From 13dd97d60d8c7a34abc36048f5bb773614fcf833 Mon Sep 17 00:00:00 2001
From: webtecnica <webtecnica@gmail.com>
Date: Sun, 2 Aug 2026 14:58:33 -0300
Subject: [PATCH] fix(renderer): preserve nested mixed ul/ol list hierarchy
 (#6700)

---
 static/ui.js                             | 111 ++++++++++++++---------
 tests/test_886_ordered_list_numbering.py |  35 ++++---
 tests/test_renderer_js_behaviour.py      |  64 ++++++++++++-
 tests/test_sprint16.py                   |  40 ++++++--
 4 files changed, 178 insertions(+), 72 deletions(-)

diff --git a/static/ui.js b/static/ui.js
index abe4634d..cd25232e 100644
--- a/static/ui.js
+++ b/static/ui.js
@@ -7801,48 +7801,72 @@ function renderMd(raw){
   s=s.replace(/^---+$/gm,'<hr>');
   // (Blockquotes are handled by the pre-pass at the top of renderMd, before
   // fence_stash. The per-line passes below never see > prefixes.)
-  function _renderListBlock(lines, ordered){
-    const marker=ordered?'\\d+\\. ':'[-*+] ';
-    let html=ordered?'<ol>':'<ul>';
-    let item=null;
-    const flush=()=>{
-      if(!item) return;
-      const body=item.parts.join('\n').trim();
-      const text=body;
-      let inner;
-      if(!ordered && /^\[x\] /i.test(text)) inner='<span class="task-done">✅</span> '+inlineMd(text.slice(4));
-      else if(!ordered && /^\[ \] /.test(text)) inner='<span class="task-todo">☐</span> '+inlineMd(text.slice(4));
-      else inner=inlineMd(text);
-      const valueAttr=item.value!==null?` value="${item.value}"`:'';
-      const styleAttr=item.indent?` style="margin-left:16px"`:'';
-      html+=`<li${valueAttr}${styleAttr}>${inner}</li>`;
-      item=null;
+  function _renderListBlock(lines){
+    // Single-pass mixed-marker list renderer (#6700). Builds a real nested
+    // <ul>/<ol> tree with a stack keyed by indentation depth + marker kind
+    // instead of rendering one marker family and re-parsing the generated
+    // HTML with the other. Only item text goes through inlineMd(); tags
+    // emitted here are never parsed as Markdown again.
+    const root={ordered:false,items:[]};   // container for top-level lists
+    const listStack=[];                    // listStack[k]: list open at depth k
+    const itemStack=[];                    // itemStack[k]: last item at depth k
+    const openList=(depth,ordered)=>{
+      const ln={ordered,items:[]};
+      // Attach to the nearest open ancestor item when indentation jumps by
+      // more than one level (e.g. 4-space indent straight from the top), so
+      // the new list nests under the right parent instead of detaching.
+      let parentItem=null;
+      for(let k=depth-1;k>=0;k--){
+        if(itemStack[k]){ parentItem=itemStack[k]; break; }
+      }
+      (parentItem?parentItem.sublists:root.items).push(ln);
+      listStack[depth]=ln;
+      return ln;
+    };
+    const openItem=(depth,ordered,value,content)=>{
+      for(let k=listStack.length-1;k>depth;k--) listStack.pop();
+      if(!listStack[depth]||listStack[depth].ordered!==ordered){
+        if(listStack[depth]) listStack.pop();
+        listStack[depth]=openList(depth,ordered);
+      }
+      const item={value,content:[content],sublists:[]};
+      listStack[depth].items.push(item);
+      itemStack[depth]=item;
+      for(let k=itemStack.length-1;k>depth;k--) itemStack.pop();
+      return item;
     };
     for(const raw of lines){
       const line=String(raw||'');
-      const nested=line.match(new RegExp(`^ {2,}(${marker})(.*)$`));
-      if(nested){
-        flush();
-        item={indent:true,value:ordered?parseInt(nested[1],10):null,parts:[nested[2]]};
-        continue;
+      const om=line.match(/^(\s*)(\d+)\.\s+(.*)$/);
+      const um=line.match(/^(\s*)([-*+])\s+(.*)$/);
+      const m=om||um;
+      if(m){
+        const depth=m[1].length<2?0:Math.floor(m[1].length/2);
+        openItem(depth,!!om,om?parseInt(om[2],10):null,m[3]);
+      } else if(itemStack.length){
