945 lines
31 KiB
Rust
945 lines
31 KiB
Rust
use std::char;
|
|
use std::cmp::max;
|
|
use std::cmp::Ordering;
|
|
use std::collections::HashSet;
|
|
use std::fmt;
|
|
use std::mem::swap;
|
|
use std::ops::Range;
|
|
|
|
#[derive(Clone, PartialEq, Eq, Hash)]
|
|
pub struct CharacterSet {
|
|
ranges: Vec<Range<u32>>,
|
|
}
|
|
|
|
#[derive(Debug, PartialEq, Eq)]
|
|
pub enum NfaState {
|
|
Advance {
|
|
chars: CharacterSet,
|
|
state_id: u32,
|
|
is_sep: bool,
|
|
precedence: i32,
|
|
},
|
|
Split(u32, u32),
|
|
Accept {
|
|
variable_index: usize,
|
|
precedence: i32,
|
|
},
|
|
}
|
|
|
|
#[derive(PartialEq, Eq)]
|
|
pub struct Nfa {
|
|
pub states: Vec<NfaState>,
|
|
}
|
|
|
|
#[derive(Debug)]
|
|
pub struct NfaCursor<'a> {
|
|
pub(crate) state_ids: Vec<u32>,
|
|
nfa: &'a Nfa,
|
|
}
|
|
|
|
#[derive(Debug, PartialEq, Eq)]
|
|
pub struct NfaTransition {
|
|
pub characters: CharacterSet,
|
|
pub is_separator: bool,
|
|
pub precedence: i32,
|
|
pub states: Vec<u32>,
|
|
}
|
|
|
|
impl Default for Nfa {
|
|
fn default() -> Self {
|
|
Self { states: Vec::new() }
|
|
}
|
|
}
|
|
|
|
const END: u32 = char::MAX as u32 + 1;
|
|
|
|
impl CharacterSet {
|
|
pub fn empty() -> Self {
|
|
CharacterSet { ranges: Vec::new() }
|
|
}
|
|
|
|
pub fn from_range(mut first: char, mut last: char) -> Self {
|
|
if first > last {
|
|
swap(&mut first, &mut last);
|
|
}
|
|
CharacterSet {
|
|
ranges: vec![(first as u32)..(last as u32 + 1)],
|
|
}
|
|
}
|
|
|
|
pub fn from_char(c: char) -> Self {
|
|
CharacterSet {
|
|
ranges: vec![(c as u32)..(c as u32 + 1)],
|
|
}
|
|
}
|
|
|
|
pub fn negate(mut self) -> CharacterSet {
|
|
let mut i = 0;
|
|
let mut previous_end = 0;
|
|
while i < self.ranges.len() {
|
|
let range = &mut self.ranges[i];
|
|
let start = previous_end;
|
|
previous_end = range.end;
|
|
if start < range.start {
|
|
self.ranges[i] = start..range.start;
|
|
i += 1;
|
|
} else {
|
|
self.ranges.remove(i);
|
|
}
|
|
}
|
|
if previous_end < END {
|
|
self.ranges.push(previous_end..END);
|
|
}
|
|
self
|
|
}
|
|
|
|
pub fn add_char(mut self, c: char) -> Self {
|
|
self.add_int_range(0, c as u32, c as u32 + 1);
|
|
self
|
|
}
|
|
|
|
pub fn add_range(mut self, start: char, end: char) -> Self {
|
|
self.add_int_range(0, start as u32, end as u32 + 1);
|
|
self
|
|
}
|
|
|
|
pub fn add(mut self, other: &CharacterSet) -> Self {
|
|
let mut index = 0;
|
|
for range in &other.ranges {
|
|
index = self.add_int_range(index, range.start as u32, range.end as u32);
|
|
}
|
|
self
|
|
}
|
|
|
|
fn add_int_range(&mut self, mut i: usize, start: u32, end: u32) -> usize {
|
|
while i < self.ranges.len() {
|
|
let range = &mut self.ranges[i];
|
|
if range.start > end {
|
|
self.ranges.insert(i, start..end);
|
|
return i;
|
|
}
|
|
if range.end >= start {
|
|
range.end = range.end.max(end);
|
|
range.start = range.start.min(start);
|
|
return i;
|
|
}
|
|
i += 1;
|
|
}
