fix(delegate): submit also takes the target workspace's update admission (Issue #72, review R1)

Decision 8: delegate submit uses the admission lock of the workspace the
target session belongs to. The wrapper resolved only its own
WORKSPACE_ROOT, so a job could be registered and delivered to an agent of
another installed workspace while that workspace was being updated.

lib_py.update target-workspace resolves a herdr:<name> row's workspace (the
.mam at or above its pane.cwd, else its control_plane.workspace_root);
mam_update_admit takes an fd, and submit holds the caller root on fd 7 and
the target on fd 8 until the record is written. Rowless and remote targets
keep the caller-root gate.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
This commit is contained in:
GodopuandClaude Opus 5.5 committed 2026-10-27 08:15:23 +00:00
1 parent 595e34b2d1
commit e3394655b3
4 files changed
+123 -16

No files matched your search

+10 -9
View File
@@ -3264,26 +3264,27 @@ mam_control_plane_prefix() {
# fd 7 (`7>&-` on such a child, as pane_lock's fd 9). The wait is bounded by
# MAM_UPDATE_ADMISSION_WAIT_S (default 120).
#
# mam_update_admit [workspace]: rc 0 admitted (fd 7 held, unless the workspace
# has no .mam), 1 refused (stderr).
# mam_update_admit [workspace [fd]]: rc 0 admitted (fd 7, or the given one,
# held unless the workspace has no .mam), 1 refused (stderr). Delegate submit
# also takes its target's workspace on fd 8.
mam_update_admit() {
local root="${1:-$WORKSPACE_ROOT}"
local root="${1:-$WORKSPACE_ROOT}" fd="${2:-7}"
local lock="$root/.mam/update-admission.lock" marker="$root/.mam/update-in-progress" st
# No .mam: not an installed workspace, which is all the update runs in (and a
# checkout running its own scripts gets no .mam from this).
[ -d "$root/.mam" ] || return 0
if ! exec 7>>"$lock"; then
if ! eval "exec $fd>>\"\$lock\""; then
echo "ERROR: cannot open the update admission lock $lock" >&2
return 1
fi
case "$(PYTHONPATH="$SKILL_DIR:${PYTHONPATH:-}" python3 -m lib_py.pane_lock acquire --fd 7 --shared --wait-env MAM_UPDATE_ADMISSION_WAIT_S)" in
case "$(PYTHONPATH="$SKILL_DIR:${PYTHONPATH:-}" python3 -m lib_py.pane_lock acquire --fd "$fd" --shared --wait-env MAM_UPDATE_ADMISSION_WAIT_S)" in
locked) ;;
timeout)
exec 7>&-
mam_update_admit_release "$fd"
echo "ERROR: an update of $root holds its admission lock (MAM_UPDATE_ADMISSION_WAIT_S passed); try again when it has finished" >&2
return 1 ;;
*)
exec 7>&-
mam_update_admit_release "$fd"
echo "ERROR: the update admission lock $lock could not be decided (lib_py.pane_lock)" >&2
return 1 ;;
esac
@@ -3296,11 +3297,11 @@ mam_update_admit() {
echo "NOTE: stale update marker $marker (its update is gone); ignored" >&2
return 0 ;;
esac
exec 7>&-
mam_update_admit_release "$fd"
echo "ERROR: an update of $root is in progress (marker $marker, holder $st); try again when it has finished" >&2
return 1
}
mam_update_admit_release() {
exec 7>&-
eval "exec ${1:-7}>&-"
}
+40
View File
@@ -16,6 +16,11 @@ this module what needs parsing. Standard library only, Python 3.9.
python -m lib_py.update monitor --workspace W
JSON {"pids": [...], "lock": "free"|"held"}: the monitor daemons
daemon.find_running recognises, and whether .mam/monitor.lock is held.
python -m lib_py.update target-workspace --workspace W --agent-session S
The installed workspace a delegate target `herdr:<name>` belongs to
(decision 8), when it is not W: the nearest directory holding a .mam
at or above its row's pane.cwd, else the row's control_plane
workspace_root. Nothing for a rowless or non-herdr target.
python -m lib_py.update backup-registry --workspace W --dest D
Copies the registry into D under the registry's own write lock
(agent-sessions.db.lock, as atomic_yaml takes it): the DB through the
