aboutsummaryrefslogtreecommitdiff
path: root/helix-core/src
diff options
context:
space:
mode:
Diffstat (limited to 'helix-core/src')
-rw-r--r--helix-core/src/graphemes.rs213
-rw-r--r--helix-core/src/lib.rs1
-rw-r--r--helix-core/src/selection.rs4
-rw-r--r--helix-core/src/state.rs86
4 files changed, 297 insertions, 7 deletions
diff --git a/helix-core/src/graphemes.rs b/helix-core/src/graphemes.rs
new file mode 100644
index 00000000..ec4e9d24
--- /dev/null
+++ b/helix-core/src/graphemes.rs
@@ -0,0 +1,213 @@
+// Based on https://github.com/cessen/led/blob/c4fa72405f510b7fd16052f90a598c429b3104a6/src/graphemes.rs
+use ropey::{iter::Chunks, str_utils::byte_to_char_idx, RopeSlice};
+use unicode_segmentation::{GraphemeCursor, GraphemeIncomplete};
+use unicode_width::UnicodeWidthStr;
+
+pub fn grapheme_width(g: &str) -> usize {
+ if g.as_bytes()[0] <= 127 {
+ // Fast-path ascii.
+ // Point 1: theoretically, ascii control characters should have zero
+ // width, but in our case we actually want them to have width: if they
+ // show up in text, we want to treat them as textual elements that can
+ // be editied. So we can get away with making all ascii single width
+ // here.
+ // Point 2: we're only examining the first codepoint here, which means
+ // we're ignoring graphemes formed with combining characters. However,
+ // if it starts with ascii, it's going to be a single-width grapeheme
+ // regardless, so, again, we can get away with that here.
+ // Point 3: we're only examining the first _byte_. But for utf8, when
+ // checking for ascii range values only, that works.
+ 1
+ } else {
+ // We use max(1) here because all grapeheme clusters--even illformed
+ // ones--should have at least some width so they can be edited
+ // properly.
+ UnicodeWidthStr::width(g).max(1)
+ }
+}
+
+pub fn nth_prev_grapheme_boundary(slice: &RopeSlice, char_idx: usize, n: usize) -> usize {
+ // TODO: implement this more efficiently. This has to do a lot of
+ // re-scanning of rope chunks. Probably move the main implementation here,
+ // and have prev_grapheme_boundary call this instead.
+ let mut char_idx = char_idx;
+ for _ in 0..n {
+ char_idx = prev_grapheme_boundary(slice, char_idx);
+ }
+ char_idx
+}
+
+/// Finds the previous grapheme boundary before the given char position.
+pub fn prev_grapheme_boundary(slice: &RopeSlice, char_idx: usize) -> usize {
+ // Bounds check
+ debug_assert!(char_idx <= slice.len_chars());
+
+ // We work with bytes for this, so convert.
+ let byte_idx = slice.char_to_byte(char_idx);
+
+ // Get the chunk with our byte index in it.
+ let (mut chunk, mut chunk_byte_idx, mut chunk_char_idx, _) = slice.chunk_at_byte(byte_idx);
+
+ // Set up the grapheme cursor.
+ let mut gc = GraphemeCursor::new(byte_idx, slice.len_bytes(), true);
+
+ // Find the previous grapheme cluster boundary.
+ loop {
+ match gc.prev_boundary(chunk, chunk_byte_idx) {
+ Ok(None) => return 0,
+ Ok(Some(n)) => {
+ let tmp = byte_to_char_idx(chunk, n - chunk_byte_idx);
+ return chunk_char_idx + tmp;
+ }
+ Err(GraphemeIncomplete::PrevChunk) => {
+ let (a, b, c, _) = slice.chunk_at_byte(chunk_byte_idx - 1);
+ chunk = a;
+ chunk_byte_idx = b;
+ chunk_char_idx = c;
+ }
+ Err(GraphemeIncomplete::PreContext(n)) => {
+ let ctx_chunk = slice.chunk_at_byte(n - 1).0;
+ gc.provide_context(ctx_chunk, n - ctx_chunk.len());
+ }
+ _ => unreachable!(),
+ }
+ }
+}
+
+pub fn nth_next_grapheme_boundary(slice: &RopeSlice, char_idx: usize, n: usize) -> usize {
+ // TODO: implement this more efficiently. This has to do a lot of
+ // re-scanning of rope chunks. Probably move the main implementation here,
+ // and have next_grapheme_boundary call this instead.