|
|
self.ranges.push(start..end);
|
|
i
|
|
}
|
|
|
|
pub fn does_intersect(&self, other: &CharacterSet) -> bool {
|
|
let mut left_ranges = self.ranges.iter();
|
|
let mut right_ranges = other.ranges.iter();
|
|
let mut left_range = left_ranges.next();
|
|
let mut right_range = right_ranges.next();
|
|
while let (Some(left), Some(right)) = (&left_range, &right_range) {
|
|
if left.end <= right.start {
|
|
left_range = left_ranges.next();
|
|
} else if left.start >= right.end {
|
|
right_range = right_ranges.next();
|
|
} else {
|
|
return true;
|
|
}
|
|
}
|
|
false
|
|
}
|
|
|
|
pub fn remove_intersection(&mut self, other: &mut CharacterSet) -> CharacterSet {
|
|
let mut intersection = Vec::new();
|
|
let mut left_i = 0;
|
|
let mut right_i = 0;
|
|
while left_i < self.ranges.len() && right_i < other.ranges.len() {
|
|
let left = &mut self.ranges[left_i];
|
|
let right = &mut other.ranges[right_i];
|
|
|
|
match left.start.cmp(&right.start) {
|
|
Ordering::Less => {
|
|
// [ L ]
|
|
// [ R ]
|
|
if left.end <= right.start {
|
|
left_i += 1;
|
|
continue;
|
|
}
|
|
|
|
match left.end.cmp(&right.end) {
|
|
// [ L ]
|
|
// [ R ]
|
|
Ordering::Less => {
|
|
intersection.push(right.start..left.end);
|
|
swap(&mut left.end, &mut right.start);
|
|
left_i += 1;
|
|
}
|
|
|
|
// [ L ]
|
|
// [ R ]
|
|
Ordering::Equal => {
|
|
intersection.push(right.clone());
|
|
left.end = right.start;
|
|
other.ranges.remove(right_i);
|
|
}
|
|
|
|
// [ L ]
|
|
// [ R ]
|
|
Ordering::Greater => {
|
|
intersection.push(right.clone());
|
|
let new_range = left.start..right.start;
|
|
left.start = right.end;
|
|
self.ranges.insert(left_i, new_range);
|
|
other.ranges.remove(right_i);
|
|
left_i += 1;
|
|
}
|
|
}
|
|
}
|
|
Ordering::Equal => {
|
|
// [ L ]
|
|
// [ R ]
|
|
if left.end < right.end {
|
|
intersection.push(left.start..left.end);
|
|
right.start = left.end;
|
|
self.ranges.remove(left_i);
|
|
}
|
|
// [ L ]
|
|
// [ R ]
|
|
else if left.end == right.end {
|
|
intersection.push(left.clone());
|
|
self.ranges.remove(left_i);
|
|
other.ranges.remove(right_i);
|
|
}
|
|
// [ L ]
|
|
// [ R ]
|
|
else if left.end > right.end {
|
|
intersection.push(right.clone());
|
|
left.start = right.end;
|
|
other.ranges.remove(right_i);
|
|
}
|
|
}
|
|
Ordering::Greater => {
|
|
// [ L ]
|
|
// [ R ]
|
|
if left.start >= right.end {
|
|
right_i += 1;
|
|
continue;
|
|
}
|
|
|
|
match left.end.cmp(&right.end) {
|
|
// [ L ]
|
|
// [ R ]
|
|
Ordering::Less => {
|
|
intersection.push(left.clone());
|
|
let new_range = right.start..left.start;
|
|
right.start = left.end;
|
|
other.ranges.insert(right_i, new_range);
|
|
self.ranges.remove(left_i);
|
|
right_i += 1;
|
|
}
|
|
|
|
// [ L ]
|
|
// [ R ]
|
|
Ordering::Equal => {
|
|
intersection.push(left.clone());
|
|
right.end = left.start;
|
|
self.ranges.remove(left_i);
|
|
}
|
|
|
|
// [ L ]
|
|
// [ R ]
|
|
