Add object selection (textobjects) (#385)

* Add textobjects for word

* Add textobjects for surround characters

* Apply clippy lints

* Remove ThisWordPrevBound in favor of PrevWordEnd

It's the same as PrevWordEnd except for taking the current char
into account, so use a "flag" to capture that usecase

* Add tests for PrevWordEnd movement

* Remove ThisWord* movements

They did not preserve anchor positions and were only used
for textobject boundary search anyway so replace them with
simple position finding functions

* Rewrite tests of word textobject

* Add tests for surround textobject

* Add textobject docs

* Refactor textobject word position functions

* Apply clippy lints on textobject

* Fix overflow error with textobjects

Author: sudormrfbin

book/src/keymap.md

@@ -150,14 +150,17 @@ Jumps to various locations.
 ## Match mode
 
 Enter this mode using `m` from normal mode. See the relavant section
-in [Usage](./usage.md#surround) for an explanation about surround usage.
+in [Usage](./usage.md) for an explanation about [surround](./usage.md#surround)
+and [textobject](./usage.md#textobject) usage.
 
 | Key              | Description                                     |
 | -----            | -----------                                     |
 | `m`              | Goto matching bracket                           |
 | `s` `<char>`     | Surround current selection with `<char>`        |
 | `r` `<from><to>` | Replace surround character `<from>` with `<to>` |
 | `d` `<char>`     | Delete surround character `<char>`              |
+| `a` `<object>`   | Select around textobject                        |
+| `i` `<object>`   | Select inside textobject                        |
 
 ## Object mode
 

book/src/usage.md

@@ -24,3 +24,19 @@ It can also act on multiple seletions (yay!). For example, to change every occur
 - `mr([` to replace the parens with square brackets
 
 Multiple characters are currently not supported, but planned.
+
+## Textobjects
+
+Currently supported: `word`, `surround`.
+
+![textobject-demo](https://user-images.githubusercontent.com/23398472/124231131-81a4bb00-db2d-11eb-9d10-8e577ca7b177.gif)
+
+- `ma` - Select around the object (`va` in vim, `<alt-a>` in kakoune)
+- `mi` - Select inside the object (`vi` in vim, `<alt-i>` in kakoune)
+
+| Key after `mi` or `ma` | Textobject selected      |
+| ---                    | ---                      |
+| `w`                    | Word                     |
+| `(`, `[`, `'`, etc     | Specified surround pairs |
+
+Textobjects based on treesitter, like `function`, `class`, etc are planned.

helix-core/src/lib.rs

@@ -18,6 +18,7 @@ pub mod selection;
 mod state;
 pub mod surround;
 pub mod syntax;
+pub mod textobject;
 mod transaction;
 
 pub mod unicode {

helix-core/src/movement.rs

@@ -113,6 +113,10 @@ pub fn move_prev_long_word_start(slice: RopeSlice, range: Range, count: usize) -
     word_move(slice, range, count, WordMotionTarget::PrevLongWordStart)
 }
 
+pub fn move_prev_word_end(slice: RopeSlice, range: Range, count: usize) -> Range {
+    word_move(slice, range, count, WordMotionTarget::PrevWordEnd)
+}
+
 fn word_move(slice: RopeSlice, range: Range, count: usize, target: WordMotionTarget) -> Range {
     (0..count).fold(range, |range, _| {
         slice.chars_at(range.head).range_to_target(target, range)
@@ -159,6 +163,7 @@ pub enum WordMotionTarget {
     NextWordStart,
     NextWordEnd,
     PrevWordStart,
+    PrevWordEnd,
     // A "Long word" (also known as a WORD in vim/kakoune) is strictly
     // delimited by whitespace, and can consist of punctuation as well
     // as alphanumerics.
@@ -181,7 +186,9 @@ impl CharHelpers for Chars<'_> {
     fn range_to_target(&mut self, target: WordMotionTarget, origin: Range) -> Range {
         // Characters are iterated forward or backwards depending on the motion direction.
         let characters: Box<dyn Iterator<Item = char>> = match target {
-            WordMotionTarget::PrevWordStart | WordMotionTarget::PrevLongWordStart => {
+            WordMotionTarget::PrevWordStart
+            | WordMotionTarget::PrevLongWordStart
+            | WordMotionTarget::PrevWordEnd => {
                 self.next();
                 Box::new(from_fn(|| self.prev()))
             }
@@ -190,9 +197,9 @@ impl CharHelpers for Chars<'_> {
 
         // Index advancement also depends on the direction.
         let advance: &dyn Fn(&mut usize) = match target {
-            WordMotionTarget::PrevWordStart | WordMotionTarget::PrevLongWordStart => {
-                &|u| *u = u.saturating_sub(1)
-            }
+            WordMotionTarget::PrevWordStart
+            | WordMotionTarget::PrevLongWordStart
+            | WordMotionTarget::PrevWordEnd => &|u| *u = u.saturating_sub(1),
             _ => &|u| *u += 1,
         };
 
