These are chat archives for nightscout/intend-to-bolus

27th
Jun 2016
Scott Leibrand
@scottleibrand
Jun 27 2016 01:24
@bewest have you ever gotten BT tethering actually working on an Edison running ubilinux? I know they're theoretically not supposed to be any different, but something is, and no one has yet been able to figure out why it doesn't work.
Ben West
@bewest
Jun 27 2016 01:26
the reason for that has to do with versioning between kernel and bluez
I haven't see an edison distro yet that has really great kernels on it
it's not clear which patches from intel are really needed to successfully build a kernel from scratch
bluez introduced, in the last two years several features related to ble, some of which are coordinated with android, and some of which are coordinated with linux kernel
Ben West
@bewest
Jun 27 2016 01:31
so it's a matter of mixing all the right ingredients together, bluez 5.33 should do it with the kernel available afaict... there's something significant about fact that ublinux ships with a copy of 5.28 in /root homedir
Paul Martin
@Spazholio
Jun 27 2016 01:32
There’s a walkthrough that tells you how to roll your own kernel for the Edison and use Ubilinux, but it requires 50-100GB free space, which I don’t have so I haven’t tried yet. Have you given that one a whirl yet? I’m probably going to try later on, as I’ve gotten battery_voltage compiled on Ubilinux, but it requires something that’s not present - I think a specific kernel option needs to be toggled in order for it to work correctly.
Ben West
@bewest
Jun 27 2016 01:33
it probably requires more than that
and that might not get your a newer version of the kernel, it might just let you reconfigure the version they shipped
each version of kernel typically requires patches specific to the hardware to make it work... typically intel would provide the details needed
the bakebit/yocto system is intended to weave all those things together
ubilinux is actually prepared same way, it's just a different config for bitbake so that it complies with debian guidelines
Paul Martin
@Spazholio
Jun 27 2016 01:35
I’m fine with that. I mean, it works on Yocto, and the only difference is the presence of /sys/devices/platform/pmic_ccsm/battery on Yocto.
Then again, I’ve never actually tried. =)
Ben West
@bewest
Jun 27 2016 01:36
there's likely to be bunch of subtle differences in how /proc and /sys are organized, how kernel is configured, and how some of the flash rom is used
the main big difference is apt-get and full suite of packages/repos behind it using standardized names
my previous attempts with bitbake required at least quarter terrabyte, half terrabyte was more realistic
Paul Martin
@Spazholio
Jun 27 2016 01:37
Good christ!
Why so much space?
Ben West
@bewest
Jun 27 2016 01:38
it's a system for performing cross-platform builds
what it's doing is basically compiling all of linux from scratch for your target hardware
so it grabs and sets up all the vm images needed to that
then grabs all the sources to build a new distro
then it builds them :-)
Paul Martin
@Spazholio
Jun 27 2016 01:39
Still seems like a lot, but I get it. But damn.
Ben West
@bewest
Jun 27 2016 01:39
the result of a yocto build is a new distro to be installed
the design is intended to assist in the situation where your device is too dumb to accept anything like "package" updates, so the only way to update is to build entire new image ready to flash
Paul Martin
@Spazholio
Jun 27 2016 01:40
We should have our own distro.
Ben West
@bewest
Jun 27 2016 01:40
yeah, probably
Paul Martin
@Spazholio
Jun 27 2016 01:40
There. I said it. It’s out there now. =)
Ben West
@bewest
Jun 27 2016 01:41
I was headed that way with rpi, I have great deb builds for bluez
Paul Martin
@Spazholio
Jun 27 2016 01:41
Especially since the future of Ubi is in doubt, and Yocto is lacking in some ways.
Ben West
@bewest
Jun 27 2016 01:41
yocto is a different beast though, it'd be great for someone to organize bintray or some resources to do the builds and distribute them
Paul Martin
@Spazholio
Jun 27 2016 01:41
How different?
Ben West
@bewest
Jun 27 2016 01:42
well, eg since the builds from a bot are often from scratch, it can take many hours to do a complete build
so they're harder to do with tools like travis which cap your builds at 60 minutes
Paul Martin
@Spazholio
Jun 27 2016 01:43
I don’t even know what Travis is. But anything that caps yoru build at 60 mins is bound to fail for something like this, yeah.
Ben West
@bewest
Jun 27 2016 01:43
travis is a free test builder we use to run oref0 tests on every commit
Paul Martin
@Spazholio
Jun 27 2016 01:44
I mean, I have the hardware and it’s sitting idle a lot of the time. And USB 3.0, so I could probably just have it run perpetually if needed to an external drive devoted to this..
Ben West
@bewest
Jun 27 2016 01:44
most build frameworks have a way to push build results somewhere else (like bintray/s3, etc)
Paul Martin
@Spazholio
Jun 27 2016 01:44
I’d love it if there were a way to get the exact steps needed to build Yocto or Ubi so we can duplicate and/or tweak. That would be perfect.
Ben West
@bewest
Jun 27 2016 01:44
that's what bitbake/yocto provides
Paul Martin
@Spazholio
Jun 27 2016 01:45
Is it identical?
Ben West
@bewest
Jun 27 2016 01:45
interoperable way to trade patches/recipes on how to build a linux distro
Paul Martin
@Spazholio
Jun 27 2016 01:46
I suppose calling it “Yocto” is inaccurate. When I say it, I’m referring to the specific image that’s on the stock Edison, but Yocto’s really more of a framework, isn’t it?
so "poky" is the base reference platform everyone works from
yocto "linux" is the thing you get if you build that base platform
Paul Martin
@Spazholio
Jun 27 2016 01:46
I have that exact page up already. =)
Ben West
@bewest
Jun 27 2016 01:46
to create "ubilinux" which is a distro of debian
they add to the recipes so that it all complies with the "debian policy" framework
it's still build from yocto using bitbake by deviating from the poky reference design
ubilinux is a version of yocto that extends poky to comply with debian
Paul Martin
@Spazholio
Jun 27 2016 01:48
Does Emutexlabs release their exact build steps to reproduce their image?