Ordering::Greater => {
|
|
intersection.push(left.start..right.end);
|
|
swap(&mut left.start, &mut right.end);
|
|
right_i += 1;
|
|
}
|
|
}
|
|
}
|
|
}
|
|
}
|
|
CharacterSet {
|
|
ranges: intersection,
|
|
}
|
|
}
|
|
|
|
pub fn iter<'a>(&'a self) -> impl Iterator<Item = u32> + 'a {
|
|
self.ranges.iter().flat_map(|r| r.clone())
|
|
}
|
|
|
|
pub fn chars<'a>(&'a self) -> impl Iterator<Item = char> + 'a {
|
|
self.iter().filter_map(char::from_u32)
|
|
}
|
|
|
|
pub fn is_empty(&self) -> bool {
|
|
self.ranges.is_empty()
|
|
}
|
|
|
|
pub fn simplify_ignoring<'a>(
|
|
&'a self,
|
|
ruled_out_characters: &'a HashSet<u32>,
|
|
) -> Vec<Range<char>> {
|
|
let mut prev_range: Option<Range<char>> = None;
|
|
self.chars()
|
|
.map(|c| (c, false))
|
|
.chain(Some(('\0', true)))
|
|
.filter_map(move |(c, done)| {
|
|
if done {
|
|
return prev_range.clone();
|
|
}
|
|
if ruled_out_characters.contains(&(c as u32)) {
|
|
return None;
|
|
}
|
|
if let Some(range) = prev_range.clone() {
|
|
let mut prev_range_successor = range.end as u32 + 1;
|
|
while prev_range_successor < c as u32 {
|
|
if !ruled_out_characters.contains(&prev_range_successor) {
|
|
prev_range = Some(c..c);
|
|
return Some(range);
|
|
}
|
|
prev_range_successor += 1;
|
|
}
|
|
prev_range = Some(range.start..c);
|
|
None
|
|
} else {
|
|
prev_range = Some(c..c);
|
|
None
|
|
}
|
|
})
|
|
.collect()
|
|
}
|
|
|
|
pub fn contains(&self, c: char) -> bool {
|
|
self.ranges.iter().any(|r| r.contains(&(c as u32)))
|
|
}
|
|
}
|
|
|
|
impl Ord for CharacterSet {
|
|
fn cmp(&self, other: &CharacterSet) -> Ordering {
|
|
let count_cmp = self
|
|
.ranges
|
|
.iter()
|
|
.map(|r| r.len())
|
|
.sum::<usize>()
|
|
.cmp(&other.ranges.iter().map(|r| r.len()).sum());
|
|
if count_cmp != Ordering::Equal {
|
|
return count_cmp;
|
|
}
|
|
|
|
for (left_range, right_range) in self.ranges.iter().zip(other.ranges.iter()) {
|
|
let cmp = left_range.len().cmp(&right_range.len());
|
|
if cmp != Ordering::Equal {
|
|
return cmp;
|
|
}
|
|
|
|
for (left, right) in left_range.clone().zip(right_range.clone()) {
|
|
let cmp = left.cmp(&right);
|
|
if cmp != Ordering::Equal {
|
|
return cmp;
|
|
}
|
|
}
|
|
}
|
|
return Ordering::Equal;
|
|
}
|
|
}
|
|
|
|
impl PartialOrd for CharacterSet {
|
|
fn partial_cmp(&self, other: &CharacterSet) -> Option<Ordering> {
|
|
Some(self.cmp(other))
|
|
}
|
|
}
|
|
|
|
impl fmt::Debug for CharacterSet {
|
|
fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
|
|
write!(f, "CharacterSet [")?;
|
|
let mut set = self.clone();
|
|
if self.contains(char::MAX) {
|
|
write!(f, "^ ")?;
|
|
set = set.negate();
|
|
}
|
|
for (i, c) in set.chars().enumerate() {
|
|
if i > 0 {
|
|
write!(f, ", ")?;
|
|
}
|
|
write!(f, "{:?}", c)?;
|
|
}
|
|
write!(f, "]")?;
|
|
Ok(())
|
|
}
|
|
}
|
|
|
|
impl Nfa {
|
|
pub fn new() -> Self {
|
|
Nfa { states: Vec::new() }
|
|
}
|
|
|
|
