Thanks! Sorry about the duplication - I'd started the work on this and then didn't get around to doing the Bangle.js bit before my holiday :)

Do you think maybe it makes sense for readAllRecordsSince just to call readDay rather than duplicating the code? I guess readDay could accept a 'since' argument, although hopefully out of all the days, only one of them will need part of the day transmitted, so it could just be an idea to check date.getTime() > since before forwarding the health record?

@gfwilliams I was trying to avoid computing the timestamp of the current health record for every record, but I can maybe try and do that in an efficient manner. The idea of this function was really to have a more optimized function for the sync functionality, since people can use it periodically.

Ahh, ok - I see, we don't create a Date object. It would be nice to use getDay if possible though, especially as I guess we'll now be de-duplicating in Gadgetbridge.

I don't think speed will be an issue here - the thing slowing us down will be bluetooth transmission speed :)

Thanks for sorting that so quickly - merging now!

Since metadata says {"name":"health","url":"lib.min.js"}, and lib.min.s wasn't updated with the new additions the fetching breaks.

Solution is either:

• pull from lib.js instead or
• update lib.min.js

just done I hope!

Ooops, I thought the minified files were automatically generated when deploying this 😐

Now it works - super nice! :)

@joserebelo The fetching would never be stopped when readAllRecordsSince was missing - if that's interesting info.

I thought the minified files were automatically generated when deploying this neutral_face

Well usually when you specify an unminified JS it gets minified on the fly. I can't remember why but for some reason it was decided that those files should be minified offline (we use the closure compiler then and it often does a better job)

The fetching would never be stopped when readAllRecordsSince was missing - if that's interesting info.

Yup - we only stop it once the Bangle signals that it finished. If that message never arrives, it never stops.

It's a more or less known issue, there used to be a similar problem in Huami devices.

I think we may want to eventually introduce a timeout on the Gadgetbridge side for operations that become stuck.