move radio sleep out to seperate function

[jacobrosenthal]

Splitting mesh sleep out from radio sleep with the end goal being able to implement a mesh synchronization protocol

power.sleep2 should be called something else but for some reason Ive noticed bitlash cant understand certain terms. I tried power.sleepRadio and power.sleepRF and neither worked.. power.sleep2 worked and I left it there

Looking for feedback

[matthijskooijman]

power.sleep2 should be called something else but for some reason Ive noticed bitlash cant understand certain terms. I tried power.sleepRadio and power.sleepRF and neither worked.. power.sleep2 worked and I left it there

Perhaps names should be lowercase? I can imagine that bitlash implements case insensitivity by lowercasing all commands before processing, but perhap not inside addBitlashFunction?

No time right now to look at the actual commit, I hope to find a bit of time for that (and your other PR's, keep up the good work!) soon...

[jacobrosenthal]

Perhaps names should be lowercase? I can imagine that bitlash implements case insensitivity by lowercasing all commands before processing, but perhap not inside addBitlashFunction?

Looking at all other bitlash names, that seems likely.

[jacobrosenthal]

An example which sleeps the radio repeatedly at 5 second intervals starting at (global) time 45 seconds

> uptime.report
{"type":"uptime","total":5,"sleep":0,"meshsleep":0,"random":15089,"reset":"External","at":5653}
> uptime
Total: 0 days, 0 hours, 0 minutes, 8.231392 seconds
Awake: 0 days, 0 hours, 0 minutes, 8.234288 seconds
Asleep: 0 days, 0 hours, 0 minutes, 0.000000 seconds
Global: 0 days, 0 hours, 0 minutes, 8.239792 seconds
> uptime.setoffset(10,0,1)
> uptime
Total: 0 days, 0 hours, 0 minutes, 12.182768 seconds
Awake: 0 days, 0 hours, 0 minutes, 12.185776 seconds
Asleep: 0 days, 0 hours, 0 minutes, 0.000000 seconds
Global: 0 days, 0 hours, 0 minutes, 22.191280 seconds
> print uptime.meshoffset.seconds
10
> power.setradioperiod(5000, 5000)
> power.schedulesleepradio(45, 0)
> uptime
Total: 0 days, 0 hours, 0 minutes, 27.921184 seconds
Awake: 0 days, 0 hours, 0 minutes, 27.924192 seconds
Asleep: 0 days, 0 hours, 0 minutes, 0.000000 seconds
Global: 0 days, 0 hours, 0 minutes, 37.929696 seconds
> 
> print power.getradiostate
1
> uptime
Total: 0 days, 0 hours, 0 minutes, 36.588288 seconds
Awake: 0 days, 0 hours, 0 minutes, 36.591280 seconds
Asleep: 0 days, 0 hours, 0 minutes, 0.000000 seconds
Global: 0 days, 0 hours, 0 minutes, 46.596800 seconds
> print power.getradiostate
4
> print power.getradiostate
1
> print uptime.meshsleeping.seconds
20
> uptime.report
{"type":"uptime","total":98,"sleep":0,"meshsleep":30,"random":-21287,"reset":"External","at":98333}
> 

[jacobrosenthal]

... I think uptime.report is crashing if called a couple times....

[jacobrosenthal]

found uptime crash, buffer was too small
frankly upstream seems a bit too small for the upstream verstion too which would be
{"type":"uptime","total":4294967295,"sleep":4294967295,"random":-2147483647,"reset":"External","at":4294967295}
something like 112
but its almost impossible to be awake for years, asleep the whole time, with an external reset in the same amount of time...
basically I add 23 characters
,"meshsleep":4294967295
so ill just add 25 to buffer

[jacobrosenthal]

Todo

• Vet the api and naming scheme
• Need a way to queue or delay radio sends attempted during sleep (calling NWK_DataReq crashes)
• Need to transition events timing to last symbol counter compare. sys timer is non functional during sleep - existing problem with power.sleep as well
• Create one (or more) offset synchronization protocols like NTP, RBS, or something that needs lower level rx/tx timestamps like tsmp fstp