Merge remote-tracking branch 'origin/master' into debug

10 files changed, 572 insertions, 15 deletions

diff --git a/helix-core/src/history.rs b/helix-core/src/history.rs
index b53c01fe..4b1c8d3b 100644
--- a/helix-core/src/history.rs
+++ b/helix-core/src/history.rs
@@ -1,4 +1,4 @@
-use crate::{ChangeSet, Rope, State, Transaction};
+use crate::{Assoc, ChangeSet, Range, Rope, State, Transaction};
 use once_cell::sync::Lazy;
 use regex::Regex;
 use std::num::NonZeroUsize;
@@ -133,6 +133,32 @@ impl History {
         Some(&self.revisions[last_child.get()].transaction)
     }
 
+    // Get the position of last change
+    pub fn last_edit_pos(&self) -> Option<usize> {
+        if self.current == 0 {
+            return None;
+        }
+        let current_revision = &self.revisions[self.current];
+        let primary_selection = current_revision
+            .inversion
+            .selection()
+            .expect("inversion always contains a selection")
+            .primary();
+        let (_from, to, _fragment) = current_revision
+            .transaction
+            .changes_iter()
+            // find a change that matches the primary selection
+            .find(|(from, to, _fragment)| Range::new(*from, *to).overlaps(&primary_selection))
+            // or use the first change
+            .or_else(|| current_revision.transaction.changes_iter().next())
+            .unwrap();
+        let pos = current_revision
+            .transaction
+            .changes()
+            .map_pos(to, Assoc::After);
+        Some(pos)
+    }
+
     fn lowest_common_ancestor(&self, mut a: usize, mut b: usize) -> usize {
         use std::collections::HashSet;
         let mut a_path_set = HashSet::new();
@@ -256,7 +282,7 @@ impl History {
 }
 
 /// Whether to undo by a number of edits or a duration of time.
-#[derive(Debug, PartialEq)]
+#[derive(Debug, PartialEq, Clone, Copy)]
 pub enum UndoKind {
     Steps(usize),
     TimePeriod(std::time::Duration),
diff --git a/helix-core/src/indent.rs b/helix-core/src/indent.rs
index df158363..88ab09b5 100644
--- a/helix-core/src/indent.rs
+++ b/helix-core/src/indent.rs
@@ -450,6 +450,7 @@ where
             language: vec![LanguageConfiguration {
                 scope: "source.rust".to_string(),
                 file_types: vec!["rs".to_string()],
+                shebangs: vec![],
                 language_id: "Rust".to_string(),
                 highlight_config: OnceCell::new(),
                 config: None,
diff --git a/helix-core/src/lib.rs b/helix-core/src/lib.rs
index f4284139..7d790406 100644
--- a/helix-core/src/lib.rs
+++ b/helix-core/src/lib.rs
@@ -10,6 +10,7 @@ pub mod line_ending;
 pub mod macros;
 pub mod match_brackets;
 pub mod movement;
+pub mod numbers;
 pub mod object;
 pub mod path;
 mod position;
diff --git a/helix-core/src/match_brackets.rs b/helix-core/src/match_brackets.rs
index a4b2fb9c..136ce320 100644
--- a/helix-core/src/match_brackets.rs
+++ b/helix-core/src/match_brackets.rs
@@ -1,6 +1,13 @@
 use crate::{Rope, Syntax};
 
-const PAIRS: &[(char, char)] = &[('(', ')'), ('{', '}'), ('[', ']'), ('<', '>')];
+const PAIRS: &[(char, char)] = &[
+    ('(', ')'),
+    ('{', '}'),
+    ('[', ']'),
+    ('<', '>'),
+    ('\'', '\''),
+    ('"', '"'),
+];
 // limit matching pairs to only ( ) { } [ ] < >
 