+        itemStack[itemStack.length-1].content.push(line.replace(/^ {2,}/,'').trim());
       }
-      const top=line.match(new RegExp(`^(?:  )?(${marker})(.*)$`));
-      if(top){
-        flush();
-        item={indent:false,value:ordered?parseInt(top[1],10):null,parts:[top[2]]};
-        continue;
-      }
-      if(!item) continue;
-      item.parts.push(line.replace(/^ {2,}/,'').trim());
     }
-    flush();
-    return html+(ordered?'</ol>':'</ul>');
+    const renderList=ln=>'<'+(ln.ordered?'ol':'ul')+'>'+ln.items.map(renderItem).join('')+'</'+(ln.ordered?'ol':'ul')+'>';
+    const renderItem=it=>{
+      const text=it.content.join('\n').trim();
+      let inner;
+      if(!it.ordered && /^\[x\] /i.test(text)) inner='<span class="task-done">✅</span> '+inlineMd(text.slice(4));
+      else if(!it.ordered && /^\[ \] /.test(text)) inner='<span class="task-todo">☐</span> '+inlineMd(text.slice(4));
+      else inner=inlineMd(text);
+      const valueAttr=it.value!==null?` value="${it.value}"`:'';
+      return `<li${valueAttr}>${inner}${it.sublists.map(renderList).join('')}</li>`;
+    };
+    return root.items.map(renderList).join('');
   }
-  function _renderLists(src, ordered){
+  function _renderLists(src){
+    // Single pass over source lines: collect a contiguous list region (any
+    // marker family, top-level or nested, plus continuation lines) and hand
+    // it to _renderListBlock, which builds the structural <ul>/<ol> tree.
     const lines=src.split('\n');
     const out=[];
-    const topRe=ordered?/^(?:  )?\d+\. /:/^(?:  )?[-*+] /;
-    const nestedRe=ordered?/^ {2,}\d+\. /:/^ {2,}[-*+] /;
+    const topRe=/^(?:  )?(?:\d+\.|[-*+]) /;
+    const nestedRe=/^ {2,}(?:\d+\.|[-*+]) /;
     const contRe=/^ {2,}\S/;
     let i=0;
     while(i<lines.length){
@@ -7869,18 +7893,16 @@ function renderMd(raw){
         }
         break;
       }
-      out.push(_renderListBlock(block,ordered));
+      out.push(_renderListBlock(block));
     }
     return out.join('\n');
   }
-  // Preserve continuation lines, nested indentation, and LaTeX placeholder lines
-  // inside list items without changing the wider markdown pipeline.
-  s=_renderLists(s,false);
-  // Ordered-list parsing intentionally runs on the post-unordered string; the
-  // unordered pass emits <ul> HTML that cannot satisfy the ordered-item regex.
-  // Keep continuation lines attached to their item and preserve explicit
-  // numbering via value= even when blank lines split the markdown.
-  s=_renderLists(s,true);
+  // Single-pass list stage (#6700): one parser handles both marker families
+  // with a stack keyed by indentation + marker kind, so mixed nested lists
+  // (ul→ol→ul) build valid <ul>/<ol> structure instead of re-parsing
+  // renderer-generated HTML as Markdown. value="N" is still emitted on every
+  // ordered <li> so explicit numbering survives blank-line splits (#886).
+  s=_renderLists(s);
   // Tables: | col | col | header row followed by | --- | --- | separator then data rows
   // NOTE: table pass runs BEFORE outer link pass so [label](url) in table cells
   // is handled by inlineMd() only — prevents double-linking.
@@ -8043,8 +8065,7 @@ function renderMd(raw){
     const a=_attrs(rawAttrs);
     if(name==='li'){
       const value=/^\d+$/.test(a.value||'')?` value="${esc(a.value)}"`:'';
-      const style=(a.style||'').replace(/\s+/g,'').toLowerCase()==='margin-left:16px'?` style="margin-left:16px"`:'';
-      return `<li${value}${style}>`;
+      return `<li${value}>`;
     }
     if(name==='span'){
       return `<span${_cls(a.class,['task-done','task-todo','katex-inline'])}${a['data-katex']==='inline'?' data-katex="inline"':''}>`;
diff --git a/tests/test_886_ordered_list_numbering.py b/tests/test_886_ordered_list_numbering.py
index d81adf7b..c8e11962 100644
--- a/tests/test_886_ordered_list_numbering.py
+++ b/tests/test_886_ordered_list_numbering.py
@@ -8,6 +8,11 @@ at 1, producing "1. 1. 1." instead of "1. 2. 3.".
 
