fix: CLI test, TUI integration harness — all non-blockers resolved

- gateway-cli: add load-time sanity check, fix FiveAM prefix for
  jailed-package compatibility
- TUI integration: switch all tests from file-grep to tmux capture-pane
  (agent-responds, cascade-failure, eval-command, connection-drop).
  Fixes file-buffering false negatives. Increase eval sleep to 3s.
- Cherry-pick: system-integration-tests.org org source updated

lisp/gateway-cli.lisp

@@ -15,17 +15,21 @@
   (ql:quickload :fiveam :silent t))
 
 (defpackage :passepartout-gateway-cli-tests
-  (:use :cl :fiveam :passepartout)
+  (:use :cl :passepartout)
   (:export #:cli-suite))
 
 (in-package :passepartout-gateway-cli-tests)
 
-(def-suite cli-suite :description "Verification of the CLI Gateway")
+(fiveam:def-suite cli-suite :description "Verification of the CLI Gateway")
-(in-suite cli-suite)
+(fiveam:in-suite cli-suite)
 
-(test test-gateway-cli-input-format
+(fiveam:test test-gateway-cli-input-format
   "Contract 1: gateway-cli-input injects a properly formed signal without error."
   (handler-case
-      (gateway-cli-input "hello")
+      (progn (gateway-cli-input "hello") (fiveam:is t))
     (error (c)
-      (fail "gateway-cli-input crashed: ~a" c))))
+      (fiveam:is nil "gateway-cli-input crashed: ~a" c))))
+
+(handler-case
+    (progn (gateway-cli-input "test-load") (log-message "CLI: Load-time test OK"))
+  (error (c) (log-message "CLI: Load-time test FAILED: ~a" c)))

org/gateway-cli.org

@@ -44,18 +44,29 @@ The CLI Gateway is the simplest interface to Passepartout — raw stdin/stdout o
   (ql:quickload :fiveam :silent t))
 
 (defpackage :passepartout-gateway-cli-tests
-  (:use :cl :fiveam :passepartout)
+  (:use :cl :passepartout)
   (:export #:cli-suite))
 
 (in-package :passepartout-gateway-cli-tests)
 
-(def-suite cli-suite :description "Verification of the CLI Gateway")
+(fiveam:def-suite cli-suite :description "Verification of the CLI Gateway")
-(in-suite cli-suite)
+(fiveam:in-suite cli-suite)
 
-(test test-gateway-cli-input-format
+(fiveam:test test-gateway-cli-input-format
   "Contract 1: gateway-cli-input injects a properly formed signal without error."
   (handler-case
-      (gateway-cli-input "hello")
+      (progn (gateway-cli-input "hello") (fiveam:is t))
     (error (c)
-      (fail "gateway-cli-input crashed: ~a" c))))
+      (fiveam:is nil "gateway-cli-input crashed: ~a" c))))
+#+end_src
+
+** Load-Time Sanity Check
+
+Verifies the function exists and can be called at load time without
+depending on FiveAM macro resolution in the jailed package.
+
+#+begin_src lisp
+(handler-case
+    (progn (gateway-cli-input "test-load") (log-message "CLI: Load-time test OK"))
+  (error (c) (log-message "CLI: Load-time test FAILED: ~a" c)))
 #+end_src

org/system-integration-tests.org

@@ -350,30 +350,26 @@ done
 
 test_agent_responds() {
   # Full round-trip: TUI → daemon → pipeline → TUI.
-  # Polls for ⬇ marker (any 3+ letter agent response).
+  # Uses tmux capture-pane to read the rendered screen.
   local before_ts
   before_ts=$(date +%s)
-
   tmux send-keys -t tui-test "hello" Enter
-
   while true; do
-    if grep -q '⬇.*[a-zA-Z]\{3,\}' "$TUI_LOG"; then
+    if tmux capture-pane -t tui-test -p -S -50 2>/dev/null | grep -q '⬇.*[a-zA-Z]\{3,\}'; then
       return 0
     fi
     local now_ts
     now_ts=$(date +%s)
     if (( now_ts - before_ts > 60 )); then
-      echo "TIMEOUT: no agent response in log after 60s" >&2
+      echo "TIMEOUT: no agent response after 60s" >&2
       return 1
     fi
-    sleep 1
+    sleep 2
   done
 }
 
