Re: PositionError Requests

I think UNKNOWN as 1 makes sense.

Should UNKNOWN actually be UNKNOWN_ERROR to match PERMISSION_ERROR? I  
wasn't too sure about what the short error names should be.

Thanks,
-- Greg

On Oct 29, 2008, at 8:34 AM, Doug Turner wrote:

> I recommend making UNKNOWN being 1 because we have no idea how long  
> this list will get.
>
>
> On Oct 29, 2008, at 6:33 AM, Andrei Popescu wrote:
>
>>
>> Hi Greg!
>>
>> On Tue, Oct 28, 2008 at 6:18 PM, Greg Bolsinga <bolsinga@apple.com>  
>> wrote:
>>>
>>> Hello --
>>>
>>> I'd like to request an error code 5, UNKNOWN. There are certain  
>>> errors that
>>> may occur that are not covered by any of the error codes listed.  
>>> Generally
>>> these can be when the native implementation can't determine the  
>>> correct
>>> things about the JS code given. You can search for UNKNOWN_ERROR  
>>> in the last
>>> patch to https://bugs.webkit.org/show_bug.cgi?id=21475 for some  
>>> examples.
>>>
>>
>> Sounds fine to me.
>>
>>> I'd also like to request that the errors have names, such as  
>>> suggested by
>>> Olli Pettay on October 15 on this email list.
>>>
>>
>> Ok, how about this:
>>
>> interface PositionError {
>>   const unsigned short PERMISSION_ERROR = 1;
>>   const unsigned short PROVIDER_ERROR = 2;
>>   const unsigned short NOT_FOUND = 3;
>>   const unsigned short TIMEOUT = 4;
>>   const unsigned short UNKNOWN = 5;
>>   readonly int code;
>>   readonly DOMString message;
>> };
>>
>>
>> All the best,
>> Andrei
>>
>