fix(jdoc): inline mk structs break detection

GitOrigin-RevId: 0cb1adbd59262523ce59767fee5f33e78234bd9f

java/java-frontback-psi-impl/src/com/intellij/lang/java/parser/BasicJavaDocParser.java

@@ -388,7 +388,7 @@ public final class BasicJavaDocParser {
     while (!builder.eof()) {
       builder.advanceLexer();
       previousToken = token;
-      token = getTokenType(builder);
+      token = getTokenType(builder, false);
       if (token == needle) {
         return token;
       }
@@ -398,7 +398,7 @@ public final class BasicJavaDocParser {
       }
 
       // Markdown specific, check for EOL
-      if (token == JavaDocTokenType.DOC_SPACE && previousToken == JavaDocTokenType.DOC_SPACE) {
+      if (token == TokenType.WHITE_SPACE && previousToken == TokenType.WHITE_SPACE) {
         break;
       }
     }
@@ -605,10 +605,15 @@ public final class BasicJavaDocParser {
 
   @Nullable
   private static IElementType getTokenType(PsiBuilder builder) {
+    return getTokenType(builder, true);
+  }
+
+  @Nullable
+  private static IElementType getTokenType(PsiBuilder builder, boolean skipWhitespace) {
     IElementType tokenType;
     while ((tokenType = builder.getTokenType()) == JavaDocTokenType.DOC_SPACE) {
       builder.remapCurrentToken(TokenType.WHITE_SPACE);
-      builder.advanceLexer();
+      if (skipWhitespace) builder.advanceLexer();
     }
     return tokenType;
   }

java/java-impl/src/com/intellij/codeInsight/javadoc/markdown/CodeTagParser.java

@@ -0,0 +1,68 @@
+// Copyright 2000-2024 JetBrains s.r.o. and contributors. Use of this source code is governed by the Apache 2.0 license.
+package com.intellij.codeInsight.javadoc.markdown;
+
+import com.intellij.openapi.util.text.StringUtil;
+import kotlin.ranges.IntRange;
+import org.intellij.markdown.MarkdownTokenTypes;
+import org.intellij.markdown.parser.sequentialparsers.RangesListBuilder;
+import org.intellij.markdown.parser.sequentialparsers.SequentialParser;
+import org.intellij.markdown.parser.sequentialparsers.TokensCache;
+import org.jetbrains.annotations.NotNull;
+
+import java.util.List;
+
+/// Inline parser dedicated to finding `<code>` tags.
+///
+/// Only purpose is to exclude the range from further markdown processing
+public class CodeTagParser implements SequentialParser {
+  @Override
+  public @NotNull ParsingResult parse(@NotNull TokensCache tokens, @NotNull List<IntRange> rangesToGlue) {
+    var result = new SequentialParser.ParsingResultBuilder();
+    var delegateIndices = new RangesListBuilder();
+    TokensCache.Iterator iterator = tokens.new RangesListIterator(rangesToGlue);
+
+    while (iterator.getType() != null) {
+      if (iterator.getType() == MarkdownTokenTypes.HTML_TAG && hasOpeningTag(getText(tokens, iterator), "code")) {
+        int startIndex = iterator.getIndex();
+        var endIterator = findEnd(tokens, iterator.advance());
+
+        if (endIterator != null) {
+          result.withNode(new SequentialParser.Node(new IntRange(startIndex, endIterator.getIndex() + 1), JavaDocMarkdownFlavourDescriptor.RAW_TYPE));
+          iterator = endIterator.advance();
+          continue;
+        }
+      }
+
+      delegateIndices.put(iterator.getIndex());
+      iterator = iterator.advance();
+    }
+
+    return result.withFurtherProcessing(delegateIndices.get());
+  }
+
+  /// @return The iterator located at the end tag, or `null` if the end isn't found
+  private static TokensCache.Iterator findEnd(@NotNull TokensCache tokens, TokensCache.Iterator iterator) {
+    while (iterator.getType() != null) {
+      if (iterator.getType() == MarkdownTokenTypes.HTML_TAG) {
+        if (hasClosingTag(getText(tokens, iterator), "code")) {
+          return iterator;
+        }
+      }
+
+      iterator = iterator.advance();
+    }
+    return null;
+  }
+
+  private static String getText(@NotNull TokensCache tokens, TokensCache.Iterator iterator) {
+    return tokens.getOriginalText().subSequence(iterator.getStart(), iterator.getEnd()).toString();
+  }
+
+  private static boolean hasOpeningTag(String tag, String tagName) {
+    return StringUtil.trimEnd(StringUtil.trimStart(tag, "<"), ">").equals(tagName);
+  }
+
+  private static boolean hasClosingTag(String tag, String tagName) {
+    return StringUtil.trimEnd(StringUtil.trimStart(tag, "</"), ">").equals(tagName);
+  }
+}

java/java-impl/src/com/intellij/codeInsight/javadoc/markdown/JavaDocMarkdownFlavourDescriptor.java

