On Fri, Jun 26, 2015 at 5:09 PM, Miika Turkia <[email protected]> wrote: > On Thu, Jun 25, 2015 at 1:50 PM, Dirk Hohndel <[email protected]> wrote: >> On Wed, Jun 24, 2015 at 09:12:46PM -0700, Dirk Hohndel wrote: >>> On Thu, Jun 25, 2015 at 07:21:53AM +0800, Miika Turkia wrote: >>> > I have tested out the GPS coordinates a bit with companion app. As I >>> > am lazy, I rather set it to automatic tracking in the morning and >>> > forget about it. However, this ends up getting me wrong coordinates. >>> >>> That's how I always use it. But I haven't been on a boat dive since all >>> the changes to the dive site handling, so it's entirely possible that this >>> is currently broken in creative ways :-( >>> >>> > Currently we drive to a buyu line and tie the boat there for the >>> > duration of the dive. Thus an option to grab the position after the >>> > dive starts would be perfect for these dives. (Now I usually get the >>> > position while still on the way to the dive site.) Of course this mode >>> > does not work for drift diving, but that is why I suggest this kind of >>> > an option for the new and shiny GPS import Linus envisioned. >>> >>> The way it SHOULD work is that if it has a fix during the dive, it takes >>> the first such fix. If not, it tries to find a fix before or after. >>> >>> I can stare at the code a bit to see if I see a problem, or I can add some >>> verbose debugging option that allows us to see which fixes it picks and >>> why. >> >> I did stare at the code and the more I did so the less the code made >> sense. So I did the obvious thing and rewrote it :-) ... and added a TON >> of debugging statements. This was pushed to master and daily builds have >> been triggered. > > Well, I just got a segmentation fault, but when running under gdb, all > went fine. > > ---8<--- > processing gpsfix @ "Fri Jun 26, 2015 2:05 PM" which is withing six > hours of dive from "Fri Jun 26, 2015 2:14 PM" until "Fri Jun 26, 2015 > 3:31 PM" > look at the next gps fix @ "Fri Jun 26, 2015 2:08 PM" > which is closer to the start of the dive, do continue with that > processing gpsfix @ "Fri Jun 26, 2015 2:08 PM" which is withing six > hours of dive from "Fri Jun 26, 2015 2:14 PM" until "Fri Jun 26, 2015 > 3:31 PM" > look at the next gps fix @ "Fri Jun 26, 2015 3:38 PM" > pick the one before as it's closer to the start > Segmentation fault > ---8<--- > >> Miika, I would love it if you could try this (but I understand if you >> don't have the time or network bandwidth to do so). I think the debug >> messages are verbose enough to make it easy to track down what happened >> and why the fixes are chosen. > > I think I got correct locations today. I'll try to test this a bit > more, purely for the benefit of all :D
I think this is related to downloading new dives from DC and getting the GPS positions. It occurred again after I had just downloaded the day's dives, while syncing GPS coordinates. I hope I can catch this on gdb tomorrow, otherwise I might have to start backtracking the log file and downloading again from DC... >> Now I need to add another piece of UI that allows us to import GPX files >> in the desktop app... we can do so in the Android app but if someone has a >> GPX file from some device we should just offer to read those and use the >> same synch algorithm... > > poor you :D > > It feels like Subsurface is quite slow to start up nowadays. I wonder > if this is related to my lightening fast (read incredibly slow) > Internet connection. > > Also the IOS companion app is very slow to realize there is no > Internet connectivity when I am attempting to start logging. > (Workaround: quit and open up the app to trick it to notice syncing is > not going to happen.) Do you have a pointer where to report this one? If you do not have that handy, I can always dig it out myself later. miika _______________________________________________ subsurface mailing list [email protected] http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface
