conductor(checkpoint): Checkpoint end of Phase 3 - Discussion & Context Structure Mutation

conductor/tests/verify_phase_3.py

@@ -0,0 +1,24 @@
+import subprocess
+import sys
+import os
+
+def verify_phase_3():
+ print("Verifying Phase 3: Discussion & Context Structure Mutation...")
+ 
+ # Run the comprehensive simulation test
+ result = subprocess.run(
+  ["uv", "run", "pytest", "tests/test_undo_redo_sim.py"],
+  capture_output=True,
+  text=True
+ )
+ 
+ if result.returncode == 0:
+  print("Phase 3 verification PASSED.")
+ else:
+  print("Phase 3 verification FAILED.")
+  print(result.stdout)
+  print(result.stderr)
+  sys.exit(1)
+
+if __name__ == "__main__":
+ verify_phase_3()

src/app_controller.py

@@ -403,7 +403,9 @@ class AppController:
    'ui_separate_tier4': 'ui_separate_tier4',
    'show_text_viewer': 'show_text_viewer',
    'text_viewer_title': 'text_viewer_title',
-   'text_viewer_type': 'text_viewer_type'
+   'text_viewer_type': 'text_viewer_type',
+   'disc_entries': 'disc_entries',
+   'ui_file_paths': 'ui_file_paths'
   }
   self._gettable_fields = dict(self._settable_fields)
   self._gettable_fields.update({
@@ -575,6 +577,24 @@ class AppController:
   except:
    pass
 
+ @property
+ def ui_file_paths(self) -> list[str]:
+  return [f.path if hasattr(f, 'path') else str(f) for f in self.files]
+
+ @ui_file_paths.setter
+ def ui_file_paths(self, value: list[str]) -> None:
+  old_files = {f.path: f for f in self.files if hasattr(f, 'path')}
+  new_files = []
+  import time
+  now = time.time()
+  for p in value:
+   if p in old_files:
+    new_files.append(old_files[p])
+   else:
+    from src import models
+    new_files.append(models.FileItem(path=p, injected_at=now))
+  self.files = new_files
+
  @property
  def operations_live_indicator(self) -> bool:
   return not self.is_viewing_prior_session

src/gui_2.py

@@ -346,19 +346,27 @@ class App:
    self._is_applying_snapshot = False
 
  def _handle_undo(self) -> None:
+  sys.stderr.write(f"[DEBUG History] _handle_undo called. can_undo={self.history.can_undo}\n")
+  sys.stderr.flush()
   if not self.history.can_undo:
    return
   current = self._take_snapshot()
   entry = self.history.undo(current, "Undo Action")
   if entry:
+   sys.stderr.write(f"[DEBUG History] Undoing to: {entry.description}\n")
+   sys.stderr.flush()
    self._apply_snapshot(entry.state)
 
  def _handle_redo(self) -> None:
+  sys.stderr.write(f"[DEBUG History] _handle_redo called. can_redo={self.history.can_redo}\n")
+  sys.stderr.flush()
   if not self.history.can_redo:
    return
   current = self._take_snapshot()
   entry = self.history.redo(current, "Redo Action")
   if entry:
+   sys.stderr.write(f"[DEBUG History] Redoing to: {entry.description}\n")
+   sys.stderr.flush()
    self._apply_snapshot(entry.state)
 
  def shutdown(self) -> None:
@@ -400,6 +408,8 @@ class App:
 
  @ui_file_paths.setter
  def ui_file_paths(self, paths: list[str]) -> None:
+  sys.stderr.write(f"[DEBUG] Setting ui_file_paths to: {paths}\n")
+  sys.stderr.flush()
   old_files = {f.path: f for f in self.files if hasattr(f, 'path')}
   new_files = []
   now = time.time()
@@ -407,6 +417,7 @@ class App:
    if p in old_files:
     new_files.append(old_files[p])
    else:
+    from src import models
     new_files.append(models.FileItem(path=p, injected_at=now))
   self.files = new_files
 
