Fix for : Remove blank lines instead of emptying it's contents (#687)

Fixes the PR #667 for feature request #664 --------- Co-authored-by: Bennet Sunder <[email protected]>
uber · Sep 2, 2024 · 7323a98 · 7323a98
1 parent c67dc55
commit 7323a98
Show file tree

Hide file tree

Showing 3 changed files with 208 additions and 3 deletions.
diff --git a/src/models/concrete_syntax.rs b/src/models/concrete_syntax.rs
@@ -65,13 +65,13 @@ pub(crate) fn get_all_matches_for_concrete_syntax(
     };
 
     match_map.insert(replace_node_key, replace_node_match.clone());
-
     matches.push(Match {
       matched_string: replace_node_match.text,
       range: replace_node_match.range,
       matches: match_map.into_iter().map(|(k, v)| (k, v.text)).collect(),
       associated_comma: None,
       associated_comments: Vec::new(),
+      associated_leading_empty_lines: Vec::new(),
     });
   }
   if recursive {

diff --git a/src/models/matches.rs b/src/models/matches.rs
@@ -51,6 +51,10 @@ pub struct Match {
   #[get_mut]
   #[serde(skip)]
   pub(crate) associated_comments: Vec<Range>,
+  #[get]
+  #[get_mut]
+  #[serde(skip)]
+  pub(crate) associated_leading_empty_lines: Vec<Range>,
 }
 
 #[pymethods]
@@ -74,6 +78,7 @@ impl Match {
       matches,
       associated_comma: None,
       associated_comments: Vec::new(),
+      associated_leading_empty_lines: Vec::new(),
     }
   }
 
@@ -85,6 +90,7 @@ impl Match {
     let associated_ranges = [
       self.associated_comma().iter().collect_vec(),
       self.associated_comments().iter().collect_vec(),
+      self.associated_leading_empty_lines().iter().collect_vec(),
     ]
     .concat()
     .iter()
@@ -117,8 +123,53 @@ impl Match {
   ) {
     self.get_associated_elements(node, code, piranha_arguments, true);
     self.get_associated_elements(node, code, piranha_arguments, false);
+    self.get_associated_leading_empty_lines(node, code);
   }
 
+  fn get_associated_leading_empty_lines(&mut self, matched_node: &Node, code: &String) {
+    if let Some(empty_range) = self.find_empty_line_range(code, matched_node) {
+      let skipped_range = Range {
+        start_byte: empty_range.start,
+        end_byte: empty_range.end,
+        start_point: position_for_offset(code.as_bytes(), empty_range.start),
+        end_point: position_for_offset(code.as_bytes(), empty_range.end),
+      };
+      self
+        .associated_leading_empty_lines_mut()
+        .push(skipped_range);
+    }
+  }
+
+  fn find_empty_line_range(&self, code: &str, node: &Node) -> Option<std::ops::Range<usize>> {
+    let mut end_byte = node.start_byte();
+    let code_bytes = code.as_bytes();
+    let mut found = false;
+    while end_byte > 0 {
+      let prev_char = code_bytes[end_byte - 1] as char;
+      if prev_char.is_whitespace() {
+        end_byte -= 1;
+        if prev_char == '\n' {
+          found = true;
+          break;
+        }
+      } else {
+        break;
+      }
+    }
+
+    if found {
+      // Get the previous sibling node's end byte if it exists, in order to prevent overdeletions
+      let prev_sibling_end_byte = if let Some(prev_sibling) = node.prev_sibling() {
+        prev_sibling.end_byte()
+      } else {
+        0
+      };
+      let max_end_byte = std::cmp::max(end_byte, prev_sibling_end_byte);
+      Some(max_end_byte..node.start_byte())
+    } else {
+      None
+    }
+  }
   fn found_comma(&self) -> bool {
     self.associated_comma().is_some()
   }
@@ -143,7 +194,6 @@ impl Match {
         current_node.prev_sibling()
       } {
         // Check if the sibling is a comma
-
         if !self.found_comma() && self.is_comma_safe_to_delete(&sibling, code, trailing) {
           // Add the comma to the associated matches
           self.associated_comma = Some(sibling.range().into());

diff --git a/src/models/unit_tests/source_code_unit_test.rs b/src/models/unit_tests/source_code_unit_test.rs
@@ -16,7 +16,7 @@ use tree_sitter::Parser;
 use crate::{
   filter,
   models::{
-    default_configs::{JAVA, UNUSED_CODE_PATH},
+    default_configs::{JAVA, RUBY, SWIFT, UNUSED_CODE_PATH},
     filter::Filter,
     language::PiranhaLanguage,
     matches::{Point, Range},
@@ -699,3 +699,158 @@ fn test_satisfies_outermost_enclosing_node() {
 