 #[must_use]
diff --git a/helix-core/src/movement.rs b/helix-core/src/movement.rs
index 9e85bd21..01a8f890 100644
--- a/helix-core/src/movement.rs
+++ b/helix-core/src/movement.rs
@@ -168,7 +168,7 @@ pub fn backwards_skip_while<F>(slice: RopeSlice, pos: usize, fun: F) -> Option<u
 where
     F: Fn(char) -> bool,
 {
-    let mut chars_starting_from_next = slice.chars_at(pos + 1);
+    let mut chars_starting_from_next = slice.chars_at(pos);
     let mut backwards = iter::from_fn(|| chars_starting_from_next.prev()).enumerate();
     backwards.find_map(|(i, c)| {
         if !fun(c) {
diff --git a/helix-core/src/numbers.rs b/helix-core/src/numbers.rs
new file mode 100644
index 00000000..e9f3c898
--- /dev/null
+++ b/helix-core/src/numbers.rs
@@ -0,0 +1,499 @@
+use std::borrow::Cow;
+
+use ropey::RopeSlice;
+
+use crate::{
+    textobject::{textobject_word, TextObject},
+    Range, Tendril,
+};
+
+#[derive(Debug, PartialEq, Eq)]
+pub struct NumberIncrementor<'a> {
+    pub range: Range,
+    pub value: i64,
+    pub radix: u32,
+
+    text: RopeSlice<'a>,
+}
+
+impl<'a> NumberIncrementor<'a> {
+    /// Return information about number under rang if there is one.
+    pub fn from_range(text: RopeSlice, range: Range) -> Option<NumberIncrementor> {
+        // If the cursor is on the minus sign of a number we want to get the word textobject to the
+        // right of it.
+        let range = if range.to() < text.len_chars()
+            && range.to() - range.from() <= 1
+            && text.char(range.from()) == '-'
+        {
+            Range::new(range.from() + 1, range.to() + 1)
+        } else {
+            range
+        };
+
+        let range = textobject_word(text, range, TextObject::Inside, 1, false);
+
+        // If there is a minus sign to the left of the word object, we want to include it in the range.
+        let range = if range.from() > 0 && text.char(range.from() - 1) == '-' {
+            range.extend(range.from() - 1, range.from())
+        } else {
+            range
+        };
+
+        let word: String = text
+            .slice(range.from()..range.to())
+            .chars()
+            .filter(|&c| c != '_')
+            .collect();
+        let (radix, prefixed) = if word.starts_with("0x") {
+            (16, true)
+        } else if word.starts_with("0o") {
+            (8, true)
+        } else if word.starts_with("0b") {
+            (2, true)
+        } else {
+            (10, false)
+        };
+
+        let number = if prefixed { &word[2..] } else { &word };
+
+        let value = i128::from_str_radix(number, radix).ok()?;
+        if (value.is_positive() && value.leading_zeros() < 64)
+            || (value.is_negative() && value.leading_ones() < 64)
+        {
+            return None;
+        }
+
+        let value = value as i64;
+        Some(NumberIncrementor {
+            range,
+            value,
+            radix,
+            text,
+        })
+    }
+
+    /// Add `amount` to the number and return the formatted text.
+    pub fn incremented_text(&self, amount: i64) -> Tendril {
+        let old_text: Cow<str> = self.text.slice(self.range.from()..self.range.to()).into();
+        let old_length = old_text.len();
+        let new_value = self.value.wrapping_add(amount);
+
+        // Get separator indexes from right to left.
+        let separator_rtl_indexes: Vec<usize> = old_text
+            .chars()
+            .rev()
+            .enumerate()
+            .filter_map(|(i, c)| if c == '_' { Some(i) } else { None })
+            .collect();
+
+        let format_length = if self.radix == 10 {
+            match (self.value.is_negative(), new_value.is_negative()) {
+                (true, false) => old_length - 1,
+                (false, true) => old_length + 1,
+                _ => old_text.len(),
+            }
+        } else {
+            old_text.len() - 2
+        } - separator_rtl_indexes.len();
+
+        let mut new_text = match self.radix {
+            2 => format!("0b{:01$b}", new_value, format_length),
+            8 => format!("0o{:01$o}", new_value, format_length),
+            10 if old_text.starts_with('0') || old_text.starts_with("-0") => {
+                format!("{:01$}", new_value, format_length)
+            }
+            10 => format!("{}", new_value),
+            16 => {
+                let (lower_count, upper_count): (usize, usize) =
+                    old_text.chars().skip(2).fold((0, 0), |(lower, upper), c| {
+                        (
+                            lower + c.is_ascii_lowercase().then(|| 1).unwrap_or(0),
+                            upper + c.is_ascii_uppercase().then(|| 1).unwrap_or(0),
+                        )
+                    });
