Skip to content

High-performance browser automation bridge and multi-instance orchestrator with advanced stealth injection and real-time dashboard.

chore: remove test-e2e.sh (replaced by tests/integration/orchestrator_test.go)

Mariob9d0a67c3eb0
diff --git a/memory/LATEST.md b/memory/LATEST.md
new file mode 100644
index 00000000..eea6685f
--- /dev/null
+++ b/memory/LATEST.md
@@ -0,0 +1,83 @@
+# Latest Session Work - 2026-02-28 Evening
+
+## What Was Done
+
+### 1. Created Phase 7 & 8 Proposals
+- **docs/proposals/phase-7-performance-optimization.md** (287 lines)
+ - Benchmarking: latency, memory, concurrent limits
+ - Optimizations: connection pooling, caching, batch navigation
+ - Scaling tests: 100+ instances, multi-agent load
+
+- **docs/proposals/phase-8-dashboard-ui-enhancements.md** (414 lines)
+ - Instance monitoring: Chrome status, memory/CPU, tab activity
+ - Batch operations: create multiple, terminate all, port config
+ - Navigation: search/filter, details modal, live logs
+
+- **docs/proposals/README.md** (176 lines)
+ - Decision framework for Phase 7 vs 8
+ - Proposal structure and usage guide
+
+### 2. Reorganized Testing
+Moved test logic from standalone `test-e2e.sh` into proper test structure:
+
+**Automated Integration Tests** → `tests/integration/orchestrator_test.go`
+- 11 comprehensive Go tests
+- Instance creation, IDs, ports, isolation, proxy routing
+- Run: `go test -tags integration ./tests/integration -run Orchestrator`
+
+**Manual Tests** → `tests/manual/orchestrator.md`
+- 16 detailed manual test scenarios
+- Visual (headed/headless), monitoring, port management, UI, edge cases
+- Checklist format for comprehensive validation
+
+**Updated** → `TESTING.md`
+- Added "Test Organization" section
+- References to automated and manual test guides
+- Maintains backwards-compatible quick-start examples
+
+### 3. Commits Made
+- `543dad0` - Phase 7 & 8 proposals
+- `754fc5b` - Test reorganization (orchestrator_test.go + orchestrator.md)
+- `ce29149` - Removed session memory file
+
+## Status
+
+✅ All 195+ unit tests passing
+✅ Pre-commit checks passing (gofmt, vet, test, docs validation)
+✅ All changes pushed to `origin/feat/make-cli-useful`
+✅ Current HEAD: `ce29149`
+
+## Key Files
+
+### Documentation
+- `docs/proposals/` - Phase 7 & 8 detailed roadmap (877 LOC)
+- `TESTING.md` - Test organization guide
+
+### Tests
+- `tests/integration/orchestrator_test.go` - 11 automated tests (459 LOC)
+- `tests/manual/orchestrator.md` - 16 manual test scenarios (518 LOC)
+
+## Next Steps (For Future Sessions)
+
+1. **Implement Phase 7** (performance optimization)
+ - Start benchmarking (low-hanging fruit)
+ - Implement connection pooling (1-2 hours)
+ - Add caching layer (1-2 hours)
+
+2. **Implement Phase 8** (dashboard UI)
+ - Instance monitoring screen (3-4 hours)
+ - Batch operations (3-4 hours)
+ - Search/filter and details modal (4-5 hours)
+
+3. **Consider merging to main**
+ - Phase 6 is feature-complete
+ - All tests passing
+ - Ready for production deployment
+ - Optional: Run full manual test suite first
+
+## Branch Info
+
+- **Branch**: `feat/make-cli-useful`
+- **Latest**: `ce29149` (session work)
+- **Tests**: All passing
+- **Documentation**: Complete and validated
diff --git a/test-e2e.sh b/test-e2e.sh
deleted file mode 100755
index d6c749e1..00000000
--- a/test-e2e.sh
+++ /dev/null
@@ -1,223 +0,0 @@
-#!/bin/bash
-
-# Phase 6: End-to-End Test for Multi-Instance Architecture
-# This script tests:
-# - Hash-based ID generation
-# - Auto-port allocation
-# - Instance isolation
-# - Orchestrator proxy routes
-# - Port cleanup and reuse
-
-set -e
-
-# Colors for output
-GREEN='\033[0;32m'
-RED='\033[0;31m'
-YELLOW='\033[1;33m'
-NC='\033[0m' # No Color
-
-echo -e "${YELLOW}Phase 6: End-to-End Testing${NC}"
-echo ""
-
-# Check if Pinchtab binary exists
-if [ ! -f "./pinchtab" ]; then
- echo -e "${RED}✗ Pinchtab binary not found. Run: go build -o pinchtab ./cmd/pinchtab${NC}"
- exit 1
-fi
-
-# Kill any existing Pinchtab instances
-pkill -f "./pinchtab" 2>/dev/null || true
-sleep 1
-
-# Start Pinchtab
-echo -e "${YELLOW}Starting Pinchtab dashboard...${NC}"
-./pinchtab &
-DASHBOARD_PID=$!
