Introduce `tab-size` to correcly calculate the line length with tabulations (#4167)

crates/ruff/resources/test/fixtures/flake8_simplify/SIM102.py

@@ -86,9 +86,16 @@ while x > 0:
         ):
             print("Bad module!")
 
-# SIM102
+# SIM102 (auto-fixable)
-if node.module:
+if node.module012345678:
-    if node.module == "multiprocessing" or node.module.startswith(
+    if node.module == "multiprocß9💣2ℝ" or node.module.startswith(
+        "multiprocessing."
+    ):
+        print("Bad module!")
+
+# SIM102 (not auto-fixable)
+if node.module0123456789:
+    if node.module == "multiprocß9💣2ℝ" or node.module.startswith(
         "multiprocessing."
     ):
         print("Bad module!")

crates/ruff/resources/test/fixtures/flake8_simplify/SIM108.py

@@ -80,17 +80,25 @@ else:
 
 # SIM108
 if a:
-    b = cccccccccccccccccccccccccccccccccccc
+    b = "cccccccccccccccccccccccccccccccccß"
 else:
-    b = ddddddddddddddddddddddddddddddddddddd
+    b = "ddddddddddddddddddddddddddddddddd💣"
 
 
 # OK (too long)
 if True:
     if a:
-        b = cccccccccccccccccccccccccccccccccccc
+        b = ccccccccccccccccccccccccccccccccccc
     else:
-        b = ddddddddddddddddddddddddddddddddddddd
+        b = ddddddddddddddddddddddddddddddddddd
+
+
+# OK (too long with tabs)
+if True:
+	if a:
+		b = ccccccccccccccccccccccccccccccccccc
+	else:
+		b = ddddddddddddddddddddddddddddddddddd
 
 
 # SIM108 (without fix due to trailing comment)

crates/ruff/resources/test/fixtures/flake8_simplify/SIM110.py

@@ -155,3 +155,19 @@ def f():
         if check(x):
             return False
     return True
+
+
+def f():
+    # SIM110
+    for x in "012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ":
+        if x.isdigit():
+            return True
+    return False
+
+
+def f():
+    # OK (too long)
+    for x in "012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9":
+        if x.isdigit():
+            return True
+    return False

crates/ruff/resources/test/fixtures/flake8_simplify/SIM111.py

@@ -171,3 +171,19 @@ def f():
         if x > y:
             return False
     return True
+
+
+def f():
+    # SIM111
+    for x in "012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9012ß9":
+        if x.isdigit():
+            return False
+    return True
+
+
+def f():
+    # OK (too long)
+    for x in "012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9012ß90":
+        if x.isdigit():
+            return False
+    return True

crates/ruff/resources/test/fixtures/flake8_simplify/SIM117.py

@@ -90,3 +90,13 @@ with (
         D() as d,
     ):
         print("hello")
+
+# SIM117 (auto-fixable)
+with A("01ß9💣2ℝ8901ß9💣2ℝ8901ß9💣2ℝ89") as a:
+    with B("01ß9💣2ℝ8901ß9💣2ℝ8901ß9💣2ℝ89") as b:
+        print("hello")
+
+# SIM117 (not auto-fixable too long)
+with A("01ß9💣2ℝ8901ß9💣2ℝ8901ß9💣2ℝ890") as a:
+    with B("01ß9💣2ℝ8901ß9💣2ℝ8901ß9💣2ℝ89") as b:
+        print("hello")

crates/ruff/resources/test/fixtures/flake8_simplify/SIM401.py

@@ -36,12 +36,18 @@ else:
 if key in a_dict:
     vars[idx] = a_dict[key]
 else:
-    vars[idx] = "default"
+    vars[idx] = "defaultß9💣2ℝ6789ß9💣2ℝ6789ß9💣2ℝ6789ß9💣2ℝ6789ß9💣2ℝ6789"
 
 ###
 # Negative cases
 ###
 
+# OK (too long)
+if key in a_dict:
+    vars[idx] = a_dict[key]
+else:
+    vars[idx] = "defaultß9💣2ℝ6789ß9💣2ℝ6789ß9💣2ℝ6789ß9💣2ℝ6789ß9💣2ℝ6789ß"
+
 # OK (false negative)
 if not key in a_dict:
     var = "default"

crates/ruff/resources/test/fixtures/isort/fit_line_length_comment.py

@@ -2,3 +2,7 @@ import a
 # Don't take this comment into account when determining whether the next import can fit on one line.
 from b import c
 from d import e  # Do take this comment into account when determining whether the next import can fit on one line.
+# The next import fits on one line.
+from f import g  # 012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ
+# The next import doesn't fit on one line.
+from h import i  # 012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9

crates/ruff/resources/test/fixtures/pycodestyle/E501_2.py

@@ -0,0 +1,11 @@
+a = """ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+a	=	"""ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+
+b = """ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+b	=	"""ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+
+c   =   """2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+c	=	"""2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+
+d       =       """💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+d	=	"""💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""

crates/ruff/resources/test/fixtures/pycodestyle/W505_utf_8.py

@@ -0,0 +1,31 @@
+#!/usr/bin/env python3
+"""Here's a top-level ß9💣2ℝing that's over theß9💣2ℝ."""
+
+
+def f1():
+    """Here's a ß9💣2ℝing that's also over theß9💣2ℝ."""
+
+    x = 1  # Here's a comment that's over theß9💣2ℝ, but it's not standalone.
+
+    # Here's a standalone comment that's over theß9💣2ℝ.
+
+    x = 2
+    # Another standalone that is preceded by a newline and indent toke and is over theß9💣2ℝ.
+
+    print("Here's a string that's over theß9💣2ℝ, but it's not a ß9💣2ℝing.")
+
+
+"This is also considered a ß9💣2ℝing, and is over theß9💣2ℝ."
+
+
+def f2():
+    """Here's a multi-line ß9💣2ℝing.
+
+    It's over theß9💣2ℝ on this line, which isn't the first line in the ß9💣2ℝing.
+    """
+
+
+def f3():
+    """Here's a multi-line ß9💣2ℝing.
+
+    It's over theß9💣2ℝ on this line, which isn't the first line in the ß9💣2ℝing."""

crates/ruff/src/checkers/physical_lines.rs

@@ -183,6 +183,7 @@ mod tests {
 
     use ruff_python_ast::source_code::{Indexer, Locator, Stylist};
 
+    use crate::line_width::LineLength;
     use crate::registry::Rule;
     use crate::settings::Settings;
 
@@ -196,7 +197,7 @@ mod tests {
         let indexer = Indexer::from_tokens(&tokens, &locator);
         let stylist = Stylist::from_tokens(&tokens, &locator);
 
-        let check_with_max_line_length = |line_length: usize| {
+        let check_with_max_line_length = |line_length: LineLength| {
             check_physical_lines(
                 Path::new("foo.py"),
                 &locator,
@@ -209,7 +210,8 @@ mod tests {
                 },
             )
         };
-        assert_eq!(check_with_max_line_length(8), vec![]);
+        let line_length = LineLength::from(8);
-        assert_eq!(check_with_max_line_length(8), vec![]);
+        assert_eq!(check_with_max_line_length(line_length), vec![]);
+        assert_eq!(check_with_max_line_length(line_length), vec![]);
     }
 }

crates/ruff/src/flake8_to_ruff/converter.rs

@@ -3,6 +3,7 @@ use std::collections::{HashMap, HashSet};
 use anyhow::Result;
 use itertools::Itertools;
 
+use crate::line_width::LineLength;
 use crate::registry::Linter;
 use crate::rule_selector::RuleSelector;
 use crate::rules::flake8_pytest_style::types::{
@@ -120,7 +121,9 @@ pub fn convert(
                     options.builtins = Some(parser::parse_strings(value.as_ref()));
                 }
                 "max-line-length" | "max_line_length" => match value.parse::<usize>() {
-                    Ok(line_length) => options.line_length = Some(line_length),
+                    Ok(line_length) => {
+                        options.line_length = Some(LineLength::from(line_length));
+                    }
                     Err(e) => {
                         warn_user!("Unable to parse '{key}' property: {e}");
                     }
@@ -403,7 +406,7 @@ pub fn convert(
     // Extract any settings from the existing `pyproject.toml`.
     if let Some(black) = &external_config.black {
         if let Some(line_length) = &black.line_length {
-            options.line_length = Some(*line_length);
+            options.line_length = Some(LineLength::from(*line_length));
         }
 
