summaryrefslogtreecommitdiff
path: root/src/source.rs
diff options
context:
space:
mode:
Diffstat (limited to 'src/source.rs')
-rw-r--r--src/source.rs430
1 files changed, 0 insertions, 430 deletions
diff --git a/src/source.rs b/src/source.rs
deleted file mode 100644
index 69e72d6b..00000000
--- a/src/source.rs
+++ /dev/null
@@ -1,430 +0,0 @@
-//! Source file management.
-
-use std::fmt::{self, Debug, Formatter};
-use std::hash::{Hash, Hasher};
-use std::ops::Range;
-use std::path::{Path, PathBuf};
-
-use comemo::Prehashed;
-use unscanny::Scanner;
-
-use crate::diag::SourceResult;
-use crate::parse::{is_newline, parse, reparse};
-use crate::syntax::ast::Markup;
-use crate::syntax::{Span, SyntaxNode};
-use crate::util::{PathExt, StrExt};
-
-/// A source file.
-///
-/// All line and column indices start at zero, just like byte indices. Only for
-/// user-facing display, you should add 1 to them.
-pub struct Source {
- id: SourceId,
- path: PathBuf,
- text: Prehashed<String>,
- lines: Vec<Line>,
- root: Prehashed<SyntaxNode>,
-}
-
-impl Source {
- /// Create a new source file.
- pub fn new(id: SourceId, path: &Path, text: String) -> Self {
- let lines = std::iter::once(Line { byte_idx: 0, utf16_idx: 0 })
- .chain(lines(0, 0, &text))
- .collect();
-
- let mut root = parse(&text);
- root.numberize(id, Span::FULL).unwrap();
-
- Self {
- id,
- path: path.normalize(),
- text: Prehashed::new(text),
- lines,
- root: Prehashed::new(root),
- }
- }
-
- /// Create a source file without a real id and path, usually for testing.
- pub fn detached(text: impl Into<String>) -> Self {
- Self::new(SourceId::detached(), Path::new(""), text.into())
- }
-
- /// Create a source file with the same synthetic span for all nodes.
- pub fn synthesized(text: impl Into<String>, span: Span) -> Self {
- let mut file = Self::detached(text);
- let mut root = file.root.into_inner();
- root.synthesize(span);
- file.root = Prehashed::new(root);
- file.id = span.source();
- file
- }
-
- /// The root node of the file's untyped syntax tree.
- pub fn root(&self) -> &SyntaxNode {
- &self.root
- }
-
- /// The root node of the file's typed abstract syntax tree.
- pub fn ast(&self) -> SourceResult<Markup> {
- let errors = self.root.errors();
- if errors.is_empty() {
- Ok(self.root.cast().expect("root node must be markup"))
- } else {
- Err(Box::new(errors))
- }
- }
-
- /// The id of the source file.
- pub fn id(&self) -> SourceId {
- self.id
- }
-
- /// The normalized path to the source file.
- pub fn path(&self) -> &Path {
- &self.path
- }
-
- /// The whole source as a string slice.
- pub fn text(&self) -> &str {
- &self.text
- }
-
- /// Slice out the part of the source code enclosed by the range.
- pub fn get(&self, range: Range<usize>) -> Option<&str> {
- self.text.get(range)
- }
-
- /// Fully replace the source text and increase the revision number.
- pub fn replace(&mut self, text: String) {
- self.text = Prehashed::new(text);
- self.lines = vec![Line { byte_idx: 0, utf16_idx: 0 }];
- self.lines.extend(lines(0, 0, &self.text));
- let mut root = parse(&self.text);
- root.numberize(self.id(), Span::FULL).unwrap();
- self.root = Prehashed::new(root);
- }
-
- /// Edit the source file by replacing the given range and increase the
- /// revision number.
- ///
- /// Returns the range in the new source that was ultimately reparsed.
- ///
- /// The method panics if the `replace` range is out of bounds.
- pub fn edit(&mut self, replace: Range<usize>, with: &str) -> Range<usize> {
- let start_byte = replace.start;
- let start_utf16 = self.byte_to_utf16(replace.start).unwrap();
- let mut text = std::mem::take(&mut self.text).into_inner();
- text.replace_range(replace.clone(), with);
- self.text = Prehashed::new(text);
-
- // Remove invalidated line starts.
- let line = self.byte_to_line(start_byte).unwrap();
- self.lines.truncate(line + 1);
-
- // Handle adjoining of \r and \n.
- if self.text[.. start_byte].ends_with('\r') && with.starts_with('\n') {
- self.lines.pop();
- }
-
- // Recalculate the line starts after the edit.
- self.lines
- .extend(lines(start_byte, start_utf16, &self.text[start_byte ..]));
-
- // Incrementally reparse the replaced range.
- let mut root = std::mem::take(&mut self.root).into_inner();
- let range = reparse(&mut root, &self.text, replace, with.len());
- self.root = Prehashed::new(root);
- range
- }
-
- /// Get the length of the file in UTF-8 encoded bytes.
