markdown fix but introduces a regresssion...

src/gui_2.py

@@ -3320,40 +3320,37 @@ def render_context_presets(app: App) -> None:
     app.ui_active_context_preset = ""
 
 def render_snapshot_tab(app: App) -> None:
- if imgui.begin_tab_bar("snapshot_tabs"):
+ with imscope.tab_bar("snapshot_tabs"):
-  if imgui.begin_tab_item("Aggregate MD")[0]:
+  with imscope.tab_item("Aggregate MD") as (exp, _):
-   display_md = app.last_aggregate_markdown
+   if exp:
-   if app.ui_focus_agent:
+    display_md = app.last_aggregate_markdown
-    tier_usage = app.mma_tier_usage.get(app.ui_focus_agent)
+    if app.ui_focus_agent:
-    if tier_usage:
+     tier_usage = app.mma_tier_usage.get(app.ui_focus_agent)
-     persona_name = tier_usage.get("persona")
+     if tier_usage:
-     if persona_name:
+      persona_name = tier_usage.get("persona")
-      persona = app.controller.personas.get(persona_name)
+      if persona_name:
-      if persona and persona.context_preset:
+       persona = app.controller.personas.get(persona_name)
-       cp_name = persona.context_preset
+       if persona and persona.context_preset:
-       if cp_name in app._focus_md_cache:
+        cp_name = persona.context_preset
-        display_md = app._focus_md_cache[cp_name]
+        if cp_name in app._focus_md_cache:
-       else:
+         display_md = app._focus_md_cache[cp_name]
-        flat = src.project_manager.flat_config(app.controller.project, app.active_discussion)
+        else:
-        cp = app.controller.project.get('context_presets', {}).get(cp_name)
+         flat = src.project_manager.flat_config(app.controller.project, app.active_discussion)
-        if cp:
+         cp = app.controller.project.get('context_presets', {}).get(cp_name)
-         flat["files"]["paths"] = cp.get("files", [])
+         if cp:
-         flat["screenshots"]["paths"] = cp.get("screenshots", [])
+          flat["files"]["paths"] = cp.get("files", [])
-         full_md, _, _ = src.aggregate.run(flat)
+          flat["screenshots"]["paths"] = cp.get("screenshots", [])
-         app._focus_md_cache[cp_name] = full_md
+          full_md, _, _ = src.aggregate.run(flat)
-         display_md = full_md
+          app._focus_md_cache[cp_name] = full_md
-   if imgui.button("Copy"): imgui.set_clipboard_text(display_md)
+          display_md = full_md
-   imgui.begin_child("last_agg_md", imgui.ImVec2(0, 0), True)
+    if imgui.button("Copy"): imgui.set_clipboard_text(display_md)
-   markdown_helper.render(display_md, context_id="snapshot_agg")
+    with imscope.child("last_agg_md", 0, 0, True):
-   imgui.end_child()
+     markdown_helper.render(display_md, context_id="snapshot_agg")
-   imgui.end_tab_item()
+  with imscope.tab_item("System Prompt") as (exp, _):
-  if imgui.begin_tab_item("System Prompt")[0]:
+   if exp:
-   if imgui.button("Copy"): imgui.set_clipboard_text(app.last_resolved_system_prompt)
+    if imgui.button("Copy"): imgui.set_clipboard_text(app.last_resolved_system_prompt)
-   imgui.begin_child("last_sys_prompt", imgui.ImVec2(0, 0), True)
+    with imscope.child("last_sys_prompt", 0, 0, True):
-   markdown_helper.render(app.last_resolved_system_prompt, context_id="snapshot_sys")
+     markdown_helper.render(app.last_resolved_system_prompt, context_id="snapshot_sys")
-   imgui.end_child()
-   imgui.end_tab_item()
-  imgui.end_tab_bar()
 