         if let Some(target_version) = &black.target_version {
@@ -460,6 +463,7 @@ mod tests {
     use crate::flake8_to_ruff::converter::DEFAULT_SELECTORS;
     use crate::flake8_to_ruff::pep621::Project;
     use crate::flake8_to_ruff::ExternalConfig;
+    use crate::line_width::LineLength;
     use crate::registry::Linter;
     use crate::rule_selector::RuleSelector;
     use crate::rules::pydocstyle::settings::Convention;
@@ -510,7 +514,7 @@ mod tests {
             Some(vec![]),
         )?;
         let expected = Pyproject::new(Options {
-            line_length: Some(100),
+            line_length: Some(LineLength::from(100)),
             ..default_options([])
         });
         assert_eq!(actual, expected);
@@ -529,7 +533,7 @@ mod tests {
             Some(vec![]),
         )?;
         let expected = Pyproject::new(Options {
-            line_length: Some(100),
+            line_length: Some(LineLength::from(100)),
             ..default_options([])
         });
         assert_eq!(actual, expected);

crates/ruff/src/lib.rs

@@ -21,6 +21,7 @@ pub mod fs;
 mod importer;
 pub mod jupyter;
 mod lex;
+pub mod line_width;
 pub mod linter;
 pub mod logging;
 pub mod message;

crates/ruff/src/line_width.rs

@@ -0,0 +1,165 @@
+use serde::{Deserialize, Serialize};
+use unicode_width::UnicodeWidthChar;
+
+use ruff_macros::CacheKey;
+
+/// The length of a line of text that is considered too long.
+#[derive(Clone, Copy, Debug, PartialEq, Eq, PartialOrd, Ord, Serialize, Deserialize, CacheKey)]
+#[cfg_attr(feature = "schemars", derive(schemars::JsonSchema))]
+pub struct LineLength(usize);
+
+impl Default for LineLength {
+    /// The default line length.
+    fn default() -> Self {
+        Self(88)
+    }
+}
+
+impl LineLength {
+    pub const fn get(&self) -> usize {
+        self.0
+    }
+}
+
+impl From<usize> for LineLength {
+    fn from(value: usize) -> Self {
+        Self(value)
+    }
+}
+
+/// A measure of the width of a line of text.
+///
+/// This is used to determine if a line is too long.
+/// It should be compared to a [`LineLength`].
+#[derive(Clone, Copy, Debug)]
+pub struct LineWidth {
+    /// The width of the line.
+    width: usize,
+    /// The column of the line.
+    /// This is used to calculate the width of tabs.
+    column: usize,
+    /// The tab size to use when calculating the width of tabs.
+    tab_size: TabSize,
+}
+
+impl Default for LineWidth {
+    fn default() -> Self {
+        Self::new(TabSize::default())
+    }
+}
+
+impl PartialEq for LineWidth {
+    fn eq(&self, other: &Self) -> bool {
+        self.width == other.width
+    }
+}
+
+impl Eq for LineWidth {}
+
+impl PartialOrd for LineWidth {
+    fn partial_cmp(&self, other: &Self) -> Option<std::cmp::Ordering> {
+        self.width.partial_cmp(&other.width)
+    }
+}
+
+impl Ord for LineWidth {
+    fn cmp(&self, other: &Self) -> std::cmp::Ordering {
+        self.width.cmp(&other.width)
+    }
+}
+
+impl LineWidth {
+    pub fn get(&self) -> usize {
+        self.width
+    }
+
+    /// Creates a new `LineWidth` with the given tab size.
+    pub fn new(tab_size: TabSize) -> Self {
+        LineWidth {
+            width: 0,
+            column: 0,
+            tab_size,
+        }
+    }
+
+    fn update(mut self, chars: impl Iterator<Item = char>) -> Self {
+        let tab_size: usize = self.tab_size.into();
+        for c in chars {
+            match c {
+                '\t' => {
+                    let tab_offset = tab_size - (self.column % tab_size);
+                    self.width += tab_offset;
+                    self.column += tab_offset;
+                }
+                '\n' | '\r' => {
+                    self.width = 0;
+                    self.column = 0;
+                }
+                _ => {
+                    self.width += c.width().unwrap_or(0);
+                    self.column += 1;
+                }
+            }
+        }
+        self
+    }
+
+    /// Adds the given text to the line width.
+    #[must_use]
+    pub fn add_str(self, text: &str) -> Self {
+        self.update(text.chars())
+    }
+
+    /// Adds the given character to the line width.
+    #[must_use]
+    pub fn add_char(self, c: char) -> Self {
+        self.update(std::iter::once(c))
+    }
+
+    /// Adds the given width to the line width.
+    /// Also adds the given width to the column.
+    /// It is generally better to use [`LineWidth::add_str`] or [`LineWidth::add_char`].
+    /// The width and column should be the same for the corresponding text.
+    /// Currently, this is only used to add spaces.
+    #[must_use]
+    pub fn add_width(mut self, width: usize) -> Self {
+        self.width += width;
+        self.column += width;
+        self
+    }
+}
+
+impl PartialEq<LineLength> for LineWidth {
+    fn eq(&self, other: &LineLength) -> bool {
+        self.width == other.0
+    }
+}
+
+impl PartialOrd<LineLength> for LineWidth {
+    fn partial_cmp(&self, other: &LineLength) -> Option<std::cmp::Ordering> {
+        self.width.partial_cmp(&other.0)
+    }
+}
+
+/// The size of a tab.
+#[derive(Clone, Copy, Debug, PartialEq, Eq, Serialize, Deserialize, CacheKey)]
+#[cfg_attr(feature = "schemars", derive(schemars::JsonSchema))]
+pub struct TabSize(pub u8);
+
+impl Default for TabSize {
+    fn default() -> Self {
+        Self(4)
+    }
+}
+
+impl From<u8> for TabSize {
+    fn from(tab_size: u8) -> Self {
+        Self(tab_size)
+    }
+}
+
+impl From<TabSize> for usize {
+    fn from(tab_size: TabSize) -> Self {
+        tab_size.0 as usize
+    }
+}

crates/ruff/src/message/text.rs

@@ -11,6 +11,7 @@ use ruff_text_size::{TextRange, TextSize};
 use ruff_python_ast::source_code::{OneIndexed, SourceLocation};
 
 use crate::fs::relativize_path;
+use crate::line_width::{LineWidth, TabSize};
 use crate::message::diff::Diff;
 use crate::message::{Emitter, EmitterContext, Message};
 use crate::registry::AsRule;
@@ -237,18 +238,21 @@ impl Display for MessageCodeFrame<'_> {
 }
 
 fn replace_whitespace(source: &str, annotation_range: TextRange) -> SourceCode {
-    static TAB_SIZE: u32 = 4; // TODO(jonathan): use `pycodestyle.tab-size`
+    static TAB_SIZE: TabSize = TabSize(4); // TODO(jonathan): use `tab-size`
 
     let mut result = String::new();
     let mut last_end = 0;
     let mut range = annotation_range;
-    let mut column = 0;
+    let mut line_width = LineWidth::new(TAB_SIZE);
 
-    for (index, c) in source.chars().enumerate() {
+    for (index, c) in source.char_indices() {
-        match c {
+        let old_width = line_width.get();
-            '\t' => {
+        line_width = line_width.add_char(c);
-                let tab_width = TAB_SIZE - column % TAB_SIZE;
+
-                column += tab_width;
+        if matches!(c, '\t') {
+            // SAFETY: The difference is a value in the range [1..TAB_SIZE] which is guaranteed to be less than `u32`.
+            #[allow(clippy::cast_possible_truncation)]
+            let tab_width = (line_width.get() - old_width) as u32;
 
             if index < usize::from(annotation_range.start()) {
                 range += TextSize::new(tab_width - 1);
@@ -264,13 +268,6 @@ fn replace_whitespace(source: &str, annotation_range: TextRange) -> SourceCode {
 
             last_end = index + 1;
         }
-            '\n' | '\r' => {
-                column = 0;
-            }
-            _ => {
-                column += 1;
-            }
-        }
     }
 
     // No tabs

crates/ruff/src/rules/flake8_simplify/rules/ast_if.rs

@@ -2,7 +2,6 @@ use log::error;
 use ruff_text_size::TextRange;
 use rustc_hash::FxHashSet;
 use rustpython_parser::ast::{self, Cmpop, Constant, Expr, ExprContext, Ranged, Stmt};
-use unicode_width::UnicodeWidthStr;
 
 use ruff_diagnostics::{AutofixKind, Diagnostic, Edit, Fix, Violation};
 use ruff_macros::{derive_message_formats, violation};
@@ -14,6 +13,7 @@ use ruff_python_ast::newlines::StrExt;
 use ruff_python_semantic::model::SemanticModel;
 
 use crate::checkers::ast::Checker;
+use crate::line_width::LineWidth;
 use crate::registry::AsRule;
 use crate::rules::flake8_simplify::rules::fix_if;
 
@@ -288,7 +288,10 @@ pub(crate) fn nested_if_statements(
                     .content()
                     .unwrap_or_default()
                     .universal_newlines()
-                    .all(|line| line.width() <= checker.settings.line_length)
+                    .all(|line| {
+                        LineWidth::new(checker.settings.tab_size).add_str(&line)
+                            <= checker.settings.line_length
+                    })
                 {
                     #[allow(deprecated)]
                     diagnostic.set_fix(Fix::unspecified(edit));
@@ -508,8 +511,9 @@ pub(crate) fn use_ternary_operator(checker: &mut Checker, stmt: &Stmt, parent: O
 