pub fn last_state_id(&self) -> u32 {
|
|
self.states.len() as u32 - 1
|
|
}
|
|
}
|
|
|
|
impl fmt::Debug for Nfa {
|
|
fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
|
|
write!(f, "Nfa {{ states: {{\n")?;
|
|
for (i, state) in self.states.iter().enumerate() {
|
|
write!(f, " {}: {:?},\n", i, state)?;
|
|
}
|
|
write!(f, "}} }}")?;
|
|
Ok(())
|
|
}
|
|
}
|
|
|
|
impl<'a> NfaCursor<'a> {
|
|
pub fn new(nfa: &'a Nfa, mut states: Vec<u32>) -> Self {
|
|
let mut result = Self {
|
|
nfa,
|
|
state_ids: Vec::new(),
|
|
};
|
|
result.add_states(&mut states);
|
|
result
|
|
}
|
|
|
|
pub fn reset(&mut self, mut states: Vec<u32>) {
|
|
self.state_ids.clear();
|
|
self.add_states(&mut states);
|
|
}
|
|
|
|
pub fn force_reset(&mut self, states: Vec<u32>) {
|
|
self.state_ids = states
|
|
}
|
|
|
|
pub fn transition_chars(&self) -> impl Iterator<Item = (&CharacterSet, bool)> {
|
|
self.raw_transitions().map(|t| (t.0, t.1))
|
|
}
|
|
|
|
pub fn transitions(&self) -> Vec<NfaTransition> {
|
|
Self::group_transitions(self.raw_transitions())
|
|
}
|
|
|
|
fn raw_transitions(&self) -> impl Iterator<Item = (&CharacterSet, bool, i32, u32)> {
|
|
self.state_ids.iter().filter_map(move |id| {
|
|
if let NfaState::Advance {
|
|
chars,
|
|
state_id,
|
|
precedence,
|
|
is_sep,
|
|
} = &self.nfa.states[*id as usize]
|
|
{
|
|
Some((chars, *is_sep, *precedence, *state_id))
|
|
} else {
|
|
None
|
|
}
|
|
})
|
|
}
|
|
|
|
fn group_transitions<'b>(
|
|
iter: impl Iterator<Item = (&'b CharacterSet, bool, i32, u32)>,
|
|
) -> Vec<NfaTransition> {
|
|
let mut result: Vec<NfaTransition> = Vec::new();
|
|
for (chars, is_sep, prec, state) in iter {
|
|
let mut chars = chars.clone();
|
|
let mut i = 0;
|
|
while i < result.len() && !chars.is_empty() {
|
|
let intersection = result[i].characters.remove_intersection(&mut chars);
|
|
if !intersection.is_empty() {
|
|
let mut intersection_states = result[i].states.clone();
|
|
match intersection_states.binary_search(&state) {
|
|
Err(j) => intersection_states.insert(j, state),
|
|
_ => {}
|
|
}
|
|
let intersection_transition = NfaTransition {
|
|
characters: intersection,
|
|
is_separator: result[i].is_separator && is_sep,
|
|
precedence: max(result[i].precedence, prec),
|
|
states: intersection_states,
|
|
};
|
|
if result[i].characters.is_empty() {
|
|
result[i] = intersection_transition;
|
|
} else {
|
|
result.insert(i, intersection_transition);
|
|
i += 1;
|
|
}
|
|
}
|
|
i += 1;
|
|
}
|
|
if !chars.is_empty() {
|
|
result.push(NfaTransition {
|
|
characters: chars,
|
|
precedence: prec,
|
|
states: vec![state],
|
|
is_separator: is_sep,
|
|
});
|
|
}
|
|
}
|
|
result.sort_unstable_by(|a, b| a.characters.cmp(&b.characters));
|
|
|
|
let mut i = 0;
|
|
'i_loop: while i < result.len() {
|
|
for j in 0..i {
|
|
if result[j].states == result[i].states
|
|
&& result[j].is_separator == result[i].is_separator
|
|
