[grazie] IJPL-248141 Unpaired symbol ) false positive

Merge-request: IJ-MR-216423
Merged-by: Ilia Permiashkin <ilia.permiashkin@jetbrains.com>

GitOrigin-RevId: 34e3d027c18b657a3e4287a4fe265e166cc025a0

plugins/grazie/src/test/kotlin/com/intellij/grazie/text/TextExtractionTest.java

@@ -511,4 +511,9 @@ public class TextExtractionTest extends BasePlatformTestCase {
   public static TextContent extractText(String fileName, String fileText, int offset, Project project) {
     return TextExtractor.findTextAt(createFile(fileName, fileText, project), offset, TextContent.TextDomain.ALL);
   }
+
+  public static Set<TextContent> extractAllTexts(String fileName, String fileText, Project project) {
+    PsiFile file = createFile(fileName, fileText, project);
+    return TextExtractor.findAllTextContents(file.getViewProvider(), TextContent.TextDomain.ALL);
+  }
 }

plugins/kotlin/grazie/intellij.kotlin.grazie.tests/test/org/jetbrains/kotlin/idea/fir/grazie/KotlinGrazieSupportTest28.kt

@@ -5,16 +5,13 @@ import com.intellij.grazie.GrazieTestBase
 import com.intellij.grazie.jlanguage.Lang
 import com.intellij.grazie.text.TextContent
 import com.intellij.grazie.text.TextContentTest
+import com.intellij.grazie.text.TextExtractionTest
 import com.intellij.grazie.text.TextExtractor
 import com.intellij.testFramework.LightProjectDescriptor
 import org.jetbrains.kotlin.idea.test.KotlinWithJdkAndRuntimeLightProjectDescriptor
 
 class KotlinGrazieSupportTest28 : GrazieTestBase() {
-    
-
-    override fun getProjectDescriptor(): LightProjectDescriptor {
-        return KotlinWithJdkAndRuntimeLightProjectDescriptor.getInstance()
-    }
+    override fun getProjectDescriptor(): LightProjectDescriptor = KotlinWithJdkAndRuntimeLightProjectDescriptor.getInstance()
 
     override val additionalEnabledRules: Set<String> = setOf("UPPERCASE_SENTENCE_START")
 
@@ -88,4 +85,59 @@ class KotlinGrazieSupportTest28 : GrazieTestBase() {
         )
         myFixture.checkHighlighting()
     }
+
+    fun `test code-like fragments are not extracted`() {
+        val texts = TextExtractionTest.extractAllTexts("a.kt", $$"""
+            /**
+             * The Markdown lexer folds a nested list item's leading indentation into its `LIST_BULLET`/`LIST_NUMBER` token
+             * (e.g. `"  - "`), so a list / list-item block would otherwise start inside the line's indentation. Trim that
+             * leading whitespace here so the block starts at its real content; otherwise offset-based consumers such as
+             * indent auto-detection (`FormatterBasedLineIndentInfoBuilder`) undercount the indent of nested list lines.
+             *
+             * With code fence:
+             * ```
+             * fun String.helloWorld() {
+             *     println("Hello World, $this")
+             * }
+             * ```
+             * With indentation: 
+             *
+             *      fun String.helloWorld() {
+             *         println("Hello World, $this")
+             *      }
+             *
+             * With tilde:
+             * ~~~
+             * fun String.helloWorld() {
+             *     println("Hello World, $this")
+             * }
+             * ~~~
+             *
+             * With backticks:
+             * `fun main() { println("Hello, Kotlin") }`
+             */
+            fun main() {}
+        """.trimIndent(), project)
+        assertEquals(1, texts.size)
+        assertEquals(texts.first().toString(), """
+            The Markdown lexer folds a nested list item's leading indentation into its / token
+            (e.g. ), so a list / list-item block would otherwise start inside the line's indentation. Trim that
+            leading whitespace here so the block starts at its real content; otherwise offset-based consumers such as
+            indent auto-detection () undercount the indent of nested list lines.
+
+            With code fence:
+
+            With indentation:
+
+
+
+
+
+            With tilde:
+
+
+            With backticks:
+
+        """.trimIndent())
+    }
 }

plugins/kotlin/grazie/src/org/jetbrains/kotlin/idea/grazie/KotlinTextExtractor.kt

@@ -28,13 +28,15 @@ import org.jetbrains.kotlin.psi.psiUtil.isSingleQuoted
 import java.util.regex.Pattern
 
 internal class KotlinTextExtractor : TextExtractor() {
-  private val kdocBuilder = TextContentBuilder.FromPsi
-    .withUnknown { e -> e.elementType == KDocTokens.MARKDOWN_LINK && e.text.startsWith("[") }
-    .excluding { e -> e.elementType == KDocTokens.MARKDOWN_LINK && !e.text.startsWith("[") }
-    .excluding { e -> val elementType = e.elementType
-        elementType == LEADING_ASTERISK || elementType == CODE_BLOCK_TEXT || elementType == CODE_SPAN_TEXT
-    }
-    .removingIndents(" \t").removingLineSuffixes(" \t")
+    private val kdocBuilder = TextContentBuilder.FromPsi
+        .withUnknown { e -> e.elementType == KDocTokens.MARKDOWN_LINK && e.text.startsWith("[") }
+        .excluding { e -> e.elementType == KDocTokens.MARKDOWN_LINK && !e.text.startsWith("[") }
+        .excluding { e ->
+            val elementType = e.elementType
+            elementType == LEADING_ASTERISK || elementType == CODE_BLOCK_TEXT
+        }
+        .withUnknown { e -> e.elementType == CODE_SPAN_TEXT }
+        .removingIndents(" \t").removingLineSuffixes(" \t")
 
   public override fun buildTextContents(root: PsiElement, allowedDomains: Set<TextContent.TextDomain>): List<TextContent> {
       if (InjectedLanguageManager.getInstance(root.project).shouldInspectionsBeLenient(root)) {
@@ -76,7 +78,7 @@ internal class KotlinTextExtractor : TextExtractor() {
     return null
   }
 
-    private val codeFragments = Pattern.compile("(?s)```.+?```|`.+?`")
+    private val codeFragments = Pattern.compile("(?s)```.+?```|~~~.+?~~~|``")
     private val markdownHeading = Pattern.compile("^[ \\t]*#{1,6}[ \\t]+[^\\n]*?(\\n|$)", Pattern.MULTILINE)
 
     private fun TextContent.removeCode(): TextContent? =

plugins/kotlin/idea/tests/testData/grazie/Docs.kt

@@ -66,6 +66,15 @@ class ForMultiLanguageSupport {
     // Das <TYPO descr="Typo: In word 'daert'">daert</TYPO> geschätzt fünf <STYLE_SUGGESTION descr="MANNSTUNDE">Mannstunden</STYLE_SUGGESTION>.
 }
 
+/**
+ * The Markdown lexer folds a nested list item's leading indentation into its `LIST_BULLET`/`LIST_NUMBER` token
+ * (e.g. `"  - "`), so a list / list-item block would otherwise start inside the line's indentation. Trim that
+ * leading whitespace here so the block starts at its real content; otherwise offset-based consumers such as
+ * indent auto-detection (`ForMultiLanguageSupport`) undercount the indent of nested list lines.
+ */
+fun ff() {}
+
+
 /**
  * Returns `an true` if expression is part of when condition expression that looks like
  * ```