Improve code block handling within lists.

core/src/main/java/com/google/googlejavaformat/java/javadoc/JavadocWriter.java

@@ -335,6 +335,10 @@ void writeLiteral(Literal token) {
   }
 
   void writeMarkdownFencedCodeBlock(MarkdownFencedCodeBlock token) {
+    if (wroteAnythingSignificant && !atStartOfLine) {
+      // A reminder that atStartOfLine is still true after `-␣` because it is a StartOfLineToken.
+      requestBlankLine();
+    }
     flushWhitespace();
     output.append(token.start());
     token

core/src/main/java/com/google/googlejavaformat/java/javadoc/MarkdownPositions.java

@@ -30,6 +30,7 @@
 import com.google.googlejavaformat.java.javadoc.Token.MarkdownFencedCodeBlock;
 import com.google.googlejavaformat.java.javadoc.Token.ParagraphCloseTag;
 import com.google.googlejavaformat.java.javadoc.Token.ParagraphOpenTag;
+import java.util.Objects;
 import java.util.regex.Matcher;
 import java.util.regex.Pattern;
 import org.commonmark.node.BulletList;
@@ -128,12 +129,16 @@ private void visitFencedCodeBlock(FencedCodeBlock fencedCodeBlock) {
       // indentation gets subtracted from FencedCodeBlock.getLiteral(), which is the actual text
       // represented by the code block.
       int start = startPosition(fencedCodeBlock) + fencedCodeBlock.getFenceIndent();
+      int closingLength =
+          Objects.requireNonNullElse(
+              fencedCodeBlock.getClosingFenceLength(), fencedCodeBlock.getOpeningFenceLength());
+      // We have observed getClosingFenceLength() returning null in some cases.
       MarkdownFencedCodeBlock token =
           new MarkdownFencedCodeBlock(
               input.substring(start, endPosition(fencedCodeBlock)),
               fencedCodeBlock.getFenceCharacter().repeat(fencedCodeBlock.getOpeningFenceLength())
                   + fencedCodeBlock.getInfo(),
-              fencedCodeBlock.getFenceCharacter().repeat(fencedCodeBlock.getClosingFenceLength()),
+              fencedCodeBlock.getFenceCharacter().repeat(closingLength),
               fencedCodeBlock.getLiteral());
       positionToToken.get(start).addLast(token);
     }

core/src/main/java/com/google/googlejavaformat/java/javadoc/Token.java

@@ -38,11 +38,12 @@ default int length() {
 
   /**
    * Tokens that are always pinned to the following token. For example, {@code <p>} in {@code <p>Foo
-   * bar} (never {@code <p> Foo bar} or {@code <p>\nFoo bar}).
+   * bar} (never {@code <p> Foo bar} or {@code <p>\nFoo bar}); or {#code
+   * <li>} or {@code -␣} in {@code <li>Foo bar} or {@code -␣Foo bar}.
    *
-   * <p>This is not the only kind of "pinning" that we do: See also the joining of Literal tokens
-   * done by the lexer. The special pinning here is necessary because these tokens are not of type
-   * Literal (because they require other special handling).
+   *     <p>This is not the only kind of "pinning" that we do: See also the joining of Literal
+   *     tokens done by the lexer. The special pinning here is necessary because these tokens are
+   *     not of type Literal (because they require other special handling).
    */
   interface StartOfLineToken {}
 

core/src/test/java/com/google/googlejavaformat/java/JavadocFormattingTest.java

@@ -1739,9 +1739,14 @@ public void markdownFencedCodeBlocks() {
 ///    ```
 ///
 /// - flibbertigibbet
+///   ```
+///   code block in a list after text with no blank line intervening (one will be inserted)
+///   ```
+///
+/// - flibbertigibbet
 ///
 ///   ```
-///   code block in a list after text
+///   code block in a list after text with a blank line intervening
 ///   ```
 ///
 /// ~~~java
@@ -1767,8 +1772,16 @@ class Test {}
 ///   in a list
 ///   ```
 ///
-/// - flibbertigibbet```
-///   code block in a list after text
+/// - flibbertigibbet
+///
+///   ```
+///   code block in a list after text with no blank line intervening (one will be inserted)
+///   ```
+///
+/// - flibbertigibbet
+///
+///   ```
+///   code block in a list after text with a blank line intervening
 ///   ```
 ///
 /// ~~~java