Update org.jetbrains.markdown from 0.3.1 to 0.5.2 (#3231)

1 files changed, 39 insertions, 0 deletions

diff --git a/subprojects/analysis-markdown-jb/src/main/kotlin/org/jetbrains/dokka/analysis/markdown/jb/MarkdownParser.kt b/subprojects/analysis-markdown-jb/src/main/kotlin/org/jetbrains/dokka/analysis/markdown/jb/MarkdownParser.kt
index 256c7b98..130c6def 100644
--- a/subprojects/analysis-markdown-jb/src/main/kotlin/org/jetbrains/dokka/analysis/markdown/jb/MarkdownParser.kt
+++ b/subprojects/analysis-markdown-jb/src/main/kotlin/org/jetbrains/dokka/analysis/markdown/jb/MarkdownParser.kt
@@ -77,6 +77,44 @@ public open class MarkdownParser(
             ).flatMap { it.children }
         )
 
+    /**
+     * Handler for [MarkdownTokenTypes.ATX_CONTENT], which is the content of the header
+     * elements like [MarkdownElementTypes.ATX_1], [MarkdownElementTypes.ATX_2] and so on.
+     *
+     * For example, a header line like `# Header text` is expected to be parsed into:
+     * - One [MarkdownTokenTypes.ATX_HEADER] with startOffset = 0, endOffset = 1 (only the `#` symbol)
+     * - Composite [MarkdownTokenTypes.ATX_CONTENT] with four children: WHITE_SPACE, TEXT, WHITE_SPACE, TEXT.
+     */
+    private fun headerContentHandler(node: ASTNode): List<DocTag> {
+        // ATX_CONTENT contains everything after the `#` symbol, so if there's a space
+        // in-between the `#` symbol and the text (like `# header`), it will be present here too.
+        // However, we don't need the leading space between the `#` symbol and the text, nor do we need trailing spaces,
+        // so we just skip it (otherwise the header text will be parsed as `<whitespace>header` instead of `header`).
+        // If there's more space between `#` and text, like `#     header`, it will still be a single WHITE_SPACE
+        // element, but it will be wider, so the solution below should still hold. The same applies to trailing spaces.
+        val trimmedChildren = node.children.trimWhitespaceToken()
+
+        val children = trimmedChildren.evaluateChildren()
+        return DocTagsFromIElementFactory.getInstance(
+            MarkdownElementTypes.PARAGRAPH, // PARAGRAPH instead of TEXT to preserve compatibility with prev. versions
+            children = children
+        )
+    }
+
+    /**
+     * @return a sublist of [this] list that does not contain
+     *         leading and trailing [MarkdownTokenTypes.WHITE_SPACE] elements
+     */
+    private fun List<ASTNode>.trimWhitespaceToken(): List<ASTNode> {
+        val firstNonWhitespaceIndex = this.indexOfFirst { it.type != MarkdownTokenTypes.WHITE_SPACE }
+        if (firstNonWhitespaceIndex == -1) {
+            return this
+        }
+        val lastNonWhitespaceIndex = this.indexOfLast { it.type != MarkdownTokenTypes.WHITE_SPACE }
+
+        return this.subList(firstNonWhitespaceIndex, lastNonWhitespaceIndex + 1)
+    }
+
     private fun horizontalRulesHandler() =
         DocTagsFromIElementFactory.getInstance(MarkdownTokenTypes.HORIZONTAL_RULE)
 
@@ -365,6 +403,7 @@ public open class MarkdownParser(
             MarkdownElementTypes.ATX_5,
             MarkdownElementTypes.ATX_6,
             -> headersHandler(node)
+            MarkdownTokenTypes.ATX_CONTENT -> headerContentHandler(node)
             MarkdownTokenTypes.HORIZONTAL_RULE -> horizontalRulesHandler()
             MarkdownElementTypes.STRONG -> strongHandler(node)
             MarkdownElementTypes.EMPH -> emphasisHandler(node)