+                if upper_count > lower_count {
+                    format!("0x{:01$X}", new_value, format_length)
+                } else {
+                    format!("0x{:01$x}", new_value, format_length)
+                }
+            }
+            _ => unimplemented!("radix not supported: {}", self.radix),
+        };
+
+        // Add separators from original number.
+        for &rtl_index in &separator_rtl_indexes {
+            if rtl_index < new_text.len() {
+                let new_index = new_text.len() - rtl_index;
+                new_text.insert(new_index, '_');
+            }
+        }
+
+        // Add in additional separators if necessary.
+        if new_text.len() > old_length && !separator_rtl_indexes.is_empty() {
+            let spacing = match separator_rtl_indexes.as_slice() {
+                [.., b, a] => a - b - 1,
+                _ => separator_rtl_indexes[0],
+            };
+
+            let prefix_length = if self.radix == 10 { 0 } else { 2 };
+            if let Some(mut index) = new_text.find('_') {
+                while index - prefix_length > spacing {
+                    index -= spacing;
+                    new_text.insert(index, '_');
+                }
+            }
+        }
+
+        new_text.into()
+    }
+}
+
+#[cfg(test)]
+mod test {
+    use super::*;
+    use crate::Rope;
+
+    #[test]
+    fn test_decimal_at_point() {
+        let rope = Rope::from_str("Test text 12345 more text.");
+        let range = Range::point(12);
+        assert_eq!(
+            NumberIncrementor::from_range(rope.slice(..), range),
+            Some(NumberIncrementor {
+                range: Range::new(10, 15),
+                value: 12345,
+                radix: 10,
+                text: rope.slice(..),
+            })
+        );
+    }
+
+    #[test]
+    fn test_uppercase_hexadecimal_at_point() {
+        let rope = Rope::from_str("Test text 0x123ABCDEF more text.");
+        let range = Range::point(12);
+        assert_eq!(
+            NumberIncrementor::from_range(rope.slice(..), range),
+            Some(NumberIncrementor {
+                range: Range::new(10, 21),
+                value: 0x123ABCDEF,
+                radix: 16,
+                text: rope.slice(..),
+            })
+        );
+    }
+
+    #[test]
+    fn test_lowercase_hexadecimal_at_point() {
+        let rope = Rope::from_str("Test text 0xfa3b4e more text.");
+        let range = Range::point(12);
+        assert_eq!(
+            NumberIncrementor::from_range(rope.slice(..), range),
+            Some(NumberIncrementor {
+                range: Range::new(10, 18),
+                value: 0xfa3b4e,
+                radix: 16,
+                text: rope.slice(..),
+            })
+        );
+    }
+
+    #[test]
+    fn test_octal_at_point() {
+        let rope = Rope::from_str("Test text 0o1074312 more text.");
+        let range = Range::point(12);
+        assert_eq!(
+            NumberIncrementor::from_range(rope.slice(..), range),
+            Some(NumberIncrementor {
+                range: Range::new(10, 19),
+                value: 0o1074312,
+                radix: 8,
+                text: rope.slice(..),
+            })
+        );
+    }
+
+    #[test]
+    fn test_binary_at_point() {
+        let rope = Rope::from_str("Test text 0b10111010010101 more text.");
+        let range = Range::point(12);
+        assert_eq!(
+            NumberIncrementor::from_range(rope.slice(..), range),
+            Some(NumberIncrementor {
+                range: Range::new(10, 26),
+                value: 0b10111010010101,
+                radix: 2,
+                text: rope.slice(..),
+            })
+        );
+    }
+
+    #[test]
+    fn test_negative_decimal_at_point() {
+        let rope = Rope::from_str("Test text -54321 more text.");
+        let range = Range::point(12);
+        assert_eq!(
+            NumberIncrementor::from_range(rope.slice(..), range),
+            Some(NumberIncrementor {
+                range: Range::new(10, 16),
+                value: -54321,
+                radix: 10,
+                text: rope.slice(..),
+            })
+        );
+    }
+
+    #[test]
+    fn test_decimal_with_leading_zeroes_at_point() {
+        let rope = Rope::from_str("Test text 000045326 more text.");
+        let range = Range::point(12);
+        assert_eq!(
+            NumberIncrementor::from_range(rope.slice(..), range),
+            Some(NumberIncrementor {
+                range: Range::new(10, 19),
+                value: 45326,
+                radix: 10,
+                text: rope.slice(..),
+            })
+        );
+    }
+
+    #[test]
+    fn test_negative_decimal_cursor_on_minus_sign() {