     // Don't flag if the resulting expression would exceed the maximum line length.
     let line_start = checker.locator.line_start(stmt.start());
-    if checker.locator.contents()[TextRange::new(line_start, stmt.start())].width()
+    if LineWidth::new(checker.settings.tab_size)
-        + contents.width()
+        .add_str(&checker.locator.contents()[TextRange::new(line_start, stmt.start())])
+        .add_str(&contents)
         > checker.settings.line_length
     {
         return;
@@ -863,8 +867,9 @@ pub(crate) fn use_dict_get_with_default(
 
     // Don't flag if the resulting expression would exceed the maximum line length.
     let line_start = checker.locator.line_start(stmt.start());
-    if checker.locator.contents()[TextRange::new(line_start, stmt.start())].width()
+    if LineWidth::new(checker.settings.tab_size)
-        + contents.width()
+        .add_str(&checker.locator.contents()[TextRange::new(line_start, stmt.start())])
+        .add_str(&contents)
         > checker.settings.line_length
     {
         return;

crates/ruff/src/rules/flake8_simplify/rules/ast_with.rs

@@ -1,7 +1,6 @@
 use log::error;
 use ruff_text_size::TextRange;
 use rustpython_parser::ast::{self, Ranged, Stmt, Withitem};
-use unicode_width::UnicodeWidthStr;
 
 use ruff_diagnostics::{AutofixKind, Violation};
 use ruff_diagnostics::{Diagnostic, Fix};
@@ -10,6 +9,7 @@ use ruff_python_ast::helpers::{first_colon_range, has_comments_in};
 use ruff_python_ast::newlines::StrExt;
 
 use crate::checkers::ast::Checker;
+use crate::line_width::LineWidth;
 use crate::registry::AsRule;
 
 use super::fix_with;
@@ -111,7 +111,10 @@ pub(crate) fn multiple_with_statements(
                         .content()
                         .unwrap_or_default()
                         .universal_newlines()
-                        .all(|line| line.width() <= checker.settings.line_length)
+                        .all(|line| {
+                            LineWidth::new(checker.settings.tab_size).add_str(&line)
+                                <= checker.settings.line_length
+                        })
                     {
                         #[allow(deprecated)]
                         diagnostic.set_fix(Fix::unspecified(edit));

crates/ruff/src/rules/flake8_simplify/rules/reimplemented_builtin.rs

@@ -2,13 +2,13 @@ use ruff_text_size::{TextRange, TextSize};
 use rustpython_parser::ast::{
     self, Cmpop, Comprehension, Constant, Expr, ExprContext, Ranged, Stmt, Unaryop,
 };
-use unicode_width::UnicodeWidthStr;
 
 use ruff_diagnostics::{AutofixKind, Diagnostic, Edit, Fix, Violation};
 use ruff_macros::{derive_message_formats, violation};
 use ruff_python_ast::source_code::Generator;
 
 use crate::checkers::ast::Checker;
+use crate::line_width::LineWidth;
 use crate::registry::{AsRule, Rule};
 
 #[violation]
@@ -224,8 +224,9 @@ pub(crate) fn convert_for_loop_to_any_all(
 
                 // Don't flag if the resulting expression would exceed the maximum line length.
                 let line_start = checker.locator.line_start(stmt.start());
-                if checker.locator.contents()[TextRange::new(line_start, stmt.start())].width()
+                if LineWidth::new(checker.settings.tab_size)
-                    + contents.width()
+                    .add_str(&checker.locator.contents()[TextRange::new(line_start, stmt.start())])
+                    .add_str(&contents)
                     > checker.settings.line_length
                 {
                     return;
@@ -316,8 +317,9 @@ pub(crate) fn convert_for_loop_to_any_all(
 
                 // Don't flag if the resulting expression would exceed the maximum line length.
                 let line_start = checker.locator.line_start(stmt.start());
-                if checker.locator.contents()[TextRange::new(line_start, stmt.start())].width()
+                if LineWidth::new(checker.settings.tab_size)
-                    + contents.width()
+                    .add_str(&checker.locator.contents()[TextRange::new(line_start, stmt.start())])
+                    .add_str(&contents)
                     > checker.settings.line_length
                 {
                     return;

crates/ruff/src/rules/flake8_simplify/snapshots/ruff__rules__flake8_simplify__tests__SIM102_SIM102.py.snap

@@ -211,14 +211,14 @@ SIM102.py:83:5: SIM102 [*] Use a single `if` statement instead of nested `if` st
    85 |+    )):
    86 |+        print("Bad module!")
 88 87 | 
-89 88 | # SIM102
+89 88 | # SIM102 (auto-fixable)
-90 89 | if node.module:
+90 89 | if node.module012345678:
 
 SIM102.py:90:1: SIM102 [*] Use a single `if` statement instead of nested `if` statements
    |
-90 |   # SIM102
+90 |   # SIM102 (auto-fixable)
-91 | / if node.module:
+91 | / if node.module012345678:
-92 | |     if node.module == "multiprocessing" or node.module.startswith(
+92 | |     if node.module == "multiprocß9💣2ℝ" or node.module.startswith(
 93 | |         "multiprocessing."
 94 | |     ):
    | |______^ SIM102
@@ -229,44 +229,56 @@ SIM102.py:90:1: SIM102 [*] Use a single `if` statement instead of nested `if` st
 ℹ Suggested fix
 87 87 |             print("Bad module!")
 88 88 | 
-89 89 | # SIM102
+89 89 | # SIM102 (auto-fixable)
-90    |-if node.module:
+90    |-if node.module012345678:
-91    |-    if node.module == "multiprocessing" or node.module.startswith(
+91    |-    if node.module == "multiprocß9💣2ℝ" or node.module.startswith(
 92    |-        "multiprocessing."
 93    |-    ):
 94    |-        print("Bad module!")
-   90 |+if node.module and (node.module == "multiprocessing" or node.module.startswith(
+   90 |+if node.module012345678 and (node.module == "multiprocß9💣2ℝ" or node.module.startswith(
    91 |+    "multiprocessing."
    92 |+)):
    93 |+    print("Bad module!")
 95 94 | 
-96 95 | 
+96 95 | # SIM102 (not auto-fixable)
-97 96 | # OK
+97 96 | if node.module0123456789:
 
-SIM102.py:117:5: SIM102 [*] Use a single `if` statement instead of nested `if` statements
+SIM102.py:97:1: SIM102 Use a single `if` statement instead of nested `if` statements
     |
-117 |   if a:
+ 97 |   # SIM102 (not auto-fixable)
-118 |       # SIM 102
+ 98 | / if node.module0123456789:
-119 |       if b:
+ 99 | |     if node.module == "multiprocß9💣2ℝ" or node.module.startswith(
+100 | |         "multiprocessing."
+101 | |     ):
+    | |______^ SIM102
+102 |           print("Bad module!")
+    |
+    = help: Combine `if` statements using `and`
+
+SIM102.py:124:5: SIM102 [*] Use a single `if` statement instead of nested `if` statements
+    |
+124 |   if a:
+125 |       # SIM 102
+126 |       if b:
     |  _____^
-120 | |         if c:
+127 | |         if c:
     | |_____________^ SIM102
-121 |               print("foo")
+128 |               print("foo")
-122 |   else:
+129 |   else:
     |
     = help: Combine `if` statements using `and`
 
 ℹ Suggested fix
-114 114 | # OK
+121 121 | # OK
-115 115 | if a:
+122 122 | if a:
-116 116 |     # SIM 102
+123 123 |     # SIM 102
-117     |-    if b:
+124     |-    if b:
-118     |-        if c:
+125     |-        if c:
-119     |-            print("foo")
+126     |-            print("foo")
-    117 |+    if b and c:
+    124 |+    if b and c:
-    118 |+        print("foo")
+    125 |+        print("foo")
-120 119 | else:
+127 126 | else:
-121 120 |     print("bar")
+128 127 |     print("bar")
-122 121 | 
+129 128 | 
 
 

crates/ruff/src/rules/flake8_simplify/snapshots/ruff__rules__flake8_simplify__tests__SIM108_SIM108.py.snap

@@ -38,57 +38,57 @@ SIM108.py:58:1: SIM108 Use ternary operator `abc = x if x > 0 else -x` instead o
    |
    = help: Replace `if`-`else`-block with `abc = x if x > 0 else -x`
 
