diff --git a/helix-stdx/src/path.rs b/helix-stdx/src/path.rs index 53081b0f0..e2ef9a58a 100644 --- a/helix-stdx/src/path.rs +++ b/helix-stdx/src/path.rs @@ -269,7 +269,7 @@ pub fn get_path_suffix(src: RopeSlice<'_>, match_single_file: bool) -> Option: Sized { fn ends_with(self, text: &str) -> bool; fn starts_with(self, text: &str) -> bool; - fn regex_input(self) -> RegexInput>; + fn regex_input(self) -> RegexInput>; fn regex_input_at_bytes>( self, byte_range: R, - ) -> RegexInput>; - fn regex_input_at>(self, char_range: R) -> RegexInput>; + ) -> RegexInput>; + #[deprecated = "use regex_input_at_bytes instead"] + fn regex_input_at>(self, char_range: R) -> RegexInput>; fn first_non_whitespace_char(self) -> Option; fn last_non_whitespace_char(self) -> Option; - /// Finds the closest byte index not exceeding `byte_idx` which lies on a character boundary. - /// - /// If `byte_idx` already lies on a character boundary then it is returned as-is. When - /// `byte_idx` lies between two character boundaries, this function returns the byte index of - /// the lesser / earlier / left-hand-side boundary. - /// - /// # Example - /// - /// ``` - /// # use ropey::RopeSlice; - /// # use helix_stdx::rope::RopeSliceExt; - /// let text = RopeSlice::from("⌚"); // three bytes: e2 8c 9a - /// assert_eq!(text.floor_char_boundary(0), 0); - /// assert_eq!(text.floor_char_boundary(1), 0); - /// assert_eq!(text.floor_char_boundary(2), 0); - /// assert_eq!(text.floor_char_boundary(3), 3); - /// ``` - fn floor_char_boundary(self, byte_idx: usize) -> usize; - /// Finds the closest byte index not below `byte_idx` which lies on a character boundary. - /// - /// If `byte_idx` already lies on a character boundary then it is returned as-is. When - /// `byte_idx` lies between two character boundaries, this function returns the byte index of - /// the greater / later / right-hand-side boundary. - /// - /// # Example - /// - /// ``` - /// # use ropey::RopeSlice; - /// # use helix_stdx::rope::RopeSliceExt; - /// let text = RopeSlice::from("⌚"); // three bytes: e2 8c 9a - /// assert_eq!(text.ceil_char_boundary(0), 0); - /// assert_eq!(text.ceil_char_boundary(1), 3); - /// assert_eq!(text.ceil_char_boundary(2), 3); - /// assert_eq!(text.ceil_char_boundary(3), 3); - /// ``` - fn ceil_char_boundary(self, byte_idx: usize) -> usize; - /// Checks whether the given `byte_idx` lies on a character boundary. - /// - /// # Example - /// - /// ``` - /// # use ropey::RopeSlice; - /// # use helix_stdx::rope::RopeSliceExt; - /// let text = RopeSlice::from("⌚"); // three bytes: e2 8c 9a - /// assert!(text.is_char_boundary(0)); - /// assert!(!text.is_char_boundary(1)); - /// assert!(!text.is_char_boundary(2)); - /// assert!(text.is_char_boundary(3)); - /// ``` - #[allow(clippy::wrong_self_convention)] - fn is_char_boundary(self, byte_idx: usize) -> bool; /// Finds the closest byte index not exceeding `byte_idx` which lies on a grapheme cluster /// boundary. /// @@ -82,9 +30,10 @@ pub trait RopeSliceExt<'a>: Sized { /// # Example /// /// ``` - /// # use ropey::RopeSlice; + /// # use ropey::{RopeSlice, Rope}; /// # use helix_stdx::rope::RopeSliceExt; - /// let text = RopeSlice::from("\r\n"); // U+000D U+000A, hex: 0d 0a + /// let text = Rope::from_str("\r\n"); // U+000D U+000A, hex: 0d 0a + /// let text = text.slice(..); /// assert_eq!