Merge 4b4740a322 into b8313da5a8

helix-core/src/chars.rs

@@ -11,6 +11,7 @@ pub enum CharCategory {
     Unknown,
 }
 
+/// Determine whether a character is a line ending, whitespace, a "word" character, a punctuation or unknown
 #[inline]
 pub fn categorize_char(ch: char) -> CharCategory {
     if char_is_line_ending(ch) {
@@ -32,6 +33,12 @@ pub fn char_is_line_ending(ch: char) -> bool {
     LineEnding::from_char(ch).is_some()
 }
 
+/// Determine whether a character is a subword text object delimiter.
+#[inline]
+pub fn char_is_subword_textobj_delimiter(ch: char) -> bool {
+    ch == '_' || ch == '-' || ch == '/'
+}
+
 /// Determine whether a character qualifies as (non-line-break)
 /// whitespace.
 #[inline]

helix-core/src/movement.rs

@@ -502,7 +502,7 @@ fn is_long_word_boundary(a: char, b: char) -> bool {
     }
 }
 
-fn is_sub_word_boundary(a: char, b: char, dir: Direction) -> bool {
+pub fn is_sub_word_boundary(a: char, b: char, dir: Direction) -> bool {
     match (categorize_char(a), categorize_char(b)) {
         (CharCategory::Word, CharCategory::Word) => {
             if (a == '_') != (b == '_') {

helix-core/src/textobject.rs

@@ -3,19 +3,33 @@
 use ropey::RopeSlice;
 use tree_sitter::{Node, QueryCursor};
 
-use crate::chars::{categorize_char, char_is_whitespace, CharCategory};
+use crate::chars::{
+    categorize_char, char_is_subword_textobj_delimiter, char_is_whitespace, CharCategory,
+};
 use crate::graphemes::{next_grapheme_boundary, prev_grapheme_boundary};
 use crate::line_ending::rope_is_line_ending;
-use crate::movement::Direction;
+use crate::movement::{is_sub_word_boundary, Direction};
 use crate::syntax::LanguageConfiguration;
 use crate::Range;
 use crate::{surround, Syntax};
 
-fn find_word_boundary(slice: RopeSlice, mut pos: usize, direction: Direction, long: bool) -> usize {
+/// # Arguments
+///
+/// * `pos` - index of the character
+/// * `long` - whether it's a word or a WORD
+fn find_word_boundary(
+    slice: RopeSlice,
+    mut pos: usize,
+    direction: Direction,
+    long: bool,
+    is_subword: bool,
+) -> usize {
     use CharCategory::{Eol, Whitespace};
 
     let iter = match direction {
+        // create forward iterator
         Direction::Forward => slice.chars_at(pos),
+        // create reverse iterator, if we iterate over it we will be advancing in the opposite direction
         Direction::Backward => {
             let mut iter = slice.chars_at(pos);
             iter.reverse();
@@ -24,17 +38,44 @@ fn find_word_boundary(slice: RopeSlice, mut pos: usize, direction: Direction, lo
     };
 
     let mut prev_category = match direction {
+        // if we are at the beginning or end of the document
         Direction::Forward if pos == 0 => Whitespace,
-        Direction::Forward => categorize_char(slice.char(pos - 1)),
         Direction::Backward if pos == slice.len_chars() => Whitespace,
+        Direction::Forward => categorize_char(slice.char(pos - 1)),
         Direction::Backward => categorize_char(slice.char(pos)),
     };
 
+    let mut prev_ch = match direction {
+        // if we are at the beginning or end of the document
+        Direction::Forward if pos == 0 => ' ',
+        Direction::Backward if pos == slice.len_chars() => ' ',
+        Direction::Forward => slice.char(pos - 1),
+        Direction::Backward => slice.char(pos),
+    };
+
     for ch in iter {
         match categorize_char(ch) {
+            // When we find the first whitespace, that's going to be our position that we jump to
             Eol | Whitespace => return pos,
+            // every character other than a whitespace
             category => {
-                if !long && category != prev_category && pos != 0 && pos != slice.len_chars() {
+                let matches_short_word = !long
+                    && !is_subword
+                    && category != prev_category
+                    && pos != 0
+                    && pos != slice.len_chars();
+
+                let matches_subword = is_subword
+                    && match direction {
+                        Direction::Forward => is_sub_word_boundary(prev_ch, ch, Direction::Forward),
+                        Direction::Backward => {
+                            is_sub_word_boundary(prev_ch, ch, Direction::Backward)
+                        }
+                    };
+
+                if matches_subword {
+                    return pos;
+                } else if matches_short_word {
                     return pos;
                 } else {
                     match direction {
@@ -42,6 +83,7 @@ fn find_word_boundary(slice: RopeSlice, mut pos: usize, direction: Direction, lo
                         Direction::Backward => pos = pos.saturating_sub(1),
                     }
                     prev_category = category;
+                    prev_ch = ch;
                 }
             }
         }
@@ -75,13 +117,14 @@ pub fn textobject_word(
     textobject: TextObject,
     _count: usize,
     long: bool,
+    is_subword: bool,
 ) -> Range {
     let pos = range.cursor(slice);
 