-SIM108.py:82:1: SIM108 [*] Use ternary operator `b = cccccccccccccccccccccccccccccccccccc if a else ddddddddddddddddddddddddddddddddddddd` instead of `if`-`else`-block
+SIM108.py:82:1: SIM108 [*] Use ternary operator `b = "cccccccccccccccccccccccccccccccccß" if a else "ddddddddddddddddddddddddddddddddd💣"` instead of `if`-`else`-block
    |
 82 |   # SIM108
 83 | / if a:
-84 | |     b = cccccccccccccccccccccccccccccccccccc
+84 | |     b = "cccccccccccccccccccccccccccccccccß"
 85 | | else:
-86 | |     b = ddddddddddddddddddddddddddddddddddddd
+86 | |     b = "ddddddddddddddddddddddddddddddddd💣"
    | |_____________________________________________^ SIM108
    |
-   = help: Replace `if`-`else`-block with `b = cccccccccccccccccccccccccccccccccccc if a else ddddddddddddddddddddddddddddddddddddd`
+   = help: Replace `if`-`else`-block with `b = "cccccccccccccccccccccccccccccccccß" if a else "ddddddddddddddddddddddddddddddddd💣"`
 
 ℹ Suggested fix
 79 79 | 
 80 80 | 
 81 81 | # SIM108
 82    |-if a:
-83    |-    b = cccccccccccccccccccccccccccccccccccc
+83    |-    b = "cccccccccccccccccccccccccccccccccß"
 84    |-else:
-85    |-    b = ddddddddddddddddddddddddddddddddddddd
+85    |-    b = "ddddddddddddddddddddddddddddddddd💣"
-   82 |+b = cccccccccccccccccccccccccccccccccccc if a else ddddddddddddddddddddddddddddddddddddd
+   82 |+b = "cccccccccccccccccccccccccccccccccß" if a else "ddddddddddddddddddddddddddddddddd💣"
 86 83 | 
 87 84 | 
 88 85 | # OK (too long)
 
-SIM108.py:97:1: SIM108 Use ternary operator `exitcode = 0 if True else 1` instead of `if`-`else`-block
+SIM108.py:105:1: SIM108 Use ternary operator `exitcode = 0 if True else 1` instead of `if`-`else`-block
     |
- 97 |   # SIM108 (without fix due to trailing comment)
+105 |   # SIM108 (without fix due to trailing comment)
- 98 | / if True:
+106 | / if True:
- 99 | |     exitcode = 0
+107 | |     exitcode = 0
-100 | | else:
+108 | | else:
-101 | |     exitcode = 1  # Trailing comment
+109 | |     exitcode = 1  # Trailing comment
     | |________________^ SIM108
     |
     = help: Replace `if`-`else`-block with `exitcode = 0 if True else 1`
 
-SIM108.py:104:1: SIM108 Use ternary operator `x = 3 if True else 5` instead of `if`-`else`-block
+SIM108.py:112:1: SIM108 Use ternary operator `x = 3 if True else 5` instead of `if`-`else`-block
     |
-104 |   # SIM108
+112 |   # SIM108
-105 | / if True: x = 3  # Foo
+113 | / if True: x = 3  # Foo
-106 | | else: x = 5
+114 | | else: x = 5
     | |___________^ SIM108
     |
     = help: Replace `if`-`else`-block with `x = 3 if True else 5`
 
-SIM108.py:109:1: SIM108 Use ternary operator `x = 3 if True else 5` instead of `if`-`else`-block
+SIM108.py:117:1: SIM108 Use ternary operator `x = 3 if True else 5` instead of `if`-`else`-block
     |
-109 |   # SIM108
+117 |   # SIM108
-110 | / if True:  # Foo
+118 | / if True:  # Foo
-111 | |     x = 3
+119 | |     x = 3
-112 | | else:
+120 | | else:
-113 | |     x = 5
+121 | |     x = 5
     | |_________^ SIM108
     |
     = help: Replace `if`-`else`-block with `x = 3 if True else 5`

crates/ruff/src/rules/flake8_simplify/snapshots/ruff__rules__flake8_simplify__tests__SIM110_SIM110.py.snap

@@ -264,5 +264,34 @@ SIM110.py:154:5: SIM110 [*] Use `return all(not check(x) for x in iterable)` ins
 156     |-            return False
 157     |-    return True
     154 |+    return all(not check(x) for x in iterable)
+158 155 | 
+159 156 | 
+160 157 | def f():
+
+SIM110.py:162:5: SIM110 [*] Use `return any(x.isdigit() for x in "012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ")` instead of `for` loop
+    |
+162 |   def f():
+163 |       # SIM110
+164 |       for x in "012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ":
+    |  _____^
+165 | |         if x.isdigit():
+166 | |             return True
+167 | |     return False
+    | |________________^ SIM110
+    |
+    = help: Replace with `return any(x.isdigit() for x in "012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ")`
+
+ℹ Suggested fix
+159 159 | 
+160 160 | def f():
+161 161 |     # SIM110
+162     |-    for x in "012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ":
+163     |-        if x.isdigit():
+164     |-            return True
+165     |-    return False
+    162 |+    return any(x.isdigit() for x in "012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ")
+166 163 | 
+167 164 | 
+168 165 | def f():
 
 

crates/ruff/src/rules/flake8_simplify/snapshots/ruff__rules__flake8_simplify__tests__SIM110_SIM111.py.snap

@@ -316,5 +316,34 @@ SIM111.py:170:5: SIM110 [*] Use `return all(x <= y for x in iterable)` instead o
 172     |-            return False
 173     |-    return True
     170 |+    return all(x <= y for x in iterable)
+174 171 | 
+175 172 | 
+176 173 | def f():
+
+SIM111.py:178:5: SIM110 [*] Use `return all(not x.isdigit() for x in "012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9012ß9")` instead of `for` loop
+    |
+178 |   def f():
+179 |       # SIM111
+180 |       for x in "012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9012ß9":
+    |  _____^
+181 | |         if x.isdigit():
+182 | |             return False
+183 | |     return True
+    | |_______________^ SIM110
+    |
+    = help: Replace with `return all(not x.isdigit() for x in "012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9012ß9")`
+
+ℹ Suggested fix
+175 175 | 
+176 176 | def f():
+177 177 |     # SIM111
+178     |-    for x in "012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9012ß9":
+179     |-        if x.isdigit():
+180     |-            return False
+181     |-    return True
+    178 |+    return all(not x.isdigit() for x in "012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9012ß9")
+182 179 | 
+183 180 | 
+184 181 | def f():
 
 

crates/ruff/src/rules/flake8_simplify/snapshots/ruff__rules__flake8_simplify__tests__SIM117_SIM117.py.snap

@@ -212,5 +212,41 @@ SIM117.py:84:1: SIM117 [*] Use a single `with` statement with multiple contexts
 91    |-    ):
 92    |-        print("hello")
    89 |+    print("hello")
+93 90 | 
+94 91 | # SIM117 (auto-fixable)
+95 92 | with A("01ß9💣2ℝ8901ß9💣2ℝ8901ß9💣2ℝ89") as a:
+
+SIM117.py:95:1: SIM117 [*] Use a single `with` statement with multiple contexts instead of nested `with` statements
+   |
+95 |   # SIM117 (auto-fixable)
+96 | / with A("01ß9💣2ℝ8901ß9💣2ℝ8901ß9💣2ℝ89") as a:
+97 | |     with B("01ß9💣2ℝ8901ß9💣2ℝ8901ß9💣2ℝ89") as b:
+   | |__________________________________________________^ SIM117
+98 |           print("hello")
+   |
+   = help: Combine `with` statements
+
+ℹ Suggested fix
+92 92 |         print("hello")
+93 93 | 
+94 94 | # SIM117 (auto-fixable)
+95    |-with A("01ß9💣2ℝ8901ß9💣2ℝ8901ß9💣2ℝ89") as a:
+96    |-    with B("01ß9💣2ℝ8901ß9💣2ℝ8901ß9💣2ℝ89") as b:
+97    |-        print("hello")
+   95 |+with A("01ß9💣2ℝ8901ß9💣2ℝ8901ß9💣2ℝ89") as a, B("01ß9💣2ℝ8901ß9💣2ℝ8901ß9💣2ℝ89") as b:
+   96 |+    print("hello")
+98 97 | 
+99 98 | # SIM117 (not auto-fixable too long)
+100 99 | with A("01ß9💣2ℝ8901ß9💣2ℝ8901ß9💣2ℝ890") as a:
+
+SIM117.py:100:1: SIM117 Use a single `with` statement with multiple contexts instead of nested `with` statements
+    |
+100 |   # SIM117 (not auto-fixable too long)
+101 | / with A("01ß9💣2ℝ8901ß9💣2ℝ8901ß9💣2ℝ890") as a:
+102 | |     with B("01ß9💣2ℝ8901ß9💣2ℝ8901ß9💣2ℝ89") as b:
+    | |__________________________________________________^ SIM117
+103 |           print("hello")
+    |
+    = help: Combine `with` statements
 
 

crates/ruff/src/rules/flake8_simplify/snapshots/ruff__rules__flake8_simplify__tests__SIM401_SIM401.py.snap

