navigator.geolocation.getCurrentPosition sometimes works sometimes doesn't...

Question

So I have a pretty simple bit of JS using the navigator.geolocation.getCurrentPosition jammy.

$(document).ready(function(){
  $("#business-locate, #people-locate").click(function() {
    navigator.geolocation.getCurrentPosition(foundLocation, noLocation);
  });

  navigator.geolocation.getCurrentPosition(foundLocation, noLocation);

  function foundLocation(position) {
    var lat = position.coords.latitude;
    var lon = position.coords.longitude;
    var userLocation = lat + ', ' + lon;
    $("#business-current-location, #people-current-location").remove();
    $("#Near-Me")
      .watermark("Current Location")
      .after("<input type='hidden' name='business-current-location' id='business-current-location' value='"+userLocation+"' />");
    $("#people-Near-Me")
      .watermark("Current Location")
      .after("<input type='hidden' name='people-current-location' id='people-current-location' value='"+userLocation+"' />");
  }
  function noLocation() {
    $("#Near-Me").watermark("Could not find location");
    $("#people-Near-Me").watermark("Could not find location");
  }
})//end DocReady

Basically what's happening here is we get the current position, if it's obtained, two "watermarks" are placed in two fields that say "Current Position" and two hidden fields are created with the lat-long data as their value (they're removed in the beginning so they don't get duplicated every time). There are also two buttons that have a click function tied to them that do the same thing. Unfortunately, every third time or so, it works. What's the problem here???

Answer 1

Hi Nick, did you find a solution? I'm having a similar situation, in which in Firefox and Safari I get positioning responses when I open a fresh browser, but after that, if I reload the page, it doesn't work at all... no succes, no error handlers!

Marc

Answer 2

Hi did you manage to find an answer? I'm running into a similar problem. As far as I can see it's something to do with the asynchronous call to navigator.geolocation.getCurrentPosition, that sometimes it just doesn't return. In my app I can refresh the page 4 or 5 times and nothing, then out of the blue it comes back with a position value. Seems to me the functionality is just a bit flaky at this stage. (Sorry I know this isn't an answer but I can't post a comment above for some reason).

Answer 3

Same here people, this works perfect btw in Chrome (stable, dev and canary) just not in FF and Safari. It also works perfect on my iPhone and iPad (Safari!). This might be due to the relative newness of this feature (i.e. it is a bug). I spend almost a week on this now and I just cannot get it to work on those browsers

Here's what I found:

The first time you call getCurrentPosition it works perfect. Any subsequent call never returns, i.e. it does not fire the successCallback or the errorCallback functions. I added a few position options to my call to prove my point:

 navigator.geolocation.getCurrentPosition(successCallback, errorCallback,  {timeout: 10000});

and it times out every time (after the first successful call). I thought I could fix it with maximumAge, but that doesn't seem to be working as it is suppose to work either:

navigator.geolocation.getCurrentPosition(successCallback, errorCallback,  {maximumAge:60000, timeout: 2000});

this should prevent actually calling the getCurrentPosition function if you call it within 60 seconds, but it ignores this (however, this could be due because I actually refresh my page to trigger the second call, not sure if this is persistent accross calls)

btw, even google's examples fail on these browsers which leads me to believe that this are indeed browser bugs, try it, load it twice in Safari and it won't work the second time.

If anybody finds a solution for this, PLEASE let me know :-)

Cheers.

Answer 4

I have been having exactly the same problem, and finding almost no information online about it. Nothing at all in the books. Finally I found this sober query on stackoverflow and (ha!) it was the final impetus I needed to set up an account here.

And I have a partial answer, but alas not a complete one.

First of all, realise that the default timeout for getCurrentPosition is infinite(!). That means that your error handler will never be called if getCurrentPosition hangs somewhere on the back end.

To ensure that you get a timeout, add the optional third parameter to your call to getCurrentPosition, for example, if you want the user to wait no more than 10 seconds before giving them a clue what is happening, use:

navigator.geolocation.getCurrentPosition(successCallback,errorCallback,{timeout:10000});

Secondly, I have experienced quite different reliability in different contexts. Here at home, I get a callback within a second or two, although the accuracy is poor.

At work however, I experience quite bizarre variations in behavior: Geolocation works on some computers all the time (IE excepted, of course), others only work in chrome and safari but not firefox (gecko issue?), others work once, then subsequently fail - and the pattern changes from hour to hour, from day to day. Sometimes you have a 'lucky' computer, sometimes not. Perhaps slaughtering goats at full moon would help?

I have not been able to fathom this, but I suspect that the back end infrastructure is more uneven than advertised in the various gung-ho books and websites that are pushing this feature. I really wish that they would be a bit more straight about how flakey this feature is, and how important that timeout setting is, if you want your error handler to work properly.

I have been trying to teach this stuff to students today, and had the embarassing situation where my own computer (on the projector and several large screens) was failing silently, whereas about 80% of the students were getting a result almost instantly (using the exact same wireless network). It's very difficult to resolve these issues when my students are also making typos and other gaffes, and when my own pc is also failing.

Anyway, I hope this helps some of you guys. Thanks for the sanity check!

Answer 5

This is just to confirm all of the above. In particular, fine in Chrome (6.0.472.63), flaky in Firefox (3.6.10) and Safari (5.0) (all under Mac OS X 10.6.4). As usual, there's a reason people call it the bleeding edge: use it, and you'll get cut.