(text.floor_grapheme_boundary(0), 0); /// assert_eq!(text.floor_grapheme_boundary(1), 0); /// assert_eq!(text.floor_grapheme_boundary(2), 2); @@ -102,9 +51,10 @@ pub trait RopeSliceExt<'a>: Sized { /// # Example /// /// ``` - /// # use ropey::RopeSlice; + /// # use ropey::{RopeSlice, Rope}; /// # use helix_stdx::rope::RopeSliceExt; - /// let text = RopeSlice::from("\r\n"); // U+000D U+000A, hex: 0d 0a + /// let text = Rope::from_str("\r\n"); // U+000D U+000A, hex: 0d 0a + /// let text = text.slice(..); /// assert_eq!(text.ceil_grapheme_boundary(0), 0); /// assert_eq!(text.ceil_grapheme_boundary(1), 2); /// assert_eq!(text.ceil_grapheme_boundary(2), 2); @@ -115,9 +65,10 @@ pub trait RopeSliceExt<'a>: Sized { /// # Example /// /// ``` - /// # use ropey::RopeSlice; + /// # use ropey::{RopeSlice, Rope}; /// # use helix_stdx::rope::RopeSliceExt; - /// let text = RopeSlice::from("\r\n"); // U+000D U+000A, hex: 0d 0a + /// let text = Rope::from_str("\r\n"); // U+000D U+000A, hex: 0d 0a + /// let text = text.slice(..); /// assert!(text.is_grapheme_boundary(0)); /// assert!(!text.is_grapheme_boundary(1)); /// assert!(text.is_grapheme_boundary(2)); @@ -129,10 +80,10 @@ pub trait RopeSliceExt<'a>: Sized { /// # Example /// /// ``` - /// # use ropey::RopeSlice; + /// # use ropey::{RopeSlice, Rope}; /// # use helix_stdx::rope::RopeSliceExt; - /// let text = RopeSlice::from("πŸ˜Άβ€πŸŒ«οΈπŸ΄β€β˜ οΈπŸ–ΌοΈ"); - /// let graphemes: Vec<_> = text.graphemes().collect(); + /// let text = Rope::from_str("πŸ˜Άβ€πŸŒ«οΈπŸ΄β€β˜ οΈπŸ–ΌοΈ"); + /// let graphemes: Vec<_> = text.slice(..).graphemes().collect(); /// assert_eq!(graphemes.as_slice(), &["πŸ˜Άβ€πŸŒ«οΈ", "πŸ΄β€β˜ οΈ", "πŸ–ΌοΈ"]); /// ``` fn graphemes(self) -> RopeGraphemes<'a>; @@ -144,10 +95,10 @@ pub trait RopeSliceExt<'a>: Sized { /// # Example /// /// ``` - /// # use ropey::RopeSlice; + /// # use ropey::{RopeSlice, Rope}; /// # use helix_stdx::rope::RopeSliceExt; - /// let text = RopeSlice::from("πŸ˜Άβ€πŸŒ«οΈπŸ΄β€β˜ οΈπŸ–ΌοΈ"); - /// let graphemes: Vec<_> = text.graphemes_rev().collect(); + /// let text = Rope::from_str("πŸ˜Άβ€πŸŒ«οΈπŸ΄β€β˜ οΈπŸ–ΌοΈ"); + /// let graphemes: Vec<_> = text.slice(..).graphemes_rev().collect(); /// assert_eq!(graphemes.as_slice(), &["πŸ–ΌοΈ", "πŸ΄β€β˜ οΈ", "πŸ˜Άβ€πŸŒ«οΈ"]); /// ``` fn graphemes_rev(self) -> RevRopeGraphemes<'a>; @@ -155,36 +106,36 @@ pub trait RopeSliceExt<'a>: Sized { impl<'a> RopeSliceExt<'a> for RopeSlice<'a> { fn ends_with(self, text: &str) -> bool { - let len = self.len_bytes(); + let len = self.len(); if len < text.len() { return false; } - self.get_byte_slice(len - text.len()..) - .is_some_and(|end| end == text) + self.try_slice(len - text.len()..) + .is_ok_and(|end| end == text) } fn starts_with(self, text: &str) -> bool { - let len = self.len_bytes(); + let len = self.len(); if len < text.len() { return false; } - self.get_byte_slice(..text.len()) - .is_some_and(|start| start == text) + self.try_slice(..text.len()) + .is_ok_and(|start| start == text) } - fn regex_input(self) -> RegexInput> { + fn regex_input(self) -> RegexInput> { RegexInput::new(self) } - fn regex_input_at>(self, char_range: R) -> RegexInput> { + fn regex_input_at>(self, char_range: R) -> RegexInput> { let start_bound = match char_range.start_bound() { - Bound::Included(&val) => Bound::Included(self.char_to_byte(val)), - Bound::Excluded(&val) => Bound::Excluded(self.char_to_byte(val)), + Bound::Included(&val) => Bound::Included(self.char_to_byte_idx(val)), + Bound::Excluded(&val) => Bound::Excluded(self.char_to_byte_idx(val)), Bound::Unbounded => Bound::Unbounded, }; let end_bound = match char_range.end_bound() { - Bound::Included(&val) => Bound::Included(self.char_to_byte(val)), - Bound::Excluded(&val) => Bound::Excluded(self.char_to_byte(val)), + Bound::Included(&val) => Bound::Included(self.char_to_byte_idx(val)), + Bound::Excluded(&val) => Bound::Excluded(self.char_to_byte_idx(val)), Bound::Unbounded => Bound::Unbounded, }; self.regex_input_at_bytes((start_bound, end_bound)) @@ -192,10 +143,10 @@ impl<'a> RopeSliceExt<'a> for RopeSlice<'a> { fn regex_input_at_bytes>( self, byte_range: R, - ) -> RegexInput> { + ) -> RegexInput> { let input = match byte_range.start_bound() { Bound::Included(&pos) | Bound::Excluded(&pos) => { - RegexInput::new(RopeyCursor::at(self, pos)) + RegexInput::new(self.chunk_cursor_at(pos)) } Bound::Unbounded => RegexInput::new(self), }; @@ -211,69 +162,22 @@ impl<'a> RopeSliceExt<'a> for RopeSlice<'a> { .map(|pos| self.len_chars() - pos - 1) } - // These three are adapted from std: - - fn floor_char_boundary(self, byte_idx: usize) -> usize { - if byte_idx >= self.len_bytes() { - self.len_bytes() - } else { - let offset = self - .bytes_at(byte_idx + 1) - .reversed() - .take(4) - .position(is_utf8_char_boundary) - // A char can only be four bytes long so we are guaranteed to find a boundary. - .unwrap(); - - byte_idx - offset - } - } - - fn ceil_char_boundary(self, byte_idx: usize) -> usize { - if byte_idx > self.len_bytes() { - self.len_bytes() - } else { - let upper_bound = self.len_bytes().min(byte_idx + 4); - self.bytes_at(byte_idx) - .position(is_utf8_char_boundary) - .map_or(upper_bound, |pos| pos + byte_idx) - } - } - - fn is_char_boundary(self, byte_idx: usize) -> bool { - if byte_idx == 0 { - return true; - } - - if byte_idx >= self.len_bytes() { - byte_idx == self.len_bytes() - } else { - is_utf8_char_boundary(self.bytes_at(byte_idx).next().unwrap()) - } - } - fn floor_grapheme_boundary(self, mut byte_idx: usize) -> usize { - if byte_idx >= self.len_bytes() { - return self.len_bytes(); + if byte_idx >= self.len() { + return self.len(); } byte_idx = self.ceil_char_boundary(byte_idx + 1); - let (mut chunk, mut chunk_byte_idx, _, _) = self.chunk_at_byte(byte_idx); - - let mut cursor = GraphemeCursor::new(byte_idx, self.len_bytes(), true); - + let mut chunk_cursor = self.chunk_cursor_at(byte_idx); + let mut cursor = GraphemeCursor::new(byte_idx, self.len(), true); loop { - match cursor.prev_boundary(chunk, chunk_byte_idx) { + match cursor.prev_boundary(chunk_cursor.chunk(), chunk_cursor.byte_offset()) { Ok(None) => return 0, Ok(Some(boundary)) => return boundary, - Err(GraphemeIncomplete::PrevChunk) => { - let (ch, ch_byte_idx, _, _) = self.chunk_at_byte(chunk_byte_idx - 1); - chunk = ch; - chunk_byte_idx = ch_byte_idx; - } + Err(GraphemeIncomplete::PrevChunk) => assert!(chunk_cursor.prev()), Err(GraphemeIncomplete::PreContext(n)) => { - let ctx_chunk = self.chunk_at_byte(n - 1).0; + let ctx_chunk = self.chunk(n - 1).0; cursor.provide_context(ctx_chunk, n - ctx_chunk.len()); } _ => unreachable!(), @@ -282,8 +186,8 @@ impl<'a> RopeSliceExt<'a> for RopeSlice<'a> { } fn ceil_grapheme_boundary(self, mut byte_idx: usize) -> usize { - if byte_idx >= self.len_bytes() { - return self.len_bytes(); + if byte_idx >= self.len() { + return self.len(); } if byte_idx == 0 { @@ -292,20 +196,15 @@ impl<'a> RopeSliceExt<'a> for RopeSlice<'a> { byte_idx = self.floor_char_boundary(byte_idx - 1); - let (mut chunk, mut chunk_byte_idx, _, _) = self.chunk_at_byte(byte_idx); - - let mut cursor = GraphemeCursor::new(byte_idx, self.len_bytes(), true); - + let mut chunk_cursor = self.chunk_cursor_at(byte_idx); + let mut cursor = GraphemeCursor::new(byte_idx, self.len(), true); loop { - match cursor.next_boundary(chunk, chunk_byte_idx) { - Ok(None) => return self.len_bytes(), + match cursor.next_boundary(chunk_cursor.chunk(), chunk_cursor.byte_offset()) { + Ok(None) => return self.len(), Ok(Some(boundary)) => return boundary, - Err(GraphemeIncomplete::NextChunk) => { - chunk_byte_idx += chunk.len(); - chunk = self.chunk_at_byte(chunk_byte_idx).0; - } + Err(GraphemeIncomplete::NextChunk) => assert!(chunk_cursor.next()), Err(GraphemeIncomplete::PreContext(n)) => { - let ctx_chunk = self.chunk_at_byte(n - 1).0; + let ctx_chunk = self.chunk(n - 1).0; cursor.provide_context(ctx_chunk, n - ctx_chunk.len()); } _ => unreachable!(), @@ -319,15 +218,13 @@ impl<'a> RopeSliceExt<'a> for RopeSlice<'a> { return false; } - let (chunk, chunk_byte_idx, _, _) = self.chunk_at_byte(byte_idx); - - let mut cursor = GraphemeCursor::new(byte_idx, self.len_bytes(), true); - + let (chunk, chunk_byte_idx) = self.chunk(byte_idx); + let mut cursor = GraphemeCursor::new(byte_idx, self.len(), true); loop { match cursor.is_boundary(chunk, chunk_byte_idx) { Ok(n) => return n, Err(GraphemeIncomplete::PreContext(n)) => { - let (ctx_chunk, ctx_byte_start, _, _) = self.chunk_at_byte(n - 1); + let (ctx_chunk, ctx_byte_start) = self.chunk(n - 1); cursor.provide_context(ctx_chunk, ctx_byte_start); } Err(_) => unreachable!(), @@ -336,61 +233,30 @@ impl<'a> RopeSliceExt<'a> for RopeSlice<'a> { } fn graphemes(self) -> RopeGraphemes<'a> { - let mut chunks = self.chunks(); - let first_chunk = chunks.next().unwrap_or(""); RopeGraphemes { + chunk_cursor: self.chunk_cursor(), text: self, - chunks, - cur_chunk: first_chunk, - cur_chunk_start: 0, - cursor: GraphemeCursor::new(0, self.len_bytes(), true), + cursor: GraphemeCursor::new(0, self.len(), true), } } fn graphemes_rev(self) -> RevRopeGraphemes<'a> { - let (mut chunks, mut cur_chunk_start, _, _) = self.chunks_at_byte(self.len_bytes()); - chunks.reverse(); - let first_chunk = chunks.next().unwrap_or(""); - cur_chunk_start -= first_chunk.len(); RevRopeGraphemes { + chunk_cursor: self.chunk_cursor_at(self.len()), text: self, - chunks, - cur_chunk: first_chunk, - cur_chunk_start, - cursor: GraphemeCursor::new(self.len_bytes(), self.len_bytes(), true), + cursor: GraphemeCursor::new(self.len(), self.len(), true), } } } -// copied from std -#[inline] -const fn is_utf8_char_boundary(b: u8) -> bool { - // This is bit magic equivalent to: b < 128 || b >= 192 - (b as i8) >= -0x40 -} - /// An iterator over the graphemes of a `RopeSlice`. -#[derive(Clone)] +#[derive(Debug, Clone)] pub struct RopeGraphemes<'a> { text: RopeSlice<'a>, - chunks: Chunks<'a>, - cur_chunk: &'a str, - cur_chunk_start: usize, + chunk_cursor: ChunkCursor<'a>, cursor: GraphemeCursor, } -impl fmt::Debug for RopeGraphemes<'_> { - fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { - f.debug_struct("RopeGraphemes") - .field("text", &self.text) - .field("chunks", &self.chunks) - .field("cur_chunk", &self.cur_chunk) - .field("cur_chunk_start", &self.cur_chunk_start) - // .field("cursor", &self.cursor) - .finish() - } -} - impl<'a> Iterator for RopeGraphemes<'a> { type Item = RopeSlice<'a>; @@ -400,7 +266,7 @@ impl<'a> Iterator for RopeGraphemes<'a> { loop { match self .cursor - .next_boundary(self.cur_chunk, self.cur_chunk_start) + .next_boundary(self.chunk_cursor.chunk(), self.chunk_cursor.byte_offset()) { Ok(None) => { return None; @@ -409,50 +275,33 @@ impl<'a> Iterator for RopeGraphemes<'a> { b = n; break; } - Err(GraphemeIncomplete::NextChunk) => { - self.cur_chunk_start += self.cur_chunk.len(); - self.cur_chunk = self.chunks.next().unwrap_or(""); - } + Err(GraphemeIncomplete::NextChunk) => assert!(self.chunk_cursor.next()), Err(GraphemeIncomplete::PreContext(idx)) => { - let (chunk, byte_idx, _, _) = self.text.chunk_at_byte(idx.saturating_sub(1)); + let (chunk, byte_idx) = self.text.chunk(idx.saturating_sub(1)); self.cursor.provide_context(chunk, byte_idx); } _ => unreachable!(), } } - if a < self.cur_chunk_start { - Some(self.text.byte_slice(a..b)) + if a < self.chunk_cursor.byte_offset() { + Some(self.text.slice(a..b)) } else { - let a2 = a - self.cur_chunk_start; - let b2 = b - self.cur_chunk_start; - Some((&self.cur_chunk[a2..b2]).into()) + let a2 = a - self.chunk_cursor.byte_offset(); + let b2 = b - self.chunk_cursor.byte_offset(); + Some((&self.chunk_cursor.chunk()[a2..b2]).into()) } } } /// An iterator over the graphemes of a `RopeSlice` in reverse. -#[derive(Clone)] +#[derive(Debug, Clone)] pub struct RevRopeGraphemes<'a> { text: RopeSlice<'a>, - chunks: Chunks<'a>, - cur_chunk: &'a str, - cur_chunk_start: usize, + chunk_cursor: ChunkCursor<'a>, cursor: GraphemeCursor, } -impl fmt::Debug for RevRopeGraphemes<'_> { - fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { - f.debug_struct("RevRopeGraphemes") - .field("text", &self.text) - .field("chunks", &self.chunks) - .field("cur_chunk", &self.cur_chunk) - .field("cur_chunk_start", &self.cur_chunk_start) - // .field("cursor", &self.cursor) - .finish() - } -} - impl<'a> Iterator for RevRopeGraphemes<'a> { type Item = RopeSlice<'a>; @@ -462,7 +311,7 @@ impl<'a> Iterator for RevRopeGraphemes<'a> { loop { match self .cursor - .prev_boundary(self.cur_chunk, self.cur_chunk_start) + .prev_boundary(self.chunk_cursor.chunk(), self.chunk_cursor.byte_offset()) { Ok(None) => { return None; @@ -471,24 +320,21 @@ impl<'a> Iterator for RevRopeGraphemes<'a> { b = n; break; } - Err(GraphemeIncomplete::PrevChunk) => { - self.cur_chunk = self.chunks.next().unwrap_or(""); - self.cur_chunk_start -= self.cur_chunk.len(); - } + Err(GraphemeIncomplete::PrevChunk) => assert!