The windows hosts file modification -or- the dnsmasq alternative all work to stabilize the settings. So, a watchdog wasn't needed for that particular case.
Even so, a standard keepalive/ping watchdog could be a nice feature to consider. Perhaps there is a simple batch file method?
Here's a batch used to run a sync-backup for a WiFi nas. It shows the errorlevel loop reset to 0 by the successful ver command and the sync which may or may not be successful over WiFi (but it will loop until success).
@echo off
:ONE
timeout /t 30 /NOBREAK
VER > nul
robocopy \\source\folder D:\destination /e /copy:DT /xo /purge /w:0 /r:0
IF %ERRORLEVEL% NEQ 0 GOTO ONE
I put that sample, because the errorlevel isn't well documented for batch.
Now for watchdog
I'm fairly sure the ping command exits with an errorlevel of some sort, and then there's the tskill command to exit a program gently or taskkill to exit an unresponsive program.
added function
It seems possible to make a safety copy of hfs.ini at the batch start.
And, then after tskill and 2 seconds timeout (sleep without loading cpu), reverse the copy before starting hfs. So, if settings caused the server to freeze, functional settings and server would be restored to working order automatically.
I haven't figured out the particulars yet
But, this type of watchdog could allow for the automatic updates and settings. It would revert changes that didn't work or keep any that did work.
That approach seems more reasonable than a few other things that I have tried.