Fix confusing exit code 1 in QEMU status check

The grep command checking for crash patterns was returning exit code 1
when no crashes were found (normal grep behavior), making it unclear
what the actual failure was.

Changes:
- Redirect grep output to /dev/null with error suppression (> /dev/null 2>&1)
- Show first 20 lines of crash patterns if found
- Add clear "ERROR: HTTP server failed to start" message before exit
- Better structured output for debugging

This makes it clear the real failure is "HTTP server not responding"
rather than the crash detection check itself.

Co-authored-by: softhack007 <[email protected]>

Author: Copilot

.github/workflows/qemu-e2e-test.yml

@@ -200,11 +200,15 @@ jobs:
           tail -200 qemu-output.log || true
           echo ""
           echo "=== Checking for ESP32 exceptions/crashes ==="
-          if grep -i "exception\|abort\|backtrace\|panic" qemu-output.log; then
+          if grep -i "exception\|abort\|backtrace\|panic" qemu-output.log > /dev/null 2>&1; then
             echo "FOUND: Firmware crash detected in QEMU output"
+            echo ""
+            grep -i "exception\|abort\|backtrace\|panic" qemu-output.log | head -20
           else
             echo "No obvious crash patterns found"
           fi
+          echo ""
+          echo "ERROR: HTTP server failed to start - check QEMU logs above"
           exit 1
       
       - name: Run Playwright tests against QEMU