 Fix: emit value="N" on every <li> so the correct ordinal is preserved even when
 items end up in separate <ol> containers after the paragraph split.
+
+The list stage was rewritten for #6700 into a single-pass, mixed-marker parser
+(a stack keyed by indentation + marker kind). These static assertions pin the
+parts of that implementation that keep #886 working: the ordered-marker digit
+capture and the value="N" emission on <li>.
 """
 import os
 import re
@@ -21,15 +26,15 @@ def get_ui_js():
 
 class TestOrderedListNumbering:
     def _ordered_list_block(self, src: str) -> str:
-        start = src.find("function _renderListBlock(lines, ordered){")
+        start = src.find("function _renderListBlock(lines){")
         assert start != -1, "_renderListBlock helper not found in ui.js"
-        end = src.find("function _renderLists(src, ordered){", start)
+        end = src.find("function _renderLists(src){", start)
         assert end != -1, "_renderLists helper not found after _renderListBlock"
         return src[start:end]
 
     def _ordered_list_dispatch_block(self, src: str) -> str:
-        start = src.find("s=_renderLists(s,true);")
-        assert start != -1, "ordered-list dispatch not found in ui.js"
+        start = src.find("s=_renderLists(s);")
+        assert start != -1, "list dispatch not found in ui.js"
         return src[max(0, start - 260):start + 80]
 
     def test_li_value_attr_present_in_ordered_list_block(self):
@@ -42,18 +47,18 @@ class TestOrderedListNumbering:
         )
 
     def test_li_value_uses_parsed_number(self):
-        """The value= must be derived from parseInt of the captured digit, not hardcoded."""
+        """The value= must be derived from parseInt of the captured digits, not hardcoded."""
         src = get_ui_js()
         ol_block = self._ordered_list_block(src)
         assert 'parseInt' in ol_block, (
             "Ordered-list block should use parseInt() to parse the list number (#886)"
         )
 
-    def test_numMatch_variable_present(self):
-        """The ordered-list branch must still capture digits from the markdown marker."""
+    def test_marker_digit_capture_present(self):
+        """The ordered branch must still capture digits from the markdown marker."""
         src = get_ui_js()
         ol_block = self._ordered_list_block(src)
-        assert "const marker=ordered?'\\\\d+\\\\. ':'[-*+] ';" in ol_block, (
+        assert "\\d+\\." in ol_block or re.search(r"\(\\d\+\)\\\.", ol_block), (
             "Ordered-list block should keep a digit marker pattern for numbered items (#886)"
         )
 
@@ -68,18 +73,18 @@ class TestOrderedListNumbering:
         )
 
     def test_ordered_list_comment_references_issue(self):
-        """A comment near the OL fix should reference the issue (#886) or the symptom."""
+        """A comment near the list dispatch should reference #886 or the symptom."""
         src = get_ui_js()
         context = self._ordered_list_dispatch_block(src)
         has_comment = '#886' in context or '1. 1. 1.' in context or 'blank lines' in context.lower()
         assert has_comment, (
-            "Expected a comment near the OL fix explaining the blank-line issue (#886)"
+            "Expected a comment near the list dispatch explaining the blank-line issue (#886)"
         )
 