@@ -105,18 +105,18 @@ SIM401.py:30:1: SIM401 [*] Use `var = dicts[idx].get(key, "default")` instead of
 35 32 | # SIM401 (complex expression in var)
 36 33 | if key in a_dict:
 
-SIM401.py:36:1: SIM401 [*] Use `vars[idx] = a_dict.get(key, "default")` instead of an `if` block
+SIM401.py:36:1: SIM401 [*] Use `vars[idx] = a_dict.get(key, "defaultß9💣2ℝ6789ß9💣2ℝ6789ß9💣2ℝ6789ß9💣2ℝ6789ß9💣2ℝ6789")` instead of an `if` block
    |
 36 |   # SIM401 (complex expression in var)
 37 | / if key in a_dict:
 38 | |     vars[idx] = a_dict[key]
 39 | | else:
-40 | |     vars[idx] = "default"
+40 | |     vars[idx] = "defaultß9💣2ℝ6789ß9💣2ℝ6789ß9💣2ℝ6789ß9💣2ℝ6789ß9💣2ℝ6789"
-   | |_________________________^ SIM401
+   | |___________________________________________________________________________^ SIM401
 41 |   
 42 |   ###
    |
-   = help: Replace with `vars[idx] = a_dict.get(key, "default")`
+   = help: Replace with `vars[idx] = a_dict.get(key, "defaultß9💣2ℝ6789ß9💣2ℝ6789ß9💣2ℝ6789ß9💣2ℝ6789ß9💣2ℝ6789")`
 
 ℹ Suggested fix
 33 33 |     var = "default"
@@ -125,8 +125,8 @@ SIM401.py:36:1: SIM401 [*] Use `vars[idx] = a_dict.get(key, "default")` instead
 36    |-if key in a_dict:
 37    |-    vars[idx] = a_dict[key]
 38    |-else:
-39    |-    vars[idx] = "default"
+39    |-    vars[idx] = "defaultß9💣2ℝ6789ß9💣2ℝ6789ß9💣2ℝ6789ß9💣2ℝ6789ß9💣2ℝ6789"
-   36 |+vars[idx] = a_dict.get(key, "default")
+   36 |+vars[idx] = a_dict.get(key, "defaultß9💣2ℝ6789ß9💣2ℝ6789ß9💣2ℝ6789ß9💣2ℝ6789ß9💣2ℝ6789")
 40 37 | 
 41 38 | ###
 42 39 | # Negative cases

crates/ruff/src/rules/isort/format.rs

@@ -1,7 +1,7 @@
-use unicode_width::UnicodeWidthStr;
-
 use ruff_python_ast::source_code::Stylist;
 
+use crate::line_width::{LineLength, LineWidth};
+
 use super::types::{AliasData, CommentSet, ImportFromData, Importable};
 
 // Guess a capacity to use for string allocation.
@@ -45,7 +45,8 @@ pub(crate) fn format_import_from(
     import_from: &ImportFromData,
     comments: &CommentSet,
     aliases: &[(AliasData, CommentSet)],
-    line_length: usize,
+    line_length: LineLength,
+    indentation_width: LineWidth,
     stylist: &Stylist,
     force_wrap_aliases: bool,
     is_first: bool,
@@ -56,8 +57,14 @@ pub(crate) fn format_import_from(
             .iter()
             .all(|(alias, _)| alias.name == "*" && alias.asname.is_none())
     {
-        let (single_line, ..) =
+        let (single_line, ..) = format_single_line(
-            format_single_line(import_from, comments, aliases, is_first, stylist);
+            import_from,
+            comments,
+            aliases,
+            is_first,
+            stylist,
+            indentation_width,
+        );
         return single_line;
     }
 
@@ -71,8 +78,14 @@ pub(crate) fn format_import_from(
             || aliases.len() == 1
             || aliases.iter().all(|(alias, _)| alias.asname.is_none()))
     {
-        let (single_line, import_width) =
+        let (single_line, import_width) = format_single_line(
-            format_single_line(import_from, comments, aliases, is_first, stylist);
+            import_from,
+            comments,
+            aliases,
+            is_first,
+            stylist,
+            indentation_width,
+        );
         if import_width <= line_length || aliases.iter().any(|(alias, _)| alias.name == "*") {
             return single_line;
         }
@@ -90,9 +103,10 @@ fn format_single_line(
     aliases: &[(AliasData, CommentSet)],
     is_first: bool,
     stylist: &Stylist,
-) -> (String, usize) {
+    indentation_width: LineWidth,
+) -> (String, LineWidth) {
     let mut output = String::with_capacity(CAPACITY);
-    let mut line_width = 0;
+    let mut line_width = indentation_width;
 
     if !is_first && !comments.atop.is_empty() {
         output.push_str(&stylist.line_ending());
@@ -106,28 +120,28 @@ fn format_single_line(
     output.push_str("from ");
     output.push_str(&module_name);
     output.push_str(" import ");
-    line_width += 5 + module_name.width() + 8;
+    line_width = line_width.add_width(5).add_str(&module_name).add_width(8);
 
     for (index, (AliasData { name, asname }, comments)) in aliases.iter().enumerate() {
         if let Some(asname) = asname {
             output.push_str(name);
             output.push_str(" as ");
             output.push_str(asname);
-            line_width += name.width() + 4 + asname.width();
+            line_width = line_width.add_str(name).add_width(4).add_str(asname);
         } else {
             output.push_str(name);
-            line_width += name.width();
+            line_width = line_width.add_str(name);
         }
         if index < aliases.len() - 1 {
             output.push_str(", ");
-            line_width += 2;
+            line_width = line_width.add_width(2);
         }
 
         for comment in &comments.inline {
             output.push(' ');
             output.push(' ');
             output.push_str(comment);
-            line_width += 2 + comment.width();
+            line_width = line_width.add_width(2).add_str(comment);
         }
     }
 
@@ -135,7 +149,7 @@ fn format_single_line(
         output.push(' ');
         output.push(' ');
         output.push_str(comment);
-        line_width += 2 + comment.width();
+        line_width = line_width.add_width(2).add_str(comment);
     }
 
     output.push_str(&stylist.line_ending());

crates/ruff/src/rules/isort/mod.rs

@@ -17,6 +17,7 @@ use sorting::cmp_either_import;
 use types::EitherImport::{Import, ImportFrom};
 use types::{AliasData, EitherImport, TrailingComma};
 
+use crate::line_width::{LineLength, LineWidth};
 use crate::rules::isort::categorize::KnownModules;
 use crate::rules::isort::types::ImportBlock;
 use crate::settings::types::PythonVersion;
@@ -65,7 +66,8 @@ pub(crate) fn format_imports(
     block: &Block,
     comments: Vec<Comment>,
     locator: &Locator,
-    line_length: usize,
+    line_length: LineLength,
+    indentation_width: LineWidth,
     stylist: &Stylist,
     src: &[PathBuf],
     package: Option<&Path>,
@@ -107,6 +109,7 @@ pub(crate) fn format_imports(
         let block_output = format_import_block(
             block,
             line_length,
+            indentation_width,
             stylist,
             src,
             package,
@@ -162,7 +165,8 @@ pub(crate) fn format_imports(
 #[allow(clippy::too_many_arguments, clippy::fn_params_excessive_bools)]
 fn format_import_block(
     block: ImportBlock,
-    line_length: usize,
+    line_length: LineLength,
+    indentation_width: LineWidth,
     stylist: &Stylist,
     src: &[PathBuf],
     package: Option<&Path>,
@@ -264,6 +268,7 @@ fn format_import_block(
                         &comments,
                         &aliases,
                         line_length,
+                        indentation_width,
                         stylist,
                         force_wrap_aliases,
                         is_first_statement,

crates/ruff/src/rules/isort/rules/organize_imports.rs

@@ -13,6 +13,7 @@ use ruff_python_ast::helpers::{
 use ruff_python_ast::source_code::{Indexer, Locator, Stylist};
 use ruff_python_ast::whitespace::leading_space;
 
+use crate::line_width::LineWidth;
 use crate::registry::AsRule;
 use crate::settings::Settings;
 
@@ -116,7 +117,8 @@ pub(crate) fn organize_imports(
         block,
         comments,
         locator,
-        settings.line_length - indentation.len(),
+        settings.line_length,
+        LineWidth::new(settings.tab_size).add_str(indentation),
         stylist,
         &settings.src,
         package,

crates/ruff/src/rules/isort/snapshots/ruff__rules__isort__tests__fit_line_length_comment.py.snap

@@ -7,6 +7,10 @@ fit_line_length_comment.py:1:1: I001 [*] Import block is un-sorted or un-formatt
 2 | | # Don't take this comment into account when determining whether the next import can fit on one line.
 3 | | from b import c
 4 | | from d import e  # Do take this comment into account when determining whether the next import can fit on one line.
+5 | | # The next import fits on one line.
+6 | | from f import g  # 012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ
+7 | | # The next import doesn't fit on one line.
+8 | | from h import i  # 012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9
   |
   = help: Organize imports
 
