M2.8: switch over to the new loop

nisse-dev now runs only the conversational loop. The gated state machine
it was built alongside is gone:

- Removed control.py, planner.py, executor.py, approvals.py, gates.py
  (command_form moves to loop/reader.py), publish.py and poller.py, the
  old assembly in agent.py (its shared helpers move to loop/build.py),
  the store's state-machine view (get_state, transition), their tests,
  the old end-to-end and live planning/milestone/gate tests, and
  specs/orchestrator.qnt with its scenarios.
- Roles are driver, helper, expert and classifier; driver is required.
  loop:, roles.planner, roles.executor and tuning.executor are refused
  with what to use instead; tuning.driver takes the turn cap and failure
  limit.
- The benchmark and the capability matrix measure the driver: the
  loop's own system prompt and tools, the task's plan as the agreed
  restatement. Task plans end with finish_milestone, so earlier runs
  show as (old); stored runs' executor tuning reads as the driver's.
- A delegation on the loop records what its helper's alias resolved to
  (L9), as the old path's did.
- The dump shows a work item's revision and what it waits for.
- Behaviour the old tests covered and the loop shares stays covered:
  driver runs that fail on a model call, stop mid-turn, reset the
  failure count; go-aheads ignoring case and punctuation; a request too
  big to fit is never sent; the assembly tests, now on build_loop_agent.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PPQ14L3LikyefCMFBqEbKE

nisse-dev/src/nisse_dev/config.py

@@ -17,7 +17,7 @@ silently ignored would look exactly like a knob that has no effect.
 import os
 import re
 from collections.abc import Mapping
-from dataclasses import dataclass, field, fields, replace
+from dataclasses import dataclass, field, fields
 from pathlib import Path
 from typing import Any
 
@@ -32,8 +32,7 @@ REQUIRED_KEYS = (
     "state_db",
     "llm.base_url",
     "llm.token",
-    "roles.executor",
-    "roles.planner",
+    "roles.driver",
     "roles.helper",
     "roles.expert",
     "matrix.homeserver",
@@ -46,12 +45,16 @@ REQUIRED_KEYS = (
     "gitea.operators",
 )
 OPTIONAL_KEYS = (
-    "roles.driver",
     "roles.classifier",
     "llm.max_output_tokens",
     "llm.request_timeout_s",
-    "loop",
 )
+# Settings the old loop had (until M2.8), and what to do instead.
+RETIRED_KEYS = {
+    "loop": "the loop is the only one now; remove it",
+    "roles.planner": "use roles.driver",
+    "roles.executor": "the driver does the work now; remove it",
+}
 # A project entry (D8), under `projects: {name: {...}}`.
 PROJECT_REQUIRED = ("repo", "workspace", "build_host.address", "build_host.key")
 PROJECT_OPTIONAL = (
@@ -85,7 +88,7 @@ class ConfigError(RuntimeError):
 
 
 @dataclass(frozen=True)
-class ExecutorTuning:
+class DriverTuning:
     max_turns: int = 40
     failure_limit: int = 2
 
@@ -125,7 +128,7 @@ class LoopTuning:
 
 @dataclass(frozen=True)
 class Tuning:
-    executor: ExecutorTuning = field(default_factory=ExecutorTuning)
+    driver: DriverTuning = field(default_factory=DriverTuning)
     expert: ExpertTuning = field(default_factory=ExpertTuning)
     delegation: DelegationTuning = field(default_factory=DelegationTuning)
     tools: ToolsTuning = field(default_factory=ToolsTuning)
@@ -148,6 +151,9 @@ def _parse_tuning(raw: Any, errors: list[str]) -> Tuning:
     sections: dict[str, Any] = {}
     for name, section in raw.items():
         cls = _TUNING_SECTIONS.get(name)
+        if name == "executor":  # until M2.8
+            errors.append("tuning.executor: retired — it is tuning.driver now")
+            continue
         if cls is None:
             errors.append(f"tuning.{name}: unknown section")
             continue
@@ -474,60 +480,8 @@ class AgentConfig:
     max_output_tokens: int = DEFAULT_MAX_OUTPUT_TOKENS
     request_timeout_s: float = DEFAULT_REQUEST_TIMEOUT_S
     tuning: Tuning = field(default_factory=Tuning)
-    # Which loop runs: "old" (control.py) until the switch-over (M2.8), or
-    # "new" (nisse_dev.loop).
-    loop: str = "old"
     config_file: Path | None = None
 