-    def test_list_without_blank_lines_unaffected(self):
-        """A compact list should still flow through the ordered-list helper."""
+    def test_lists_route_through_single_shared_helper(self):
+        """All list rendering must flow through the single-pass shared helper."""
         src = get_ui_js()
-        ol_block = self._ordered_list_dispatch_block(src)
-        assert "s=_renderLists(s,true);" in ol_block, (
-            "Ordered-list rendering should still route through the shared helper"
+        dispatch = self._ordered_list_dispatch_block(src)
+        assert "s=_renderLists(s);" in dispatch, (
+            "List rendering should route through the single-pass _renderLists helper"
         )
diff --git a/tests/test_renderer_js_behaviour.py b/tests/test_renderer_js_behaviour.py
index 3056a8d6..9c3c587a 100644
--- a/tests/test_renderer_js_behaviour.py
+++ b/tests/test_renderer_js_behaviour.py
@@ -338,10 +338,13 @@ class TestMarkdownListsWithLatex:
         assert "<li>next item</li>" in out
 
     def test_nested_indentation_stays_in_list(self, driver_path):
+        """Same-marker nesting builds a structural nested <ul> instead of a
+        styled sibling <li> — the margin-left convention was removed by the
+        single-pass mixed-marker parser (#6700)."""
         out = _render(driver_path, "- parent\n  - child")
         assert "<ul>" in out
-        assert "<li>parent</li>" in out
-        assert '<li style="margin-left:16px">child</li>' in out
+        assert "<li>parent<ul><li>child</li></ul></li>" in out
+        assert "margin-left:16px" not in out
 
     def test_display_math_line_stays_inside_list_item(self, driver_path):
         src = "- intro\n\n  $$x^2$$\n\n  continuation"
@@ -361,6 +364,63 @@ class TestMarkdownListsWithLatex:
         assert '<li value="3">tail</li>' in out
 