@@ -6,11 +6,18 @@ import org.intellij.markdown.MarkdownElementTypes;
 import org.intellij.markdown.MarkdownTokenTypes;
 import org.intellij.markdown.ast.ASTNode;
 import org.intellij.markdown.flavours.gfm.GFMFlavourDescriptor;
+import org.intellij.markdown.flavours.gfm.GFMTokenTypes;
+import org.intellij.markdown.flavours.gfm.StrikeThroughDelimiterParser;
 import org.intellij.markdown.html.GeneratingProvider;
 import org.intellij.markdown.html.HtmlGenerator;
+import org.intellij.markdown.html.TransparentInlineHolderProvider;
 import org.intellij.markdown.html.TrimmingInlineHolderProvider;
 import org.intellij.markdown.parser.LinkMap;
 import org.intellij.markdown.parser.MarkerProcessorFactory;
+import org.intellij.markdown.parser.sequentialparsers.EmphasisLikeParser;
+import org.intellij.markdown.parser.sequentialparsers.SequentialParser;
+import org.intellij.markdown.parser.sequentialparsers.SequentialParserManager;
+import org.intellij.markdown.parser.sequentialparsers.impl.*;
 import org.jetbrains.annotations.NotNull;
 import org.jetbrains.annotations.Nullable;
 
@@ -23,10 +30,31 @@ import java.util.Map;
  * Flavour descriptor which describes how to generate html text from markdown documentation
  */
 public class JavaDocMarkdownFlavourDescriptor extends GFMFlavourDescriptor {
+  public static final IElementType RAW_TYPE = new IElementType("RAW_TYPE");
+
   public JavaDocMarkdownFlavourDescriptor() {
     super(true, false, true);
   }
 
+  @Override
+  public @NotNull SequentialParserManager getSequentialParserManager() {
+    return new SequentialParserManager() {
+      @Override
+      public @NotNull List<SequentialParser> getParserSequence() {
+        return List.of(
+          new AutolinkParser(List.of(MarkdownTokenTypes.AUTOLINK, GFMTokenTypes.GFM_AUTOLINK)),
+          new BacktickParser(),
+          new CodeTagParser(), // Custom parser excluding <code> blocks content to be processed as markdown
+          new MathParser(),
+          new ImageParser(),
+          new InlineLinkParser(),
+          new ReferenceLinkParser(),
+          new EmphasisLikeParser(new EmphStrongDelimiterParser(), new StrikeThroughDelimiterParser())
+        );
+      }
+    };
+  }
+
   @NotNull
   @Override
   public Map<IElementType, GeneratingProvider> createHtmlGeneratingProviders(@NotNull LinkMap linkMap, @Nullable URI baseURI) {
@@ -38,6 +66,9 @@ public class JavaDocMarkdownFlavourDescriptor extends GFMFlavourDescriptor {
     // Paragraphs may need to be inlined
     result.put(MarkdownElementTypes.PARAGRAPH, new JavaDocParagraphProvider());
 
+    // Custom type, no processing
+    result.put(RAW_TYPE, new TransparentInlineHolderProvider());
+
     return result;
   }
 

java/java-tests/testData/codeInsight/javadocIG/markdownInlineCodeBlock.html

@@ -1,3 +1,5 @@
 <html><head><base href="placeholder"></head><body><div class='definition'><pre><span style="color:#000080;font-weight:bold;">class</span> <span style="color:#000000;">MarkdownCodeBlock</span></pre></div><div class='content'><p>Single liner code block:
  <code><span style="">Hello world</span></code></p><p>No tags are interpreted inside them
- <code><span style="">{@link java.lang.String niceLink}</span></code></p></div><table class='sections'><p></table>
+ <code><span style="">{@link java.lang.String niceLink}</span></code></p><p>No markdown markup is interpreted inside them
+ <code><span style="">_Hello_ &lt;code&gt;</span></code></p><p>This is a broken inline link
+ `Start of broken link</p>end of broken link`</div><table class='sections'><p></table>

java/java-tests/testData/codeInsight/javadocIG/markdownInlineCodeBlock.java

@@ -4,4 +4,12 @@
 ///
 /// No tags are interpreted inside them
 /// `{@link java.lang.String niceLink}`
+///
+/// No markdown markup is interpreted inside them
+/// `_Hello_ <code>`
+///
+/// This is a broken inline link
+/// `Start of broken link
+///
+/// end of broken link`
 class MarkdownCodeBlock {}

java/java-tests/testData/codeInsight/javadocIG/preTagStrictBeforeCode.html

@@ -3,4 +3,5 @@
     Some text with inline <code><span style="">code</span></code>
   </pre>
   <pre><code><span style="">String&#32;fullLine&#32;=&#32;</span><span style="color:#008000;font-weight:bold;">""</span><span style="">;</span></code></pre>  </pre>
+  <pre><code><span style="">String&#32;secondFullLine&#32;=&#32;</span><span style="color:#008000;font-weight:bold;">""</span><span style="">;</span></code></pre>
  </div><table class='sections'><p></table>

java/java-tests/testData/codeInsight/javadocIG/preTagStrictBeforeCode.java

@@ -7,6 +7,7 @@
  * String fullLine = "";
  * }
  * </pre>
+ * <pre> {@code String secondFullLine = ""; }</pre>
  */
 class Test {
   public String field = null;