&& result[j].precedence == result[i].precedence
|
|
{
|
|
let mut characters = CharacterSet::empty();
|
|
swap(&mut characters, &mut result[j].characters);
|
|
result[j].characters = characters.add(&result[i].characters);
|
|
result.remove(i);
|
|
continue 'i_loop;
|
|
}
|
|
}
|
|
i += 1;
|
|
}
|
|
|
|
result
|
|
}
|
|
|
|
pub fn completions(&self) -> impl Iterator<Item = (usize, i32)> + '_ {
|
|
self.state_ids.iter().filter_map(move |state_id| {
|
|
if let NfaState::Accept {
|
|
variable_index,
|
|
precedence,
|
|
} = self.nfa.states[*state_id as usize]
|
|
{
|
|
Some((variable_index, precedence))
|
|
} else {
|
|
None
|
|
}
|
|
})
|
|
}
|
|
|
|
pub fn add_states(&mut self, new_state_ids: &mut Vec<u32>) {
|
|
let mut i = 0;
|
|
while i < new_state_ids.len() {
|
|
let state_id = new_state_ids[i];
|
|
let state = &self.nfa.states[state_id as usize];
|
|
if let NfaState::Split(left, right) = state {
|
|
let mut has_left = false;
|
|
let mut has_right = false;
|
|
for new_state_id in new_state_ids.iter() {
|
|
if *new_state_id == *left {
|
|
has_left = true;
|
|
}
|
|
if *new_state_id == *right {
|
|
has_right = true;
|
|
}
|
|
}
|
|
if !has_left {
|
|
new_state_ids.push(*left);
|
|
}
|
|
if !has_right {
|
|
new_state_ids.push(*right);
|
|
}
|
|
} else if let Err(i) = self.state_ids.binary_search(&state_id) {
|
|
self.state_ids.insert(i, state_id);
|
|
}
|
|
i += 1;
|
|
}
|
|
}
|
|
}
|
|
|
|
#[cfg(test)]
|
|
mod tests {
|
|
use super::*;
|
|
|
|
#[test]
|
|
fn test_group_transitions() {
|
|
let table = [
|
|
// overlapping character classes
|
|
(
|
|
vec![
|
|
(CharacterSet::empty().add_range('a', 'f'), false, 0, 1),
|
|
(CharacterSet::empty().add_range('d', 'i'), false, 1, 2),
|
|
],
|
|
vec![
|
|
NfaTransition {
|
|
characters: CharacterSet::empty().add_range('a', 'c'),
|
|
is_separator: false,
|
|
precedence: 0,
|
|
states: vec![1],
|
|
},
|
|
NfaTransition {
|
|
characters: CharacterSet::empty().add_range('d', 'f'),
|
|
is_separator: false,
|
|
precedence: 1,
|
|
states: vec![1, 2],
|
|
},
|
|
NfaTransition {
|
|
characters: CharacterSet::empty().add_range('g', 'i'),
|
|
is_separator: false,
|
|
precedence: 1,
|
|
states: vec![2],
|
|
},
|
|
],
|
|
),
|
|
// large character class followed by many individual characters
|
|
(
|
|
vec![
|
|
(CharacterSet::empty().add_range('a', 'z'), false, 0, 1),
|
|
(CharacterSet::empty().add_char('d'), false, 0, 2),
|
|
(CharacterSet::empty().add_char('i'), false, 0, 3),
|
|
(CharacterSet::empty().add_char('f'), false, 0, 4),
|
|
],
|
|
vec![
|
|
NfaTransition {
|
|
characters: CharacterSet::empty().add_char('d'),
|
|
is_separator: false,
|
|
precedence: 0,
|
|
states: vec![1, 2],
|
|
},
|
|
NfaTransition {
|
|
characters: CharacterSet::empty().add_char('f'),
|
|
is_separator: false,
|
|
precedence: 0,
|
|
states: vec![1, 4],
|
|
},
|
|
NfaTransition {
|
|
characters: CharacterSet::empty().add_char('i'),
|
|