(self.chunk_cursor.prev()), Err(GraphemeIncomplete::PreContext(idx)) => { - let (chunk, byte_idx, _, _) = self.text.chunk_at_byte(idx.saturating_sub(1)); + let (chunk, byte_idx) = self.text.chunk(idx.saturating_sub(1)); self.cursor.provide_context(chunk, byte_idx); } _ => unreachable!(), } } - if a >= self.cur_chunk_start + self.cur_chunk.len() { - Some(self.text.byte_slice(b..a)) + if a >= self.chunk_cursor.byte_offset() + self.chunk_cursor.chunk().len() { + Some(self.text.slice(b..a)) } else { - let a2 = a - self.cur_chunk_start; - let b2 = b - self.cur_chunk_start; - Some((&self.cur_chunk[b2..a2]).into()) + let a2 = a - self.chunk_cursor.byte_offset(); + let b2 = b - self.chunk_cursor.byte_offset(); + Some((&self.chunk_cursor.chunk()[b2..a2]).into()) } } } @@ -509,46 +355,11 @@ mod tests { assert!(RopeSlice::from("asdf").ends_with("f")); } - #[test] - fn char_boundaries() { - let ascii = RopeSlice::from("ascii"); - // When the given index lies on a character boundary, the index should not change. - for byte_idx in 0..=ascii.len_bytes() { - assert_eq!(ascii.floor_char_boundary(byte_idx), byte_idx); - assert_eq!(ascii.ceil_char_boundary(byte_idx), byte_idx); - assert!(ascii.is_char_boundary(byte_idx)); - } - - // This is a polyfill of a method of this trait which was replaced by ceil_char_boundary. - // It returns the _character index_ of the given byte index, rounding up if it does not - // already lie on a character boundary. - fn byte_to_next_char(slice: RopeSlice, byte_idx: usize) -> usize { - slice.byte_to_char(slice.ceil_char_boundary(byte_idx)) - } - - for i in 0..=6 { - assert_eq!(byte_to_next_char(RopeSlice::from("foobar"), i), i); - } - for char_idx in 0..10 { - let len = "πŸ˜†".len(); - assert_eq!( - byte_to_next_char(RopeSlice::from("πŸ˜†πŸ˜†πŸ˜†πŸ˜†πŸ˜†πŸ˜†πŸ˜†πŸ˜†πŸ˜†πŸ˜†"), char_idx * len), - char_idx - ); - for i in 1..=len { - assert_eq!( - byte_to_next_char(RopeSlice::from("πŸ˜†πŸ˜†πŸ˜†πŸ˜†πŸ˜†πŸ˜†πŸ˜†πŸ˜†πŸ˜†πŸ˜†"), char_idx * len + i), - char_idx + 1 - ); - } - } - } - #[test] fn grapheme_boundaries() { let ascii = RopeSlice::from("ascii"); // When the given index lies on a grapheme boundary, the index should not change. - for byte_idx in 0..=ascii.len_bytes() { + for byte_idx in 0..=ascii.len() { assert_eq!(ascii.floor_char_boundary(byte_idx), byte_idx); assert_eq!(ascii.ceil_char_boundary(byte_idx), byte_idx); assert!(ascii.is_grapheme_boundary(byte_idx)); @@ -558,7 +369,7 @@ mod tests { // 13 bytes, hex: f0 9f 8f b4 + e2 80 8d + e2 98 a0 + ef b8 8f let g = RopeSlice::from("πŸ΄β€β˜ οΈ\r\n"); let emoji_len = "πŸ΄β€β˜ οΈ".len(); - let end = g.len_bytes(); + let end = g.len(); for byte_idx in 0..emoji_len { assert_eq!(g.floor_grapheme_boundary(byte_idx), 0);