+        let rope = Rope::from_str("Test text -54321 more text.");
+        let range = Range::point(10);
+        assert_eq!(
+            NumberIncrementor::from_range(rope.slice(..), range),
+            Some(NumberIncrementor {
+                range: Range::new(10, 16),
+                value: -54321,
+                radix: 10,
+                text: rope.slice(..),
+            })
+        );
+    }
+
+    #[test]
+    fn test_number_under_range_start_of_rope() {
+        let rope = Rope::from_str("100");
+        let range = Range::point(0);
+        assert_eq!(
+            NumberIncrementor::from_range(rope.slice(..), range),
+            Some(NumberIncrementor {
+                range: Range::new(0, 3),
+                value: 100,
+                radix: 10,
+                text: rope.slice(..),
+            })
+        );
+    }
+
+    #[test]
+    fn test_number_under_range_end_of_rope() {
+        let rope = Rope::from_str("100");
+        let range = Range::point(2);
+        assert_eq!(
+            NumberIncrementor::from_range(rope.slice(..), range),
+            Some(NumberIncrementor {
+                range: Range::new(0, 3),
+                value: 100,
+                radix: 10,
+                text: rope.slice(..),
+            })
+        );
+    }
+
+    #[test]
+    fn test_number_surrounded_by_punctuation() {
+        let rope = Rope::from_str(",100;");
+        let range = Range::point(1);
+        assert_eq!(
+            NumberIncrementor::from_range(rope.slice(..), range),
+            Some(NumberIncrementor {
+                range: Range::new(1, 4),
+                value: 100,
+                radix: 10,
+                text: rope.slice(..),
+            })
+        );
+    }
+
+    #[test]
+    fn test_not_a_number_point() {
+        let rope = Rope::from_str("Test text 45326 more text.");
+        let range = Range::point(6);
+        assert_eq!(NumberIncrementor::from_range(rope.slice(..), range), None);
+    }
+
+    #[test]
+    fn test_number_too_large_at_point() {
+        let rope = Rope::from_str("Test text 0xFFFFFFFFFFFFFFFFF more text.");
+        let range = Range::point(12);
+        assert_eq!(NumberIncrementor::from_range(rope.slice(..), range), None);
+    }
+
+    #[test]
+    fn test_number_cursor_one_right_of_number() {
+        let rope = Rope::from_str("100 ");
+        let range = Range::point(3);
+        assert_eq!(NumberIncrementor::from_range(rope.slice(..), range), None);
+    }
+
+    #[test]
+    fn test_number_cursor_one_left_of_number() {
+        let rope = Rope::from_str(" 100");
+        let range = Range::point(0);
+        assert_eq!(NumberIncrementor::from_range(rope.slice(..), range), None);
+    }
+
+    #[test]
+    fn test_increment_basic_decimal_numbers() {
+        let tests = [
+            ("100", 1, "101"),
+            ("100", -1, "99"),
+            ("99", 1, "100"),
+            ("100", 1000, "1100"),
+            ("100", -1000, "-900"),
+            ("-1", 1, "0"),
+            ("-1", 2, "1"),
+            ("1", -1, "0"),
+            ("1", -2, "-1"),
+        ];
+
+        for (original, amount, expected) in tests {
+            let rope = Rope::from_str(original);
+            let range = Range::point(0);
+            assert_eq!(
+                NumberIncrementor::from_range(rope.slice(..), range)
+                    .unwrap()
+                    .incremented_text(amount),
+                expected.into()
+            );
+        }
+    }
+
+    #[test]
+    fn test_increment_basic_hexadedimal_numbers() {
+        let tests = [
+            ("0x0100", 1, "0x0101"),
+            ("0x0100", -1, "0x00ff"),
+            ("0x0001", -1, "0x0000"),
+            ("0x0000", -1, "0xffffffffffffffff"),
+            ("0xffffffffffffffff", 1, "0x0000000000000000"),
+            ("0xffffffffffffffff", 2, "0x0000000000000001"),
+            ("0xffffffffffffffff", -1, "0xfffffffffffffffe"),
+            ("0xABCDEF1234567890", 1, "0xABCDEF1234567891"),
+            ("0xabcdef1234567890", 1, "0xabcdef1234567891"),
+        ];
+
+        for (original, amount, expected) in tests {
+            let rope = Rope::from_str(original);
+            let range = Range::point(0);
+            assert_eq!(
+                NumberIncrementor::from_range(rope.slice(..), range)
+                    .unwrap()
+                    .incremented_text(amount),
+                expected.into()
+            );
+        }
+    }
+
+    #[test]
+    fn test_increment_basic_octal_numbers() {
+        let tests = [
+            ("0o0107", 1, "0o0110"),
+            ("0o0110", -1, "0o0107"),
+            ("0o0001", -1, "0o0000"),
+            ("0o7777", 1, "0o10000"),