 test_agent_not_cascade_failure() {
-  # Check if the ⬇ response is a cascade failure meaning no LLM backend is configured.
+  if tmux capture-pane -t tui-test -p -S -50 2>/dev/null | grep '⬇' | grep -qi 'cascade.*fail\|exhausted\|neural cascade'; then
-  # This is a WARNING, not a failure — the daemon is alive but needs an API key.
-  if grep '⬇' "$TUI_LOG" | grep -qi 'cascade.*fail\|exhausted\|neural cascade'; then
     echo "NOTE: LLM cascade failure — no API key configured (warning only)" >&2
     WARN=$((WARN + 1))
   fi
@@ -382,8 +378,8 @@ test_agent_not_cascade_failure() {
 
 test_eval_command() {
   tmux send-keys -t tui-test "/eval (+ 1 2)" Enter
-  sleep 2
+  sleep 3
-  grep -q '=> 3' "$TUI_LOG"
+  tmux capture-pane -t tui-test -p -S -10 2>/dev/null | grep -q '=> 3'
 }
 
 test_status_bar() {
@@ -391,14 +387,8 @@ test_status_bar() {
 }
 
 test_connection_drop() {
-  # Drop the daemon's connection to the TUI (the TUI notices and logs an error).
-  # We don't kill the daemon itself — just close its TCP listener temporarily.
-  # If the daemon and TUI share the same host, restarting the TUI's session
-  # forces a reconnect. This test is best-effort; in pure headless mode,
-  # the connection may already be lost from the TUI startup delay.
   sleep 1
-  grep -qi 'connection.*lost\|ERROR.*Connection\|error.*connect' "$TUI_LOG" || true
+  tmux capture-pane -t tui-test -p -S -10 2>/dev/null | grep -qi 'connection.*lost\|ERROR.*Connection\|error.*connect' || true
-  # Not a hard failure — connection monitoring depends on TUI rendering speed
   return 0
 }
 

test/integration-tui.sh

@@ -42,30 +42,26 @@ done
 
 test_agent_responds() {
   # Full round-trip: TUI → daemon → pipeline → TUI.
-  # Polls for ⬇ marker (any 3+ letter agent response).
+  # Uses tmux capture-pane to read the rendered screen.
   local before_ts
   before_ts=$(date +%s)
-
   tmux send-keys -t tui-test "hello" Enter
-
   while true; do
-    if grep -q '⬇.*[a-zA-Z]\{3,\}' "$TUI_LOG"; then
+    if tmux capture-pane -t tui-test -p -S -50 2>/dev/null | grep -q '⬇.*[a-zA-Z]\{3,\}'; then
       return 0
     fi
     local now_ts
     now_ts=$(date +%s)
     if (( now_ts - before_ts > 60 )); then
-      echo "TIMEOUT: no agent response in log after 60s" >&2
+      echo "TIMEOUT: no agent response after 60s" >&2
       return 1
     fi
-    sleep 1
+    sleep 2
   done
 }
 
 test_agent_not_cascade_failure() {
-  # Check if the ⬇ response is a cascade failure meaning no LLM backend is configured.
+  if tmux capture-pane -t tui-test -p -S -50 2>/dev/null | grep '⬇' | grep -qi 'cascade.*fail\|exhausted\|neural cascade'; then
-  # This is a WARNING, not a failure — the daemon is alive but needs an API key.
-  if grep '⬇' "$TUI_LOG" | grep -qi 'cascade.*fail\|exhausted\|neural cascade'; then
     echo "NOTE: LLM cascade failure — no API key configured (warning only)" >&2
     WARN=$((WARN + 1))
   fi
@@ -74,8 +70,8 @@ test_agent_not_cascade_failure() {
 
 test_eval_command() {
   tmux send-keys -t tui-test "/eval (+ 1 2)" Enter
-  sleep 2
+  sleep 3
-  grep -q '=> 3' "$TUI_LOG"
+  tmux capture-pane -t tui-test -p -S -10 2>/dev/null | grep -q '=> 3'
 }
 
 test_status_bar() {
@@ -83,14 +79,8 @@ test_status_bar() {
 }
 
 test_connection_drop() {
-  # Drop the daemon's connection to the TUI (the TUI notices and logs an error).
-  # We don't kill the daemon itself — just close its TCP listener temporarily.
-  # If the daemon and TUI share the same host, restarting the TUI's session
-  # forces a reconnect. This test is best-effort; in pure headless mode,
-  # the connection may already be lost from the TUI startup delay.
   sleep 1
-  grep -qi 'connection.*lost\|ERROR.*Connection\|error.*connect' "$TUI_LOG" || true
+  tmux capture-pane -t tui-test -p -S -10 2>/dev/null | grep -qi 'connection.*lost\|ERROR.*Connection\|error.*connect' || true
-  # Not a hard failure — connection monitoring depends on TUI rendering speed
   return 0
 }