-sleep 2
-
-echo -e "${GREEN}✓ Dashboard started (PID: $DASHBOARD_PID)${NC}"
-echo ""
-
-# Test 1: Create Instance 1 (Headed)
-echo -e "${YELLOW}Test 1: Create headed instance...${NC}"
-INST1=$(curl -s -X POST http://localhost:9867/instances/launch \
- -H "Content-Type: application/json" \
- -d '{"name":"work","headless":false}')
-
-INST1_ID=$(echo $INST1 | jq -r '.id')
-INST1_PORT=$(echo $INST1 | jq -r '.port')
-INST1_PROFILE=$(echo $INST1 | jq -r '.profileId')
-
-if [[ ! $INST1_ID =~ ^inst_ ]]; then
- echo -e "${RED}✗ FAILED: Instance ID should have inst_ prefix, got: $INST1_ID${NC}"
- kill $DASHBOARD_PID 2>/dev/null || true
- exit 1
-fi
-
-if [[ ! $INST1_PROFILE =~ ^prof_ ]]; then
- echo -e "${RED}✗ FAILED: Profile ID should have prof_ prefix, got: $INST1_PROFILE${NC}"
- kill $DASHBOARD_PID 2>/dev/null || true
- exit 1
-fi
-
-echo -e "${GREEN}✓ Instance 1 created: $INST1_ID (port: $INST1_PORT)${NC}"
-echo ""
-
-# Test 2: Create Instance 2 (Headless)
-echo -e "${YELLOW}Test 2: Create headless instance...${NC}"
-INST2=$(curl -s -X POST http://localhost:9867/instances/launch \
- -H "Content-Type: application/json" \
- -d '{"name":"scrape","headless":true}')
-
-INST2_ID=$(echo $INST2 | jq -r '.id')
-INST2_PORT=$(echo $INST2 | jq -r '.port')
-
-if [ "$INST1_PORT" = "$INST2_PORT" ]; then
- echo -e "${RED}✗ FAILED: Instances should have different ports${NC}"
- kill $DASHBOARD_PID 2>/dev/null || true
- exit 1
-fi
-
-echo -e "${GREEN}✓ Instance 2 created: $INST2_ID (port: $INST2_PORT)${NC}"
-echo -e "${GREEN}✓ Port allocation working (9868 and 9869)${NC}"
-echo ""
-
-# Test 3: Wait for Chrome initialization
-echo -e "${YELLOW}Test 3: Wait for Chrome initialization...${NC}"
-sleep 3
-
-# Check health
-HEALTH1=$(curl -s http://localhost:$INST1_PORT/health | jq -r '.status')
-HEALTH2=$(curl -s http://localhost:$INST2_PORT/health | jq -r '.status')
-
-if [ "$HEALTH1" != "ok" ] || [ "$HEALTH2" != "ok" ]; then
- echo -e "${RED}✗ FAILED: Health check failed (status 1: $HEALTH1, status 2: $HEALTH2)${NC}"
- kill $DASHBOARD_PID 2>/dev/null || true
- exit 1
-fi
-
-echo -e "${GREEN}✓ Both instances healthy${NC}"
-echo -e "${GREEN}✓ Chrome initialized successfully${NC}"
-echo ""
-
-# Test 4: Navigate via orchestrator proxy
-echo -e "${YELLOW}Test 4: Navigate via orchestrator proxy...${NC}"
-NAV1=$(curl -s -X POST "http://localhost:9867/instances/$INST1_ID/navigate" \
- -H "Content-Type: application/json" \
- -d '{"url":"https://example.com"}')
-
-TAB1_ID=$(echo $NAV1 | jq -r '.tabId')
-
-if [[ ! $TAB1_ID =~ ^tab_ ]]; then
- echo -e "${RED}✗ FAILED: Tab ID should have tab_ prefix, got: $TAB1_ID${NC}"
- kill $DASHBOARD_PID 2>/dev/null || true
- exit 1
-fi
-
-echo -e "${GREEN}✓ Instance 1 navigated to example.com${NC}"
-echo -e "${GREEN}✓ Tab ID generated: $TAB1_ID${NC}"
-echo ""
-
-# Test 5: Navigate instance 2
-echo -e "${YELLOW}Test 5: Navigate second instance...${NC}"
-NAV2=$(curl -s -X POST "http://localhost:9867/instances/$INST2_ID/navigate" \