+ let mut char_idx = char_idx;
+ for _ in 0..n {
+ char_idx = next_grapheme_boundary(slice, char_idx);
+ }
+ char_idx
+}
+
+/// Finds the next grapheme boundary after the given char position.
+pub fn next_grapheme_boundary(slice: &RopeSlice, char_idx: usize) -> usize {
+ // Bounds check
+ debug_assert!(char_idx <= slice.len_chars());
+
+ // We work with bytes for this, so convert.
+ let byte_idx = slice.char_to_byte(char_idx);
+
+ // Get the chunk with our byte index in it.
+ let (mut chunk, mut chunk_byte_idx, mut chunk_char_idx, _) = slice.chunk_at_byte(byte_idx);
+
+ // Set up the grapheme cursor.
+ let mut gc = GraphemeCursor::new(byte_idx, slice.len_bytes(), true);
+
+ // Find the next grapheme cluster boundary.
+ loop {
+ match gc.next_boundary(chunk, chunk_byte_idx) {
+ Ok(None) => return slice.len_chars(),
+ Ok(Some(n)) => {
+ let tmp = byte_to_char_idx(chunk, n - chunk_byte_idx);
+ return chunk_char_idx + tmp;
+ }
+ Err(GraphemeIncomplete::NextChunk) => {
+ chunk_byte_idx += chunk.len();
+ let (a, _, c, _) = slice.chunk_at_byte(chunk_byte_idx);
+ chunk = a;
+ chunk_char_idx = c;
+ }
+ Err(GraphemeIncomplete::PreContext(n)) => {
+ let ctx_chunk = slice.chunk_at_byte(n - 1).0;
+ gc.provide_context(ctx_chunk, n - ctx_chunk.len());
+ }
+ _ => unreachable!(),
+ }
+ }
+}
+
+/// Returns whether the given char position is a grapheme boundary.
+pub fn is_grapheme_boundary(slice: &RopeSlice, char_idx: usize) -> bool {
+ // Bounds check
+ debug_assert!(char_idx <= slice.len_chars());
+
+ // We work with bytes for this, so convert.
+ let byte_idx = slice.char_to_byte(char_idx);
+
+ // Get the chunk with our byte index in it.
+ let (chunk, chunk_byte_idx, _, _) = slice.chunk_at_byte(byte_idx);
+
+ // Set up the grapheme cursor.
+ let mut gc = GraphemeCursor::new(byte_idx, slice.len_bytes(), true);
+
+ // Determine if the given position is a grapheme cluster boundary.
+ loop {
+ match gc.is_boundary(chunk, chunk_byte_idx) {
+ Ok(n) => return n,
+ Err(GraphemeIncomplete::PreContext(n)) => {
+ let (ctx_chunk, ctx_byte_start, _, _) = slice.chunk_at_byte(n - 1);
+ gc.provide_context(ctx_chunk, ctx_byte_start);
+ }
+ _ => unreachable!(),
+ }
+ }
+}
+
+/// An iterator over the graphemes of a RopeSlice.
+#[derive(Clone)]
+pub struct RopeGraphemes<'a> {
+ text: RopeSlice<'a>,
+ chunks: Chunks<'a>,
+ cur_chunk: &'a str,
+ cur_chunk_start: usize,
+ cursor: GraphemeCursor,
+}
+
+impl<'a> RopeGraphemes<'a> {
+ pub fn new<'b>(slice: &RopeSlice<'b>) -> RopeGraphemes<'b> {
+ let mut chunks = slice.chunks();
+ let first_chunk = chunks.next().unwrap_or("");
+ RopeGraphemes {
+ text: *slice,
+ chunks: chunks,
+ cur_chunk: first_chunk,
+ cur_chunk_start: 0,
+ cursor: GraphemeCursor::new(0, slice.len_bytes(), true),
+ }
+ }
+}
+
+impl<'a> Iterator for RopeGraphemes<'a> {
+ type Item = RopeSlice<'a>;
+
+ fn next(&mut self) -> Option<RopeSlice<'a>> {
+ let a = self.cursor.cur_cursor();
+ let b;
+ loop {
+ match self
+ .cursor
+ .next_boundary(self.cur_chunk, self.cur_chunk_start)
+ {
+ Ok(None) => {
+ return None;
+ }
+ Ok(Some(n)) => {
+ b = n;
+ break;
+ }
+ Err(GraphemeIncomplete::NextChunk) => {
+ self.cur_chunk_start += self.cur_chunk.len();
+ self.cur_chunk = self.chunks.next().unwrap_or("");
+ }
+ _ => unreachable!(),
+ }
+ }
+
+ if a < self.cur_chunk_start {
+ let a_char = self.text.byte_to_char(a);
+ let b_char = self.text.byte_to_char(b);
+
+ Some(self.text.slice(a_char..b_char))
+ } else {
+ let a2 = a - self.cur_chunk_start;
+ let b2 = b - self.cur_chunk_start;
+ Some((&self.cur_chunk[a2..b2]).into())
+ }
+ }
+}
diff --git a/helix-core/src/lib.rs b/helix-core/src/lib.rs
index 421d8f3c..d2c78d3f 100644
--- a/helix-core/src/lib.rs
+++ b/helix-core/src/lib.rs
@@ -1,4 +1,5 @@
mod buffer;
+mod graphemes;
mod selection;
mod state;
mod transaction;
diff --git a/helix-core/src/selection.rs b/helix-core/src/selection.rs
index 98bbdb7f..b02560a8 100644
--- a/helix-core/src/selection.rs
+++ b/helix-core/src/selection.rs
@@ -99,8 +99,8 @@ impl Range {
/// A selection consists of one or more selection ranges.
