Re: [gpsd-users] GPSCTL do more than you ask him to do.

[Steve Bourland]

First I hope this works well, I have been lurking a long time on this list 
and not posting much at all, let alone in response to something...

> ------------------------------
>
> Message: 6
> Date: Thu, 15 Oct 2015 14:41:47 -0700
> From: "Gary E. Miller" <address@hidden>
> To: "Eric S. Raymond" <address@hidden>
> Cc: Cuningan Re Set <address@hidden>,     gpsd-users
>         <address@hidden>
> Subject: Re: [gpsd-users] GPSCTL do more than you ask him to do.
> Message-ID: <address@hidden>
> Content-Type: text/plain; charset="us-ascii"
>
> Yo Eric!
>
> On Thu, 15 Oct 2015 17:09:39 -0400
> "Eric S. Raymond" <address@hidden> wrote:
>
> > > Everytime I started to look at gpsctl I choked.
> >
> > Why, in particular?
>
> Hard to say what I don't understand about a subject I don't understand.  :-)
>
> I would like a simple function, dependent on GPS type, that I could pass
> a simple control string and it would do the checksum/framing/etc. and send
> the string.  No more.
>
> So I could code something like:
>
>         gpsctl -t u-blox --set-pps-width 100
>
> Then in main() maybe call that calls a function like:
>         set_pps_width( 'u-blox', 100)
>
> That would know the string to send to some GPS types, hopefully u-blox,
> to set the PPS width to 100 mSec.
>
> send_pps_width() would call a function like:
>         send_ublox( 'XXYYZZ' );
>
> send_ublox() would know how to checksum and frame XXYYZZ and sent to the
> GPS.
>
> For some GPS, like garmin serial, a wait for ACK would also be needed, so
> some primitive scheduler.
>
> But not high on my TODo list.
>
>
>
> RGDS
> GARY

Yo Gary!
  I pretend to be a C programming at work at times, and get incredible 
utility out of gpsd so I would love to find a way to help.

  My thinking on re-doing gpsctl would be to have it structured by devices 
then by command instead of command then device.  My mental image is 
instead of having commands to send to u-blox in multiple locations (one is 
in the set-pps-width, another in set-nmea-output, another in some other 
command that the device may not have....), have main call a device handler 
and pass the command request and parameters to the handler in a generic 
manner (all device handlers have the same form).  The handler will then 
know what the device can do (in one place) and have one set of code for 
formatting and sending the various commands.  If devices don't offer an 
option, return that the option isn't available. 
Seems cleaner to me (on first read of your message to Eric)?

I haven't looked at the gpsctl code at all, so I have no idea what it 
looks like now.  I have no idea if I can even offer up any help to try and 
restructure/clean it up.  I would like to toss out the idea that I would 
love to try and give back to the project at some point...and this could be 
the entry point?

                                        Thanks for all your work,
                                                Steve