) = (0x9, 0x20..=0x26, 0x28..=0x7E, NON_ASCII);
// ;; Multiline Literal String
// ml-literal-string = ml-literal-string-delim [ newline ] ml-literal-body
// ml-literal-string-delim
fn ml_literal_string<'i>(input: &mut Input<'i>) -> ModalResult<Cow<'i, str>> {
trace(
"ml-literal-string",
delimited(
(ML_LITERAL_STRING_DELIM, opt(newline)),
cut_err(ml_literal_body.map(|t| {
if t.contains("\r\n") {
Cow::Owned(t.replace("\r\n", "\n"))
} else {
Cow::Borrowed(t)
}
}))
.context(StrContext::Label("multiline literal string")),
cut_err(ML_LITERAL_STRING_DELIM).context(StrContext::Label("multiline literal string")),
),
)
.parse_next(input)
}
// ml-literal-string-delim = 3apostrophe
const ML_LITERAL_STRING_DELIM: &[u8] = b"'''";
// ml-literal-body = *mll-content *( mll-quotes 1*mll-content ) [ mll-quotes ]
fn ml_literal_body<'i>(input: &mut Input<'i>) -> ModalResult<&'i str> {
(
repeat(0.., mll_content).map(|()| ()),
repeat(
0..,
(
mll_quotes(none_of(APOSTROPHE).value(())),
repeat(1.., mll_content).map(|()| ()),
),
)
.map(|()| ()),
opt(mll_quotes(ML_LITERAL_STRING_DELIM.void())),
)
.take()
.try_map(std::str::from_utf8)
.parse_next(input)
}
// mll-content = mll-char / newline
fn mll_content(input: &mut Input<'_>) -> ModalResult<u8> {
alt((one_of(MLL_CHAR), newline.value(b'\n'))).parse_next(input)
}
// mll-char = %x09 / %x20-26 / %x28-7E / non-ascii
const MLL_CHAR: (
u8,
RangeInclusive<u8>,
RangeInclusive<u8>,
RangeInclusive<u8>,
) = (0x9, 0x20..=0x26, 0x28..=0x7E, NON_ASCII);
// mll-quotes = 1*2apostrophe
fn mll_quotes<'i>(
mut term: impl ModalParser<Input<'i>, (), ContextError>,
) -> impl ModalParser<Input<'i>, &'i str, ContextError> {
move |input: &mut Input<'i>| {
let start = input.checkpoint();
let res = terminated(b"''", peek(term.by_ref()))
.map(|b| unsafe { from_utf8_unchecked(b, "`bytes` out non-ASCII") })
.parse_next(input);
match res {
Err(winnow::error::ErrMode::Backtrack(_)) => {
input.reset(&start);
terminated(b"'", peek(term.by_ref()))
.map(|b| unsafe { from_utf8_unchecked(b, "`bytes` out non-ASCII") })
.parse_next(input)
}
res => res,
}
}
}
#[cfg(test)]
#[cfg(feature = "parse")]
#[cfg(feature = "display")]
mod test {
use super::*;
#[test]
fn basic_string() {
let input =
r#""I'm a string. \"You can quote me\". Name\tJos\u00E9\nLocation\tSF. \U0002070E""#;
let expected = "I\'m a string. \"You can quote me\". Name\tJosé\nLocation\tSF. \u{2070E}";
let parsed = string.parse(new_input(input));
assert_eq!(parsed.as_deref(), Ok(expected), "Parsing {input:?}");
}
#[test]
fn ml_basic_string() {
let cases = [
(
r#""""
Roses are red
Violets are blue""""#,
r#"Roses are red
Violets are blue"#,
),
(r#"""" \""" """"#, " \"\"\" "),
(r#"""" \\""""#, " \\"),
];
for &(input, expected) in &cases {
let parsed = string.parse(new_input(input));
assert_eq!(parsed.as_deref(), Ok(expected), "Parsing {input:?}");