- -H "Content-Type: application/json" \
- -d '{"url":"https://github.com"}')
-
-TAB2_ID=$(echo $NAV2 | jq -r '.tabId')
-
-if [ "$TAB1_ID" = "$TAB2_ID" ]; then
- echo -e "${RED}✗ FAILED: Tab IDs should be different (isolation)${NC}"
- kill $DASHBOARD_PID 2>/dev/null || true
- exit 1
-fi
-
-echo -e "${GREEN}✓ Instance 2 navigated to github.com${NC}"
-echo -e "${GREEN}✓ Tab isolation verified (different IDs)${NC}"
-echo ""
-
-# Test 6: List instances
-echo -e "${YELLOW}Test 6: List all instances...${NC}"
-INSTANCES=$(curl -s http://localhost:9867/instances | jq '.[] | .id' | wc -l)
-
-if [ "$INSTANCES" -ne 2 ]; then
- echo -e "${RED}✗ FAILED: Expected 2 instances, got $INSTANCES${NC}"
- kill $DASHBOARD_PID 2>/dev/null || true
- exit 1
-fi
-
-echo -e "${GREEN}✓ Both instances listed correctly${NC}"
-echo ""
-
-# Test 7: Stop instance 1
-echo -e "${YELLOW}Test 7: Stop instance 1...${NC}"
-curl -s -X POST "http://localhost:9867/instances/$INST1_ID/stop" > /dev/null
-sleep 1
-
-REMAINING=$(curl -s http://localhost:9867/instances | jq '.[] | .id' | wc -l)
-
-if [ "$REMAINING" -ne 1 ]; then
- echo -e "${RED}✗ FAILED: Expected 1 instance after stopping, got $REMAINING${NC}"
- kill $DASHBOARD_PID 2>/dev/null || true
- exit 1
-fi
-
-echo -e "${GREEN}✓ Instance 1 stopped${NC}"
-echo -e "${GREEN}✓ Port 9868 released${NC}"
-echo ""
-
-# Test 8: Port reuse
-echo -e "${YELLOW}Test 8: Test port reuse...${NC}"
-INST3=$(curl -s -X POST http://localhost:9867/instances/launch \
- -H "Content-Type: application/json" \
- -d '{"name":"reuse","headless":true}')
-
-INST3_PORT=$(echo $INST3 | jq -r '.port')
-
-if [ "$INST3_PORT" != "9868" ]; then
- echo -e "${RED}✗ FAILED: Port should be reused (9868), got $INST3_PORT${NC}"
- kill $DASHBOARD_PID 2>/dev/null || true
- exit 1
-fi
-
-echo -e "${GREEN}✓ Port 9868 reused correctly${NC}"
-echo ""
-
-# Test 9: Cleanup
-echo -e "${YELLOW}Test 9: Cleanup...${NC}"
-INST3_ID=$(echo $INST3 | jq -r '.id')
-
-curl -s -X POST "http://localhost:9867/instances/$INST2_ID/stop" > /dev/null
-curl -s -X POST "http://localhost:9867/instances/$INST3_ID/stop" > /dev/null
-sleep 1
-
-FINAL=$(curl -s http://localhost:9867/instances | jq '.[] | .id' | wc -l)
-
-if [ "$FINAL" -ne 0 ]; then
- echo -e "${RED}✗ FAILED: Expected 0 instances after cleanup, got $FINAL${NC}"
- kill $DASHBOARD_PID 2>/dev/null || true
- exit 1
-fi
-
-echo -e "${GREEN}✓ All instances cleaned up${NC}"
-echo ""
-
-# Cleanup
-kill $DASHBOARD_PID 2>/dev/null || true
-sleep 1
-
-# Summary
-echo ""
-echo -e "${GREEN}═══════════════════════════════════════${NC}"
-echo -e "${GREEN}✅ ALL TESTS PASSED!${NC}"
-echo -e "${GREEN}═══════════════════════════════════════${NC}"
-echo ""
-echo "Tested:"
-echo " ✓ Hash-based ID generation (prof_X, inst_X, tab_X)"
-echo " ✓ Auto-port allocation (9868, 9869, ...)"
-echo " ✓ Port release and reuse"
-echo " ✓ Instance isolation (different tab IDs)"
-echo " ✓ Orchestrator proxy routes"
-echo " ✓ Health checks on instances"
-echo " ✓ Instance creation and cleanup"
-echo ""