Ben West
@bewest
Jun 27 2016 01:48
debian is essentially a suite of strict policies that say "there shalt be an /etc/apt/sources.list", and "apt packages must have $xyz" etc
sure they do, it's all GPL
they can't support it mostly because of intel's "committent" to the hardware
if you look at the forums, you'll see many from our community, like @oskarpearson weighing in often that intel hasn't really provided the community as a whole with what's needed to continue working on this thing
so everyone is kind of frozen with the kernel intel originally released
it's all open and whatnot
but updating the patches to newer linux versions requires some knowledge of the edison and knowledge of linux that is particular, and easiest for intel to update, but apparently they've had resources spent in other areas
Paul Martin
@Spazholio
Jun 27 2016 01:51
"intel hasn't really provided the community as a whole with what's needed to continue working on this thing” <— that’s what I mean. Yeah, it’s GPL and they’re providing all of their sources, but I don’t think they’re providing their actual build steps to recreate. Which I think it still following the GPL, no?
Ugh. Just a mess.
Ben West
@bewest
Jun 27 2016 01:51
no, they've provided their BSP
the BSP is the "recipe" that bitbake uses to build things
it's the exact deliverable that makes the system interoperable and repeatable
Paul Martin
@Spazholio
Jun 27 2016 01:52
So provided we have this BSP, we could theoretically continue to use the Edison as we do now, we just need to roll our own? Does it allow for updating the kernel and such easily, or is that just a nightmare scenario?
Ben West
@bewest
Jun 27 2016 01:52
but customizing the build flags is different from updating the patches to accommodate a new version of kernel
sure, it does, but the updated kernel requires new patches
that's the part typically someone like intel would do but haven't
the old patches won't apply to the new kernel
in theory, we have the BSP, so we could extend it for a new kernel
anyone could
Paul Martin
@Spazholio
Jun 27 2016 01:55
In theory, it’s pretty simple, yeah. Practice is likely WAY different.
Ben West
@bewest
Jun 27 2016 01:55
but it's the type of task that is like: "yes anyone can go to space (provided they have a big enough rocket, no one will prevent you)"
in practice, it involves patching the kernel in precise ways to make sure it works on this hardware
certainly anyone in linux community could do this... it's more typical for efforts to go into repackaging distros to provide things like debian offshoots
Paul Martin
@Spazholio
Jun 27 2016 01:56
And the stock Yocto image works fine for most people, as does Ubilinux, so someone stepping up to maintain this would be doing it at personal time and cost.
Ben West
@bewest
Jun 27 2016 01:56
there's certainly a market of developers out there who could be paid to make it work, in theory intel has many of those on staff/retainer
right, so the signal that's worse than ubilinux closing shp
is that the only reason ubilinux is going by wayside is because intel basically did a single release on this and then walked away
Paul Martin
@Spazholio
Jun 27 2016 01:58
Do you think they did it thinking that someone would take up some sort of sponsorship and then didn’t?
Ben West
@bewest
Jun 27 2016 01:58
so anyone is going to have same issues ubilinux and everyone else has, intel has special knowledge about creating new linux kernel, could most easily do it
eg, the rpi stuff is getting updated all the time, which has allowed several debian flavors to become popular
it's because they get updated that people feel it's worth the time investment to contribute to it further
so if intel were to update their BSP to include updated kernels, I'm sure we'd see ubilinux pop back to ife
as it is, things like BLE require recent good versions of bluez which all require much more modern kernels
Paul Martin
@Spazholio
Jun 27 2016 02:01
I wonder if they’re waiting to see if there’s some sort of uproar before they decide to throw us a bone, or just let it die on the vine.
Ben West
@bewest
Jun 27 2016 02:04
or just looking at the sales numbers
I think this subg explorer board we've been working on will be interesting
it should appeal to a bunch of devs beyond our openaps group
it would be great to set up a community run bitbake somewhere
Paul Martin
@Spazholio
Jun 27 2016 02:07
What would that take? An AWS instance or something?
Ben West
@bewest
Jun 27 2016 02:08
yeah, some linux machine with some cpus for compiling and generous disk
Paul Martin
@Spazholio
Jun 27 2016 02:08
What would something like that cost? I understand that AWS is fairly inexpensive, but I don’t know specifically what that means...
Ben West
@bewest
Jun 27 2016 02:08
services like circleci, wercker, and travis do this automatically in response to github
I haven't determined whether it's cheaper to pay one of them to do it vs spinning up on own
Paul Martin
@Spazholio
Jun 27 2016 02:15
I’d be interested in some numbers if you throw some together. I was looking at prices, but don’t know what would actually be required to build an image other than space and time.
I figure you’d know that sort of information way better. =)
ryanchen26
@ryanchen26
Jun 27 2016 02:33
I'm starting my Edison rig. I flashed it with ubilinux. All seemed well, waited more than 2min, switched over to console and logged in...but it's still "Poky Yocto 1.7.2", not ubilinux. Did I not wait long enough after flashing?
Paul Martin
@Spazholio
Jun 27 2016 02:39
What’s the output of uname -a?
ryanchen26
@ryanchen26
Jun 27 2016 02:40
Linux edison 3.10.17-yocto-standard-r2 #7 SMP PREEMPT Thu Feb 26 09:57:06 UTC 2015 i686 GNU/Linux
Paul Martin
@Spazholio
Jun 27 2016 02:42
That’s what my Ubilinux system says. I think you’re good.
My Yocto rig shows Linux testrig 3.10.98-poky-edison+ #1 SMP PREEMPT Mon Jun 6 14:32:08 PDT 2016 i686 GNU/Linux
So, my Edison/Yocto testing rig seems to have…lost ssh access. I’m serial’d in right now, and can’t seem to find it. That’s weird, right?
Paul Martin
@Spazholio
Jun 27 2016 02:48
I have a /lib/systemd/system/sshd.service file, but that’s about it. Nothing to actually enable or start the service.
ryanchen26
@ryanchen26
Jun 27 2016 02:55
If everything is cool with my ubilinux flashing, how come nano and sudo are not found?
Paul Martin
@Spazholio
Jun 27 2016 02:56
That’s a new one to me. It couldn’t hurt to re-flash it, I suppose. What happens if you run /usr/bin/nano?
ryanchen26
@ryanchen26
Jun 27 2016 02:57
no such file or directory
Paul Martin
@Spazholio
Jun 27 2016 02:57
That’s pretty weird. I haven’t run into that before.
ryanchen26
@ryanchen26
Jun 27 2016 02:58
When you flash, you connect to the OTG. Do you also connect the console to a battery for extra power...as indicated by some?
