Stats basic function: Tap on the button top left to open stats. Swipe left to close, or swipe down to expand. Double tap on a cell to get the list of options.
Stats now has the following options:
Accuracy - Reports GPS accuracy.
Altitude - GPS altitude
Direction - Direction of travel
Distance - length of current tracklog. Currently this does NOT also add on the distance from the last trackpoint in the tracklog to the current GPS location. This is on my list to get altered. So distance will only update as each trackpoint is added to the track currently.
GPS Reference - the location reported by the GPS
Speed - speed reported by the GPS
Time - hoping I don't need to explain this one.
Vertical Accuracy - again reported by the GPS, but I have only seen zero on my phone. Its either very very accurate or I have a bug somewhere! Is anyone getting a sensible value?
I think there is still some finessing required. Turn off the GPS, things don't blank currently. I also believe there is some 'stacking' up of data, in that if you stop walking, it takes a few seconds for speed to drop to zero. I think we aren't processing data as fast as its coming in, so we probably just need to drop some values.
Overall though this is taking shape now, and the updates to the screen now feel smooth.
Some large changes which I am sure are for the better. Will they fix the reported issues, on that I just don't know....
First up, I've done some automated testing of sleeping and waking the app, and I have detected and fixed a small memory leak. Certainly good to fix it, but was it enough to crash someones app after 2 hours.... its possible, but I wouldn't have thought so.
So the app works in two parts - the tracklog recording (recorder) and the main app (app). The system was set that the recorder would call the app about updates, and if the app was killed off, we get a warning, and it would disconnect. I've changed this to use the Android broadcast system, so now the recorder just SHOUTS changes, and if the app is alive and running, it hears them and processes them. The upshot is, that a crash in the app should not crash the recorder. Frankly I thought the code before should have avoided this situation, but this certainly furthers the separation. When the app starts up, if will catch up any updates its missed. This happens as the next trackpoint is added. I will try to improve this - you will currently also see this if you put the app to the background, when you bring it to the foreground, the tracklog will update as the system records the next trackpoint.
If the recorder crashes it will restart by itself. Android decides when exactly, and on my phone its about 30seconds to a minute. So if the recorder does crash, it should start back up, although you might get a small gap in the track depending on your settings or you might not even notice. Due to the current separation, the recorder takes distance / time settings when a tracklog is commenced, and won't update the settings whilst recording. This is on my list to alter.
You will also find that the recorder notification now updates telling you which trackpoint is being added.
You can download from:
Google Play: https://play.google.com/apps/testing/com.anquet.omn2
** this takes a little while to come out. HockeyApp email notifications will come out once the build completes.
P.S. 58 got released as Menu->Edit Mode would crash the system - I had been using that to purposefully crash the system and test the recovery. Its back to normal in 58.
Hi, Managed a 15 miles cycle ride this morning, a case of starting tracklog , putting phone in pocket and not looking at it until I had finished. For almost the f1rst time it did not freeze so when checking at the end it behaved as expected, gave accurate measurements and was able to stop tracklog, pan around screen to check route, save the route without the OMN has stopped responding message from Google. Seems to be really good progress. However I did freeze, any chance of building in hypothermia prevention.
Looking Good Colin
@Colin - that made me smile! In two parts, a) that its working, and b) the final comment!
@Roger - thanks.
@David - the unacceptable time will be a lot better in the next beta, although I'm still seeing one case where things are slow to start up that I haven't tracked down yet.
@Ted - we are currently concentrating on getting it all working and stable, then we can start to work on other options in the list. Hopefully the next beta will be a further improvement.
Chaps, I'm having a good run on the code today. Lunch and gym now, but I'm hopeful of a further beta late UK time.
Started a tracklog on a walk at 9:52. Once on the walk after about 7km looked at the phone and touched the track to see the current distance. Tracklog continued to record until I looked at the phone at 13:12 and again touched track to see the current distance (using the info display towards the bottom of the screen). I next looked at the phone about 13:40. At one time the top half of the screen had a map displayed but the bottom was black. I then got an ANR and whilst trying to provide feedback got another ANR so I don't know if the logs will have got through. The tracklog had stopped recording detailed points at 13:12. Two final points are in the log after, 27 minutes and 1.6km, and after 2 minutes and 0.12km. [I am set to record after 25m or 15 seconds].
Trying to restart was a complete mess. White screens; OMN Blue screen - which never seemed to move onto a map display; ANRs; Tracklogs that recorded a handful of points and then failed. The sequence of these errors is lost but the takeaway is that the mechanism after a failure (or user error) to restart both the map display and the tracklog recording tasks is not sufficiently robust.
When I eventually got OMN to work I stopped tracklogging and exited the application. The phone had a relaxing drive home and now OMN is behaving itself again!
The next one is now up:
I'm working tomorrow, so I might pop up, but don't feel too bad for me, as I have a few nights away up in the Brazilian mountains, I'm back late Tuesday.
@David - how you are using it shouldn't cause issues, but you can of course now set 'Distance' in the new stats bar, and you should get a figure that updates as you walk. I've seen similar re-start issues, although the next beta should be better on all fronts, all though I've not chased after this issue. If you were hitting the 'Start Tracklog' it should be a clean tracklog. I can put a check in for that to make sure I clean out anything that is stuck there.
@Roger - if you can give the next one a try please. You will see in the notes, quite a few changes that may or may not be a resolution for what you have seen.
@Ted - Are you zoomed out quite a long way? I *think* its the cache not holding enough tiles, and then re-loading/re-drawing some. I don't think you will see it on 25khd - possibly on 25k. But most likely its just on OSVMLC.
Many thanks chaps,