+            ("0o1000", -1, "0o0777"),
+            ("0o0107", 10, "0o0121"),
+            ("0o0000", -1, "0o1777777777777777777777"),
+            ("0o1777777777777777777777", 1, "0o0000000000000000000000"),
+            ("0o1777777777777777777777", 2, "0o0000000000000000000001"),
+            ("0o1777777777777777777777", -1, "0o1777777777777777777776"),
+        ];
+
+        for (original, amount, expected) in tests {
+            let rope = Rope::from_str(original);
+            let range = Range::point(0);
+            assert_eq!(
+                NumberIncrementor::from_range(rope.slice(..), range)
+                    .unwrap()
+                    .incremented_text(amount),
+                expected.into()
+            );
+        }
+    }
+
+    #[test]
+    fn test_increment_basic_binary_numbers() {
+        let tests = [
+            ("0b00000100", 1, "0b00000101"),
+            ("0b00000100", -1, "0b00000011"),
+            ("0b00000100", 2, "0b00000110"),
+            ("0b00000100", -2, "0b00000010"),
+            ("0b00000001", -1, "0b00000000"),
+            ("0b00111111", 10, "0b01001001"),
+            ("0b11111111", 1, "0b100000000"),
+            ("0b10000000", -1, "0b01111111"),
+            (
+                "0b0000",
+                -1,
+                "0b1111111111111111111111111111111111111111111111111111111111111111",
+            ),
+            (
+                "0b1111111111111111111111111111111111111111111111111111111111111111",
+                1,
+                "0b0000000000000000000000000000000000000000000000000000000000000000",
+            ),
+            (
+                "0b1111111111111111111111111111111111111111111111111111111111111111",
+                2,
+                "0b0000000000000000000000000000000000000000000000000000000000000001",
+            ),
+            (
+                "0b1111111111111111111111111111111111111111111111111111111111111111",
+                -1,
+                "0b1111111111111111111111111111111111111111111111111111111111111110",
+            ),
+        ];
+
+        for (original, amount, expected) in tests {
+            let rope = Rope::from_str(original);
+            let range = Range::point(0);
+            assert_eq!(
+                NumberIncrementor::from_range(rope.slice(..), range)
+                    .unwrap()
+                    .incremented_text(amount),
+                expected.into()
+            );
+        }
+    }
+
+    #[test]
+    fn test_increment_with_separators() {
+        let tests = [
+            ("999_999", 1, "1_000_000"),
+            ("1_000_000", -1, "999_999"),
+            ("-999_999", -1, "-1_000_000"),
+            ("0x0000_0000_0001", 0x1_ffff_0000, "0x0001_ffff_0001"),
+            ("0x0000_0000_0001", 0x1_ffff_0000, "0x0001_ffff_0001"),
+            ("0x0000_0000_0001", 0x1_ffff_0000, "0x0001_ffff_0001"),
+            ("0x0000_0000", -1, "0xffff_ffff_ffff_ffff"),
+            ("0x0000_0000_0000", -1, "0xffff_ffff_ffff_ffff"),
+            ("0b01111111_11111111", 1, "0b10000000_00000000"),
+            ("0b11111111_11111111", 1, "0b1_00000000_00000000"),
+        ];
+
+        for (original, amount, expected) in tests {
+            let rope = Rope::from_str(original);
+            let range = Range::point(0);
+            assert_eq!(
+                NumberIncrementor::from_range(rope.slice(..), range)
+                    .unwrap()
+                    .incremented_text(amount),
+                expected.into()
+            );
+        }
+    }
+}
diff --git a/helix-core/src/path.rs b/helix-core/src/path.rs
index 6c37cfa1..a6644465 100644
--- a/helix-core/src/path.rs
+++ b/helix-core/src/path.rs
@@ -40,7 +40,6 @@ pub fn expand_tilde(path: &Path) -> PathBuf {
 /// needs to improve on.
 /// Copied from cargo: <https://github.com/rust-lang/cargo/blob/070e459c2d8b79c5b2ac5218064e7603329c92ae/crates/cargo-util/src/paths.rs#L81>
 pub fn get_normalized_path(path: &Path) -> PathBuf {
-    let path = expand_tilde(path);
     let mut components = path.components().peekable();
     let mut ret = if let Some(c @ Component::Prefix(..)) = components.peek().cloned() {
         components.next();
@@ -72,10 +71,11 @@ pub fn get_normalized_path(path: &Path) -> PathBuf {
 /// This function is used instead of `std::fs::canonicalize` because we don't want to verify
 /// here if the path exists, just normalize it's components.
 pub fn get_canonicalized_path(path: &Path) -> std::io::Result<PathBuf> {
+    let path = expand_tilde(path);
     let path = if path.is_relative() {
         std::env::current_dir().map(|current_dir| current_dir.join(path))?
     } else {
-        path.to_path_buf()
+        path
     };
 