 #endregion: Context Management
 
@@ -3375,9 +3372,8 @@ def render_discussion_hub(app: App) -> None:
     if exp:
      if imgui.button("Copy to Clipboard"):
       imgui.set_clipboard_text(app.context_preview_text)
-     imgui.begin_child("ctx_preview_scroll_tab", imgui.ImVec2(0, 0), True)
+     with imscope.child("ctx_preview_scroll_tab", 0, 0, True):
-     markdown_helper.render(app.context_preview_text, context_id="ctx_preview_tab")
+      markdown_helper.render(app.context_preview_text, context_id="ctx_preview_tab")
-     imgui.end_child()
   with imscope.tab_item("Snapshot") as (exp, opened):
    if exp: render_snapshot_tab(app)
   with imscope.tab_item("Takes") as (exp, opened):
@@ -4461,9 +4457,8 @@ def render_approve_script_modal(app: App) -> None:
    
    avail_y = imgui.get_content_region_avail().y - 40 # reserve space for buttons
    if app.ui_approve_modal_preview:
-    imgui.begin_child("preview_child", imgui.ImVec2(-1, avail_y), True)
+    with imscope.child("preview_child", -1, avail_y, True):
-    markdown_helper.render(f"```powershell\n{dlg._script}\n```", context_id="approve_script_preview")
+     markdown_helper.render(f"```powershell\n{dlg._script}\n```", context_id="approve_script_preview")
-    imgui.end_child()
    else:
     ch, dlg._script = imgui.input_text_multiline("##confirm_script", dlg._script, imgui.ImVec2(-1, avail_y))
    imgui.separator()
@@ -5482,7 +5477,6 @@ def render_context_preview_window(app: App) -> None:
    imgui.same_line()
    if imgui.button("Copy to Clipboard"):
     imgui.set_clipboard_text(app.context_preview_text)
-   imgui.begin_child("ctx_preview_scroll", imgui.ImVec2(0, 0), True)
+   with imscope.child("ctx_preview_scroll", 0, 0, True):
-   markdown_helper.render(app.context_preview_text, context_id="ctx_preview")
+    markdown_helper.render(app.context_preview_text, context_id="ctx_preview")
-   imgui.end_child()
 

src/markdown_helper.py

@@ -134,9 +134,9 @@ class MarkdownRenderer:
    if md_buf:
     chunk = "".join(md_buf)
     if chunk.strip():
-     imgui_md.render(chunk)
+     self._render_md_no_bullet_overlap(chunk)
-     imgui.spacing()
     md_buf.clear()
+
   def flush_code() -> None:
    nonlocal block_idx
    if code_buf:
@@ -182,6 +182,41 @@ class MarkdownRenderer:
 
   flush_md()
   flush_code()
+
+ def _render_md_no_bullet_overlap(self, chunk: str) -> None:
+  """Render markdown, but split out bulleted-list sections and render them
+  as plain text via imgui.text to avoid imgui_md's known bullet-overlap bug.
+  """
+  import re
+  list_pattern = re.compile(r"^(?P<indent>[ \t]*)(?:[-*+])\s+", re.MULTILINE)
+  current_pos = 0
+  for m in list_pattern.finditer(chunk):
+   if m.start() > current_pos:
+    pre = chunk[current_pos:m.start()]
+    if pre.strip(): imgui_md.render(pre)
+   list_start = m.start()
+   indent_len = len(m.group("indent"))
+   i = list_start
+   while i < len(chunk):
+    line_end = chunk.find("\n", i)
+    if line_end == -1: line_end = len(chunk)
+    line = chunk[i:line_end]
+    stripped = line.lstrip(" \t")
+    leading_ws = len(line) - len(stripped)
+    if leading_ws == indent_len and (stripped.startswith(("- ", "* ", "+ ")) or stripped == ""):
+     i = line_end + 1
+     continue
+    if leading_ws < indent_len:
+     break
+    i = line_end + 1
+   list_block = chunk[list_start:i]
+   for line in list_block.splitlines():
+    imgui.text(line) if line.strip() else imgui.spacing()
+   current_pos = i
+  if current_pos < len(chunk):
+   tail = chunk[current_pos:]
+   if tail.strip(): imgui_md.render(tail)
+
  def render_unindented(self, text: str) -> None:
   """Render Markdown text with automatic unindentation."""
   imgui_md.render_unindented(text)