Posted by: | dragorn |
---|---|
Subject: | Driver access difference between devl and newcore |
Date: | 16:08:15 29/09/2008 |
> > Okay, I think I've figured this out. Very strange to say the least. In all my testing and configuration I've been setting gps=false since I don't have a gps device. This morning I left the value as true and started up the new-core server. Guess what, it works! Now at least I have something to work from. Since I haven't looked at that part of your code so I'm not sure what is going on, but I'll take a look.
>
> That's very weird. I'll take a look.
OK, two things:
1) I've rolled all the new source definition and privsep changes, see the devel blog. This changes how sources and drone sources are defined (for the better, I think)
2) I found and fixed the gps=false bug, I think.
If you don't want to use the all-new code yet, you can make the following change in gps_wrapper.h... in the GPSNull class, change:
virtual unsigned int MergeSet(unsigned int in_max_fd, fd_set *out_rset,
fd_set *out_wset) {
return 0;
}
to
return in_max_fd;
This should solve the gps issue killing things.
I'd appreciate testing of the new source handling code, of course, too.
-m