Paul Martin
@Spazholio
Jun 27 2016 02:59
Well, whenever I do, I have both USBs connected, just so I can monitor it. But the console port provides no power.
ryanchen26
@ryanchen26
Jun 27 2016 02:59
both connected to the pi?
Paul Martin
@Spazholio
Jun 27 2016 03:02
Correct.
How did you have it done?
Do you have the output from the flashing process handy? Can you paste it?
ryanchen26
@ryanchen26
Jun 27 2016 03:04
flashing again....
Paul Martin
@Spazholio
Jun 27 2016 03:05
I had an issue when flashing from my Pi (which I’ve done plenty of times before) where it’d error out on the last step after 13%, 33%, 54% - different points each time. No idea why.
So it appears as though it finished, unless you looked at the output. I don’t think I was ever able to get to a command prompt when that happened though.
ryanchen26
@ryanchen26
Jun 27 2016 03:07
Flashing is now proceeding, only connected to Pi with OTG... everything quickly went to 100%
Using U-Boot target: edison-blankcdc Now waiting for dfu device 8087:0a99 Please plug and reboot the board Flashing IFWI Download [=========================] 100% 4194304 bytes Download [=========================] 100% 4194304 bytes Flashing U-Boot Download [=========================] 100% 245760 bytes Flashing U-Boot Environment Download [=========================] 100% 65536 bytes Flashing U-Boot Environment Backup Download [=========================] 100% 65536 bytes Rebooting to apply partition changes Now waiting for dfu device 8087:0a99 Flashing boot partition (kernel) Download [=========================] 100% 5980160 bytes Flashing rootfs, (it can take up to 10 minutes... Please be patient) Rebooting U-boot & Kernel System Flash Success... Your board needs to reboot to complete the flashing procedure, please do not unplug it for 2 minutes.
Paul Martin
@Spazholio
Jun 27 2016 03:08
I swear that I recall there being a status bar after the Flashing rootfs, (it can take up to 10 minutes... Please be patient) line on mine.
ryanchen26
@ryanchen26
Jun 27 2016 03:09
I'm using ubilinux-edison-150309.tar.gz
ryanchen26
@ryanchen26
Jun 27 2016 03:16
Shit, there is suppose to be a progress bar after rootfs. I can see it in the example screenshot, and it's a significant download. Crap.
Paul Martin
@Spazholio
Jun 27 2016 03:18
On a Pi3, it takes upwards of 45 mins to complete.
ryanchen26
@ryanchen26
Jun 27 2016 03:21
Right, I guess I'll just start over. From the beginning...
ryanchen26
@ryanchen26
Jun 27 2016 03:32
Yeah, I was following the learn.sparkfun page for installing Ubilinux. It didn't have the swapfile stop/start in the github doc. I think it's going to work now. :)
Paul Martin
@Spazholio
Jun 27 2016 03:32
It should definitely help.
ddaniels1
@ddaniels1
Jun 27 2016 04:15
Strange and too much to be a coincidence. My loop fails much (though not all) of the time right after running monitor/meal and before running settings/settings:
Jun 26 21:11:02 localhost openaps-loop: reporting monitor/meal.json
Jun 26 21:11:12 localhost SubgRfspyLink opening serial port
Jun 26 21:11:12 localhost subg_rfspy status: OK
Jun 26 21:11:12 localhost Version: subg_rfspy 0.8
Jun 26 21:11:12 localhost Timed out or other comms error - Received an error response Timeout - retrying: 1 of 3
Jun 26 21:11:13 localhost Timed out or other comms error - Received an error response Timeout - retrying: 2 of 3
Jun 26 21:11:14 localhost Timed out or other comms error - Received an error response Timeout - retrying: 3 of 3
Jun 26 21:11:16 localhost openaps-loop: settings/settings.json  raised  need more than 0 values to unpack
Jun 26 21:11:16 localhost openaps-loop: pump://JSON/read_settings/settings/settings.json
Jun 26 21:11:16 localhost openaps-loop: Traceback (most recent call last):
Jun 26 21:11:16 localhost openaps-loop:   File "/usr/local/bin/openaps-report", line 6, in <module>
Jun 26 21:11:16 localhost openaps-loop:     exec(compile(open(__file__).read(), __file__, 'exec'))
Jun 26 21:11:16 localhost openaps-loop:   File "/home/edison/src/openaps/bin/openaps-report", line 82, in <module>
Jun 26 21:11:16 localhost openaps-loop:     app( )
Jun 26 21:11:16 localhost openaps-loop:   File "/home/edison/src/openaps/openaps/cli/__init__.py", line 51, in __call__
Jun 26 21:11:16 localhost openaps-loop:     self.run(self.args)
Jun 26 21:11:16 localhost openaps-loop:   File "/home/edison/src/openaps/bin/openaps-report", line 75, in run
Jun 26 21:11:16 localhost openaps-loop:     output = app(args, self)
Jun 26 21:11:16 localhost openaps-loop:   File "/home/edison/src/openaps/openaps/cli/subcommand.py", line 52, in __call__
Jun 26 21:11:16 localhost openaps-loop:     return self.method.main(args, app)
Jun 26 21:11:16 localhost openaps-loop:   File "/home/edison/src/openaps/openaps/reports/invoke.py", line 40, in main
Jun 26 21:11:16 localhost openaps-loop:     output = task.method(args, app)
Jun 26 21:11:16 localhost openaps-loop:   File "/home/edison/src/openaps/openaps/uses/use.py", line 45, in __call__
Jun 26 21:11:16 localhost openaps-loop:     output = self.main(args, app)
Jun 26 21:11:16 localhost openaps-loop:   File "/home/edison/src/openaps/openaps/vendors/medtronic.py", line 333, in main
Jun 26 21:11:16 localhost openaps-loop:     return getattr(self.pump.model, name)(**self.get_params(args))
Jun 26 21:11:16 localhost openaps-loop:   File "/home/edison/src/decoding-carelink/decocare/models/__init__.py", line 26, in __call__
Jun 26 21:11:16 localhost openaps-loop:     self.validate( )
Jun 26 21:11:16 localhost openaps-loop:   File "/home/edison/src/decoding-carelink/decocare/models/__init__.py", line 17, in validate
Jun 26 21:11:16 localhost openaps-loop:     data = self.response.getData( )
Jun 26 21:11:16 localhost openaps-loop:   File "/home/edison/src/decoding-carelink/decocare/commands.py", line 1309, in getData
Jun 26 21:11:16 localhost openaps-loop:     values['maxBasal'] = lib.BangInt(data[7:9]) / 40.0
Jun 26 21:11:16 localhost openaps-loop:   File "/home/edison/src/decoding-carelink/decocare/lib.py", line 235, in BangInt
Jun 26 21:11:16 localhost openaps-loop:     ( x, y ) = ints
Jun 26 21:11:16 localhost openaps-loop: ValueError: need more than 0 values to unpack
Running AMA on an Edison, any ideas?
cdaccarett
@cdaccarett
Jun 27 2016 05:30
Hello everyone i'm trying to set up the nightscout upload step and im getting an "command not found error" bash: autoconfigure-device-crd: command not found. I'm wondering if i forgot to add anything specific that i missed.
