[pycodestyle] Implement blank-line-at-end-of-file (W391)

.gitattributes

@@ -2,6 +2,8 @@
 
 crates/ruff_linter/resources/test/fixtures/isort/line_ending_crlf.py text eol=crlf
 crates/ruff_linter/resources/test/fixtures/pycodestyle/W605_1.py text eol=crlf
+crates/ruff_linter/resources/test/fixtures/pycodestyle/W391_2.py text eol=crlf
+crates/ruff_linter/resources/test/fixtures/pycodestyle/W391_3.py text eol=crlf
 
 crates/ruff_python_formatter/resources/test/fixtures/ruff/docstring_code_examples_crlf.py text eol=crlf
 crates/ruff_python_formatter/tests/snapshots/format@docstring_code_examples_crlf.py.snap text eol=crlf

crates/ruff_linter/resources/test/fixtures/pycodestyle/W391_0.py

@@ -0,0 +1,16 @@
+# Unix style
+def foo() -> None:
+    pass
+
+
+def bar() -> None:
+    pass
+
+
+
+if __name__ == '__main__':
+    foo()
+    bar()
+
+
+

crates/ruff_linter/resources/test/fixtures/pycodestyle/W391_1.py

@@ -0,0 +1,13 @@
+# Unix style
+def foo() -> None:
+    pass
+
+
+def bar() -> None:
+    pass
+
+
+
+if __name__ == '__main__':
+    foo()
+    bar()

crates/ruff_linter/resources/test/fixtures/pycodestyle/W391_2.py

@@ -0,0 +1,17 @@
+# Windows style
+def foo() -> None:
+    pass
+
+
+def bar() -> None:
+    pass
+
+
+
+if __name__ == '__main__':
+    foo()
+    bar()
+
+
+
+

crates/ruff_linter/resources/test/fixtures/pycodestyle/W391_3.py

@@ -0,0 +1,13 @@
+# Windows style
+def foo() -> None:
+    pass
+
+
+def bar() -> None:
+    pass
+
+
+
+if __name__ == '__main__':
+    foo()
+    bar()

crates/ruff_linter/src/checkers/physical_lines.rs

@@ -10,7 +10,7 @@ use crate::registry::Rule;
 use crate::rules::flake8_copyright::rules::missing_copyright_notice;
 use crate::rules::pycodestyle::rules::{
     doc_line_too_long, line_too_long, mixed_spaces_and_tabs, no_newline_at_end_of_file,
-    trailing_whitespace,
+    too_many_newlines_at_end_of_file, trailing_whitespace,
 };
 use crate::rules::pylint;
 use crate::settings::LinterSettings;
@@ -27,6 +27,8 @@ pub(crate) fn check_physical_lines(
     let enforce_doc_line_too_long = settings.rules.enabled(Rule::DocLineTooLong);
     let enforce_line_too_long = settings.rules.enabled(Rule::LineTooLong);
     let enforce_no_newline_at_end_of_file = settings.rules.enabled(Rule::MissingNewlineAtEndOfFile);
+    let enforce_too_many_newlines_at_end_of_file =
+        settings.rules.enabled(Rule::TooManyNewlinesAtEndOfFile);
     let enforce_mixed_spaces_and_tabs = settings.rules.enabled(Rule::MixedSpacesAndTabs);
     let enforce_bidirectional_unicode = settings.rules.enabled(Rule::BidirectionalUnicode);
     let enforce_trailing_whitespace = settings.rules.enabled(Rule::TrailingWhitespace);
@@ -77,6 +79,12 @@ pub(crate) fn check_physical_lines(
         }
     }
 
