fix: make integration services fail fast

tests/integration/start.sh

@@ -28,6 +28,19 @@ ghcr.io/hoytech/strfry:latest
 rm -Rf lnbits_itest_data
 unzip data.zip
 
+# The fixture was created with the standalone tpos extension installed. tpos
+# is no longer part of the LNbits dev tree, so leaving its database metadata in
+# the fixture makes current LNbits attempt to import a module that is absent.
+# The integration suite only exercises nwcprovider.
+python3 - <<'PY'
+import sqlite3
+
+with sqlite3.connect("lnbits_itest_data/database.sqlite3") as conn:
+    conn.execute("DELETE FROM installed_extensions WHERE id = 'tpos'")
+    conn.execute("DELETE FROM dbversions WHERE db = 'tpos'")
+PY
+rm -f lnbits_itest_data/ext_tpos.sqlite3 lnbits_itest_data/zips/tpos.zip
+
 id=$(id -u)
 gid=$(id -g)
 
@@ -45,24 +58,50 @@ docker run --name=lnbits_nwcprovider_ext_lnbits_test \
 -v ${PWD}/../../.devcontainer/pre-setup.sh:/pre-setup.sh:ro \
 mcr.microsoft.com/devcontainers/python:1-3.12 bash -c "while true; do sleep 1000; done"
 
-docker network create lnbits_nwcprovider_ext_test_network || true
-docker network connect lnbits_nwcprovider_ext_test_network lnbits_nwcprovider_ext_nostr_test --alias nostr|| true
-docker network connect lnbits_nwcprovider_ext_test_network lnbits_nwcprovider_ext_lnbits_test --alias lnbits|| true
+if ! docker network inspect lnbits_nwcprovider_ext_test_network >/dev/null 2>&1; then
+    docker network create lnbits_nwcprovider_ext_test_network
+fi
+docker network connect lnbits_nwcprovider_ext_test_network lnbits_nwcprovider_ext_nostr_test --alias nostr
+docker network connect lnbits_nwcprovider_ext_test_network lnbits_nwcprovider_ext_lnbits_test --alias lnbits
 
 docker exec -u root lnbits_nwcprovider_ext_lnbits_test bash -c "id -u $id &>/dev/null || useradd -m -u $id tester"
 docker exec -u root lnbits_nwcprovider_ext_lnbits_test bash -c "bash /pre-setup.sh"
 docker exec --user $id:$gid lnbits_nwcprovider_ext_lnbits_test  bash -c "curl -sSL https://install.python-poetry.org | python3 -"
 
-set +e
 docker exec --user $id:$gid lnbits_nwcprovider_ext_lnbits_test bash -c "export PATH=\"\$HOME/.local/bin:\$PATH\" && bash /setup.sh /nwcprovider"
 docker exec --user $id:$gid lnbits_nwcprovider_ext_lnbits_test bash -c "ln -s /app/.env \$HOME/lnbits/.env"
 
 
-ARGS=""
 if [ "$HEADLESS" != "" ];
 then
-    ARGS="-d"
-fi
+    # Keep the server log inside the container so a failed health check can
+    # show the actual startup error instead of silently swallowing it.
+    docker exec --user $id:$gid -d lnbits_nwcprovider_ext_lnbits_test bash -c "export PATH=\"\$HOME/.local/bin:\$PATH\" && cd \$HOME/lnbits && poetry run lnbits > /tmp/lnbits.log 2>&1"
 
-docker exec --user $id:$gid $ARGS lnbits_nwcprovider_ext_lnbits_test bash -c "export PATH=\"\$HOME/.local/bin:\$PATH\" && cd \$HOME/lnbits && poetry run lnbits"
- 
+    wait_for_http() {
+        local service="$1"
+        local url="$2"
+        local timeout_seconds="$3"
+        local deadline=$((SECONDS + timeout_seconds))
+
+        until curl --fail --silent --show-error --max-time 2 "$url" >/dev/null 2>&1; do
+            if [ "$SECONDS" -ge "$deadline" ]; then
+                echo "Timed out waiting for $service at $url" >&2
+                docker ps -a >&2
+                if [ "$service" = "LNbits" ]; then
+                    docker exec lnbits_nwcprovider_ext_lnbits_test tail -n 100 /tmp/lnbits.log >&2 || true
+                else
+                    docker logs --tail 100 lnbits_nwcprovider_ext_nostr_test >&2 || true
+                fi
+                return 1
+            fi
+            sleep 1
+        done
+    }
+
+    # LNbits may need a few minutes for a fresh database migration.
+    wait_for_http "nostr relay" "http://localhost:7777" 180
+    wait_for_http "LNbits" "http://localhost:5002" 180
+else
+    docker exec --user $id:$gid lnbits_nwcprovider_ext_lnbits_test bash -c "export PATH=\"\$HOME/.local/bin:\$PATH\" && cd \$HOME/lnbits && poetry run lnbits"
+fi