-    let word_start = find_word_boundary(slice, pos, Direction::Backward, long);
+    let word_start = find_word_boundary(slice, pos, Direction::Backward, long, is_subword);
     let word_end = match slice.get_char(pos).map(categorize_char) {
         None | Some(CharCategory::Whitespace | CharCategory::Eol) => pos,
-        _ => find_word_boundary(slice, pos + 1, Direction::Forward, long),
+        _ => find_word_boundary(slice, pos + 1, Direction::Forward, long, is_subword),
     };
 
     // Special case.
@@ -89,9 +132,28 @@ pub fn textobject_word(
         return Range::new(word_start, word_end);
     }
 
-    match textobject {
-        TextObject::Inside => Range::new(word_start, word_end),
-        TextObject::Around => {
+    match (textobject, is_subword) {
+        (TextObject::Inside, true) => Range::new(word_start, word_end),
+        (TextObject::Around, true) => {
+            let underscores_count_right = slice
+                .chars_at(word_end)
+                .take_while(|c| char_is_subword_textobj_delimiter(*c))
+                .count();
+
+            if underscores_count_right > 0 {
+                Range::new(word_start, word_end + underscores_count_right)
+            } else {
+                let underscore_count_left = {
+                    let mut iter = slice.chars_at(word_start);
+                    iter.reverse();
+                    iter.take_while(|c| char_is_subword_textobj_delimiter(*c))
+                        .count()
+                };
+                Range::new(word_start - underscore_count_left, word_end)
+            }
+        }
+        (TextObject::Inside, false) => Range::new(word_start, word_end),
+        (TextObject::Around, false) => {
             let whitespace_count_right = slice
                 .chars_at(word_end)
                 .take_while(|c| char_is_whitespace(*c))
@@ -108,7 +170,8 @@ pub fn textobject_word(
                 Range::new(word_start - whitespace_count_left, word_end)
             }
         }
-        TextObject::Movement => unreachable!(),
+        (TextObject::Movement, false) => unreachable!(),
+        (TextObject::Movement, true) => unreachable!(),
     }
 }
 
@@ -401,7 +464,7 @@ fn test_textobject_word() {
                 let (pos, objtype, expected_range) = case;
                 // cursor is a single width selection
                 let range = Range::new(pos, pos + 1);
-                let result = textobject_word(slice, range, objtype, 1, false);
+                let result = textobject_word(slice, range, objtype, 1, false, false);
                 assert_eq!(
                     result,
                     expected_range.into(),

helix-term/src/commands.rs

@@ -5590,8 +5590,15 @@ fn select_textobject(cx: &mut Context, objtype: textobject::TextObject) {
 
                 let selection = doc.selection(view.id).clone().transform(|range| {
                     match ch {
-                        'w' => textobject::textobject_word(text, range, objtype, count, false),
-                        'W' => textobject::textobject_word(text, range, objtype, count, true),
+                        'w' => {
+                            textobject::textobject_word(text, range, objtype, count, false, false)
+                        }
+                        'W' => {
+                            textobject::textobject_word(text, range, objtype, count, true, false)
+                        }
+                        's' => {
+                            textobject::textobject_word(text, range, objtype, count, false, true)
+                        }
                         't' => textobject_treesitter("class", range),
                         'f' => textobject_treesitter("function", range),
                         'a' => textobject_treesitter("parameter", range),

helix-term/src/commands/lsp.rs

@@ -1068,7 +1068,7 @@ fn get_prefill_from_word_boundary(editor: &Editor) -> String {
             primary_selection
         } else {
             use helix_core::textobject::{textobject_word, TextObject};
-            textobject_word(text, primary_selection, TextObject::Inside, 1, false)
+            textobject_word(text, primary_selection, TextObject::Inside, 1, false, false)
         }
         .fragment(text)
         .into()

helix-term/src/ui/prompt.rs

@@ -586,6 +586,7 @@ fn handle_event(&mut self, event: &Event, cx: &mut Context) -> EventResult {
                     textobject::TextObject::Inside,
                     1,
                     false,
+                    false,
                 );
                 let line = text.slice(range.from()..range.to()).to_string();
                 if !line.is_empty() {