+    if enforce_too_many_newlines_at_end_of_file {
+        if let Some(diagnostic) = too_many_newlines_at_end_of_file(locator) {
+            diagnostics.push(diagnostic);
+        }
+    }
+
     if enforce_copyright_notice {
         if let Some(diagnostic) = missing_copyright_notice(locator, settings) {
             diagnostics.push(diagnostic);

crates/ruff_linter/src/codes.rs

@@ -167,6 +167,7 @@ pub fn code_to_rule(linter: Linter, code: &str) -> Option<(RuleGroup, Rule)> {
         (Pycodestyle, "W291") => (RuleGroup::Stable, rules::pycodestyle::rules::TrailingWhitespace),
         (Pycodestyle, "W292") => (RuleGroup::Stable, rules::pycodestyle::rules::MissingNewlineAtEndOfFile),
         (Pycodestyle, "W293") => (RuleGroup::Stable, rules::pycodestyle::rules::BlankLineWithWhitespace),
+        (Pycodestyle, "W391") => (RuleGroup::Preview, rules::pycodestyle::rules::TooManyNewlinesAtEndOfFile),
         (Pycodestyle, "W505") => (RuleGroup::Stable, rules::pycodestyle::rules::DocLineTooLong),
         (Pycodestyle, "W605") => (RuleGroup::Stable, rules::pycodestyle::rules::InvalidEscapeSequence),
 

crates/ruff_linter/src/registry.rs

@@ -254,6 +254,7 @@ impl Rule {
             | Rule::MissingCopyrightNotice
             | Rule::MissingNewlineAtEndOfFile
             | Rule::MixedSpacesAndTabs
+            | Rule::TooManyNewlinesAtEndOfFile
             | Rule::TrailingWhitespace => LintSource::PhysicalLines,
             Rule::AmbiguousUnicodeCharacterComment
             | Rule::AvoidableEscapedQuote

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

@@ -71,6 +71,10 @@ mod tests {
     #[test_case(Rule::IsLiteral, Path::new("constant_literals.py"))]
     #[test_case(Rule::TypeComparison, Path::new("E721.py"))]
     #[test_case(Rule::ModuleImportNotAtTopOfFile, Path::new("E402_2.py"))]
+    #[test_case(Rule::TooManyNewlinesAtEndOfFile, Path::new("W391_0.py"))]
+    #[test_case(Rule::TooManyNewlinesAtEndOfFile, Path::new("W391_1.py"))]
+    #[test_case(Rule::TooManyNewlinesAtEndOfFile, Path::new("W391_2.py"))]
+    #[test_case(Rule::TooManyNewlinesAtEndOfFile, Path::new("W391_3.py"))]
     fn preview_rules(rule_code: Rule, path: &Path) -> Result<()> {
         let snapshot = format!(
             "preview__{}_{}",

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

@@ -17,6 +17,7 @@ pub(crate) use module_import_not_at_top_of_file::*;
 pub(crate) use multiple_imports_on_one_line::*;
 pub(crate) use not_tests::*;
 pub(crate) use tab_indentation::*;
+pub(crate) use too_many_newlines_at_end_of_file::*;
 pub(crate) use trailing_whitespace::*;
 pub(crate) use type_comparison::*;
 
@@ -39,5 +40,6 @@ mod module_import_not_at_top_of_file;
 mod multiple_imports_on_one_line;
 mod not_tests;
 mod tab_indentation;
+mod too_many_newlines_at_end_of_file;
 mod trailing_whitespace;
 mod type_comparison;

crates/ruff_linter/src/rules/pycodestyle/rules/too_many_newlines_at_end_of_file.rs

