fix(completions): multi-byte char sanitization

crates/pgls_completions/src/sanitization.rs

@@ -38,7 +38,9 @@ pub(crate) fn is_sanitized_token_with_quote(node_under_cursor_txt: &str) -> bool
     // Node under cursor text will be "REPLACED_TOKEN_WITH_QUOTE".
     // The SANITIZED_TOKEN_WITH_QUOTE does not have the leading ".
     // We need to omit it from the txt.
-    &node_under_cursor_txt[1..] == SANITIZED_TOKEN_WITH_QUOTE
+    node_under_cursor_txt
+        .strip_prefix('"')
+        .is_some_and(|it| it == SANITIZED_TOKEN_WITH_QUOTE)
 }
 
 impl<'larger, 'smaller> From<CompletionParams<'larger>> for SanitizedCompletionParams<'smaller>
@@ -47,6 +49,13 @@ where
 {
     fn from(mut params: CompletionParams<'larger>) -> Self {
         params.text = params.text.to_ascii_lowercase();
+
+        while !params.text.is_char_boundary(params.position.into())
+            && params.position < TextSize::new(params.text.len().try_into().unwrap())
+        {
+            params.position = params.position.checked_add(1.into()).unwrap();
+        }
+
         if cursor_inbetween_nodes(&params.text, params.position)
             || cursor_prepared_to_write_token_after_last_node(&params.text, params.position)
             || cursor_before_semicolon(params.tree, params.position)
@@ -275,7 +284,11 @@ fn cursor_between_parentheses(sql: &str, position: TextSize) -> bool {
         .unwrap_or_default();
 
     // (.. and |)
-    let after_and_keyword = &sql[position.saturating_sub(4)..position] == "and " && after == ')';
+    let and_check_position = position.saturating_sub(4);
+    let after_and_keyword = sql.is_char_boundary(and_check_position)
+        && &sql[and_check_position..position] == "and "
+        && after == ')';
+
     let after_eq_sign = before == '=' && after == ')';
 
     let head_of_list = before == '(' && after == ',';
@@ -312,7 +325,7 @@ mod tests {
     use pgls_text_size::TextSize;
 
     use crate::{
-        CompletionParams, SanitizedCompletionParams,
+        CompletionParams, SanitizedCompletionParams, is_sanitized_token_with_quote,
         sanitization::{
             cursor_after_opened_quote, cursor_before_semicolon, cursor_between_parentheses,
             cursor_inbetween_nodes, cursor_on_a_dot,
@@ -635,4 +648,49 @@ mod tests {
             TextSize::new(20)
         ));
     }
+
+    #[test]
+    fn multibyte_characters() {
+        is_sanitized_token_with_quote("é"); // should not panic
+
+        {
+            // cursor in the middle of multi-byte char
+            // select * from "auth"."é|é; <-- cursor in the middle of the é multi-byte char
+            let input = r#"select * from "auth"."é;"#;
+            let position = TextSize::new(23);
+
+            let params = get_test_params(input, position);
+
+            let sanitized = SanitizedCompletionParams::from(params);
+
+            assert_eq!(sanitized.text, r#"select * from "auth"."é;"#);
+            assert_eq!(sanitized.position, TextSize::new(24)); // adjusts the cursor position to end of multibyte char
+        }
+
+        {
+            // cursor in front of multibyte char
+            // select * from "auth"."|é; <-- cursor front of the é multi-byte char
+            let input = r#"select * from "auth"."é;"#;
+            let position = TextSize::new(22);
+
+            let params = get_test_params(input, position);
+
+            let sanitized = SanitizedCompletionParams::from(params);
+
+            assert_eq!(
+                sanitized.text,
+                r#"select * from "auth"."REPLACED_TOKEN_WITH_QUOTE"é; "#
+            );
+            assert_eq!(sanitized.position, TextSize::new(22)); // leaves cursor position as is
+        }
+
+        {
+            let input = "insert into instruments (name, id, €,  )";
+            let position = TextSize::new(36);
+
+            let params = get_test_params(input, position);
+
+            let _ = SanitizedCompletionParams::from(params); // should not panic
+        }
+    }
 }