     Ok(get_normalized_path(path.as_path()))
diff --git a/helix-core/src/selection.rs b/helix-core/src/selection.rs
index f3b5d2c8..b4d1dffa 100644
--- a/helix-core/src/selection.rs
+++ b/helix-core/src/selection.rs
@@ -360,7 +360,7 @@ impl Selection {
         self.normalize()
     }
 
-    /// Adds a new range to the selection and makes it the primary range.
+    /// Removes a range from the selection.
     pub fn remove(mut self, index: usize) -> Self {
         assert!(
             self.ranges.len() > 1,
@@ -528,14 +528,15 @@ impl<'a> IntoIterator for &'a Selection {
 
 // TODO: checkSelection -> check if valid for doc length && sorted
 
-pub fn keep_matches(
+pub fn keep_or_remove_matches(
     text: RopeSlice,
     selection: &Selection,
     regex: &crate::regex::Regex,
+    remove: bool,
 ) -> Option<Selection> {
     let result: SmallVec<_> = selection
         .iter()
-        .filter(|range| regex.is_match(&range.fragment(text)))
+        .filter(|range| regex.is_match(&range.fragment(text)) ^ remove)
         .copied()
         .collect();
 
diff --git a/helix-core/src/syntax.rs b/helix-core/src/syntax.rs
index 18504c21..f136ecd0 100644
--- a/helix-core/src/syntax.rs
+++ b/helix-core/src/syntax.rs
@@ -40,18 +40,21 @@ where
 }
 
 #[derive(Debug, Serialize, Deserialize)]
+#[serde(deny_unknown_fields)]
 pub struct Configuration {
     pub language: Vec<LanguageConfiguration>,
 }
 
 // largely based on tree-sitter/cli/src/loader.rs
 #[derive(Debug, Serialize, Deserialize)]
-#[serde(rename_all = "kebab-case")]
+#[serde(rename_all = "kebab-case", deny_unknown_fields)]
 pub struct LanguageConfiguration {
     #[serde(rename = "name")]
     pub language_id: String,
     pub scope: String,           // source.rust
     pub file_types: Vec<String>, // filename ends_with? <Gemfile, rb, etc>
+    #[serde(default)]
+    pub shebangs: Vec<String>, // interpreter(s) associated with language
     pub roots: Vec<String>,      // these indicate project roots <.git, Cargo.toml>
     pub comment_token: Option<String>,
 