@@ -0,0 +1,63 @@
+use regex::Regex;
+
+use ruff_text_size::{TextRange, TextSize};
+
+use ruff_diagnostics::{AlwaysFixableViolation, Diagnostic, Edit, Fix};
+use ruff_macros::{derive_message_formats, violation};
+use ruff_source_file::Locator;
+
+/// ## What it does
+/// Checks for files with too many new lines at the end of the file.
+///
+/// ## Why is this bad?
+/// Trailing blank lines are superfluous.
+/// However the last line should end with a new line.
+///
+/// ## Example
+/// ```python
+/// spam(1)\n\n\n
+/// ```
+///
+/// Use instead:
+/// ```python
+/// spam(1)\n
+/// ```
+#[violation]
+pub struct TooManyNewlinesAtEndOfFile;
+
+impl AlwaysFixableViolation for TooManyNewlinesAtEndOfFile {
+    #[derive_message_formats]
+    fn message(&self) -> String {
+        format!("Too many newlines at the end of file")
+    }
+
+    fn fix_title(&self) -> String {
+        "Remove extraneous trailing newlines".to_string()
+    }
+}
+
+/// W391
+pub(crate) fn too_many_newlines_at_end_of_file(locator: &Locator) -> Option<Diagnostic> {
+    let source = locator.contents();
+
+    // Ignore empty and BOM only files
+    if source.is_empty() || source == "\u{feff}" {
+        return None;
+    }
+
+    // Regex to match multiple newline characters at the end of the file
+    let newline_regex = Regex::new(r"(\r\n){2,}$|\n{2,}$|\r{2,}$").unwrap();
+
+    if let Some(mat) = newline_regex.find(source) {
+        let start_pos = TextSize::new((mat.start() + 1).try_into().unwrap());
+        let end_pos = TextSize::new(mat.end().try_into().unwrap());
+
+        // Calculate the TextRange to keep only one newline at the end
+        let range = TextRange::new(start_pos, end_pos);
+        let mut diagnostic = Diagnostic::new(TooManyNewlinesAtEndOfFile, range);
+        diagnostic.set_fix(Fix::safe_edit(Edit::deletion(start_pos, end_pos)));
+        return Some(diagnostic);
+    }
+
+    None
+}

crates/ruff_linter/src/rules/pycodestyle/snapshots/ruff_linter__rules__pycodestyle__tests__preview__W391_W391_0.py.snap

@@ -0,0 +1,20 @@
+---
+source: crates/ruff_linter/src/rules/pycodestyle/mod.rs
+---
+W391_0.py:14:1: W391 [*] Too many newlines at the end of file
+   |
+12 |       foo()
+13 |       bar()
+14 | / 
+15 | | 
+16 | | 
+   |
+   = help: Remove extraneous trailing newlines
+
+ℹ Safe fix
+11 11 | if __name__ == '__main__':
+12 12 |     foo()
+13 13 |     bar()
+14    |-
+15    |-
+16    |-

crates/ruff_linter/src/rules/pycodestyle/snapshots/ruff_linter__rules__pycodestyle__tests__preview__W391_W391_1.py.snap

@@ -0,0 +1,4 @@
+---
+source: crates/ruff_linter/src/rules/pycodestyle/mod.rs
+---
+

crates/ruff_linter/src/rules/pycodestyle/snapshots/ruff_linter__rules__pycodestyle__tests__preview__W391_W391_2.py.snap

@@ -0,0 +1,25 @@
+---
+source: crates/ruff_linter/src/rules/pycodestyle/mod.rs
+---
+W391_2.py:13:11: W391 [*] Too many newlines at the end of file
+   |
+11 |   if __name__ == '__main__':
+12 |       foo()
+13 |       bar()
+14 | | 
+15 | | 
+16 | | 
+17 | | 
+   |
+   = help: Remove extraneous trailing newlines
+
+ℹ Safe fix
+10 10 | 
+11 11 | if __name__ == '__main__':
+12 12 |     foo()
+13    |-    bar()
+14    |-
+15    |-
+16    |-
+17    |-
+   13 |+    bar()

crates/ruff_linter/src/rules/pycodestyle/snapshots/ruff_linter__rules__pycodestyle__tests__preview__W391_W391_3.py.snap

@@ -0,0 +1,4 @@
+---
+source: crates/ruff_linter/src/rules/pycodestyle/mod.rs
+---
+

scripts/check_docs_formatted.py

@@ -104,6 +104,7 @@
     "tab-after-operator",
     "tab-before-keyword",
     "tab-before-operator",
+    "too-many-newlines-at-end-of-file",
     "trailing-whitespace",
     "unexpected-indentation",
 ]