   assert!(source_code_unit.is_satisfied(*node, &rule_positive, &HashMap::new(), &mut rule_store,));
 }
+
+#[test]
+fn test_removes_blank_lines_after_inline_cleanup() {
+  let inline_cleanup_rule = piranha_rule! {
+    name= "inline_cleanup_rule",
+    query= "
+    (
+      (if_modifier
+          body : ((_) @body)
+          [
+            condition: (false)
+            condition: (parenthesized_statements (false))
+          ]
+      )@if_modifier
+    )          
+    ",
+    replace_node = "if_modifier",
+    replace = ""
+  };
+
+  let inline_rule = InstantiatedRule::new(&inline_cleanup_rule, &HashMap::new());
+
+  let source_code = r#"
+      def method_name
+        do_something if false
+      end
+  "#
+  .trim();
+
+  let piranha_arguments = PiranhaArgumentsBuilder::default()
+    .paths_to_codebase(vec![UNUSED_CODE_PATH.to_string()])
+    .language(PiranhaLanguage::from(RUBY))
+    .build();
+
+  let mut rule_store = RuleStore::new(&piranha_arguments);
+  let mut parser = piranha_arguments.language().parser();
+
+  let source_code_unit = SourceCodeUnit::new(
+    &mut parser,
+    source_code.to_string(),
+    &HashMap::new(),
+    PathBuf::new().as_path(),
+    &piranha_arguments,
+  );
+  let matches = source_code_unit.get_matches(
+    &inline_rule,
+    &mut rule_store,
+    source_code_unit.root_node(),
+    true,
+  );
+  assert_eq!(matches.len(), 1);
+  assert_eq!(
+    matches
+      .first()
+      .unwrap()
+      .associated_leading_empty_lines
+      .len(),
+    1
+  );
+  assert_eq!(
+    *matches
+      .first()
+      .unwrap()
+      .associated_leading_empty_lines
+      .first()
+      .unwrap(),
+    Range {
+      start_byte: 15,
+      end_byte: 24,
+      start_point: Point { row: 0, column: 15 },
+      end_point: Point { row: 1, column: 8 }
+    }
+  );
+}
+
+#[test]
+fn test_switch_entry_blank_lines() {
+  let inline_cleanup_rule = piranha_rule! {
+    name= "inline_cleanup_rule",
+    query= "
+    (
+        (switch_entry
+            (switch_pattern) @p1
+            (switch_pattern) @p2
+            (switch_pattern) @p3
+            (switch_pattern) @p4
+            (switch_pattern) @p5
+            (switch_pattern) @p6
+        ) @custom_entry
+        (#eq? @p1 \".case_zeta_first\")
+    )
+    ",
+    replace_node = "custom_entry",
+    replace = ""
+  };
+
+  let inline_rule = InstantiatedRule::new(&inline_cleanup_rule, &HashMap::new());
+
+  let source_code = r#"
+    public var namespace: ParameterNamespace {
+        switch self {
+        case .case_random_word_alpha,
+             .case_random_word_beta,
+             .case_random_word_gamma,
+             .case_random_word_delta:
+            return .namespace_group_alpha
+        case .case_zeta_first,
+             .case_zeta_second,
+             .case_zeta_third,
+             .case_zeta_fourth,
+             .case_zeta_fifth,
+             .case_zeta_sixth:
+            return .namespace_group_beta
+        case .case_random_word_omega:
+            return .namespace_group_gamma
+        }
+    }
+  "#
+  .trim();
+
+  let piranha_arguments = PiranhaArgumentsBuilder::default()
+    .paths_to_codebase(vec![UNUSED_CODE_PATH.to_string()])
+    .language(PiranhaLanguage::from(SWIFT))
+    .build();
+
+  let mut rule_store = RuleStore::new(&piranha_arguments);
+  let mut parser = piranha_arguments.language().parser();
+
+  let mut source_code_unit = SourceCodeUnit::new(
+    &mut parser,
+    source_code.to_string(),
+    &HashMap::new(),
+    PathBuf::new().as_path(),
+    &piranha_arguments,
+  );
+
+  source_code_unit.apply_rule(inline_rule, &mut rule_store, &mut parser, &None);
+  let transformed_code = source_code_unit.code();
+
+  let expected_code = r#"
+    public var namespace: ParameterNamespace {
+        switch self {
+        case .case_random_word_alpha,
+             .case_random_word_beta,
+             .case_random_word_gamma,
+             .case_random_word_delta:
+            return .namespace_group_alpha
+        case .case_random_word_omega:
+            return .namespace_group_gamma
+        }
+    }
+  "#
+  .trim();
+  assert_eq!(transformed_code, expected_code);
+}