Ben West
@bewest
Jun 27 2016 05:39
crd vs crud?
also, should begin with nightscout so: nightscout autoconfigure-device-crud
cdaccarett
@cdaccarett
Jun 27 2016 05:46
@bewest looks like the "$" was the issue. Thank you
ryanchen26
@ryanchen26
Jun 27 2016 07:22
Something I've noticed with my 3 year old is that for him to stay level at night the basal stays at zero for hours at time. The problem is that in the morning we test for keytones and have been getting numbers between 1 and 4 (which is off the chart). We've been told by our medical advisors that if the BG is in range that we shouldn't worry about the keytones. Of course throughout the day my son is eating and getting boluses and that takes care of the keytones eventually. To reduce the keytones we've lowered the target, and reduced the basal program to the lowest possible at midnight (0.025) but the system never goes lower than 0.05. It's 0.05 or zero. Anyway, its a tricky business.
Chris Oattes
@cjo20
Jun 27 2016 08:47
@ryanchen26 there was a merge to dev last night that might chamge it never doing 0.025
Jeremy M Lucas (GetRileyLink.org)
@jlucasvt
Jun 27 2016 12:37
All, I wanted to announce that the RileyLink Order Site is up and running. We will be taking preorders for RileyLinks starting today and plan on shipping July 10. https://getrileylink.azurewebsites.net/. Please forward this on to those that may want to get an order in.
Brooke Armstrong Pazoles
@bpazoles
Jun 27 2016 13:21
Thanks @jlucasvt!!
Jeremy M Lucas (GetRileyLink.org)
@jlucasvt
Jun 27 2016 13:22
@bpazoles Just doing what I can. We will beat this affliction. >#wearenotwaiting
Scott Leibrand
@scottleibrand
Jun 27 2016 14:16
All: Should we be recommending Pi 2 or Pi 3 in the docs? https://github.com/openaps/docs/pull/269#issuecomment-228757931
David Cintron
@loudestnoise
Jun 27 2016 14:17
I have had no issues with my RPi3. I’m getting 12 to 14 hours battery life on a 9000mAh battery
Scott Leibrand
@scottleibrand
Jun 27 2016 14:18
Please also comment on the PR.
David Cintron
@loudestnoise
Jun 27 2016 14:31
:+1:
Sarah Hodson Grady
@shgrady
Jun 27 2016 14:51
@ryanchen26 my non-diabetic husband has more ketones in the morning than I do, sounds disconcerting though, I went to zero basal on a bike ride on Saturday for several hours, my purple strips disappeared once I got rehydrated and fed.
Paul Martin
@Spazholio
Jun 27 2016 14:54
I was always under the impression that the presence of ketones wasn't concerning if it wasn't coupled with a lack of insulin and requisite high blood sugars?
Sarah Hodson Grady
@shgrady
Jun 27 2016 15:01
but I believe if your ketones are high(from fasting or exercise), then you suddenly have a transient high blood sugar you can get into trouble quicker, because you already are a little metabolically altered, just my experience, though they say it takes very little insulin to avoid out of control ketone production, just saying the ketones can dehydrate you, complicating management a little bit.
Paul Martin
@Spazholio
Jun 27 2016 15:08
Ahhh, I see.
David Cintron
@loudestnoise
Jun 27 2016 15:21
Ok, so let’s say I want to build an Edison rig and I use dexcom G5. Is there a setup that would allow me to do that without physically having the receiver connected?
I know G4 folk use openxshareable or whatever
I guess the obvious answer is pull BGs from Nightscout
but an offline option would be nice
I’ve been playing around with a RileyLink & Loop, but after using auto-sens and meal assist I just don’t know.
@scottleibrand you gotta get oref0 working with Loop :wink2:
Dana Lewis
@danamlewis
Jun 27 2016 15:27
@loudestnoise sounds like you’re volunteering to help/do it ;)
David Cintron
@loudestnoise
Jun 27 2016 15:28
Yeah, but I lack the knowledge on writing code.
Dana Lewis
@danamlewis
Jun 27 2016 15:29
so did we when we first started :D
David Cintron
@loudestnoise
Jun 27 2016 15:29
But that’s just an excuse. Maybe I should start learning Swift, Apple just announced at WWDC an app for teaching kids how to code. :smile:
GabeA23
@GabeA23
Jun 27 2016 15:32
@loudestnoise I had the same question... looking for an offline option for my G5 as well. I have a G4 laying around too though... does openxshareable allow people to grab their BGs directly from their G4 transmitter without a receiver?
Dana Lewis
@danamlewis
Jun 27 2016 15:34
@GabeA23 no. but it allows to grab them without going to the cloud or plugging the receiver in
GabeA23
@GabeA23
Jun 27 2016 15:34
@danamlewis Gotcha... so they still have to have their receiver on their person, just not plugged in?
Dana Lewis
@danamlewis
Jun 27 2016 15:34
correct
GabeA23
@GabeA23
Jun 27 2016 15:35
Thanks Dana!
Dana Lewis
@danamlewis
Jun 27 2016 15:36
:+1:
GabeA23
@GabeA23
Jun 27 2016 15:36
@loudestnoise what's your experience been like messing with the RileyLink? I'm currently on an Edison + TI stick rig and wondering whether to fork out the $145 for a RileyLink
Dana Lewis
@danamlewis
Jun 27 2016 15:37
@GabeA23 my $.02 - if you think you’ll want to head toward a Loop setup, a Rileylink will make sense. but if you’re happy with your Edison, I don’t see any benefits to a RL over a TI
David Cintron
@loudestnoise
Jun 27 2016 15:40
I think RL is really cool, but Nate has a different perspective on how often pump communications should happen. I also have issues with my loop running consistently, but I’m not sure why
Nate being @loudnate
also, it helps if you have the x23 model pump
Dana Lewis
@danamlewis
Jun 27 2016 15:40
@loudestnoise is that with Loop, or all of your rigs?
