refactor(markdown-parser): promote blank lines between list items to MdNewline nodes (#9313)

Co-authored-by: autofix-ci[bot] <114827586+autofix-ci[bot]@users.noreply.github.com>

Author: jfmcdowell

crates/biome_markdown_factory/src/generated/node_factory.rs

@@ -1679,6 +1679,31 @@ impl IntoFormat<MarkdownFormatContext> for biome_markdown_syntax::AnyMdBlock {
         )
     }
 }
+impl AsFormat<MarkdownFormatContext> for biome_markdown_syntax::AnyMdBulletListMember {
+    type Format<'a> = FormatRefWithRule<
+        'a,
+        biome_markdown_syntax::AnyMdBulletListMember,
+        crate::markdown::any::bullet_list_member::FormatAnyMdBulletListMember,
+    >;
+    fn format(&self) -> Self::Format<'_> {
+        FormatRefWithRule::new(
+            self,
+            crate::markdown::any::bullet_list_member::FormatAnyMdBulletListMember::default(),
+        )
+    }
+}
+impl IntoFormat<MarkdownFormatContext> for biome_markdown_syntax::AnyMdBulletListMember {
+    type Format = FormatOwnedWithRule<
+        biome_markdown_syntax::AnyMdBulletListMember,
+        crate::markdown::any::bullet_list_member::FormatAnyMdBulletListMember,
+    >;
+    fn into_format(self) -> Self::Format {
+        FormatOwnedWithRule::new(
+            self,
+            crate::markdown::any::bullet_list_member::FormatAnyMdBulletListMember::default(),
+        )
+    }
+}
 impl AsFormat<MarkdownFormatContext> for biome_markdown_syntax::AnyMdCodeBlock {
     type Format<'a> = FormatRefWithRule<
         'a,

crates/biome_markdown_factory/src/generated/syntax_factory.rs

@@ -0,0 +1,15 @@
+//! This is a generated file. Don't modify it by hand! Run 'cargo codegen formatter' to re-generate the file.
+
+use crate::prelude::*;
+use biome_markdown_syntax::AnyMdBulletListMember;
+#[derive(Debug, Clone, Default)]
+pub(crate) struct FormatAnyMdBulletListMember;
+impl FormatRule<AnyMdBulletListMember> for FormatAnyMdBulletListMember {
+    type Context = MarkdownFormatContext;
+    fn fmt(&self, node: &AnyMdBulletListMember, f: &mut MarkdownFormatter) -> FormatResult<()> {
+        match node {
+            AnyMdBulletListMember::MdBullet(node) => node.format().fmt(f),
+            AnyMdBulletListMember::MdNewline(node) => node.format().fmt(f),
+        }
+    }
+}

crates/biome_markdown_formatter/src/generated.rs

@@ -1,6 +1,7 @@
 //! This is a generated file. Don't modify it by hand! Run 'cargo codegen formatter' to re-generate the file.
 
 pub(crate) mod block;
+pub(crate) mod bullet_list_member;
 pub(crate) mod code_block;
 pub(crate) mod container_block;
 pub(crate) mod inline;

crates/biome_markdown_formatter/src/markdown/any/bullet_list_member.rs

@@ -331,8 +331,10 @@ fn skip_blank_lines_between_items(
         // Blank lines between items make the list loose
         *is_tight = false;
         *last_item_ends_with_blank = true;
-        // Skip the blank line as trivia (no tree node created)
-        p.parse_as_skipped_trivia_tokens(|p| p.bump(NEWLINE));
+        // Emit blank line as an explicit MdNewline CST node
+        let newline_m = p.start();
+        p.bump(NEWLINE);
+        newline_m.complete(p, MD_NEWLINE);
     }
 }
 

crates/biome_markdown_formatter/src/markdown/any/mod.rs

@@ -1161,6 +1161,7 @@ fn set_inline_emphasis_context(p: &mut MarkdownParser) -> Option<EmphasisContext
 fn inline_list_source_len(p: &mut MarkdownParser) -> usize {
     p.lookahead(|p| {
         let mut len = 0usize;
+        let mut has_content = false;
 
         loop {
             if p.at(T![EOF]) {
@@ -1186,11 +1187,19 @@ fn inline_list_source_len(p: &mut MarkdownParser) -> usize {
                     consume_quote_prefix_without_virtual(p, quote_depth);
                 }
 
+                if has_content && p.at(MD_SETEXT_UNDERLINE_LITERAL) && allow_setext_heading(p) {
+                    break;
+                }
+
+                if has_content && p.at(MD_THEMATIC_BREAK_LITERAL) && is_dash_only_thematic_break(p)
+                {
+                    break;
+                }
                 if quote_depth > 0 && p.at(R_ANGLE) && !has_quote_prefix(p, quote_depth) {
                     consume_partial_quote_prefix_lookahead(p, quote_depth, &mut len);
                 }
 
-                if at_paragraph_break(p, true) {
+                if at_paragraph_break(p, has_content) {
                     break;
                 }
 
