Because this is a core app it 100% needs some good testing before it goes in. I've lost count of the amount of times someone's managed to break everyone's device because something wasn't tested :)

What was the 'unintended code'? What bug are you trying to fix?

edit: sorry! I'm slow today - I didn't realise that was code you'd added until just now :)

As it is, given you don't have it running all the time, if:

• the hrm reading has non-zero confidence at the time of running: it won't turn on the HRM sensor
• the hrm reading has 80+ confidence, but not more than the stored reading: it'll turn off the HRM sensor early. So the stored reading will be used instead, because the stored reading has more confidence.

This PR fixes that. I'll go ahead and try to test it in the IDE.

Just tested it. I don't see any errors, and with some added logging, things seem to be working as they should.

Contemplating whether I should raise the confidence threshold to a 90 - while there are some 40BPM measurements with 80-90% confidence, there are also some with 100%.

Thanks - yes, I reckon raising it to 90% is probably a good idea - thanks for checking this out!

Once it's 90 I'll merge?

Looks like I should update boot.min.js too, what minifier are you using? Is https://skalman.github.io/UglifyJS-online/ okay?

Ahh, yes - sorry.

It's best to clone https://github.com/espruino/EspruinoDocs and use bin/minify.js which uses Closure with some special settings. That's what I use.

How are you feeling about this now? Still seems to work well, and it's producing better HRM results?

It doesn't really produce better HRM results - it does fix a bug that might cause the wrong ones to be recorded though, so it's worth merging.

Ok, thanks!