@@ -19,5 +23,14 @@ fit_line_length_comment.py:1:1: I001 [*] Import block is un-sorted or un-formatt
    5  |+from d import (
    6  |+    e,  # Do take this comment into account when determining whether the next import can fit on one line.
    7  |+)
+   8  |+
+5  9  | # The next import fits on one line.
+6  10 | from f import g  # 012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ
+   11 |+
+7  12 | # The next import doesn't fit on one line.
+8     |-from h import i  # 012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9
+   13 |+from h import (
+   14 |+    i,  # 012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9012ß9💣2ℝ9
+   15 |+)
 
 

crates/ruff/src/rules/pycodestyle/helpers.rs

@@ -1,10 +1,12 @@
 use ruff_text_size::{TextLen, TextRange};
 use rustpython_parser::ast::{self, Cmpop, Expr};
-use unicode_width::{UnicodeWidthChar, UnicodeWidthStr};
+use unicode_width::UnicodeWidthStr;
 
 use ruff_python_ast::newlines::Line;
 use ruff_python_ast::source_code::Generator;
 
+use crate::line_width::{LineLength, LineWidth, TabSize};
+
 pub(crate) fn is_ambiguous_name(name: &str) -> bool {
     name == "l" || name == "I" || name == "O"
 }
@@ -26,19 +28,19 @@ pub(crate) fn compare(
 
 pub(super) fn is_overlong(
     line: &Line,
-    limit: usize,
+    limit: LineLength,
     ignore_overlong_task_comments: bool,
     task_tags: &[String],
+    tab_size: TabSize,
 ) -> Option<Overlong> {
     let mut start_offset = line.start();
-    let mut width = 0;
+    let mut width = LineWidth::new(tab_size);
 
     for c in line.chars() {
         if width < limit {
             start_offset += c.text_len();
         }
-
+        width = width.add_char(c);
-        width += c.width().unwrap_or(0);
     }
 
     if width <= limit {
@@ -64,14 +66,14 @@ pub(super) fn is_overlong(
     // begins before the limit.
     let last_chunk = chunks.last().unwrap_or(second_chunk);
     if last_chunk.contains("://") {
-        if width - last_chunk.width() <= limit {
+        if width.get() - last_chunk.width() <= limit.get() {
             return None;
         }
     }
 
     Some(Overlong {
         range: TextRange::new(start_offset, line.end()),
-        width,
+        width: width.get(),
     })
 }
 

crates/ruff/src/rules/pycodestyle/mod.rs

@@ -11,6 +11,7 @@ mod tests {
     use anyhow::Result;
     use test_case::test_case;
 
+    use crate::line_width::LineLength;
     use crate::registry::Rule;
     use crate::test::test_path;
     use crate::{assert_messages, settings};
@@ -166,7 +167,7 @@ mod tests {
             Path::new("pycodestyle/W505.py"),
             &settings::Settings {
                 pycodestyle: Settings {
-                    max_doc_length: Some(50),
+                    max_doc_length: Some(LineLength::from(50)),
                     ..Settings::default()
                 },
                 ..settings::Settings::for_rule(Rule::DocLineTooLong)
@@ -175,4 +176,37 @@ mod tests {
         assert_messages!(diagnostics);
         Ok(())
     }
+
+    #[test]
+    fn max_doc_length_with_utf_8() -> Result<()> {
+        let diagnostics = test_path(
+            Path::new("pycodestyle/W505_utf_8.py"),
+            &settings::Settings {
+                pycodestyle: Settings {
+                    max_doc_length: Some(LineLength::from(50)),
+                    ..Settings::default()
+                },
+                ..settings::Settings::for_rule(Rule::DocLineTooLong)
+            },
+        )?;
+        assert_messages!(diagnostics);
+        Ok(())
+    }
+
+    #[test_case(1)]
+    #[test_case(2)]
+    #[test_case(4)]
+    #[test_case(8)]
+    fn tab_size(tab_size: u8) -> Result<()> {
+        let snapshot = format!("tab_size_{tab_size}");
+        let diagnostics = test_path(
+            Path::new("pycodestyle/E501_2.py"),
+            &settings::Settings {
+                tab_size: tab_size.into(),
+                ..settings::Settings::for_rule(Rule::LineTooLong)
+            },
+        )?;
+        assert_messages!(snapshot, diagnostics);
+        Ok(())
+    }
 }

crates/ruff/src/rules/pycodestyle/rules/doc_line_too_long.rs

@@ -48,6 +48,12 @@ pub(crate) fn doc_line_too_long(line: &Line, settings: &Settings) -> Option<Diag
         limit,
         settings.pycodestyle.ignore_overlong_task_comments,
         &settings.task_tags,
+        settings.tab_size,
     )
-    .map(|overlong| Diagnostic::new(DocLineTooLong(overlong.width(), limit), overlong.range()))
+    .map(|overlong| {
+        Diagnostic::new(
+            DocLineTooLong(overlong.width(), limit.get()),
+            overlong.range(),
+        )
+    })
 }

crates/ruff/src/rules/pycodestyle/rules/line_too_long.rs

@@ -43,6 +43,7 @@ pub(crate) fn line_too_long(line: &Line, settings: &Settings) -> Option<Diagnost
         limit,
         settings.pycodestyle.ignore_overlong_task_comments,
         &settings.task_tags,
+        settings.tab_size,
     )
-    .map(|overlong| Diagnostic::new(LineTooLong(overlong.width(), limit), overlong.range()))
+    .map(|overlong| Diagnostic::new(LineTooLong(overlong.width(), limit.get()), overlong.range()))
 }

crates/ruff/src/rules/pycodestyle/settings.rs

@@ -4,6 +4,8 @@ use serde::{Deserialize, Serialize};
 
 use ruff_macros::{CacheKey, CombineOptions, ConfigurationOptions};
 
+use crate::line_width::LineLength;
+
 #[derive(
     Debug, PartialEq, Eq, Serialize, Deserialize, Default, ConfigurationOptions, CombineOptions,
 )]
@@ -19,7 +21,7 @@ pub struct Options {
     )]
     /// The maximum line length to allow for line-length violations within
     /// documentation (`W505`), including standalone comments.
-    pub max_doc_length: Option<usize>,
+    pub max_doc_length: Option<LineLength>,
     #[option(
         default = "false",
         value_type = "bool",
@@ -35,7 +37,7 @@ pub struct Options {
 
 #[derive(Debug, Default, CacheKey)]
 pub struct Settings {
-    pub max_doc_length: Option<usize>,
+    pub max_doc_length: Option<LineLength>,
     pub ignore_overlong_task_comments: bool,
 }
 

crates/ruff/src/rules/pycodestyle/snapshots/ruff__rules__pycodestyle__tests__max_doc_length_with_utf_8.snap

@@ -0,0 +1,62 @@
+---
+source: crates/ruff/src/rules/pycodestyle/mod.rs
+---
+W505_utf_8.py:2:50: W505 Doc line too long (57 > 50 characters)
+  |
+2 | #!/usr/bin/env python3
+3 | """Here's a top-level ß9💣2ℝing that's over theß9💣2ℝ."""
+  |                                                    ^^^^^^ W505
+  |
+
+W505_utf_8.py:6:49: W505 Doc line too long (56 > 50 characters)
+  |
+6 | def f1():
+7 |     """Here's a ß9💣2ℝing that's also over theß9💣2ℝ."""
+  |                                                   ^^^^^^ W505
+8 | 
+9 |     x = 1  # Here's a comment that's over theß9💣2ℝ, but it's not standalone.
+  |
+
+W505_utf_8.py:10:51: W505 Doc line too long (56 > 50 characters)
+   |
+10 |     x = 1  # Here's a comment that's over theß9💣2ℝ, but it's not standalone.
+11 | 
+12 |     # Here's a standalone comment that's over theß9💣2ℝ.
+   |                                                   ^^^^^^ W505
+13 | 
+14 |     x = 2
+   |
+
+W505_utf_8.py:13:51: W505 Doc line too long (93 > 50 characters)
+   |
+13 |     x = 2
+14 |     # Another standalone that is preceded by a newline and indent toke and is over theß9💣2ℝ.
+   |                                                   ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ W505
+15 | 
+16 |     print("Here's a string that's over theß9💣2ℝ, but it's not a ß9💣2ℝing.")
+   |
+
+W505_utf_8.py:18:50: W505 Doc line too long (61 > 50 characters)
+   |
+18 | "This is also considered a ß9💣2ℝing, and is over theß9💣2ℝ."
+   |                                                   ^^^^^^^^^^^ W505
+   |
+
+W505_utf_8.py:24:50: W505 Doc line too long (82 > 50 characters)
+   |
+24 |     """Here's a multi-line ß9💣2ℝing.
+25 | 
+26 |     It's over theß9💣2ℝ on this line, which isn't the first line in the ß9💣2ℝing.
+   |                                                   ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ W505
+27 |     """
+   |
+
+W505_utf_8.py:31:50: W505 Doc line too long (85 > 50 characters)
+   |
+31 |     """Here's a multi-line ß9💣2ℝing.
+32 | 
+33 |     It's over theß9💣2ℝ on this line, which isn't the first line in the ß9💣2ℝing."""
+   |                                                   ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ W505
+   |
+
+

crates/ruff/src/rules/pycodestyle/snapshots/ruff__rules__pycodestyle__tests__tab_size_1.snap