}
let invalid_cases = [r#"""" """#, r#"""" \""""#];
for input in &invalid_cases {
let parsed = string.parse(new_input(input));
assert!(parsed.is_err());
}
}
#[test]
fn ml_basic_string_escape_ws() {
let inputs = [
r#""""
The quick brown \
fox jumps over \
the lazy dog.""""#,
r#""""\
The quick brown \
fox jumps over \
the lazy dog.\
""""#,
];
for input in &inputs {
let expected = "The quick brown fox jumps over the lazy dog.";
let parsed = string.parse(new_input(input));
assert_eq!(parsed.as_deref(), Ok(expected), "Parsing {input:?}");
}
let empties = [
r#""""\
""""#,
r#""""
\
\
""""#,
];
for input in &empties {
let expected = "";
let parsed = string.parse(new_input(input));
assert_eq!(parsed.as_deref(), Ok(expected), "Parsing {input:?}");
}
}
#[test]
fn literal_string() {
let inputs = [
r"'C:\Users\nodejs\templates'",
r"'\\ServerX\admin$\system32\'",
r#"'Tom "Dubs" Preston-Werner'"#,
r"'<\i\c*\s*>'",
];
for input in &inputs {
let expected = &input[1..input.len() - 1];
let parsed = string.parse(new_input(input));
assert_eq!(parsed.as_deref(), Ok(expected), "Parsing {input:?}");
}
}
#[test]
fn ml_literal_string() {
let inputs = [
r"'''I [dw]on't need \d{2} apples'''",
r#"''''one_quote''''"#,
];
for input in &inputs {
let expected = &input[3..input.len() - 3];
let parsed = string.parse(new_input(input));
assert_eq!(parsed.as_deref(), Ok(expected), "Parsing {input:?}");
}
let input = r#"'''
The first newline is
trimmed in raw strings.
All other whitespace
is preserved.
'''"#;
let expected = &input[4..input.len() - 3];
let parsed = string.parse(new_input(input));
assert_eq!(parsed.as_deref(), Ok(expected), "Parsing {input:?}");
}
}
+89
View File
@@ -0,0 +1,89 @@
use std::cell::RefCell;
#[allow(unused_imports)]
use std::ops::DerefMut;
use winnow::combinator::cut_err;
use winnow::combinator::delimited;
use winnow::combinator::peek;
use winnow::token::take;
// https://github.com/rust-lang/rust/issues/41358
use crate::parser::key::key;
use crate::parser::prelude::*;
use crate::parser::state::ParseState;
use crate::parser::trivia::line_trailing;
// std-table-open = %x5B ws ; [ Left square bracket
pub(crate) const STD_TABLE_OPEN: u8 = b'[';
// std-table-close = ws %x5D ; ] Right square bracket
const STD_TABLE_CLOSE: u8 = b']';
// array-table-open = %x5B.5B ws ; [[ Double left square bracket
const ARRAY_TABLE_OPEN: &[u8] = b"[[";
// array-table-close = ws %x5D.5D ; ]] Double right square bracket
const ARRAY_TABLE_CLOSE: &[u8] = b"]]";
// ;; Standard Table
// std-table = std-table-open key *( table-key-sep key) std-table-close
fn std_table<'s, 'i>(
state: &'s RefCell<ParseState>,
) -> impl ModalParser<Input<'i>, (), ContextError> + 's {
move |i: &mut Input<'i>| {
(
delimited(
STD_TABLE_OPEN,
cut_err(key),
cut_err(STD_TABLE_CLOSE)
.context(StrContext::Expected(StrContextValue::CharLiteral('.')))
