Move `utf8-encoding-declaration` to token-based rules (#6110)

Closes #5979.

crates/ruff/resources/test/fixtures/pyupgrade/UP009_5.py

@@ -0,0 +1,4 @@
+print('Hello world')
+
+#!/usr/bin/python
+# -*- coding: utf-8 -*-

crates/ruff/src/checkers/physical_lines.rs

@@ -13,7 +13,6 @@ use crate::rules::pycodestyle::rules::{
     tab_indentation, trailing_whitespace,
 };
 use crate::rules::pylint;
-use crate::rules::pyupgrade::rules::unnecessary_coding_comment;
 use crate::settings::Settings;
 
 pub(crate) fn check_physical_lines(
@@ -28,7 +27,6 @@ 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_unnecessary_coding_comment = settings.rules.enabled(Rule::UTF8EncodingDeclaration);
     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);
@@ -37,27 +35,9 @@ pub(crate) fn check_physical_lines(
     let enforce_tab_indentation = settings.rules.enabled(Rule::TabIndentation);
     let enforce_copyright_notice = settings.rules.enabled(Rule::MissingCopyrightNotice);
 
-    let fix_unnecessary_coding_comment = settings.rules.should_fix(Rule::UTF8EncodingDeclaration);
-
-    let mut commented_lines_iter = indexer.comment_ranges().iter().peekable();
     let mut doc_lines_iter = doc_lines.iter().peekable();
 
-    for (index, line) in locator.contents().universal_newlines().enumerate() {
-        while commented_lines_iter
-            .next_if(|comment_range| line.range().contains_range(**comment_range))
-            .is_some()
-        {
-            if enforce_unnecessary_coding_comment {
-                if index < 2 {
-                    if let Some(diagnostic) =
-                        unnecessary_coding_comment(&line, fix_unnecessary_coding_comment)
-                    {
-                        diagnostics.push(diagnostic);
-                    }
-                }
-            }
-        }
-
+    for line in locator.contents().universal_newlines() {
         while doc_lines_iter
             .next_if(|doc_line_start| line.range().contains_inclusive(**doc_line_start))
             .is_some()

crates/ruff/src/checkers/tokens.rs

@@ -69,6 +69,10 @@ pub(crate) fn check_tokens(
         eradicate::rules::commented_out_code(&mut diagnostics, locator, indexer, settings);
     }
 
+    if settings.rules.enabled(Rule::UTF8EncodingDeclaration) {
+        pyupgrade::rules::unnecessary_coding_comment(&mut diagnostics, locator, indexer, settings);
+    }
+
     if settings.rules.enabled(Rule::InvalidEscapeSequence) {
         for (tok, range) in tokens.iter().flatten() {
             if tok.is_string() {

crates/ruff/src/registry.rs

@@ -246,8 +246,7 @@ impl Rule {
             | Rule::MissingNewlineAtEndOfFile
             | Rule::MixedSpacesAndTabs
             | Rule::TabIndentation
-            | Rule::TrailingWhitespace
-            | Rule::UTF8EncodingDeclaration => LintSource::PhysicalLines,
+            | Rule::TrailingWhitespace => LintSource::PhysicalLines,
             Rule::AmbiguousUnicodeCharacterComment
             | Rule::AmbiguousUnicodeCharacterDocstring
             | Rule::AmbiguousUnicodeCharacterString
@@ -289,7 +288,8 @@ impl Rule {
             | Rule::SingleLineImplicitStringConcatenation
             | Rule::TrailingCommaOnBareTuple
             | Rule::TypeCommentInStub
-            | Rule::UselessSemicolon => LintSource::Tokens,
+            | Rule::UselessSemicolon
+            | Rule::UTF8EncodingDeclaration => LintSource::Tokens,
             Rule::IOError => LintSource::Io,
             Rule::UnsortedImports | Rule::MissingRequiredImport => LintSource::Imports,
             Rule::ImplicitNamespacePackage | Rule::InvalidModuleName => LintSource::Filesystem,