@@ -309,6 +312,7 @@ pub struct Loader {
     // highlight_names ?
     language_configs: Vec<Arc<LanguageConfiguration>>,
     language_config_ids_by_file_type: HashMap<String, usize>, // Vec<usize>
+    language_config_ids_by_shebang: HashMap<String, usize>,
 }
 
 impl Loader {
@@ -316,6 +320,7 @@ impl Loader {
         let mut loader = Self {
             language_configs: Vec::new(),
             language_config_ids_by_file_type: HashMap::new(),
+            language_config_ids_by_shebang: HashMap::new(),
         };
 
         for config in config.language {
@@ -328,6 +333,11 @@ impl Loader {
                     .language_config_ids_by_file_type
                     .insert(file_type.clone(), language_id);
             }
+            for shebang in &config.shebangs {
+                loader
+                    .language_config_ids_by_shebang
+                    .insert(shebang.clone(), language_id);
+            }
 
             loader.language_configs.push(Arc::new(config));
         }
@@ -353,6 +363,17 @@ impl Loader {
         // TODO: content_regex handling conflict resolution
     }
 
+    pub fn language_config_for_shebang(&self, source: &Rope) -> Option<Arc<LanguageConfiguration>> {
+        let line = Cow::from(source.line(0));
+        static SHEBANG_REGEX: Lazy<Regex> =
+            Lazy::new(|| Regex::new(r"^#!\s*(?:\S*[/\\](?:env\s+)?)?([^\s\.\d]+)").unwrap());
+        let configuration_id = SHEBANG_REGEX
+            .captures(&line)
+            .and_then(|cap| self.language_config_ids_by_shebang.get(&cap[1]));
+
+        configuration_id.and_then(|&id| self.language_configs.get(id).cloned())
+    }
+
     pub fn language_config_for_scope(&self, scope: &str) -> Option<Arc<LanguageConfiguration>> {
         self.language_configs
             .iter()
diff --git a/helix-core/src/textobject.rs b/helix-core/src/textobject.rs
index 975ed115..24f063d4 100644
--- a/helix-core/src/textobject.rs
+++ b/helix-core/src/textobject.rs
@@ -10,7 +10,7 @@ use crate::surround;
 use crate::syntax::LanguageConfiguration;
 use crate::Range;
 
-fn find_word_boundary(slice: RopeSlice, mut pos: usize, direction: Direction) -> usize {
+fn find_word_boundary(slice: RopeSlice, mut pos: usize, direction: Direction, long: bool) -> usize {
     use CharCategory::{Eol, Whitespace};
 
     let iter = match direction {
@@ -33,7 +33,7 @@ fn find_word_boundary(slice: RopeSlice, mut pos: usize, direction: Direction) ->
         match categorize_char(ch) {
             Eol | Whitespace => return pos,
             category => {
-                if category != prev_category && pos != 0 && pos != slice.len_chars() {
+                if !long && category != prev_category && pos != 0 && pos != slice.len_chars() {
                     return pos;
                 } else {
                     match direction {
@@ -70,13 +70,14 @@ pub fn textobject_word(
     range: Range,
     textobject: TextObject,
     _count: usize,
+    long: bool,
 ) -> Range {
     let pos = range.cursor(slice);
 
-    let word_start = find_word_boundary(slice, pos, Direction::Backward);
+    let word_start = find_word_boundary(slice, pos, Direction::Backward, long);
     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),
+        _ => find_word_boundary(slice, pos + 1, Direction::Forward, long),
     };
 
     // Special case.
@@ -268,7 +269,7 @@ mod test {
             let slice = doc.slice(..);
             for &case in scenario {
                 let (pos, objtype, expected_range) = case;
-                let result = textobject_word(slice, Range::point(pos), objtype, 1);
+                let result = textobject_word(slice, Range::point(pos), objtype, 1, false);
                 assert_eq!(
                     result,
                     expected_range.into(),