- pub fn len_bytes(&self) -> usize {
- self.text.len()
- }
-
- /// Get the length of the file in UTF-16 code units.
- pub fn len_utf16(&self) -> usize {
- let last = self.lines.last().unwrap();
- last.utf16_idx + self.text[last.byte_idx ..].len_utf16()
- }
-
- /// Get the length of the file in lines.
- pub fn len_lines(&self) -> usize {
- self.lines.len()
- }
-
- /// Map a span that points into this source file to a byte range.
- ///
- /// Panics if the span does not point into this source file.
- pub fn range(&self, span: Span) -> Range<usize> {
- self.root
- .range(span, 0)
- .expect("span does not point into this source file")
- }
-
- /// Return the index of the UTF-16 code unit at the byte index.
- pub fn byte_to_utf16(&self, byte_idx: usize) -> Option<usize> {
- let line_idx = self.byte_to_line(byte_idx)?;
- let line = self.lines.get(line_idx)?;
- let head = self.text.get(line.byte_idx .. byte_idx)?;
- Some(line.utf16_idx + head.len_utf16())
- }
-
- /// Return the index of the line that contains the given byte index.
- pub fn byte_to_line(&self, byte_idx: usize) -> Option<usize> {
- (byte_idx <= self.text.len()).then(|| {
- match self.lines.binary_search_by_key(&byte_idx, |line| line.byte_idx) {
- Ok(i) => i,
- Err(i) => i - 1,
- }
- })
- }
-
- /// Return the index of the column at the byte index.
- ///
- /// The column is defined as the number of characters in the line before the
- /// byte index.
- pub fn byte_to_column(&self, byte_idx: usize) -> Option<usize> {
- let line = self.byte_to_line(byte_idx)?;
- let start = self.line_to_byte(line)?;
- let head = self.get(start .. byte_idx)?;
- Some(head.chars().count())
- }
-
- /// Return the byte index at the UTF-16 code unit.
- pub fn utf16_to_byte(&self, utf16_idx: usize) -> Option<usize> {
- let line = self.lines.get(
- match self.lines.binary_search_by_key(&utf16_idx, |line| line.utf16_idx) {
- Ok(i) => i,
- Err(i) => i - 1,
- },
- )?;
-
- let mut k = line.utf16_idx;
- for (i, c) in self.text[line.byte_idx ..].char_indices() {
- if k >= utf16_idx {
- return Some(line.byte_idx + i);
- }
- k += c.len_utf16();
- }
-
- (k == utf16_idx).then(|| self.text.len())
- }
-
-
- /// Return the byte position at which the given line starts.
- pub fn line_to_byte(&self, line_idx: usize) -> Option<usize> {
- self.lines.get(line_idx).map(|line| line.byte_idx)
- }
-
- /// Return the range which encloses the given line.
- pub fn line_to_range(&self, line_idx: usize) -> Option<Range<usize>> {
- let start = self.line_to_byte(line_idx)?;
- let end = self.line_to_byte(line_idx + 1).unwrap_or(self.text.len());
- Some(start .. end)
- }
-
- /// Return the byte index of the given (line, column) pair.
- ///
- /// The column defines the number of characters to go beyond the start of
- /// the line.
- pub fn line_column_to_byte(
- &self,
- line_idx: usize,
- column_idx: usize,
- ) -> Option<usize> {
- let range = self.line_to_range(line_idx)?;
- let line = self.get(range.clone())?;
- let mut chars = line.chars();
- for _ in 0 .. column_idx {
- chars.next();
- }
- Some(range.start + (line.len() - chars.as_str().len()))
- }
-}
-
-impl Debug for Source {
- fn fmt(&self, f: &mut Formatter) -> fmt::Result {
- write!(f, "Source({})", self.path.display())
- }
-}
-
-impl Hash for Source {
- fn hash<H: Hasher>(&self, state: &mut H) {
- self.id.hash(state);
- self.path.hash(state);
- self.text.hash(state);
- self.root.hash(state);
- }
-}
-
-/// A unique identifier for a loaded source file.
-#[derive(Debug, Copy, Clone, Eq, PartialEq, Hash)]
-pub struct SourceId(u16);
-
-impl SourceId {
- /// Create a new source id for a file that is not part of a store.
- pub const fn detached() -> Self {
- Self(u16::MAX)
- }
-
- /// Create a source id from a number.
- pub const fn from_u16(v: u16) -> Self {
- Self(v)
- }
-
- /// Extract the underlying number.
- pub const fn into_u16(self) -> u16 {
- self.0
- }
-}
-
-/// Metadata about a line.
-#[derive(Debug, Copy, Clone, Eq, PartialEq)]
-struct Line {
- /// The UTF-8 byte offset where the line starts.
- byte_idx: usize,
- /// The UTF-16 codepoint offset where the line starts.
- utf16_idx: usize,
-}
-
-/// Iterate over the lines in the string.