.context(StrContext::Expected(StrContextValue::StringLiteral("]"))),
)
.with_span(),
cut_err(line_trailing)
.context(StrContext::Expected(StrContextValue::CharLiteral('\n')))
.context(StrContext::Expected(StrContextValue::CharLiteral('#'))),
)
.try_map(|((h, span), t)| state.borrow_mut().deref_mut().on_std_header(h, t, span))
.parse_next(i)
}
}
// ;; Array Table
// array-table = array-table-open key *( table-key-sep key) array-table-close
fn array_table<'s, 'i>(
state: &'s RefCell<ParseState>,
) -> impl ModalParser<Input<'i>, (), ContextError> + 's {
move |i: &mut Input<'i>| {
(
delimited(
ARRAY_TABLE_OPEN,
cut_err(key),
cut_err(ARRAY_TABLE_CLOSE)
.context(StrContext::Expected(StrContextValue::CharLiteral('.')))
.context(StrContext::Expected(StrContextValue::StringLiteral("]]"))),
)
.with_span(),
cut_err(line_trailing)
.context(StrContext::Expected(StrContextValue::CharLiteral('\n')))
.context(StrContext::Expected(StrContextValue::CharLiteral('#'))),
)
.try_map(|((h, span), t)| state.borrow_mut().deref_mut().on_array_header(h, t, span))
.parse_next(i)
}
}
// ;; Table
// table = std-table / array-table
pub(crate) fn table<'s, 'i>(
state: &'s RefCell<ParseState>,
) -> impl ModalParser<Input<'i>, (), ContextError> + 's {
move |i: &mut Input<'i>| {
dispatch!(peek::<_, &[u8],_,_>(take(2usize));
b"[[" => array_table(state),
_ => std_table(state),
)
.context(StrContext::Label("table header"))
.parse_next(i)
}
}
+166
View File
@@ -0,0 +1,166 @@
use std::ops::RangeInclusive;
use winnow::combinator::alt;
use winnow::combinator::empty;
use winnow::combinator::eof;
use winnow::combinator::fail;
use winnow::combinator::opt;
use winnow::combinator::peek;
use winnow::combinator::repeat;
use winnow::combinator::terminated;
use winnow::prelude::*;
use winnow::token::any;
use winnow::token::one_of;
use winnow::token::take_while;
use crate::parser::prelude::*;
pub(crate) unsafe fn from_utf8_unchecked<'b>(
bytes: &'b [u8],
safety_justification: &'static str,
) -> &'b str {
unsafe {
if cfg!(debug_assertions) {
// Catch problems more quickly when testing
std::str::from_utf8(bytes).expect(safety_justification)
} else {
std::str::from_utf8_unchecked(bytes)
}
}
}
// wschar = ( %x20 / ; Space
// %x09 ) ; Horizontal tab
pub(crate) const WSCHAR: (u8, u8) = (b' ', b'\t');
// ws = *wschar
pub(crate) fn ws<'i>(input: &mut Input<'i>) -> ModalResult<&'i str> {
take_while(0.., WSCHAR)
.map(|b| unsafe { from_utf8_unchecked(b, "`is_wschar` filters out on-ASCII") })
.parse_next(input)
}
// non-ascii = %x80-D7FF / %xE000-10FFFF
// - ASCII is 0xxxxxxx
// - First byte for UTF-8 is 11xxxxxx
// - Subsequent UTF-8 bytes are 10xxxxxx
pub(crate) const NON_ASCII: RangeInclusive<u8> = 0x80..=0xff;
// non-eol = %x09 / %x20-7E / non-ascii
pub(crate) const NON_EOL: (u8, RangeInclusive<u8>, RangeInclusive<u8>) =
(0x09, 0x20..=0x7E, NON_ASCII);
// comment-start-symbol = %x23 ; #
pub(crate) const COMMENT_START_SYMBOL: u8 = b'#';
// comment = comment-start-symbol *non-eol
pub(crate) fn comment(input: &mut Input<'_>) -> ModalResult<()> {
(COMMENT_START_SYMBOL, take_while(0.., NON_EOL))
.void()
.parse_next(input)
}
// newline = ( %x0A / ; LF
// %x0D.0A ) ; CRLF