@@ -0,0 +1,56 @@
+---
+source: crates/ruff/src/rules/pycodestyle/mod.rs
+---
+E501_2.py:1:81: E501 Line too long (89 > 88 characters)
+  |
+1 | a = """ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+  |                                                                                         ^ E501
+2 | a   =   """ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+  |
+
+E501_2.py:2:81: E501 Line too long (89 > 88 characters)
+  |
+2 | a = """ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+3 | a   =   """ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+  |                                                                                             ^ E501
+4 | 
+5 | b = """ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+  |
+
+E501_2.py:4:81: E501 Line too long (89 > 88 characters)
+  |
+4 | a   =   """ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+5 | 
+6 | b = """ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+  |                                                                                         ^ E501
+7 | b   =   """ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+  |
+
+E501_2.py:5:81: E501 Line too long (89 > 88 characters)
+  |
+5 | b = """ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+6 | b   =   """ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+  |                                                                                             ^ E501
+7 | 
+8 | c   =   """2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+  |
+
+E501_2.py:7:82: E501 Line too long (89 > 88 characters)
+   |
+ 7 | b   =   """ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+ 8 | 
+ 9 | c   =   """2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+   |                                                                                         ^ E501
+10 | c   =   """2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+   |
+
+E501_2.py:10:82: E501 Line too long (89 > 88 characters)
+   |
+10 | c   =   """2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+11 | 
+12 | d       =       """💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+   |                                                                                         ^ E501
+13 | d   =   """💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+   |
+
+

crates/ruff/src/rules/pycodestyle/snapshots/ruff__rules__pycodestyle__tests__tab_size_2.snap

@@ -0,0 +1,56 @@
+---
+source: crates/ruff/src/rules/pycodestyle/mod.rs
+---
+E501_2.py:1:81: E501 Line too long (89 > 88 characters)
+  |
+1 | a = """ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+  |                                                                                         ^ E501
+2 | a   =   """ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+  |
+
+E501_2.py:2:81: E501 Line too long (89 > 88 characters)
+  |
+2 | a = """ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+3 | a   =   """ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+  |                                                                                             ^ E501
+4 | 
+5 | b = """ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+  |
+
+E501_2.py:4:81: E501 Line too long (89 > 88 characters)
+  |
+4 | a   =   """ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+5 | 
+6 | b = """ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+  |                                                                                         ^ E501
+7 | b   =   """ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+  |
+
+E501_2.py:5:81: E501 Line too long (89 > 88 characters)
+  |
+5 | b = """ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+6 | b   =   """ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+  |                                                                                             ^ E501
+7 | 
+8 | c   =   """2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+  |
+
+E501_2.py:7:82: E501 Line too long (89 > 88 characters)
+   |
+ 7 | b   =   """ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+ 8 | 
+ 9 | c   =   """2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+   |                                                                                         ^ E501
+10 | c   =   """2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+   |
+
+E501_2.py:10:82: E501 Line too long (89 > 88 characters)
+   |
+10 | c   =   """2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+11 | 
+12 | d       =       """💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+   |                                                                                         ^ E501
+13 | d   =   """💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+   |
+
+

crates/ruff/src/rules/pycodestyle/snapshots/ruff__rules__pycodestyle__tests__tab_size_4.snap

@@ -0,0 +1,65 @@
+---
+source: crates/ruff/src/rules/pycodestyle/mod.rs
+---
+E501_2.py:1:81: E501 Line too long (89 > 88 characters)
+  |
+1 | a = """ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+  |                                                                                         ^ E501
+2 | a   =   """ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+  |
+
+E501_2.py:2:77: E501 Line too long (93 > 88 characters)
+  |
+2 | a = """ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+3 | a   =   """ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+  |                                                                                         ^^^^^ E501
+4 | 
+5 | b = """ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+  |
+
+E501_2.py:4:81: E501 Line too long (89 > 88 characters)
+  |
+4 | a   =   """ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+5 | 
+6 | b = """ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+  |                                                                                         ^ E501
+7 | b   =   """ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+  |
+
+E501_2.py:5:77: E501 Line too long (93 > 88 characters)
+  |
+5 | b = """ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+6 | b   =   """ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+  |                                                                                         ^^^^^ E501
+7 | 
+8 | c   =   """2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+  |
+
+E501_2.py:7:82: E501 Line too long (89 > 88 characters)
+   |
+ 7 | b   =   """ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+ 8 | 
+ 9 | c   =   """2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+   |                                                                                         ^ E501
+10 | c   =   """2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+   |
+
+E501_2.py:8:78: E501 Line too long (89 > 88 characters)
+   |
+ 8 | c   =   """2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+ 9 | c   =   """2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+   |                                                                                         ^ E501
+10 | 
+11 | d       =       """💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+   |
+
+E501_2.py:10:82: E501 Line too long (89 > 88 characters)
+   |
+10 | c   =   """2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+11 | 
+12 | d       =       """💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+   |                                                                                         ^ E501
+13 | d   =   """💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+   |
+
+

crates/ruff/src/rules/pycodestyle/snapshots/ruff__rules__pycodestyle__tests__tab_size_8.snap

@@ -0,0 +1,72 @@
+---
+source: crates/ruff/src/rules/pycodestyle/mod.rs
+---
+E501_2.py:1:81: E501 Line too long (89 > 88 characters)
+  |
+1 | a = """ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+  |                                                                                         ^ E501
+2 | a   =   """ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+  |
+
+E501_2.py:2:70: E501 Line too long (101 > 88 characters)
+  |
+2 | a = """ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+3 | a   =   """ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+  |                                                                                 ^^^^^^^^^^^^^ E501
+4 | 
+5 | b = """ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+  |
+
+E501_2.py:4:81: E501 Line too long (89 > 88 characters)
+  |
+4 | a   =   """ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+5 | 
+6 | b = """ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+  |                                                                                         ^ E501
+7 | b   =   """ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+  |
+
+E501_2.py:5:70: E501 Line too long (101 > 88 characters)
+  |
+5 | b = """ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+6 | b   =   """ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+  |                                                                                 ^^^^^^^^^^^^^ E501
+7 | 
+8 | c   =   """2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+  |
+
+E501_2.py:7:82: E501 Line too long (89 > 88 characters)
+   |
+ 7 | b   =   """ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+ 8 | 
+ 9 | c   =   """2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+   |                                                                                         ^ E501
+10 | c   =   """2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+   |
+
+E501_2.py:8:71: E501 Line too long (97 > 88 characters)
+   |
+ 8 | c   =   """2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+ 9 | c   =   """2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+   |                                                                                  ^^^^^^^^ E501
+10 | 
+11 | d       =       """💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+   |
+
+E501_2.py:10:82: E501 Line too long (89 > 88 characters)
+   |
+10 | c   =   """2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+11 | 
+12 | d       =       """💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+   |                                                                                         ^ E501
+13 | d   =   """💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+   |
+
+E501_2.py:11:70: E501 Line too long (89 > 88 characters)
+   |
+11 | d       =       """💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+12 | d   =   """💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A67ß9💣2ℝ4A6"""
+   |                                                                                 ^ E501
+   |
+
+

crates/ruff/src/settings/configuration.rs

@@ -13,6 +13,7 @@ use shellexpand;
 use shellexpand::LookupError;
 
 use crate::fs;
+use crate::line_width::{LineLength, TabSize};
 use crate::rule_selector::RuleSelector;
 use crate::rules::{
     flake8_annotations, flake8_bandit, flake8_bugbear, flake8_builtins, flake8_comprehensions,
@@ -56,7 +57,8 @@ pub struct Configuration {
     pub format: Option<SerializationFormat>,
     pub ignore_init_module_imports: Option<bool>,
     pub include: Option<Vec<FilePattern>>,
-    pub line_length: Option<usize>,
+    pub line_length: Option<LineLength>,
+    pub tab_size: Option<TabSize>,
     pub namespace_packages: Option<Vec<PathBuf>>,
     pub required_version: Option<Version>,
     pub respect_gitignore: Option<bool>,
@@ -194,6 +196,7 @@ impl Configuration {
                     .collect()
             }),
             line_length: options.line_length,
+            tab_size: options.tab_size,
             namespace_packages: options
                 .namespace_packages
                 .map(|namespace_package| resolve_src(&namespace_package, project_root))
@@ -281,6 +284,7 @@ impl Configuration {
                 .ignore_init_module_imports
                 .or(config.ignore_init_module_imports),
             line_length: self.line_length.or(config.line_length),
+            tab_size: self.tab_size.or(config.tab_size),
             namespace_packages: self.namespace_packages.or(config.namespace_packages),
             per_file_ignores: self.per_file_ignores.or(config.per_file_ignores),
             required_version: self.required_version.or(config.required_version),

crates/ruff/src/settings/defaults.rs

@@ -6,6 +6,7 @@ use regex::Regex;
 use rustc_hash::FxHashSet;
 
 use crate::codes::{self, RuleCodePrefix};
+use crate::line_width::{LineLength, TabSize};
 use crate::registry::Linter;
 use crate::rule_selector::{prefix_to_selector, RuleSelector};
 use crate::rules::{
@@ -26,8 +27,6 @@ pub const PREFIXES: &[RuleSelector] = &[
 
 pub const TARGET_VERSION: PythonVersion = PythonVersion::Py310;
 