is_separator: false,
|
|
precedence: 0,
|
|
states: vec![1, 3],
|
|
},
|
|
NfaTransition {
|
|
characters: CharacterSet::empty()
|
|
.add_range('a', 'c')
|
|
.add_char('e')
|
|
.add_range('g', 'h')
|
|
.add_range('j', 'z'),
|
|
is_separator: false,
|
|
precedence: 0,
|
|
states: vec![1],
|
|
},
|
|
],
|
|
),
|
|
// negated character class followed by an individual character
|
|
(
|
|
vec![
|
|
(CharacterSet::empty().add_char('0'), false, 0, 1),
|
|
(CharacterSet::empty().add_char('b'), false, 0, 2),
|
|
(
|
|
CharacterSet::empty().add_range('a', 'f').negate(),
|
|
false,
|
|
0,
|
|
3,
|
|
),
|
|
(CharacterSet::empty().add_char('c'), false, 0, 4),
|
|
],
|
|
vec![
|
|
NfaTransition {
|
|
characters: CharacterSet::empty().add_char('0'),
|
|
precedence: 0,
|
|
states: vec![1, 3],
|
|
is_separator: false,
|
|
},
|
|
NfaTransition {
|
|
characters: CharacterSet::empty().add_char('b'),
|
|
precedence: 0,
|
|
states: vec![2],
|
|
is_separator: false,
|
|
},
|
|
NfaTransition {
|
|
characters: CharacterSet::empty().add_char('c'),
|
|
precedence: 0,
|
|
states: vec![4],
|
|
is_separator: false,
|
|
},
|
|
NfaTransition {
|
|
characters: CharacterSet::empty()
|
|
.add_range('a', 'f')
|
|
.add_char('0')
|
|
.negate(),
|
|
precedence: 0,
|
|
states: vec![3],
|
|
is_separator: false,
|
|
},
|
|
],
|
|
),
|
|
// multiple negated character classes
|
|
(
|
|
vec![
|
|
(CharacterSet::from_char('a'), false, 0, 1),
|
|
(CharacterSet::from_range('a', 'c').negate(), false, 0, 2),
|
|
(CharacterSet::from_char('g'), false, 0, 6),
|
|
(CharacterSet::from_range('d', 'f').negate(), false, 0, 3),
|
|
(CharacterSet::from_range('g', 'i').negate(), false, 0, 4),
|
|
(CharacterSet::from_char('g'), false, 0, 5),
|
|
],
|
|
vec![
|
|
NfaTransition {
|
|
characters: CharacterSet::from_char('a'),
|
|
precedence: 0,
|
|
states: vec![1, 3, 4],
|
|
is_separator: false,
|
|
},
|
|
NfaTransition {
|
|
characters: CharacterSet::from_char('g'),
|
|
precedence: 0,
|
|
states: vec![2, 3, 5, 6],
|
|
is_separator: false,
|
|
},
|
|
NfaTransition {
|
|
characters: CharacterSet::from_range('b', 'c'),
|
|
precedence: 0,
|
|
states: vec![3, 4],
|
|
is_separator: false,
|
|
},
|
|
NfaTransition {
|
|
characters: CharacterSet::from_range('h', 'i'),
|
|
precedence: 0,
|
|
states: vec![2, 3],
|
|
is_separator: false,
|
|
},
|
|
NfaTransition {
|
|
characters: CharacterSet::from_range('d', 'f'),
|
|
precedence: 0,
|
|
states: vec![2, 4],
|
|
is_separator: false,
|
|
},
|
|
NfaTransition {
|
|
characters: CharacterSet::from_range('a', 'i').negate(),
|
|
precedence: 0,
|
|
states: vec![2, 3, 4],
|
|
is_separator: false,
|
|
},
|
|
],
|
|
),
|
|
// disjoint characters with same state
|
|
(
|
|
vec![
|
|
(CharacterSet::from_char('a'), false, 0, 1),
|
|
(CharacterSet::from_char('b'), false, 0, 2),
|
|
(CharacterSet::from_char('c'), false, 0, 1),
|
|
(CharacterSet::from_char('d'), false, 0, 1),
|