@@ -265,7 +272,7 @@ fn reached_target(target: WordMotionTarget, peek: char, next_peek: Option<&char>
     };
 
     match target {
-        WordMotionTarget::NextWordStart => {
+        WordMotionTarget::NextWordStart | WordMotionTarget::PrevWordEnd => {
             is_word_boundary(peek, *next_peek)
                 && (char_is_line_ending(*next_peek) || !next_peek.is_whitespace())
         }
@@ -913,6 +920,88 @@ mod test {
         }
     }
 
+    #[test]
+    fn test_behaviour_when_moving_to_end_of_previous_words() {
+        let tests = array::IntoIter::new([
+            ("Basic backward motion from the middle of a word",
+                vec![(1, Range::new(9, 9), Range::new(9, 5))]),
+            ("Starting from after boundary retreats the anchor",
+                vec![(1, Range::new(0, 13), Range::new(12, 8))]),
+            ("Jump     to end of a word succeeded by whitespace",
+                vec![(1, Range::new(10, 10), Range::new(10, 4))]),
+            ("    Jump to start of line from end of word preceded by whitespace",
+                vec![(1, Range::new(7, 7), Range::new(7, 0))]),
+            ("Previous anchor is irrelevant for backward motions",
+                vec![(1, Range::new(26, 12), Range::new(12, 8))]),
+            ("    Starting from whitespace moves to first space in sequence",
+                vec![(1, Range::new(0, 3), Range::new(3, 0))]),
+            ("Test identifiers_with_underscores are considered a single word",
+                vec![(1, Range::new(0, 25), Range::new(25, 4))]),
+            ("Jumping\n    \nback through a newline selects whitespace",
+                vec![(1, Range::new(0, 13), Range::new(11, 8))]),
+            ("Jumping to start of word from the end selects the whole word",
+                vec![(1, Range::new(15, 15), Range::new(15, 10))]),
+            ("alphanumeric.!,and.?=punctuation are considered 'words' for the purposes of word motion",
+                vec![
+                    (1, Range::new(30, 30), Range::new(30, 21)),
+                    (1, Range::new(30, 21), Range::new(20, 18)),
+                    (1, Range::new(20, 18), Range::new(17, 15))
+                ]),
+
+            ("...   ... punctuation and spaces behave as expected",
+                vec![
+                    (1, Range::new(0, 10), Range::new(9, 9)),
+                    (1, Range::new(9, 6), Range::new(5, 3)),
+                ]),
+            (".._.._ punctuation is not joined by underscores into a single block",
+                vec![(1, Range::new(0, 5), Range::new(4, 3))]),
+            ("Newlines\n\nare bridged seamlessly.",
+                vec![
+                    (1, Range::new(0, 10), Range::new(7, 0)),
+                ]),
+            ("Jumping    \n\n\n\n\nback from within a newline group selects previous block",
+                vec![
+                    (1, Range::new(0, 13), Range::new(10, 7)),
+                ]),
+            ("Failed motions do not modify the range",
+                vec![
+                    (0, Range::new(3, 0), Range::new(3, 0)),
+                ]),
+            ("Multiple motions at once resolve correctly",
+                vec![
+                    (3, Range::new(23, 23), Range::new(15, 8)),
+                ]),
+            ("Excessive motions are performed partially",
+                vec![
+                    (999, Range::new(40, 40), Range::new(8, 0)),
+                ]),
+            ("", // Edge case of moving backwards in empty string
+                vec![
+                    (1, Range::new(0, 0), Range::new(0, 0)),
+                ]),
+            ("\n\n\n\n\n", // Edge case of moving backwards in all newlines
+                vec![
+                    (1, Range::new(0, 0), Range::new(0, 0)),
+                ]),
+            ("   \n   \nJumping back through alternated space blocks and newlines selects the space blocks",
+                vec![
+                    (1, Range::new(0, 7), Range::new(6, 4)),
+                    (1, Range::new(6, 4), Range::new(2, 0)),
+                ]),
+            ("Test ヒーリクス multibyte characters behave as normal characters",
+                vec![
+                    (1, Range::new(0, 9), Range::new(9, 4)),
+                ]),
+        ]);
+
+        for (sample, scenario) in tests {
+            for (count, begin, expected_end) in scenario.into_iter() {
+                let range = move_prev_word_end(Rope::from(sample).slice(..), begin, count);
+                assert_eq!(range, expected_end, "Case failed: [{}]", sample);
+            }
+        }
+    }
+
     #[test]
     fn test_behaviour_when_moving_to_end_of_next_long_words() {
         let tests = array::IntoIter::new([

helix-core/src/selection.rs

@@ -130,6 +130,16 @@ impl Range {
     }
 }
 
+impl From<(usize, usize)> for Range {
+    fn from(tuple: (usize, usize)) -> Self {
+        Self {
+            anchor: tuple.0,
+            head: tuple.1,
+            horiz: None,
+        }
+    }
+}
+
 /// A selection consists of one or more selection ranges.
 /// invariant: A selection can never be empty (always contains at least primary range).
 #[derive(Debug, Clone, PartialEq, Eq)]