Pier-Yves Lessard
Embedded software engineer working at NXP semiconductor on embedded AI optimization for the automotive industry. Past experience in EV/motor control software. Author of 2 (relatively) widely used open source library dedicated to ECU communications and the main developer of Scrutiny Debugger, a project soon to be released. Father of two who develops open source stuff between 21h and 00h
Sessions
Debugging and testing an embedded application is always painful. A serial printf
might not be enough, a high end JTAG with 1000+ pages of documentation might be too costly or complex.
Scrutiny Debugger is a new open source project that offers an alternative by enabling remote control of the memory through any communication channel (Serial, UDP, etc.). How does that work? A Python server continuously communicates with an embedded application that runs a small instrumentation library. Using the debugging symbols, extracted at compile time, the server exposes all the variables and memory structure to client applications through a websocket API. 2 clients are available: an Electron GUI and a Python SDK for programmatic interaction with the server.
Clients can read/write variables or raw memory. They can do graphs of variables; being continuous time logging or embedded graphs that triggers on a specific variable change, like an electronic scope does. Not the best for low-level driver development; but ideal for high-level embedded application.
The Python SDK is fully synchronized with the target device, meaning that a Python script can remotely run and behave like it was an internal thread inside the device; but with slow memory access time. That powerfully enables HIL (Hardware-in-the-loop) testing.
Q&A Discussion for the hardware block