@@ -1210,9 +1221,10 @@ class App:
   if self.perf_profiling_enabled: self.perf_monitor.end_component("_gui_func")
 
  def _handle_history_logic(self) -> None:
-  if self._is_applying_snapshot:
+  # Skip history tracking only during active AI thinking/tool execution
+  is_thinking = self.ai_status in ["sending...", "streaming...", "running powershell...", "fetching url...", "searching web..."]
+  if self._is_applying_snapshot or is_thinking:
    return  
-  
   io = imgui.get_io()
   
   # 1. Hotkey handling (Undo: Ctrl+Z, Redo: Ctrl+Y or Ctrl+Shift+Z)
@@ -1241,19 +1253,27 @@ class App:
    current.global_system_prompt != self._last_ui_snapshot.global_system_prompt or
    current.base_system_prompt != self._last_ui_snapshot.base_system_prompt or
    current.use_default_base_prompt != self._last_ui_snapshot.use_default_base_prompt or
-   current.temperature != self._last_ui_snapshot.temperature or
+   abs(current.temperature - self._last_ui_snapshot.temperature) > 1e-5 or
-   current.top_p != self._last_ui_snapshot.top_p or
+   abs(current.top_p - self._last_ui_snapshot.top_p) > 1e-5 or
    current.max_tokens != self._last_ui_snapshot.max_tokens or
    current.auto_add_history != self._last_ui_snapshot.auto_add_history or
-   len(current.disc_entries) != len(self._last_ui_snapshot.disc_entries)
+   len(current.disc_entries) != len(self._last_ui_snapshot.disc_entries) or
+   len(current.files) != len(self._last_ui_snapshot.files) or
+   len(current.screenshots) != len(self._last_ui_snapshot.screenshots)
   )
   
+  if not changed and len(current.disc_entries) > 0:
+   if current.disc_entries[-1].get('content') != self._last_ui_snapshot.disc_entries[-1].get('content'):
+    changed = True
+
   if changed:
    if not self._pending_snapshot:
     self._pending_snapshot = True
     self._snapshot_timer = time.time()
+    # Capture state BEFORE current change
     self._state_to_push = self._last_ui_snapshot
    else:
+    # Reset timer for settle debounce
     self._snapshot_timer = time.time()
    
    self._last_ui_snapshot = current
@@ -1266,7 +1286,6 @@ class App:
  def _render_base_prompt_diff_modal(self) -> None:
   if not getattr(self.controller, "_show_base_prompt_diff_modal", False):
    return
-  
   imgui.open_popup("Base Prompt Diff")
   if imgui.begin_popup_modal("Base Prompt Diff", True, imgui.WindowFlags_.always_auto_resize)[0]:
    imgui.text_colored(C_IN, "Difference between Default and Custom Base System Prompt")

tests/test_undo_redo_sim.py

@@ -3,6 +3,7 @@ import time
 import sys
 import os
 import json
+from pathlib import Path
 
 sys.path.append(os.path.abspath(os.path.join(os.path.dirname(__file__), "..")))
 sys.path.append(os.path.abspath(os.path.join(os.path.dirname(__file__), "..", "src")))
@@ -15,46 +16,116 @@ def test_undo_redo_lifecycle(live_gui):
  assert client.wait_for_server(timeout=15), "Hook server did not start"
  
  # 1. Set initial state
+ print("Setting initial state...")
  client.set_value('temperature', 0.5)
  client.set_value('ai_input', "Initial Input")
- time.sleep(0.5) # Wait for debounce timer to start if it was triggered
  
- # Trigger a snapshot by waiting for debounce
+ # Wait for settle and first push (S_init -> S0)
- time.sleep(2.0) 
+ time.sleep(3.0) 
  
  # 2. Change state
+ print("Modifying state...")
  client.set_value('temperature', 1.5)
  client.set_value('ai_input', "Modified Input")
  
- # Wait for debounce to push "Modified" state to history
+ # Wait for settle and second push (S0 -> S1)
- time.sleep(2.0)
+ time.sleep(3.0)
  
  # Verify current state
- assert client.get_value('temperature') == 1.5
+ temp = client.get_value('temperature')
- assert client.get_value('ai_input') == "Modified Input"
+ ai_in = client.get_value('ai_input')
+ print(f"Current state: temp={temp}, ai_input={ai_in}")
+ assert temp == 1.5
+ assert ai_in == "Modified Input"
  