|
(CharacterSet::from_char('e'), false, 0, 2),
|
|
],
|
|
vec![
|
|
NfaTransition {
|
|
characters: CharacterSet::empty().add_char('a').add_range('c', 'd'),
|
|
precedence: 0,
|
|
states: vec![1],
|
|
is_separator: false,
|
|
},
|
|
NfaTransition {
|
|
characters: CharacterSet::empty().add_char('b').add_char('e'),
|
|
precedence: 0,
|
|
states: vec![2],
|
|
is_separator: false,
|
|
},
|
|
],
|
|
),
|
|
];
|
|
|
|
for (i, row) in table.iter().enumerate() {
|
|
assert_eq!(
|
|
NfaCursor::group_transitions(
|
|
row.0
|
|
.iter()
|
|
.map(|(chars, is_sep, prec, state)| (chars, *is_sep, *prec, *state))
|
|
),
|
|
row.1,
|
|
"row {}",
|
|
i
|
|
);
|
|
}
|
|
}
|
|
|
|
#[test]
|
|
fn test_character_set_remove_intersection() {
|
|
struct Row {
|
|
left: CharacterSet,
|
|
right: CharacterSet,
|
|
left_only: CharacterSet,
|
|
right_only: CharacterSet,
|
|
intersection: CharacterSet,
|
|
}
|
|
|
|
let rows = [
|
|
// [ L ]
|
|
// [ R ]
|
|
Row {
|
|
left: CharacterSet::from_range('a', 'f'),
|
|
right: CharacterSet::from_range('g', 'm'),
|
|
left_only: CharacterSet::from_range('a', 'f'),
|
|
right_only: CharacterSet::from_range('g', 'm'),
|
|
intersection: CharacterSet::empty(),
|
|
},
|
|
// [ L ]
|
|
// [ R ]
|
|
Row {
|
|
left: CharacterSet::from_range('a', 'f'),
|
|
right: CharacterSet::from_range('c', 'i'),
|
|
left_only: CharacterSet::from_range('a', 'b'),
|
|
right_only: CharacterSet::from_range('g', 'i'),
|
|
intersection: CharacterSet::from_range('c', 'f'),
|
|
},
|
|
// [ L ]
|
|
// [ R ]
|
|
Row {
|
|
left: CharacterSet::from_range('a', 'f'),
|
|
right: CharacterSet::from_range('d', 'f'),
|
|
left_only: CharacterSet::from_range('a', 'c'),
|
|
right_only: CharacterSet::empty(),
|
|
intersection: CharacterSet::from_range('d', 'f'),
|
|
},
|
|
// [ L ]
|
|
// [ R ]
|
|
Row {
|
|
left: CharacterSet::from_range('a', 'm'),
|
|
right: CharacterSet::from_range('d', 'f'),
|
|
left_only: CharacterSet::empty()
|
|
.add_range('a', 'c')
|
|
.add_range('g', 'm'),
|
|
right_only: CharacterSet::empty(),
|
|
intersection: CharacterSet::from_range('d', 'f'),
|
|
},
|
|
// [ L1 ] [ L2 ]
|
|
// [ R ]
|
|
Row {
|
|
left: CharacterSet::empty()
|
|
.add_range('a', 'e')
|
|
.add_range('h', 'l'),
|
|
right: CharacterSet::from_range('c', 'i'),
|
|
left_only: CharacterSet::empty()
|
|
.add_range('a', 'b')
|
|
.add_range('j', 'l'),
|
|
right_only: CharacterSet::from_range('f', 'g'),
|
|
intersection: CharacterSet::empty()
|
|
.add_range('c', 'e')
|
|
.add_range('h', 'i'),
|
|
},
|
|
];
|
|
|
|
for (i, row) in rows.iter().enumerate() {
|
|
let mut left = row.left.clone();
|
|
let mut right = row.right.clone();
|
|
assert_eq!(
|
|
left.remove_intersection(&mut right),
|
|
row.intersection,
|
|
"row {}a: {:?} && {:?}",
|
|
i,
|
|
row.left,
|
|
row.right
|
|
);
|
|
assert_eq!(
|
|
left, row.left_only,
|
|
"row {}a: {:?} - {:?}",
|
|
i, row.left, row.right
|
|
);
|
|