GabeA23
@GabeA23
Jun 27 2016 15:42
Thanks @loudestnoise. Is the Rileylink smaller/more portable than your edison setup? And can it run offline?
David Cintron
@loudestnoise
Jun 27 2016 15:42
Here’s some photos my wife took of the RL - http://imgur.com/a/XUmAV
it’s pretty tiny
Sarah Hodson Grady
@shgrady
Jun 27 2016 15:43
@loudestnoise Udacity has a free class for learning Swift, I like the style of their other classes.
David Cintron
@loudestnoise
Jun 27 2016 15:43
The one edison rig I’ve seen was about the same size, especially if you’re not doing USB to connect TI stick
GabeA23
@GabeA23
Jun 27 2016 15:45

@loudestnoise gotcha. Thanks for the pics. Looks awesome! I'm currently using USB but will eventually go serial.

And can you run it offline? Because if I understand correctly, I can use xdripG5 to get my BG's directly from my G5 transmitter, which outputs a BLE signal to my Riley link, which outputs a radio signal to my pump, correct? And it seems like I can do all of that offline. I'm not sure if I'll have internet all the time at work this year.

David Cintron
@loudestnoise
Jun 27 2016 15:46
Yep, you can loop offline
assuming G5
@shgrady I’ll have to check that out
David Cintron
@loudestnoise
Jun 27 2016 15:50
Nice!
Dana Lewis
@danamlewis
Jun 27 2016 15:53
@loudestnoise consistency issues - is that Loop, or all of your rigs? (Asking because we just re-spun my rigs last night and they’re running a heck of a lot faster)
David Cintron
@loudestnoise
Jun 27 2016 15:54
just the RL + Loop. My RPi rig with a TI stick is pretty solid
Jeremy M Lucas (GetRileyLink.org)
@jlucasvt
Jun 27 2016 15:56
ITunesU has a Free Swift Course from Stanford U
David Cintron
@loudestnoise
Jun 27 2016 15:58
Nice, I see a couple actually in iTunesU
Jeremy M Lucas (GetRileyLink.org)
@jlucasvt
Jun 27 2016 16:04
@GabeA23 (iPhone + Loop App + RileyLink) = Offline Closed Loop. The current build, is starting to support Nightscout as well.
eyim
@eyim
Jun 27 2016 16:20
@ryanchen26 We got different advice from our endo. Regardless of bg, if you have med to large keytones it is an issue. Especially if he has gone without insulin for a few hours. Like you, we sometimes go several hours of low temping to keep our son stable at night. The majority of the time, this is no problem. If there is any keytones, we give him carbs to keep him up and restart insulin. If he wakes up with Med or Large keytones even with okay bg, he says he feels awful.
Rocky DuBose
@rdubose55_twitter
Jun 27 2016 16:38
hi! newbie here. but first time in shows I might have some chance of contributing.
Dana Lewis
@danamlewis
Jun 27 2016 16:39
welcome @rdubose55_twitter!
Jeremy M Lucas (GetRileyLink.org)
@jlucasvt
Jun 27 2016 16:42
@rdubose55_twitter :clap:
Dana Lewis
@danamlewis
Jun 27 2016 16:44
Medtronic has submitted their PMA to the FDA for the 670G! yay, progress on the commercial market.
cdaccarett
@cdaccarett
Jun 27 2016 16:53
Hello, im currently in the process of setting up my first pi2 rig. The doc say "A G5 can also be used, but may require some extra work beyond this setup in order to configure." I'm wondering if there are docs on how to go about setting it up with the g5.
Jeremy M Lucas (GetRileyLink.org)
@jlucasvt
Jun 27 2016 16:53
If you have an older G4 Reciever you can use that and just plug it into the RPi2.
Dana Lewis
@danamlewis
Jun 27 2016 16:54
@cdaccarett they’re not in the main docs yet (someone needs to write them up/put in the PR :)), but search back here in Gitter (web interface is better than apps for searching Gitter), there’s been several troubleshooting conversations on setting up the G5
cdaccarett
@cdaccarett
Jun 27 2016 16:55
Awesome, thank you :)
Rocky DuBose
@rdubose55_twitter
Jun 27 2016 16:56
dana! you are all over this stuff, aren't you. :) My son, now 3 yr-old and T1, and I have been following a ketogenic diet for 8 months. Very long story very short. I can't imagine a T1 not wanting to, at a minimum, at least try going ketogenic. We have NEVER had to take him to the hospital for a hypoglycemic event, even though twice he was in the 20's! Once the brain has "keto-adapted" to using ketone bodies for 80% of its fuel, it has a level of protection against low BG events. The 1st time he was so very low, my 1st though was "My God, have I killed my son?" Instead, he simply awoke and asked, "What's wrong Daddy?" I handed him a 1/4 cup of juice. He grabbed it with both hands. Smiled. Drank deeply. Thanked me and went back to sleep. Since then, he has climbed back on his weight and growth curves and seems to only want to eat if his sugars go low. None of the disorientation and confusion we hear about. The "feeling awful" spoken of above is know as "keto-flu" and is to be expected. It happens when the brain and body has not yet adapted to the ketone bodies as a fuel but hasn't sufficient blood glucose to fully run them. Hope this helps.
David Cintron
@loudestnoise
Jun 27 2016 16:56
@cdaccarett I think the main thing you do is openaps use YOURCGMDEVICE config -5
and that lets it know it’s a G5
and then try something like openaps use YOURCGMDEVICE iter_glucose and see if it returns values
Kim St. Dennis
@kimstdennis_twitter
Jun 27 2016 17:00
@cdaccarett it’s the same setup. The G5 communication is different than the G4’s and there is slightly different data received (from what I read), but there have been PRs put in to accommodate those things.
I have a G5 with share -> usb on Rpi2, plus use the iphone app -> bridge -> nightscout and can grab data either way.
Dana Lewis
@danamlewis
Jun 27 2016 17:12
::highly recommends someone put a PR into the main OpenAPS docs to reflect the G5 setup process:: :) (needs to go somewhere in the setup process, but also change the wording on the hardware page)
David Cintron
@loudestnoise
Jun 27 2016 17:15
:+1:
cdaccarett
@cdaccarett
Jun 27 2016 17:36
Thanks for the help!