@@ -231,6 +236,31 @@ def precheck(workspace: str, state: dict, herdr: str, env: Mapping[str, str]) ->
return blockers
# --------------------------------------------------------------------------- delegate target
def target_workspace(workspace: str, state: dict, agent_session: str) -> str:
if not agent_session.startswith("herdr:"):
return ""
name = agent_session[len("herdr:"):]
row = next((s for s in state.get("herdr_sessions") or []
if isinstance(s, dict) and s.get("name") == name), None)
if row is None:
return ""
found = ""
cwd = (row.get("pane") or {}).get("cwd") or ""
d = _real(cwd) if cwd else ""
while d and not found:
if os.path.isdir(os.path.join(d, ".mam")):
found = d
elif os.path.dirname(d) == d:
break
d = os.path.dirname(d)
if not found:
root = (row.get("control_plane") or {}).get("workspace_root") or ""
found = _real(root) if root else ""
return "" if found == _real(workspace) else found
# --------------------------------------------------------------------------- monitor
def monitor_state(workspace: str) -> dict:
@@ -288,6 +318,9 @@ def main(argv: Optional[List[str]] = None) -> int:
s.add_argument("--herdr", required=True)
s = sub.add_parser("monitor")
s.add_argument("--workspace", required=True)
s = sub.add_parser("target-workspace")
s.add_argument("--workspace", required=True)
s.add_argument("--agent-session", required=True)
s = sub.add_parser("backup-registry")
s.add_argument("--workspace", required=True)
s.add_argument("--dest", required=True)
@@ -303,6 +336,13 @@ def main(argv: Optional[List[str]] = None) -> int:
for line in lines:
print(line)
return 1 if lines else 0
if a.cmd == "target-workspace":
try:
state = _state()
except ValueError:
state = {}
print(target_workspace(ws, state, a.agent_session))
return 0
if a.cmd == "monitor":
print(json.dumps(monitor_state(ws)))
return 0
@@ -301,10 +301,19 @@ cmd_submit() {
local reg_dry=()
if [[ "$DRY_RUN" == "1" ]]; then reg_dry=(--dry-run); fi
# Issue #72: no job is registered while an update of the target's workspace
# runs; held until the record is written (lib.sh mam_update_admit). A
# dry-run record is never dispatched and is not counted by the update.
if [[ "$DRY_RUN" != "1" ]]; then mam_update_admit "$WORKSPACE_ROOT" || exit 1; fi
# Issue #72: no job is registered while an update of this registry's
# workspace or of the target's (its row's workspace, lib_py.update
# target-workspace) runs; held until the record is written (lib.sh
# mam_update_admit). A dry-run record is never dispatched and is not
# counted by the update.
if [[ "$DRY_RUN" != "1" ]]; then
local target_ws
target_ws="$(MAM_STATE_JSON="$(load_state_json)" PYTHONPATH="$SKILL_DIR:${PYTHONPATH:-}" "$PY" -m lib_py.update \
target-workspace --workspace "$WORKSPACE_ROOT" --agent-session "$AGENT_SESSION")" \
|| { echo "ERROR: cannot resolve the workspace of $AGENT_SESSION" >&2; exit 1; }
mam_update_admit "$WORKSPACE_ROOT" || exit 1
if [[ -n "$target_ws" ]]; then mam_update_admit "$target_ws" 8 || exit 1; fi
fi
# 1) register job (prints the new job id)
JOB_ID="$("$PY" "$SCRIPT_DIR/scripts/registry.py" --registry-dir "$REGISTRY_DIR" register \
"${reg_prompt[@]}" --agent "$AGENT" --agent-session "$AGENT_SESSION" --role "$DELEGATE_ROLE" \
@@ -314,7 +323,7 @@ cmd_submit() {
--job-type "$TYPE" --reviewer "$REVIEWER" --reviewer-session "$REVIEWER_SESSION" \
--max-iterations "$MAX_ITERATIONS" ${reg_parent[@]+"${reg_parent[@]}"} ${reg_dry[@]+"${reg_dry[@]}"})"
echo "registered job: $JOB_ID"
[[ "$DRY_RUN" == "1" ]] || mam_update_admit_release
[[ "$DRY_RUN" == "1" ]] || { mam_update_admit_release; mam_update_admit_release 8; }
# 1-1) Provision job directory and write direct brief.md (MAM Job Restructuring)
local job_dir="$REGISTRY_DIR/$JOB_ID"