-    # -- one project, for the old loop ----------------------------------------
-    # control.py (until M2.8) knows one project; these read it, and refuse a
-    # configuration with more.
-
-    def only_project(self) -> Project:
-        if len(self.projects) != 1:
-            names = ", ".join(p.name for p in self.projects)
-            raise ConfigError(
-                f"cannot start: the old loop runs exactly one project, but "
-                f"{len(self.projects)} are configured ({names}); use `loop: new`"
-            )
-        return self.projects[0]
-
-    def with_project(self, **changes: Any) -> "AgentConfig":
-        """This configuration with its one project changed (for tests)."""
-        return replace(self, projects=(replace(self.only_project(), **changes),))
-
-    @property
-    def project(self) -> str:
-        return self.only_project().name
-
-    @property
-    def gitea_repo(self) -> str:
-        return self.only_project().repo
-
-    @property
-    def workspace(self) -> Path:
-        return self.only_project().workspace
-
-    @property
-    def build_host(self) -> BuildHost:
-        return self.only_project().build_host
-
-    @property
-    def base_branch(self) -> str:
-        return self.only_project().base_branch
-
-    @property
-    def push_key(self) -> Path | None:
-        return self.only_project().push_key
-
-    @property
-    def push_user(self) -> str | None:
-        return self.only_project().push_user
-
-    @property
-    def push_token(self) -> str | None:
-        return self.only_project().push_token
-
     @classmethod
     def load(cls, env: Mapping[str, str] = os.environ) -> "AgentConfig":
         """The file named by NISSE_CONFIG, or ./nisse-dev.yaml."""
@@ -573,7 +527,10 @@ class AgentConfig:
 
         known = set(REQUIRED_KEYS) | set(OPTIONAL_KEYS)
         for key in sorted(set(flat) - known):
-            errors.append(f"{key}: unknown setting")
+            if key in RETIRED_KEYS:
+                errors.append(f"{key}: retired — {RETIRED_KEYS[key]}")
+            else:
+                errors.append(f"{key}: unknown setting")
         values = {
             key: _resolve(value, key, env, errors)
             for key, value in flat.items()
@@ -607,9 +564,6 @@ class AgentConfig:
                 raise ValueError
         except (TypeError, ValueError):
             errors.append("llm.request_timeout_s: must be a positive number of seconds")
-        loop = str(values.get("loop") or "old")
-        if loop not in ("old", "new"):
-            errors.append("loop: must be old or new")
         if errors:
             raise ConfigError("cannot start: " + "; ".join(errors))
 
@@ -618,11 +572,9 @@ class AgentConfig:
             llm_base_url=str(values["llm.base_url"]),
             llm_token=str(values["llm.token"]),
             roles=RoleMap(
-                executor=str(values["roles.executor"]),
-                planner=str(values["roles.planner"]),
+                driver=str(values["roles.driver"]),
                 helper=str(values["roles.helper"]),
                 expert=str(values["roles.expert"]),
-                driver=str(values.get("roles.driver") or ""),
                 classifier=str(values.get("roles.classifier") or ""),
             ),
             matrix=MatrixConfig(
@@ -640,7 +592,6 @@ class AgentConfig:
             max_output_tokens=max_output,
             request_timeout_s=request_timeout,
             tuning=tuning,
-            loop=loop,
             config_file=config_file,
         )
         config._refuse_secrets_inside_the_clone(env)
@@ -685,7 +636,6 @@ class AgentConfig:
                 f"{p.name} ({p.repo}, {p.build_host.address}:{p.workspace})"
                 for p in self.projects
             ),
-            "planner": self.roles.planner,
-            "executor": self.roles.executor,
+            "driver": self.roles.driver,
             "helper": self.roles.helper,
         }