David Cintron
@loudestnoise
Jun 27 2016 17:48
and PR-ed G5 for Hardware section in Phase 0! Someone already added the config info to Phase 2, so really that’s all that’s left.
Dana Lewis
@danamlewis
Jun 27 2016 17:49
thanks @loudestnoise :)
Rocky DuBose
@rdubose55_twitter
Jun 27 2016 18:16
Newbie here. Hmmm... Raspberry pi online. Answers to putty, etc. Github's "EZsetup" appeared to go well. But the plugged in, Dexcom 4 unresponsive to <python read_data>. Rather, "read_data" returns "DexCom 4 not found." So, program executes without known errors - just can't find the plugged in device. And, <dmesg> returns the dexcom's values and shows the 1st 5 lines of the device's profile. Suggestions? Or do I just need to "explore" on my own some more. (Don't know much and don't want to be a burden to you guys.)
Jeremy M Lucas (GetRileyLink.org)
@jlucasvt
Jun 27 2016 18:20
what does
openaps device show
output?
(I never can get markdown to work in Gitter)
Dana Lewis
@danamlewis
Jun 27 2016 18:22
it’s because you’re not using backticks, @jlucasvt
backtick single in front and behind this text
three backticks before and after this text
Jeremy M Lucas (GetRileyLink.org)
@jlucasvt
Jun 27 2016 18:23
\\tick tick tick\\
Dana Lewis
@danamlewis
Jun 27 2016 18:23
the backtick is option+the tilde key on my mac laptop
Jeremy M Lucas (GetRileyLink.org)
@jlucasvt
Jun 27 2016 18:23
oh i see that’s why it looked like a single quote
now i get it not single quotes
Dana Lewis
@danamlewis
Jun 27 2016 18:24
yep :) :clap:
Jeremy M Lucas (GetRileyLink.org)
@jlucasvt
Jun 27 2016 18:25
(old dog just taught new trick)
Dana Lewis
@danamlewis
Jun 27 2016 18:27
:dog: :dog2: :tada:
Kim St. Dennis
@kimstdennis_twitter
Jun 27 2016 18:50
I'm only missing emojis now.
Rocky DuBose
@rdubose55_twitter
Jun 27 2016 18:50
thx for the comeback. at work now. will check upon getting home 2nite.
Kim St. Dennis
@kimstdennis_twitter
Jun 27 2016 19:31
@danamlewis question; today my dex was off by ~70 mg/dl (dex was above fingerstick). What have you done / seen others do to accommodate this situation with openaps?
Dana Lewis
@danamlewis
Jun 27 2016 19:32
like all closed loops, we’re at the mercy of having reasonably trending CGMs…so what I’ve observed for me (and what I’m seeing many people say) is we’re less likely to go more than a few hours with trashy data.
ie if you’re on day 13 of a sensor…you might’ve used to want to get it to 3 weeks before openaps, but now many people are taking a shorter sensor life and focusing on calibrating and making sure the sensor is behaving reasonably
can’t always control that, but many people are adapting their calibration and sensor change patterns differently than how they did pre-looping
@kimstdennis_twitter curious if you don’t mind sharing, what’s your calibration patterns? and how old is the sensor?
Sarah Hodson Grady
@shgrady
Jun 27 2016 19:34
@kimstdennis_twitter was your number in motion?
like your finger stick said 30, when your CGM still said 100?
Kim St. Dennis
@kimstdennis_twitter
Jun 27 2016 19:53
@shgrady yes, dex was 200 finger stick was 130ish.
@danamlewis thanks for the feedback. This sensor just started it's second session last night. so I got my first two cals in and fell asleep. The wide variation happened this morning, but it made me think it's going to happen again at some point and I started to ponder how I would handle it. My usual calibration pattern is to cal when flat and at a decent level while trying to hit the every 12 hour mark. I've had the reoccurring idea to give the loop feedback somehow (web gui, an app or something). That needs more thought though.
David Cintron
@loudestnoise
Jun 27 2016 19:59
I’ve been getting some big jumps when I calibrate in the morning. This morning for example, I disconnected for exactly 15 minutes while showering, reconencted and resumed. Checked BG via meter and it was 40 points higher than what dexcom said. I inputted it anyway into dexcom and then it wanted another calibration. Checked again and inputted second number and then it gave new sensor reading. I wasn’t exactly flat, but I also wasn’t rising like crazy either.
2016-06-27_14-48-40.png
I normally like to calibrate in the morning before I eat anything, but maybe I should have waited for it to level out first
Dana Lewis
@danamlewis
Jun 27 2016 20:02
@kimstdennis_twitter and are you g4 or g5?
I find a significant difference when I pre-soak a sensor vs not
had to do a sensor start with no soak for the first time in a long time, and the difference for the first two days was significnat
GabeA23
@GabeA23
Jun 27 2016 20:06
@danamlewis what does "pre-soak" mean?
Dana Lewis
@danamlewis
Jun 27 2016 20:07
I put a new sensor in 12-24 hours before the end of the old sensor. I put an old transmitter in the new one just to protect it. But that way when the old sensor is done, and I put the live transmitter on the new sensor (that’s been in my body for many hours), the readings I get after the 2 hour calibration are more like days 3-7 (ie a lot better)
and then you get 7 days of good data from a sensor
Rocky DuBose
@rdubose55_twitter
Jun 27 2016 20:07
clever girl!
Dana Lewis
@danamlewis
Jun 27 2016 20:08
it works if you can plan ahead/know when you’re going to end your sensor; but if your sensor rips out, etc. it obviously doesn’t work. but when you can, it’s great
Matthias Granberry
@mgranberry
Jun 27 2016 20:09
most of my sensors get taken out by a toddler now.
Apparently arm-mounted sensors make an attractive handle when being carried.
Kim St. Dennis
@kimstdennis_twitter
Jun 27 2016 20:13
@danamlewis G5
Kim St. Dennis
@kimstdennis_twitter
Jun 27 2016 20:18
@danamlewis I'll have to try pre-soaking in the future. This was just a fluke (first day, second session, same transmitter), but triggered the though of how to handle it in the future.