  # 3. Undo
  print("Sending Undo...")
- # Since we don't have a direct 'undo' hook, we can use 'click' if we added a button, 
- # or we can simulate the hotkey if the hook API supports it.
- # Actually, I'll add a 'btn_undo' and 'btn_redo' to clickable actions for testing.
  client.click('btn_undo')
  
  # Wait for state to revert
- time.sleep(0.5)
+ time.sleep(1.0)
  
- # Should be back to initial state
+ ai_in_undo = client.get_value('ai_input')
- # Note: Undo moves 'current' to redo stack and pops from undo.
+ temp_undo = client.get_value('temperature')
- # If we were at 'Modified', and we undo, we should get 'Initial'.
+ print(f"After undo: ai_input={ai_in_undo}, temp={temp_undo}")
- assert client.get_value('ai_input') == "Initial Input"
+ 
- assert client.get_value('temperature') == 0.5
+ assert ai_in_undo == "Initial Input"
+ assert temp_undo == 0.5
  
  # 4. Redo
  print("Sending Redo...")
  client.click('btn_redo')
+ time.sleep(1.0)
+ 
+ ai_in_redo = client.get_value('ai_input')
+ temp_redo = client.get_value('temperature')
+ print(f"After redo: ai_input={ai_in_redo}, temp={temp_redo}")
+ 
+ assert ai_in_redo == "Modified Input"
+ assert temp_redo == 1.5
+ 
+ print("Undo/Redo basic lifecycle PASSED.")
+
+@pytest.mark.integration
+def test_undo_redo_discussion_mutation(live_gui):
+ client = api_hook_client.ApiHookClient()
+ assert client.wait_for_server(timeout=15)
+ 
+ # Get initial entries count
+ initial_entries = client.get_value('disc_entries')
+ initial_count = len(initial_entries)
+ print(f"Initial entries: {initial_count}")
+ 
+ # 1. Add an entry (we simulate this by appending to disc_entries)
+ # Wait for settle
+ time.sleep(2.0)
+ 
+ new_entries = initial_entries + [{"role": "User", "content": "New Entry", "collapsed": False, "ts": "2026-03-11 12:00:00"}]
+ client.set_value('disc_entries', new_entries)
+ 
+ # Wait for debounce
+ time.sleep(2.0)
+ assert len(client.get_value('disc_entries')) == initial_count + 1
+ 
+ # 2. Undo addition
+ print("Undoing entry addition...")
+ client.click('btn_undo')
  time.sleep(0.5)
+ assert len(client.get_value('disc_entries')) == initial_count
  
- assert client.get_value('ai_input') == "Modified Input"
+ # 3. Redo addition
- assert client.get_value('temperature') == 1.5
+ print("Redoing entry addition...")
+ client.click('btn_redo')
+ time.sleep(0.5)
+ assert len(client.get_value('disc_entries')) == initial_count + 1
  
- print("Undo/Redo lifecycle test PASSED.")
+ print("Undo/Redo discussion mutation PASSED.")
+
+@pytest.mark.integration
+def test_undo_redo_context_mutation(live_gui):
+ client = api_hook_client.ApiHookClient()
+ assert client.wait_for_server(timeout=15)
+ 
+ # Wait for settle
+ time.sleep(2.0)
+ 
+ # 1. Add a file
+ client.set_value('ui_file_paths', ['test_undo.py'])
+ 
+ # Wait for debounce
+ time.sleep(2.0)
+ assert 'test_undo.py' in client.get_value('ui_file_paths')
+ 
+ # 2. Undo addition
+ print("Undoing file addition...")
+ client.click('btn_undo')
+ time.sleep(0.5)
+ assert 'test_undo.py' not in client.get_value('ui_file_paths')
+ 
+ # 3. Redo addition
+ print("Redoing file addition...")
+ client.click('btn_redo')
+ time.sleep(0.5)
+ assert 'test_undo.py' in client.get_value('ui_file_paths')
+ 
+ print("Undo/Redo context mutation PASSED.")