Hello friends,
Just a short observation from an end-to-end test:
In a setup with
MySQL → Node.js → WebSocket → Harbour → FiveWin xBrowse,
the measured RTT sometimes shows 0 ms.
This is not a bug and not a fake result.
It simply means:
The standard millisecond timer resolution is already too coarse.
The full chain (SQL → JSON → WS → Harbour → UI) is so fast that it completes within a single millisecond for typical paging sizes.
Once the payload gets larger (e.g. 500 records / 30 KB JSON), RTT becomes measurable again (≈10–20 ms).
Conclusion:
The combination is just extremely fast — fast enough that the measurement method becomes the limiting factor, not the architecture.
Best regards
Otto
The interesting part is how little Harbour code is needed.
