Geolocation test-suite

Hello Geolocation WG,

I've a few questions/comments on the geolocation test-suite  
(http://dev.w3.org/geo/api/test-suite/) which are detailed below.

1. The spec. says that accuracy should be non-negative but many of the  
tests use a negative value for accuracy.  Is this a test-case issue or are  
there any other reasons? The tests in question are 71, 34, 74, 72, 70 and  
73. Also when the accuracy is negative, wouldn't that trigger an error?

2. Test 91 checks that a negative timeout value is inferred as 0. This  
combined with maximumAge:0 would result in a timeout error almost  
immediately. Hence I think the checking for pass/fail should be included  
in the error callback instead of success callback. Also checking for time  
elapsed doesn't work very well since the user could take time to answer  
the geolocation privacy dialog and this elapse of time isn't supposed to  
affect the timeout specified, is it? So probably these tests should not  
check the exact interval of time but the result only. This affects test 83  
also.

3. Test 8 is a little bit dependent on the global test-case timeout value  
set in common.js. Maybe it could just check that both callbacks were  
called? Also it seems possible that the callbacks happen in the same  
fraction of time so that also need to adjusted for.

4. Test 109 needs granting access so that we can check that the server  
returns error and the error code is POSITION_UNAVAILABLE. The script  
should probably be using 'askAccept' instead of 'askWait'? Same issue with  
test 93.

5. In the tests where heading values are checked, isn't a valid speed also  
required to be present? The spec. says that when speed is 0, heading  
should be NaN; So when speed is null, can heading be valid? (Tests  
affected are 68, 37, 103)

6. Test 144 has a typo: 'post.coords' should be 'pos.coords'.

7. Test 92 checks for a timeout error, so shouldn't it be using the error  
callback to determine pass/fail?

8. Test 142 would trigger error callback as the user would deny  
permission; Shouldn't it be using the error callback?

9. In test 38, shouldn't the value of speed be non-negative?

10. Wouldn't test 119 trigger the error callback unless there is a cached  
position, since the timeout value is 0? Same goes for test 120.

Any comments/clarifications are much appreciated.


-- 
Hari Kumar G
Quality Assurance
Opera Software

Received on Monday, 22 November 2010 17:04:52 UTC