Fix uncessary-coding-comment fix when there's leading content (#6775)

Closes #6756

Including whitespace, code, and continuations.

crates/ruff/resources/test/fixtures/pyupgrade/UP009_10.py

@@ -0,0 +1,7 @@
+"""
+# coding=utf8""" # empty comment
+
+"""
+Invalid coding declaration since it is nested inside a docstring
+The following empty comment tests for false positives as our implementation visits comments
+"""

crates/ruff/resources/test/fixtures/pyupgrade/UP009_6.py

@@ -0,0 +1,7 @@
+ # coding=utf8
+print("Hello world")
+
+"""
+Regression test for https://github.com/astral-sh/ruff/issues/6756
+The leading space must be removed to prevent invalid syntax.
+"""

crates/ruff/resources/test/fixtures/pyupgrade/UP009_7.py

@@ -0,0 +1,7 @@
+	# coding=utf8
+print("Hello world")
+
+"""
+Regression test for https://github.com/astral-sh/ruff/issues/6756
+The leading tab must be removed to prevent invalid syntax.
+"""

crates/ruff/resources/test/fixtures/pyupgrade/UP009_8.py

@@ -0,0 +1,6 @@
+print("foo") # coding=utf8
+print("Hello world")
+
+"""
+Invalid coding declaration due to a statement before the comment
+"""

crates/ruff/resources/test/fixtures/pyupgrade/UP009_9.py

@@ -0,0 +1,7 @@
+x = 1 \
+  # coding=utf8
+x = 2
+
+"""
+Invalid coding declaration due to continuation on preceding line
+"""

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

@@ -67,6 +67,11 @@ mod tests {
     #[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::UTF8EncodingDeclaration, Path::new("UP009_6.py"))]
+    #[test_case(Rule::UTF8EncodingDeclaration, Path::new("UP009_7.py"))]
+    #[test_case(Rule::UTF8EncodingDeclaration, Path::new("UP009_8.py"))]
+    #[test_case(Rule::UTF8EncodingDeclaration, Path::new("UP009_9.py"))]
+    #[test_case(Rule::UTF8EncodingDeclaration, Path::new("UP009_10.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

@@ -5,6 +5,7 @@ use ruff_diagnostics::{AlwaysAutofixableViolation, Diagnostic, Edit, Fix};
 use ruff_macros::{derive_message_formats, violation};
 use ruff_python_index::Indexer;
 use ruff_source_file::Locator;
+use ruff_text_size::TextRange;
 
 use crate::registry::AsRule;
 use crate::settings::Settings;
@@ -55,20 +56,45 @@ pub(crate) fn unnecessary_coding_comment(
 ) {
     // 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) {
+    for comment_range in indexer.comment_ranges().iter().take(2) {
+        // If leading content is not whitespace then it's not a valid coding comment e.g.
+        // ```
+        // print(x) # coding=utf8
+        // ```
+        let line_range = locator.full_line_range(comment_range.start());
+        if !locator
+            .slice(TextRange::new(line_range.start(), comment_range.start()))
+            .trim()
+            .is_empty()
+        {
+            continue;
+        }
+
+        // If the line is after a continuation then it's not a valid coding comment e.g.
+        // ```
+        // x = 1 \
+        //    # coding=utf8
+        // x = 2
+        // ```
+        if indexer
+            .preceded_by_continuations(line_range.start(), locator)
+            .is_some()
+        {
+            continue;
+        }
+
+        if CODING_COMMENT_REGEX.is_match(locator.slice(line_range)) {
             #[allow(deprecated)]
-            let line = locator.compute_line_index(range.start());
-            if line.to_zero_indexed() > 1 {
+            let index = locator.compute_line_index(line_range.start());
+            if index.to_zero_indexed() > 1 {
                 continue;
             }
 
-            let mut diagnostic = Diagnostic::new(UTF8EncodingDeclaration, *range);
+            let mut diagnostic = Diagnostic::new(UTF8EncodingDeclaration, *comment_range);
             if settings.rules.should_fix(diagnostic.kind.rule()) {
                 diagnostic.set_fix(Fix::automatic(Edit::deletion(
-                    range.start(),
-                    locator.full_line_end(range.end()),
+                    line_range.start(),
+                    line_range.end(),
                 )));
             }
             diagnostics.push(diagnostic);

crates/ruff/src/rules/pyupgrade/snapshots/ruff__rules__pyupgrade__tests__UP009_10.py.snap

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

crates/ruff/src/rules/pyupgrade/snapshots/ruff__rules__pyupgrade__tests__UP009_6.py.snap

@@ -0,0 +1,18 @@
+---
+source: crates/ruff/src/rules/pyupgrade/mod.rs
+---
+UP009_6.py:1:2: UP009 [*] UTF-8 encoding declaration is unnecessary
+  |
+1 |  # coding=utf8
+  |  ^^^^^^^^^^^^^ UP009
+2 | print("Hello world")
+  |
+  = help: Remove unnecessary coding comment
+
+ℹ Fix
+1   |- # coding=utf8
+2 1 | print("Hello world")
+3 2 | 
+4 3 | """
+
+

crates/ruff/src/rules/pyupgrade/snapshots/ruff__rules__pyupgrade__tests__UP009_7.py.snap

@@ -0,0 +1,18 @@
+---
+source: crates/ruff/src/rules/pyupgrade/mod.rs
+---
+UP009_7.py:1:2: UP009 [*] UTF-8 encoding declaration is unnecessary
+  |
+1 |     # coding=utf8
+  |     ^^^^^^^^^^^^^ UP009
+2 | print("Hello world")
+  |
+  = help: Remove unnecessary coding comment
+
+ℹ Fix
+1   |-	# coding=utf8
+2 1 | print("Hello world")
+3 2 | 
+4 3 | """
+
+

crates/ruff/src/rules/pyupgrade/snapshots/ruff__rules__pyupgrade__tests__UP009_8.py.snap

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

crates/ruff/src/rules/pyupgrade/snapshots/ruff__rules__pyupgrade__tests__UP009_9.py.snap

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