Detailed Description

RAII watchdog guard for the canonical loop()-top use case.

First construction (per-process) lazily arms the watchdog with the supplied timeout (default 15 000 ms), prints any prior-boot reset/crash diagnostic through fl::println, and pauses for 3 s on a crash so the developer can read the message before loop() repaints serial output. Construction feeds the watchdog immediately so the current loop iteration starts with a fresh deadline; destruction feeds it again so the next iteration also starts fresh.

Single-instance rule. Only one ScopedWatchdog should be alive at a time. The class detects nested / overlapping live instances at runtime and prints a one-shot warning via fl::println (the WDT is still fed in both ctors and dtors so the program keeps running). Two guards on the same source line are a compile error (the FL_WATCHDOG_AUTO macro stamps the local variable name from __LINE__, so a duplicate would collide).

Prefer the FL_WATCHDOG_AUTO(...) macro below — it stamps a unique stack variable name and is the canonical entry point.

Definition at line 205 of file watchdog.h.

#include <watchdog.h>

Public Member Functions
	ScopedWatchdog () FL_NOEXCEPT
	Default construct with the library default timeout (15 000 ms).

	ScopedWatchdog (const ScopedWatchdog &) FL_NOEXCEPT=delete

	ScopedWatchdog (fl::u32 timeout_ms) FL_NOEXCEPT
	Construct with an explicit timeout.

	~ScopedWatchdog () FL_NOEXCEPT
	Feed the watchdog at end-of-scope so the next `loop()` iteration has a clean deadline window.

ScopedWatchdog &	operator= (const ScopedWatchdog &) FL_NOEXCEPT=delete

Static Public Member Functions
static int	activeScopeCount () FL_NOEXCEPT
	Observability: number of simultaneously-alive ScopedWatchdog instances.

The documentation for this class was generated from the following files:

fl/wdt/watchdog.h
fl/wdt/watchdog.cpp.hpp

Detailed Description

Public Member Functions

Static Public Member Functions