 
+class TestMixedNestedLists:
+    """#6700: mixed ul/ol nesting must build a valid structural hierarchy.
+
+    Regression for the old two-pass list renderer, where the ordered pass
+    re-parsed the <ul> HTML emitted by the unordered pass as Markdown. That
+    leaked escaped fragments like `&lt;/li&gt;&lt;li style=&quot;margin-left:
+    16px&quot;&gt;` into the chat and flattened the inverse (ol→ul) shape.
+    """
+
+    def test_ul_ol_ul_nested_then_top_level_return(self, driver_path):
+        src = (
+            "- normal item\n"
+            "  1. numbered child\n"
+            "  2. second numbered child\n"
+            "  - unordered child again\n"
+            "- next normal item"
+        )
+        out = _render(driver_path, src)
+        # No escaped renderer-generated fragments may leak into the output
+        assert "&lt;/li&gt;" not in out, out
+        assert "&lt;ul" not in out, out
+        assert "&lt;ol" not in out, out
+        assert "margin-left:16px" not in out, out
+        # Balanced containers
+        assert out.count("<ul>") == out.count("</ul>"), out
+        assert out.count("<ol>") == out.count("</ol>"), out
+        # Exact structural hierarchy: ul > li > (ol, ul) > li
+        assert (
+            '<ul><li>normal item'
+            '<ol><li value="1">numbered child</li>'
+            '<li value="2">second numbered child</li></ol>'
+            '<ul><li>unordered child again</li></ul></li>'
+            '<li>next normal item</li></ul>'
+        ) in out, out
+
+    def test_ol_ul_nested_keeps_bullets(self, driver_path):
+        src = "1. first step\n   - detail A\n   - detail B\n2. second step"
+        out = _render(driver_path, src)
+        assert "&lt;/li&gt;" not in out, out
+        assert "&lt;ul" not in out, out
+        assert "&lt;ol" not in out, out
+        assert out.count("<ol>") == out.count("</ol>"), out
+        assert out.count("<ul>") == out.count("</ul>"), out
+        assert (
+            '<ol><li value="1">first step'
+            '<ul><li>detail A</li><li>detail B</li></ul></li>'
+            '<li value="2">second step</li></ol>'
+        ) in out, out
+
+    def test_top_level_marker_switch_starts_sibling_list(self, driver_path):
+        out = _render(driver_path, "- bullet\n1. numbered")
+        assert (
+            '<ul><li>bullet</li></ul><ol><li value="1">numbered</li></ol>'
+        ) in out, out
+        assert "&lt;/li&gt;" not in out, out
+
+
 # ─────────────────────────────────────────────────────────────────────────────
 # Block-level constructs INSIDE blockquotes — the six bugs documented in
 # blockquote-rendering-bugs.md. Each test feeds the exact input from the
diff --git a/tests/test_sprint16.py b/tests/test_sprint16.py
index fd9882bc..446796f1 100644
--- a/tests/test_sprint16.py
+++ b/tests/test_sprint16.py
@@ -100,22 +100,40 @@ def render_md(raw):
     def handle_ul(block):
         lines = block.strip().split("\n")
         out = "<ul>"
+        prev = 0
         for l in lines:
-            indent = bool(re.match(r"^ {2,}", l))
-            text = re.sub(r"^ {0,4}[-*+] ", "", l)
-            style = ' style="margin-left:16px"' if indent else ""
-            out += f"<li{style}>{inline_md(text)}</li>"
-        return out + "</ul>"
+            m = re.match(r"^( *)([-*+]) (.*)$", l)
+            if not m:
+                continue
+            depth = 0 if len(m.group(1)) < 2 else len(m.group(1)) // 2
+            if depth > prev:
+                out += "<ul>"
+            elif depth < prev:
+                out += "</ul>" * (prev - depth)
+            prev = depth
+            out += f"<li>{inline_md(m.group(3))}</li>"
+        out += "</ul>" * (prev + 1)
+        return out
 
     s = re.sub(r"((?:^(?:  )?[-*+] .+\n?)+)", lambda m: handle_ul(m.group()), s, flags=re.M)
 
     def handle_ol(block):
         lines = block.strip().split("\n")
         out = "<ol>"
+        prev = 0
         for l in lines:
-            text = re.sub(r"^ {0,4}\d+\. ", "", l)
-            out += f"<li>{inline_md(text)}</li>"
-        return out + "</ol>"
+            m = re.match(r"^( *)(\d+)\. (.*)$", l)
+            if not m:
+                continue
+            depth = 0 if len(m.group(1)) < 2 else len(m.group(1)) // 2
+            if depth > prev:
+                out += "<ol>"
+            elif depth < prev:
+                out += "</ol>" * (prev - depth)
+            prev = depth
+            out += f"<li>{inline_md(m.group(3))}</li>"
+        out += "</ol>" * (prev + 1)
+        return out
 
     s = re.sub(r"((?:^(?:  )?\d+\. .+\n?)+)", lambda m: handle_ol(m.group()), s, flags=re.M)
 
@@ -559,10 +577,12 @@ def test_ordered_list_code_spans(cleanup_test_sessions):
     assert "<code>npm start</code>" in out
 
 def test_indented_list_item_bold(cleanup_test_sessions):
-    """Bold inside indented (nested) list item."""
+    """Bold inside indented (nested) list item — nested items now build a
+    structural nested <ul> instead of a margin-left sibling (#6700)."""
     out = render_md("- top level\n  - **nested bold**")
     assert "<strong>nested bold</strong>" in out
-    assert "margin-left:16px" in out
+    assert "<ul><li><strong>nested bold</strong></li></ul>" in out
+    assert "margin-left:16px" not in out
 
 # --- Blockquote variants ---