View Single Post
Posts: 214 | Thanked: 140 times | Joined on Aug 2010
#149
Originally Posted by morti View Post
Got a problem lately, where the daemon seems to be satisfied with cell location even though mode is on --gps. When starting the daemon, GPS icon flashes for a few seconds until coarse accuracy appears, then disappears and latitude is updated.

Verbose output of an example case:
Code:
Starting ZapLoc daemon... 
Location method: GPS only 
60 minutes between samples 
22:07:38: Starting GPS... 
22:07:38: Actual GPS interval 60 minutes 
22:07:38: Got loc data <snip> (acc nan), Mode 1 - IGNORED 
22:07:40: Got loc data <snip> (acc 53.763000), head nan, speed nan, Mode 2, Try 0 
22:07:40: Latitude update <snip> (acc 53.763000) 
22:07:41: Checked you in at <snip> on Latitude 
22:07:41: General location update <snip> (acc 53.763000) 
22:07:41: Searching for closest spot: 
22:07:41: The closest AutoCheck spot is <snip>
22:07:41: Spot ignored because it is further than 50 meters away 
22:07:41: Sleeping... Zzzzz 
22:07:45: Got loc data <snip> (acc 53.763000), Mode 0 - IGNORED 
22:08:08: A good location fix was acquired within the timeout: Good!
(build 43, happened with earlier builds as well)

As you can see from the timestamps, the daemon doesn't even try to wait for a better fix.

Ideas?
Yep. Your tower claims to be closer to you than 53 meters... which I kind of doubt it is... I actually seen this myself, towers that obviously claim their accuracy is way higher than reality.

Basically, I'm now looking at accuracy, rather than the sample mode, but I think that in "--gps" mode, I should change it to still be looking at the mode=3 (which means "this value came from the GPS"). This is a trivial fix.

Curious; how far away would you guesstimate that tower actually is? I noticed that the tower I normally get connected to, which is about 2000m from me, claims to have "253 meter" accuracy. Dunno what's with the "53" at the end at all these

/Z

Last edited by MasterZap; 2011-05-27 at 05:43.