pub(crate) fn newline(input: &mut Input<'_>) -> ModalResult<()> {
dispatch! {any;
b'\n' => empty,
b'\r' => one_of(LF).void(),
_ => fail,
}
.parse_next(input)
}
pub(crate) const LF: u8 = b'\n';
pub(crate) const CR: u8 = b'\r';
// ws-newline = *( wschar / newline )
pub(crate) fn ws_newline(input: &mut Input<'_>) -> ModalResult<()> {
repeat(
0..,
alt((newline.value(&b"\n"[..]), take_while(1.., WSCHAR))),
)
.map(|()| ())
.parse_next(input)
}
// ws-newlines = newline *( wschar / newline )
pub(crate) fn ws_newlines(input: &mut Input<'_>) -> ModalResult<()> {
(newline, ws_newline).void().parse_next(input)
}
// note: this rule is not present in the original grammar
// ws-comment-newline = *( ws-newline-nonempty / comment )
pub(crate) fn ws_comment_newline(input: &mut Input<'_>) -> ModalResult<()> {
let mut start = input.checkpoint();
loop {
let _ = ws.parse_next(input)?;
let next_token = opt(peek(any)).parse_next(input)?;
match next_token {
Some(b'#') => (comment, newline).void().parse_next(input)?,
Some(b'\n') => (newline).void().parse_next(input)?,
Some(b'\r') => (newline).void().parse_next(input)?,
_ => break,
}
let end = input.checkpoint();
if start == end {
break;
}
start = end;
}
Ok(())
}
// note: this rule is not present in the original grammar
// line-ending = newline / eof
pub(crate) fn line_ending(input: &mut Input<'_>) -> ModalResult<()> {
alt((newline.value("\n"), eof.value("")))
.void()
.parse_next(input)
}
// note: this rule is not present in the original grammar
// line-trailing = ws [comment] skip-line-ending
pub(crate) fn line_trailing(input: &mut Input<'_>) -> ModalResult<std::ops::Range<usize>> {
terminated((ws, opt(comment)).span(), line_ending).parse_next(input)
}
#[cfg(test)]
#[cfg(feature = "parse")]
#[cfg(feature = "display")]
mod test {
use super::*;
#[test]
fn trivia() {
let inputs = [
"",
r#" "#,
r#"
"#,
r#"
# comment
# comment2
"#,
r#"
"#,
r#"# comment
# comment2
"#,
];
for input in inputs {
dbg!(input);
let parsed = ws_comment_newline.take().parse(new_input(input));
assert!(parsed.is_ok(), "{parsed:?}");
let parsed = parsed.unwrap();
assert_eq!(parsed, input.as_bytes());
}
}
}
+154
View File
@@ -0,0 +1,154 @@
use winnow::combinator::alt;
use winnow::combinator::fail;
use winnow::combinator::peek;
use winnow::token::any;
use crate::parser::array::array;
use crate::parser::datetime::date_time;
use crate::parser::inline_table::inline_table;
use crate::parser::numbers::{float, integer};
use crate::parser::prelude::*;
use crate::parser::strings::string;
use crate::repr::{Formatted, Repr};
use crate::RawString;
use crate::Value;
// val = string / boolean / array / inline-table / date-time / float / integer
pub(crate) fn value(input: &mut Input<'_>) -> ModalResult<Value> {
dispatch! {peek(any);
crate::parser::strings::QUOTATION_MARK |
crate::parser::strings::APOSTROPHE => string.map(|s| {
Value::String(Formatted::new(
s.into_owned()
))
}),
crate::parser::array::ARRAY_OPEN => check_recursion(array).map(Value::Array),
crate::parser::inline_table::INLINE_TABLE_OPEN => check_recursion(inline_table).map(Value::InlineTable),
// Date/number starts
b'+' | b'-' | b'0'..=b'9' => {
// Uncommon enough not to be worth optimizing at this time