Rocky DuBose
@rdubose55_twitter
Jun 27 2016 20:24
any place I could get some old g4 transmitters? I have written a fair amount of CNC code that would mill off the material necessary to replace the batteries with a great deal of precision, +- 0.003". but would need a few to experiment on. and some time!
Sarah Hodson Grady
@shgrady
Jun 27 2016 20:25
@rdubose55_twitter dead ones or almost dead ones?
Dana Lewis
@danamlewis
Jun 27 2016 20:27
@rdubose55_twitter if you don’t get any here, ask on CGM In The Cloud Facebook group
Paul Martin
@Spazholio
Jun 27 2016 20:49
@danamlewis If I don’t have a second transmitter, what do you recommend to protect the plastic flap on a soaking sensor?
David Cintron
@loudestnoise
Jun 27 2016 20:51
LOL, you just haven’t been dex-ing enough then. You gotta pay your dues and wait to have old transmitters lying around
Ben West
@bewest
Jun 27 2016 20:51
I've used piece of tape or post it note
David Cintron
@loudestnoise
Jun 27 2016 20:51
just be sure to write dexcom on the post it note
otherwise it won’t work
Ben West
@bewest
Jun 27 2016 20:51
you just don't want it flapping around a lot
Kim St. Dennis
@kimstdennis_twitter
Jun 27 2016 20:51
lol @loudestnoise
Paul Martin
@Spazholio
Jun 27 2016 20:52
Actually, I can’t get the transmitter OUT unless it’s off my body. I tried like hell, and just couldn’t do it. =)
Dana Lewis
@danamlewis
Jun 27 2016 20:52
@loudestnoise I’ve heard tape, etc.
@Spazholio you just gently lift the edge of the “leg” to pop it out, but might be hard depending on if you have it way far back somewhere on your body and can’t reach it
David Cintron
@loudestnoise
Jun 27 2016 20:53
or get your SO to do it as that’s what I do :wink2:
Rocky DuBose
@rdubose55_twitter
Jun 27 2016 20:59
concerning expired transmitters dead or almost dead - Either or both. I'm sure I'll go through a few before the procedure gets perfected. lol! Plus (being new to the CGM thing, it appears there are, at least, 2 styles of G4 transmitters. And I have no idea as to the other and many varieties this community serves. My immediate focus would be on G4's as that is what my toddler son has.
David Cintron
@loudestnoise
Jun 27 2016 21:01
@rdubose55_twitter I might have one of the slim G4s still at my house. I’ll check and if I do I’ll shoot you a PM and I can send it to you
Rocky DuBose
@rdubose55_twitter
Jun 27 2016 21:07
just for the record, the CNC code would be open-sourced to this community. not interested in monetizing this place.
Ben West
@bewest
Jun 27 2016 21:07
yes, there are basically two styles
there's some good measurements posted somewhere, lemme see
Rocky DuBose
@rdubose55_twitter
Jun 27 2016 21:09
k. thx in advance.
guess if you had some nice x-ray pictures, that'd be ideal
Rocky DuBose
@rdubose55_twitter
Jun 27 2016 21:19
fun! am very impressed with your selfless effort and tremendous talent in the coding I investigated. Afraid I have a TON of catching up to be very helpful.
Dana Lewis
@danamlewis
Jun 27 2016 21:28
In lieu of hurried phone typing (which is what usually what happens when I'm trying to explain soaking a sensor), I wrote it up so I can just link back to it. That being said, I wanted to make sure it makes sense, so would love for folks here to read it and ask Qs of anything that needs clarifying! https://diyps.org/2016/06/27/how-to-soak-a-new-cgm-sensor-for-better-first-day-bgs/
Lynne Butts
@flutelynne_twitter
Jun 27 2016 21:31
@danamlewis Just read through it. Looks great to me!
Ben West
@bewest
Jun 27 2016 21:32
I can confirm using a simple piece of tape works
I've used post it notes when there was no tape available
Dana Lewis
@danamlewis
Jun 27 2016 21:36
thx @flutelynne_twitter and :+1: @bewest
Garrett Webb
@garetis
Jun 27 2016 21:36
@danamlewis I didn't read the leadup to your comment, but I know for the Medtronic CGM that is discouraged, because whatever material is on the sensor, if it is in the body without a charge then hydrogen peroxide will build up and potentially cause damage... But, as you noted, if you have an old transmitter that con contribute some charge then it might not be an issue. Regarding Dexcom and their materials, I don't know... YMMV, obviously, and if it works, then it works. Just fyi that you may want some caveat in there
Dana Lewis
@danamlewis
Jun 27 2016 21:37
good point, I’ll specify that I’m talking about Dexcom sensors in this post! @garetis
Garrett Webb
@garetis
Jun 27 2016 21:52
Does anyone know why treatments wouldn't show up, or "no treatments to upload" would show up with openaps upload-recent-treatments?
Ben West
@bewest
Jun 27 2016 22:00
it's expected for there sometimes to be nothing to upload
if nothing happened on the pump in the last 5 minutes, there won't be any treatment info to upload
that can easily happen if the current settings are fine with no changes needed for next half hour
Dana Lewis
@danamlewis
Jun 27 2016 22:03
if you think that’s the case ^ and want to nudge it, you could always set a temp slightly higher or lower than what it is now, and watch it read and react to that in order to test it
Paul Martin
@Spazholio
Jun 27 2016 22:11
Has anyone here tried their hand at making a Pebble watchface? Wondering how easy it would be to take an existing one, and add color to it...
Dana Lewis
@danamlewis
Jun 27 2016 22:15
Nah. We were just using cards for pebble and piping some json into it.
Paul Martin
@Spazholio
Jun 27 2016 22:16
Oh? Huh. But as an app, not a watch face? Or….
Also, I thought you were using Urchin?
Paul Martin
@Spazholio
Jun 27 2016 22:22
Another silly question (for anyone using Urchin) - can you explain what these are:
IMG_6062.JPG
I assume the “1.9 x 26” is saying “1.9u/hr for the next 26 mins”...
Dana Lewis
@danamlewis
Jun 27 2016 22:25
I use cards for pebble as an app and also urchin now, for various reasons.
And yes, that's what it should be. You can customize what you want to go in there/how to display it...
Paul Martin
@Spazholio
Jun 27 2016 22:26
What’s the “2: -1.0u -> 99” mean?
I’ve customized it quite a bit (basically turned everything on) but wasn’t entirely certain what some of those were.
