Re: [gpsd-dev] Getting the driver-switching logic right

[Eric S. Raymond]

Michael R. Davis <address@hidden>:
> From a packagers point of view, we need to ship gpsd with all
> drivers compiled in the binary.  However, I'd like to be able to
> turn off certain drivers in configuration files.  I also proposed to
> never to switch a confirmed driver without a restart of the service.

We absolutely *cannot* adopt that rule.  It doesn't even really make sense.

I start the daemon.

I plug a SiRF into a USB port.  Use it.  Unplug it.

Now I plug in a USB uBlox. The daemon must not be stuck to the
"confirmed" SiRF driver so hard that it can't switch to the uBlox
driver, even if it thinks that's on the same device as the SiRF was
previously.

Why would you want to "turn off" a driver?  What would that even mean?

Configuration files are wrong, *wrong*, **wrong**.  Automatically 
adapting to inputs as they present is the right thing.  If you think
a configuration file is the answer, your question is incorrect.
-- 
                <a href="http://www.catb.org/~esr/";>Eric S. Raymond</a>