alt((
date_time
.map(Value::from),
float
.map(Value::from),
integer
.map(Value::from),
))
},
// Report as if they were numbers because its most likely a typo
b'_' => {
integer
.map(Value::from)
.context(StrContext::Expected(StrContextValue::Description("leading digit")))
},
// Report as if they were numbers because its most likely a typo
b'.' => {
float
.map(Value::from)
.context(StrContext::Expected(StrContextValue::Description("leading digit")))
},
b't' => {
crate::parser::numbers::true_.map(Value::from)
.context(StrContext::Label("string"))
.context(StrContext::Expected(StrContextValue::CharLiteral('"')))
.context(StrContext::Expected(StrContextValue::CharLiteral('\'')))
},
b'f' => {
crate::parser::numbers::false_.map(Value::from)
.context(StrContext::Label("string"))
.context(StrContext::Expected(StrContextValue::CharLiteral('"')))
.context(StrContext::Expected(StrContextValue::CharLiteral('\'')))
},
b'i' => {
crate::parser::numbers::inf.map(Value::from)
.context(StrContext::Label("string"))
.context(StrContext::Expected(StrContextValue::CharLiteral('"')))
.context(StrContext::Expected(StrContextValue::CharLiteral('\'')))
},
b'n' => {
crate::parser::numbers::nan.map(Value::from)
.context(StrContext::Label("string"))
.context(StrContext::Expected(StrContextValue::CharLiteral('"')))
.context(StrContext::Expected(StrContextValue::CharLiteral('\'')))
},
_ => {
fail
.context(StrContext::Label("string"))
.context(StrContext::Expected(StrContextValue::CharLiteral('"')))
.context(StrContext::Expected(StrContextValue::CharLiteral('\'')))
},
}
.with_span()
.map(|(value, span)| apply_raw(value, span))
.parse_next(input)
}
fn apply_raw(mut val: Value, span: std::ops::Range<usize>) -> Value {
match val {
Value::String(ref mut f) => {
let raw = RawString::with_span(span);
f.set_repr_unchecked(Repr::new_unchecked(raw));
}
Value::Integer(ref mut f) => {
let raw = RawString::with_span(span);
f.set_repr_unchecked(Repr::new_unchecked(raw));
}
Value::Float(ref mut f) => {
let raw = RawString::with_span(span);
f.set_repr_unchecked(Repr::new_unchecked(raw));
}
Value::Boolean(ref mut f) => {
let raw = RawString::with_span(span);
f.set_repr_unchecked(Repr::new_unchecked(raw));
}
Value::Datetime(ref mut f) => {
let raw = RawString::with_span(span);
f.set_repr_unchecked(Repr::new_unchecked(raw));
}
Value::Array(ref mut arr) => {
arr.span = Some(span);
}
Value::InlineTable(ref mut table) => {
table.span = Some(span);
}
};
val.decorate("", "");
val
}
#[cfg(test)]
#[cfg(feature = "parse")]
#[cfg(feature = "display")]
mod test {
use super::*;
#[test]
fn values() {
let inputs = [
"1979-05-27T00:32:00.999999",
"-239",
"1e200",
"9_224_617.445_991_228_313",
r"'''I [dw]on't need \d{2} apples'''",
r#"'''
The first newline is
trimmed in raw strings.
All other whitespace
is preserved.
'''"#,
r#""Jos\u00E9\n""#,
r#""\\\"\b/\f\n\r\t\u00E9\U000A0000""#,
r#"{ hello = "world", a = 1}"#,
r#"[ { x = 1, a = "2" }, {a = "a",b = "b", c = "c"} ]"#,
];
for input in inputs {
dbg!(input);
let mut parsed = value.parse(new_input(input));
if let Ok(parsed) = &mut parsed {
parsed.despan(input);
}
assert_eq!(parsed.map(|a| a.to_string()), Ok(input.to_owned()));
}
}
}