-pub const LINE_LENGTH: usize = 88;
-
 pub const TASK_TAGS: &[&str] = &["TODO", "FIXME", "XXX"];
 
 pub static DUMMY_VARIABLE_RGX: Lazy<Regex> =
@@ -76,7 +75,8 @@ impl Default for Settings {
             force_exclude: false,
             ignore_init_module_imports: false,
             include: FilePatternSet::try_from_vec(INCLUDE.clone()).unwrap(),
-            line_length: LINE_LENGTH,
+            line_length: LineLength::default(),
+            tab_size: TabSize::default(),
             namespace_packages: vec![],
             per_file_ignores: vec![],
             respect_gitignore: true,

crates/ruff/src/settings/mod.rs

@@ -26,6 +26,7 @@ use crate::settings::types::{FilePatternSet, PerFileIgnore, PythonVersion, Seria
 use crate::warn_user_once_by_id;
 
 use self::rule_table::RuleTable;
+use super::line_width::{LineLength, TabSize};
 
 pub mod configuration;
 pub mod defaults;
@@ -98,7 +99,8 @@ pub struct Settings {
     pub dummy_variable_rgx: Regex,
     pub external: FxHashSet<String>,
     pub ignore_init_module_imports: bool,
-    pub line_length: usize,
+    pub line_length: LineLength,
+    pub tab_size: TabSize,
     pub namespace_packages: Vec<PathBuf>,
     pub src: Vec<PathBuf>,
     pub task_tags: Vec<String>,
@@ -160,7 +162,8 @@ impl Settings {
                 config.include.unwrap_or_else(|| defaults::INCLUDE.clone()),
             )?,
             ignore_init_module_imports: config.ignore_init_module_imports.unwrap_or_default(),
-            line_length: config.line_length.unwrap_or(defaults::LINE_LENGTH),
+            line_length: config.line_length.unwrap_or_default(),
+            tab_size: config.tab_size.unwrap_or_default(),
             namespace_packages: config.namespace_packages.unwrap_or_default(),
             per_file_ignores: resolve_per_file_ignores(
                 config

crates/ruff/src/settings/options.rs

@@ -5,6 +5,7 @@ use serde::{Deserialize, Serialize};
 
 use ruff_macros::ConfigurationOptions;
 
+use crate::line_width::{LineLength, TabSize};
 use crate::rule_selector::RuleSelector;
 use crate::rules::{
     flake8_annotations, flake8_bandit, flake8_bugbear, flake8_builtins, flake8_comprehensions,
@@ -309,7 +310,16 @@ pub struct Options {
     )]
     /// The line length to use when enforcing long-lines violations (like
     /// `E501`).
-    pub line_length: Option<usize>,
+    pub line_length: Option<LineLength>,
+    #[option(
+        default = "4",
+        value_type = "int",
+        example = r#"
+            tab_size = 8
+        "#
+    )]
+    /// The tabulation size to calculate line length.
+    pub tab_size: Option<TabSize>,
     #[option(
         default = "None",
         value_type = "str",

crates/ruff/src/settings/pyproject.rs

@@ -151,6 +151,7 @@ mod tests {
     use rustc_hash::FxHashMap;
 
     use crate::codes::{self, RuleCodePrefix};
+    use crate::line_width::LineLength;
     use crate::rules::flake8_quotes::settings::Quote;
     use crate::rules::flake8_tidy_imports::banned_api::ApiBan;
     use crate::rules::flake8_tidy_imports::relative_imports::Strictness;
@@ -200,7 +201,7 @@ line-length = 79
             pyproject.tool,
             Some(Tools {
                 ruff: Some(Options {
-                    line_length: Some(79),
+                    line_length: Some(LineLength::from(79)),
                     ..Options::default()
                 })
             })
@@ -301,7 +302,7 @@ other-attribute = 1
             config,
             Options {
                 allowed_confusables: Some(vec!['−', 'ρ', '∗']),
-                line_length: Some(88),
+                line_length: Some(LineLength::from(88)),
                 extend_exclude: Some(vec![
                     "excluded_file.py".to_string(),
                     "migrations".to_string(),

crates/ruff_cli/src/args.rs

@@ -5,6 +5,7 @@ use clap::{command, Parser};
 use regex::Regex;
 use rustc_hash::FxHashMap;
 
+use ruff::line_width::LineLength;
 use ruff::logging::LogLevel;
 use ruff::registry::Rule;
 use ruff::resolver::ConfigProcessor;
@@ -545,7 +546,7 @@ impl ConfigProcessor for &Overrides {
             config.force_exclude = Some(*force_exclude);
         }
         if let Some(line_length) = &self.line_length {
-            config.line_length = Some(*line_length);
+            config.line_length = Some(LineLength::from(*line_length));
         }
         if let Some(per_file_ignores) = &self.per_file_ignores {
             config.per_file_ignores = Some(collect_per_file_ignores(per_file_ignores.clone()));

crates/ruff_wasm/src/lib.rs

@@ -5,6 +5,7 @@ use serde::{Deserialize, Serialize};
 use wasm_bindgen::prelude::*;
 
 use ruff::directives;
+use ruff::line_width::{LineLength, TabSize};
 use ruff::linter::{check_path, LinterResult};
 use ruff::registry::AsRule;
 use ruff::rules::{
@@ -102,7 +103,8 @@ pub fn defaultSettings() -> Result<JsValue, JsValue> {
         extend_unfixable: Some(Vec::default()),
         external: Some(Vec::default()),
         ignore: Some(Vec::default()),
-        line_length: Some(defaults::LINE_LENGTH),
+        line_length: Some(LineLength::default()),
+        tab_size: Some(TabSize::default()),
         select: Some(defaults::PREFIXES.to_vec()),
         target_version: Some(defaults::TARGET_VERSION),
         // Ignore a bunch of options that don't make sense in a single-file editor.

ruff.schema.json

@@ -366,12 +366,14 @@
     },
     "line-length": {
       "description": "The line length to use when enforcing long-lines violations (like `E501`).",
-      "type": [
+      "anyOf": [
-        "integer",
+        {
-        "null"
+          "$ref": "#/definitions/LineLength"
-      ],
+        },
-      "format": "uint",
+        {
-      "minimum": 0.0
+          "type": "null"
+        }
+      ]
     },
     "mccabe": {
       "description": "Options for the `mccabe` plugin.",
@@ -503,6 +505,17 @@
         "type": "string"
       }
     },
+    "tab-size": {
+      "description": "The tabulation size to calculate line length.",
+      "anyOf": [
+        {
+          "$ref": "#/definitions/TabSize"
+        },
+        {
+          "type": "null"
+        }
+      ]
+    },
     "target-version": {
       "description": "The Python version to target, e.g., when considering automatic code upgrades, like rewriting type annotations.\n\nIf omitted, the target version will be inferred from the `project.requires-python` field in the relevant `pyproject.toml` (e.g., `requires-python = \">=3.8\"`), if present.",
       "anyOf": [
@@ -1276,6 +1289,12 @@
       },
       "additionalProperties": false
     },
+    "LineLength": {
+      "description": "The length of a line of text that is considered too long.",
+      "type": "integer",
+      "format": "uint",
+      "minimum": 0.0
+    },
     "McCabeOptions": {
       "type": "object",
       "properties": {
@@ -1361,12 +1380,14 @@
         },
         "max-doc-length": {
           "description": "The maximum line length to allow for line-length violations within documentation (`W505`), including standalone comments.",
-          "type": [
+          "anyOf": [
-            "integer",
+            {
-            "null"
+              "$ref": "#/definitions/LineLength"
-          ],
+            },
-          "format": "uint",
+            {
-          "minimum": 0.0
+              "type": "null"
+            }
+          ]
         }
       },
       "additionalProperties": false
@@ -2489,6 +2510,12 @@
         }
       ]
     },
+    "TabSize": {
+      "description": "The size of a tab.",
+      "type": "integer",
+      "format": "uint8",
+      "minimum": 0.0
+    },
     "Version": {
       "type": "string"
     }