pub struct Selection {
// TODO: decide how many ranges to inline SmallVec<[Range; 1]>
- ranges: SmallVec<[Range; 1]>,
- primary_index: usize,
+ pub(crate) ranges: SmallVec<[Range; 1]>,
+ pub(crate) primary_index: usize,
}
impl Selection {
diff --git a/helix-core/src/state.rs b/helix-core/src/state.rs
index 22de6ca7..682d298a 100644
--- a/helix-core/src/state.rs
+++ b/helix-core/src/state.rs
@@ -1,17 +1,30 @@
-use crate::{Buffer, Selection};
+use crate::graphemes::{nth_next_grapheme_boundary, nth_prev_grapheme_boundary};
+use crate::{Buffer, Selection, SelectionRange};
/// A state represents the current editor state of a single buffer.
pub struct State {
- // TODO: maybe doc: ?
- buffer: Buffer,
+ doc: Buffer,
selection: Selection,
}
+#[derive(Copy, Clone, PartialEq, Eq)]
+pub enum Direction {
+ Forward,
+ Backward,
+}
+#[derive(Copy, Clone, PartialEq, Eq)]
+pub enum Granularity {
+ Character,
+ Word,
+ Line,
+ // LineBoundary
+}
+
impl State {
#[must_use]
- pub fn new(buffer: Buffer) -> Self {
+ pub fn new(doc: Buffer) -> Self {
Self {
- buffer,
+ doc,
selection: Selection::single(0, 0),
}
}
@@ -40,4 +53,67 @@ impl State {
// syntax
// foldable
// changeFilter/transactionFilter
+
+ pub fn move_pos(
+ &self,
+ pos: usize,
+ dir: Direction,
+ granularity: Granularity,
+ n: usize,
+ ) -> usize {
+ let text = &self.doc.contents;
+ match (dir, granularity) {
+ (Direction::Backward, Granularity::Character) => {
+ nth_prev_grapheme_boundary(&text.slice(..), pos, n)
+ }
+ (Direction::Forward, Granularity::Character) => {
+ nth_next_grapheme_boundary(&text.slice(..), pos, n)
+ }
+ _ => pos,
+ }
+ }
+
+ pub fn move_selection(
+ &self,
+ sel: Selection,
+ dir: Direction,
+ granularity: Granularity,
+ // TODO: n
+ ) -> Selection {
+ // TODO: move all selections according to normal cursor move semantics by collapsing it
+ // into cursors and moving them vertically
+
+ let ranges = sel.ranges.into_iter().map(|range| {
+ // let pos = if !range.is_empty() {
+ // // if selection already exists, bump it to the start or end of current select first
+ // if dir == Direction::Backward {
+ // range.from()
+ // } else {
+ // range.to()
+ // }
+ // } else {
+ let pos = self.move_pos(range.head, dir, granularity, 1)
+ // };
+ SelectionRange::new(pos, pos)
+ });
+
+ Selection::new(ranges.collect(), sel.primary_index)
+ // TODO: update selection in state via transaction
+ }
+
+ pub fn extend_selection(
+ &self,
+ sel: Selection,
+ dir: Direction,
+ granularity: Granularity,
+ n: usize,
+ ) -> Selection {
+ let ranges = sel.ranges.into_iter().map(|range| {
+ let pos = self.move_pos(range.head, dir, granularity, n);
+ SelectionRange::new(range.anchor, pos)
+ });
+
+ Selection::new(ranges.collect(), sel.primary_index)
+ // TODO: update selection in state via transaction
+ }
}