Welcome Guest,Register Now
Log In

ANT Forum

Welcome guest, please Login or Register

   

Inconsistent profile file content

Rank

Total Posts: 10

Joined 2016-10-30

PM

Hi,

I am writing a FIT decoder for MATLAB and I came across a couple things in the profile file that I believe are inconsistent.

Type bike_light_beam_angle_mode has base type: unit8, which I think it should be uint8

On the session message, swim_stroke has a ref field, but it also has an id (43), shouldn't the 43 cell be empty?

Thanks,
Joan

     
RankRankRank

Total Posts: 68

Joined 0

PM

It appears that you have found a mistake in the profile document. We will be sure to correct that for our next release.

session.swim_stroke should not have a ref_field value.
bike_light_beam_angle_mode should have a base type of uint8
     
Rank

Total Posts: 10

Joined 2016-10-30

PM

Hi Caufield,

Thanks for looking into this.

In the profile file, in the fit_base_type, any reason why Bool was not included? If it is not considered a base type, then it is not listed as a type of its own either. Any reason for its special treatment?

Joan
     
Rank

Total Posts: 10

Joined 2016-10-30

PM

Also, the type localtime_into_day, should it not be better to name it local_time_into_day?      
RankRankRank

Total Posts: 68

Joined 0

PM

There was no reason for bool to be excluded. We'll get that fixed.

With respect to localtime_into_day, I can see the argument here, but, as I am sure you can understand, it is not possible for us to change the names of types without breaking backward compatibility of our libraries.

Thanks      
Rank

Total Posts: 10

Joined 2016-10-30

PM

Perfect on the bool thing, fair enough on the localtime_into_day.

Thanks,
Joan      
Rank

Total Posts: 10

Joined 2016-10-30

PM

A couple more observations:

In device_info and software, the scale of software_version is 100, but in in file_creator it seems to not have any.

When trying to make sense of the GPS data, by looking at the numbers and knowing where I was, I have managed to get it close to where I expect by dividing by (2^32/2). It feels quite exciting to finally be able to plot all the running parameters out of my watch, but there might be an off by one error there. I think that the Profile file should have the right scale in it.