@@ -1228,9 +1237,10 @@ fn inline_list_source_len(p: &mut MarkdownParser) -> usize {
                     // We intentionally skip the heavier post-indent block-interrupt
                     // check here; the following non-NEWLINE pass still catches
                     // interrupts for emphasis-context length calculation.
-                    if p.at(MD_SETEXT_UNDERLINE_LITERAL)
-                        || (p.at(MD_THEMATIC_BREAK_LITERAL)
-                            && is_dash_only_thematic_break_text(p.cur_text()))
+                    if has_content
+                        && (p.at(MD_SETEXT_UNDERLINE_LITERAL)
+                            || (p.at(MD_THEMATIC_BREAK_LITERAL)
+                                && is_dash_only_thematic_break_text(p.cur_text())))
                     {
                         break;
                     }
@@ -1239,18 +1249,22 @@ fn inline_list_source_len(p: &mut MarkdownParser) -> usize {
                 continue;
             }
 
-            if p.at(MD_SETEXT_UNDERLINE_LITERAL) && allow_setext_heading(p) {
+            if has_content && p.at(MD_SETEXT_UNDERLINE_LITERAL) && allow_setext_heading(p) {
                 break;
             }
 
-            if p.at(MD_THEMATIC_BREAK_LITERAL) && is_dash_only_thematic_break(p) {
+            if has_content && p.at(MD_THEMATIC_BREAK_LITERAL) && is_dash_only_thematic_break(p) {
                 break;
             }
 
             if p.has_preceding_line_break() && at_block_interrupt(p) {
                 break;
             }
 
+            if !p.cur_text().chars().all(|c| c == ' ' || c == '\t') {
+                has_content = true;
+            }
+
             len += p.cur_text().len();
             p.bump(p.cur());
         }

crates/biome_markdown_parser/src/syntax/list.rs

@@ -43,13 +43,13 @@
 //! be decided with full context.
 
 use biome_markdown_syntax::{
-    AnyMdBlock, AnyMdCodeBlock, AnyMdInline, AnyMdLeafBlock, MarkdownLanguage, MdAutolink,
-    MdBlockList, MdBullet, MdBulletListItem, MdDocument, MdEntityReference, MdFencedCodeBlock,
-    MdHardLine, MdHeader, MdHtmlBlock, MdIndentCodeBlock, MdInlineCode, MdInlineEmphasis,
-    MdInlineHtml, MdInlineImage, MdInlineItalic, MdInlineItemList, MdInlineLink, MdLinkBlock,
-    MdLinkDestination, MdLinkLabel, MdLinkReferenceDefinition, MdLinkTitle, MdOrderedListItem,
-    MdParagraph, MdQuote, MdReferenceImage, MdReferenceLink, MdReferenceLinkLabel, MdSetextHeader,
-    MdSoftBreak, MdTextual, MdThematicBreakBlock,
+    AnyMdBlock, AnyMdBulletListMember, AnyMdCodeBlock, AnyMdInline, AnyMdLeafBlock,
+    MarkdownLanguage, MdAutolink, MdBlockList, MdBullet, MdBulletListItem, MdDocument,
+    MdEntityReference, MdFencedCodeBlock, MdHardLine, MdHeader, MdHtmlBlock, MdIndentCodeBlock,
+    MdInlineCode, MdInlineEmphasis, MdInlineHtml, MdInlineImage, MdInlineItalic, MdInlineItemList,
+    MdInlineLink, MdLinkBlock, MdLinkDestination, MdLinkLabel, MdLinkReferenceDefinition,
+    MdLinkTitle, MdOrderedListItem, MdParagraph, MdQuote, MdReferenceImage, MdReferenceLink,
+    MdReferenceLinkLabel, MdSetextHeader, MdSoftBreak, MdTextual, MdThematicBreakBlock,
 };
 use biome_rowan::{AstNode, AstNodeList, Direction, SyntaxNode, TextRange, WalkEvent};
 use percent_encoding::{AsciiSet, CONTROLS, utf8_percent_encode};
@@ -565,7 +565,11 @@ impl<'a> HtmlRenderer<'a> {
 
             let start = list
                 .md_bullet_list()
-                .first()
+                .iter()
+                .find_map(|member| match member {
+                    AnyMdBulletListMember::MdBullet(bullet) => Some(bullet),
+                    _ => None,
+                })
                 .and_then(|bullet| bullet.prefix().ok())
                 .and_then(|prefix| prefix.marker().ok())
                 .map_or(1, |marker| {

crates/biome_markdown_parser/src/syntax/mod.rs

@@ -0,0 +1,29 @@
+# Blank lines between bullet items
+
+- foo
+
+- bar
+
+- baz
+
+# Blank lines between ordered items
+
+1. first
+
+2. second
+
+3. third
+
+# Multiple blank lines between items
+
+- alpha
+
+
+- beta
+
+# Mixed: some items separated, some not
+
+- one
+- two
+
+- three