Re: [gpsd-dev] What is our current bug state?

[Eric S. Raymond]

Gary E. Miller <address@hidden>:
> > Well, that was true for a short while.  Now I'm seeing a garbage
> > offset again.
> 
> Where are you seeing that?  Hal reports a garbage offset in gpsmon in
> direct mode.  I see some unitialized variables in gpsmon so obvious
> fixes there.

Yes, it's in direct mode.  Apply whatever fixes you know about, please.

I know you talked earlier about ripping the direct-mode PPS support
out, but I don't want to do that because I think it's a useful smoke test
for PPS-capable hardware.  That's why I put it in.  Not involving gpsd
and minimizing the number of moving parts is useful when the user just
wants to know "Does this thing work?"

In more postive news, I've found the trick required to make the regression
tests work when gpsd uses blocking I/O.  It turns out all you have to do 
is increase WRITE_PAD and CLOSE_DELAY some.  I'm working on finding a
new pair of minima for Linux now.

This is good news because with O_NONBLOCK *not* set at tty open time
we can actually set VTIME and have it work portably. That in turn means
a very simple and robust fix for the buzzing-select problem.
-- 
                <a href="http://www.catb.org/~esr/";>Eric S. Raymond</a>