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

@@ -68,6 +68,7 @@ mod tests {
     #[test_case(Rule::UTF8EncodingDeclaration, Path::new("UP009_2.py"))]
     #[test_case(Rule::UTF8EncodingDeclaration, Path::new("UP009_3.py"))]
     #[test_case(Rule::UTF8EncodingDeclaration, Path::new("UP009_4.py"))]
+    #[test_case(Rule::UTF8EncodingDeclaration, Path::new("UP009_5.py"))]
     #[test_case(Rule::UnicodeKindPrefix, Path::new("UP025.py"))]
     #[test_case(Rule::UnnecessaryBuiltinImport, Path::new("UP029.py"))]
     #[test_case(Rule::UnnecessaryClassParentheses, Path::new("UP039.py"))]

crates/ruff/src/rules/pyupgrade/rules/unnecessary_coding_comment.rs

@@ -3,7 +3,11 @@ use regex::Regex;
 
 use ruff_diagnostics::{AlwaysAutofixableViolation, Diagnostic, Edit, Fix};
 use ruff_macros::{derive_message_formats, violation};
-use ruff_source_file::Line;
+use ruff_python_index::Indexer;
+use ruff_source_file::Locator;
+
+use crate::registry::AsRule;
+use crate::settings::Settings;
 
 /// ## What it does
 /// Checks for unnecessary UTF-8 encoding declarations.
@@ -43,18 +47,31 @@ static CODING_COMMENT_REGEX: Lazy<Regex> =
     Lazy::new(|| Regex::new(r"^[ \t\f]*#.*?coding[:=][ \t]*utf-?8").unwrap());
 
 /// UP009
-pub(crate) fn unnecessary_coding_comment(line: &Line, autofix: bool) -> Option<Diagnostic> {
-    // PEP3120 makes utf-8 the default encoding.
-    if CODING_COMMENT_REGEX.is_match(line.as_str()) {
-        let mut diagnostic = Diagnostic::new(UTF8EncodingDeclaration, line.full_range());
-        if autofix {
-            diagnostic.set_fix(Fix::automatic(Edit::deletion(
-                line.start(),
-                line.full_end(),
-            )));
+pub(crate) fn unnecessary_coding_comment(
+    diagnostics: &mut Vec<Diagnostic>,
+    locator: &Locator,
+    indexer: &Indexer,
+    settings: &Settings,
+) {
+    // The coding comment must be on one of the first two lines. Since each comment spans at least
+    // one line, we only need to check the first two comments at most.
+    for range in indexer.comment_ranges().iter().take(2) {
+        let line = locator.slice(*range);
+        if CODING_COMMENT_REGEX.is_match(line) {
+            #[allow(deprecated)]
+            let line = locator.compute_line_index(range.start());
+            if line.to_zero_indexed() > 1 {
+                continue;
+            }
+
+            let mut diagnostic = Diagnostic::new(UTF8EncodingDeclaration, *range);
+            if settings.rules.should_fix(diagnostic.kind.rule()) {
+                diagnostic.set_fix(Fix::automatic(Edit::deletion(
+                    range.start(),
+                    locator.full_line_end(range.end()),
+                )));
+            }
+            diagnostics.push(diagnostic);
         }
-        Some(diagnostic)
-    } else {
-        None
     }
 }

crates/ruff/src/rules/pyupgrade/snapshots/ruff__rules__pyupgrade__tests__UP009_0.py.snap

@@ -3,10 +3,10 @@ source: crates/ruff/src/rules/pyupgrade/mod.rs
 ---
 UP009_0.py:1:1: UP009 [*] UTF-8 encoding declaration is unnecessary
   |
-1 | / # coding=utf8
-2 | | 
-  | |_^ UP009
-3 |   print("Hello world")
+1 | # coding=utf8
+  | ^^^^^^^^^^^^^ UP009
+2 | 
+3 | print("Hello world")
   |
   = help: Remove unnecessary coding comment
 

crates/ruff/src/rules/pyupgrade/snapshots/ruff__rules__pyupgrade__tests__UP009_1.py.snap

@@ -3,11 +3,11 @@ source: crates/ruff/src/rules/pyupgrade/mod.rs
 ---
 UP009_1.py:2:1: UP009 [*] UTF-8 encoding declaration is unnecessary
   |
-1 |   #!/usr/bin/python
-2 | / # -*- coding: utf-8 -*-
-3 | | 
-  | |_^ UP009
-4 |   print('Hello world')
+1 | #!/usr/bin/python
+2 | # -*- coding: utf-8 -*-
+  | ^^^^^^^^^^^^^^^^^^^^^^^ UP009
+3 | 
+4 | print('Hello world')
   |
   = help: Remove unnecessary coding comment
 

crates/ruff/src/rules/pyupgrade/snapshots/ruff__rules__pyupgrade__tests__UP009_5.py.snap

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