assert_eq!(
|
|
right, row.right_only,
|
|
"row {}a: {:?} - {:?}",
|
|
i, row.right, row.left
|
|
);
|
|
|
|
let mut left = row.left.clone();
|
|
let mut right = row.right.clone();
|
|
assert_eq!(
|
|
right.remove_intersection(&mut left),
|
|
row.intersection,
|
|
"row {}b: {:?} && {:?}",
|
|
i,
|
|
row.left,
|
|
row.right
|
|
);
|
|
assert_eq!(
|
|
left, row.left_only,
|
|
"row {}b: {:?} - {:?}",
|
|
i, row.left, row.right
|
|
);
|
|
assert_eq!(
|
|
right, row.right_only,
|
|
"row {}b: {:?} - {:?}",
|
|
i, row.right, row.left
|
|
);
|
|
}
|
|
}
|
|
|
|
#[test]
|
|
fn test_character_set_does_intersect() {
|
|
let (a, b) = (CharacterSet::empty(), CharacterSet::empty());
|
|
assert!(!a.does_intersect(&b));
|
|
assert!(!b.does_intersect(&a));
|
|
|
|
let (a, b) = (
|
|
CharacterSet::empty().add_char('a'),
|
|
CharacterSet::empty().add_char('a'),
|
|
);
|
|
assert!(a.does_intersect(&b));
|
|
assert!(b.does_intersect(&a));
|
|
|
|
let (a, b) = (
|
|
CharacterSet::empty().add_char('b'),
|
|
CharacterSet::empty().add_char('a').add_char('c'),
|
|
);
|
|
assert!(!a.does_intersect(&b));
|
|
assert!(!b.does_intersect(&a));
|
|
|
|
let (a, b) = (
|
|
CharacterSet::from_char('b'),
|
|
CharacterSet::from_range('a', 'c'),
|
|
);
|
|
assert!(a.does_intersect(&b));
|
|
assert!(b.does_intersect(&a));
|
|
|
|
let (a, b) = (
|
|
CharacterSet::from_char('b'),
|
|
CharacterSet::from_range('a', 'c').negate(),
|
|
);
|
|
assert!(!a.does_intersect(&b));
|
|
assert!(!b.does_intersect(&a));
|
|
|
|
let (a, b) = (
|
|
CharacterSet::from_char('a').negate(),
|
|
CharacterSet::from_char('a').negate(),
|
|
);
|
|
assert!(a.does_intersect(&b));
|
|
assert!(b.does_intersect(&a));
|
|
|
|
let (a, b) = (
|
|
CharacterSet::from_char('c'),
|
|
CharacterSet::from_char('a').negate(),
|
|
);
|
|
assert!(a.does_intersect(&b));
|
|
assert!(b.does_intersect(&a));
|
|
}
|
|
|
|
#[test]
|
|
fn test_character_set_get_ranges() {
|
|
struct Row {
|
|
chars: Vec<char>,
|
|
ruled_out_chars: Vec<char>,
|
|
expected_ranges: Vec<Range<char>>,
|
|
}
|
|
|
|
let table = [
|
|
Row {
|
|
chars: vec!['a'],
|
|
ruled_out_chars: vec![],
|
|
expected_ranges: vec!['a'..'a'],
|
|
},
|
|
Row {
|
|
chars: vec!['a', 'b', 'c', 'e', 'z'],
|
|
ruled_out_chars: vec![],
|
|
expected_ranges: vec!['a'..'c', 'e'..'e', 'z'..'z'],
|
|
},
|
|
Row {
|
|
chars: vec!['a', 'b', 'c', 'e', 'h', 'z'],
|
|
ruled_out_chars: vec!['d', 'f', 'g'],
|
|
expected_ranges: vec!['a'..'h', 'z'..'z'],
|
|
},
|
|
];
|
|
|
|
for Row {
|
|
chars,
|
|
ruled_out_chars,
|
|
expected_ranges,
|
|
} in table.iter()
|
|
{
|
|
let ruled_out_chars = ruled_out_chars
|
|
.into_iter()
|
|
.map(|c: &char| *c as u32)
|
|
.collect();
|
|
let mut set = CharacterSet::empty();
|
|
for c in chars {
|
|
set = set.add_char(*c);
|
|
}
|
|
let ranges = set.simplify_ignoring(&ruled_out_chars);
|
|
assert_eq!(ranges, *expected_ranges);
|
|
}
|
|
}
|
|
}
|