-fn lines(
- byte_offset: usize,
- utf16_offset: usize,
- text: &str,
-) -> impl Iterator<Item = Line> + '_ {
- let mut s = Scanner::new(text);
- let mut utf16_idx = utf16_offset;
-
- std::iter::from_fn(move || {
- s.eat_until(|c: char| {
- utf16_idx += c.len_utf16();
- is_newline(c)
- });
-
- if s.done() {
- return None;
- }
-
- if s.eat() == Some('\r') && s.eat_if('\n') {
- utf16_idx += 1;
- }
-
- Some(Line {
- byte_idx: byte_offset + s.cursor(),
- utf16_idx,
- })
- })
-}
-
-#[cfg(test)]
-mod tests {
- use super::*;
-
- const TEST: &str = "ä\tcde\nf💛g\r\nhi\rjkl";
-
- #[test]
- fn test_source_file_new() {
- let source = Source::detached(TEST);
- assert_eq!(source.lines, [
- Line { byte_idx: 0, utf16_idx: 0 },
- Line { byte_idx: 7, utf16_idx: 6 },
- Line { byte_idx: 15, utf16_idx: 12 },
- Line { byte_idx: 18, utf16_idx: 15 },
- ]);
- }
-
- #[test]
- fn test_source_file_pos_to_line() {
- let source = Source::detached(TEST);
- assert_eq!(source.byte_to_line(0), Some(0));
- assert_eq!(source.byte_to_line(2), Some(0));
- assert_eq!(source.byte_to_line(6), Some(0));
- assert_eq!(source.byte_to_line(7), Some(1));
- assert_eq!(source.byte_to_line(8), Some(1));
- assert_eq!(source.byte_to_line(12), Some(1));
- assert_eq!(source.byte_to_line(21), Some(3));
- assert_eq!(source.byte_to_line(22), None);
- }
-
- #[test]
- fn test_source_file_pos_to_column() {
- let source = Source::detached(TEST);
- assert_eq!(source.byte_to_column(0), Some(0));
- assert_eq!(source.byte_to_column(2), Some(1));
- assert_eq!(source.byte_to_column(6), Some(5));
- assert_eq!(source.byte_to_column(7), Some(0));
- assert_eq!(source.byte_to_column(8), Some(1));
- assert_eq!(source.byte_to_column(12), Some(2));
- }
-
- #[test]
- fn test_source_file_utf16() {
- #[track_caller]
- fn roundtrip(source: &Source, byte_idx: usize, utf16_idx: usize) {
- let middle = source.byte_to_utf16(byte_idx).unwrap();
- let result = source.utf16_to_byte(middle).unwrap();
- assert_eq!(middle, utf16_idx);
- assert_eq!(result, byte_idx);
- }
-
- let source = Source::detached(TEST);
- roundtrip(&source, 0, 0);
- roundtrip(&source, 2, 1);
- roundtrip(&source, 3, 2);
- roundtrip(&source, 8, 7);
- roundtrip(&source, 12, 9);
- roundtrip(&source, 21, 18);
- assert_eq!(source.byte_to_utf16(22), None);
- assert_eq!(source.utf16_to_byte(19), None);
- }
-
- #[test]
- fn test_source_file_roundtrip() {
- #[track_caller]
- fn roundtrip(source: &Source, byte_idx: usize) {
- let line = source.byte_to_line(byte_idx).unwrap();
- let column = source.byte_to_column(byte_idx).unwrap();
- let result = source.line_column_to_byte(line, column).unwrap();
- assert_eq!(result, byte_idx);
- }
-
- let source = Source::detached(TEST);
- roundtrip(&source, 0);
- roundtrip(&source, 7);
- roundtrip(&source, 12);
- roundtrip(&source, 21);
- }
-
- #[test]
- fn test_source_file_edit() {
- #[track_caller]
- fn test(prev: &str, range: Range<usize>, with: &str, after: &str) {
- let mut source = Source::detached(prev);
- let result = Source::detached(after);
- source.edit(range, with);
- assert_eq!(source.text, result.text);
- assert_eq!(source.lines, result.lines);
- assert_eq!(*source.root, *result.root);
- }
-
- // Test inserting at the begining.
- test("abc\n", 0 .. 0, "hi\n", "hi\nabc\n");
- test("\nabc", 0 .. 0, "hi\r", "hi\r\nabc");
-
- // Test editing in the middle.
- test(TEST, 4 .. 16, "❌", "ä\tc❌i\rjkl");
-
- // Test appending.
- test("abc\ndef", 7 .. 7, "hi", "abc\ndefhi");
- test("abc\ndef\n", 8 .. 8, "hi", "abc\ndef\nhi");
-
- // Test appending with adjoining \r and \n.
- test("abc\ndef\r", 8 .. 8, "\nghi", "abc\ndef\r\nghi");
-
- // Test removing everything.
- test(TEST, 0 .. 21, "", "");
- }
-}
generated by cgit v1.2.3 (git 2.25.1) at 2026-03-16 19:12:10 +0000