I think I’ll try that “Cards For Pebble” thing. Seems interesting. Any chance I could get a pic of what it looks like when you get a moment?
Sarah Hodson Grady
@shgrady
Jun 27 2016 22:29
@Spazholio I was wondering if you had a moment to walk me through why my preflight is failing?
Dana Lewis
@danamlewis
Jun 27 2016 22:29
This is my output after using cards for pebble. https://diyps.org/2015/11/06/the-power-of-visualizing-your-data-your-way/
Paul Martin
@Spazholio
Jun 27 2016 22:32
Sweet yeti of the Serengeti!
@shgrady Sure thing. Hit me. What’s going on?
Dana Lewis
@danamlewis
Jun 27 2016 22:34
@Spazholio what, too much data for your tastes? ;)
Paul Martin
@Spazholio
Jun 27 2016 22:34
That’s just….that’s something alright. =)
I’ll admit, I’m partial to data-heavy as well though.
Sarah Hodson Grady
@shgrady
Jun 27 2016 22:36
@Spazholio actually, I think I may be onto it already, my gather alias has an error because get-profile is configured later in the setup, meaning I used it before building it.
Paul Martin
@Spazholio
Jun 27 2016 22:37
Yay, I’m helping!
(also, I’ve absolutely done that before when I get really into mucking with my loop) :smile:
@danamlewis In that screenshot, do you know what the “2: -1.0u -> 99” is in reference to? I imagine that the “99” is my predicted BG, but what is the “2: -1.0” indicating?
Dana Lewis
@danamlewis
Jun 27 2016 22:40
2 minutes ago perhaps? And do you have -1.0 net iob?
Or did you, during the screenshot time?
Sarah Hodson Grady
@shgrady
Jun 27 2016 22:41
Yes @Spazholio you are
Paul Martin
@Spazholio
Jun 27 2016 22:43
@danamlewis Nailed it. Thank you.
Sarah Hodson Grady
@shgrady
Jun 27 2016 22:51
the requirements when adding the virtual device processes are components of the reports/openaps functions right? they don't exactly match my report names.
Paul Martin
@Spazholio
Jun 27 2016 22:52
If you’re asking what I think you are, then…yes? Can you give me an example?
Sarah Hodson Grady
@shgrady
Jun 27 2016 22:52
This message was deleted
openaps device add get-profile process --require "settings bg_targets insulin_sensitivities basal_profile preferences" oref0 get-profile
see, some of those are my json report names.....
but then these are not $ openaps device add calculate-iob process --require "pumphistory profile clock" oref0 calculate-iob
well clock is...
Paul Martin
@Spazholio
Jun 27 2016 22:57
This message was deleted
I’m actually going to defer these questions to @bewest, as he’s the SME for oref0.
Sarah Hodson Grady
@shgrady
Jun 27 2016 22:59
meaning I just put them there with the command? building a database for future deposit?
Ben West
@bewest
Jun 27 2016 23:00
those aren't file names, those are the names of the arguments, eg for the help message
yes, it puts those in the config file so openaps knows that it should expect at least 4 arguments with some indication as to what the arguments might mean
that way if you only supply 3 arguments, openaps can throw an error explaining you don't have enough arguments and what the expected arguments are
so what "device add" is telling openaps for these is that the implementation is provided by some external script and the script expects bunch of arguments in order for the use, so those field names are used to format the help output and the parser for the corresponding use's and by extension the reports
the report names themselves can be whatever, these are "argument names" -- a description of the inputs
the easier way to do that
Paul Martin
@Spazholio
Jun 27 2016 23:03
How does get-profile know that the bg_targets mentioned in the argument list is actually referring to the settings/bg_targets.json file (for example)?
Sarah Hodson Grady
@shgrady
Jun 27 2016 23:03
gotcha, its not really how I have defined a function before, but it makes sense, thanks
Paul Martin
@Spazholio
Jun 27 2016 23:04
I wasn’t 100% sure of that, so didn’t want to answer @shgrady in case I was wrong.
Ben West
@bewest
Jun 27 2016 23:04
I believe it's in dev now, is simply oref0 template mint device oref0 | openaps import
it knows by the order you pass the arguments in
so, "first argument is settings"
etc
Sarah Hodson Grady
@shgrady
Jun 27 2016 23:05
like if I say do(age, name, dob), I can run do with (herage, hername, herdob)
Ben West
@bewest
Jun 27 2016 23:05
that's exactly what's happening here
it's defining the signature
you can use whatever arguments you like with the use
device is defining the use's in the same way your function signature defines the signature, and then you can apply any arguments you like to it
basically, under the hood, what happens is the shell's use has an argument parser that gets parser.add_argument(required_field_from_sig) for each of those fields defined by the device
Sarah Hodson Grady
@shgrady
Jun 27 2016 23:07
danke
Ben West
@bewest
Jun 27 2016 23:10
highly recommend trying oref0 template mint device oref0 | openaps import it sets up all the oref0 devices correctly in one shot :-)
Sarah Hodson Grady
@shgrady
Jun 27 2016 23:12
and that overwrites the ones I just put in right?
pi@artificialpancreas:~ $ open
aps device show
process://ns-status/ns-status/
process://ns/nightscout/ns https://LALALA.azurewebsites.net GOOD GRIEFY SECRETAPIHASREMOVED
dexcom://g4nonshare
process://curl/bash/-c "curl -s https://LALALA.azurewebsites.net/api/v1/entries.json | json -e 'this.glucose = this.sgv'"
medtronic://my723
process://oref0/oref0/
process://get-profile/oref0/get-profile
process://calculate-iob/oref0/calculate-iob
process://determine-basal/oref0/determine-basal
process://detect-sensitivity/oref0/detect-sensitivity
Chris Oattes
@cjo20
Jun 27 2016 23:16
you only removed the url once?
Ben West
@bewest
Jun 27 2016 23:16
just the oref0 ones, yeah
with the ns device you don't need the curl stuff
looks fine
Sarah Hodson Grady
@shgrady
Jun 27 2016 23:18
can you clarify the ns not needing the curl stuff? how to fix?
like it does not need the word curl or that the entire line's function is occurring in the NS Virtual Device?
Chris Oattes
@cjo20
Jun 27 2016 23:20
the ns device has functions which grab the data from the nightscout site
Ben West
@bewest
Jun 27 2016 23:20
hmmm
so the way I like to think about this
is to determine 1.) how/where glucose data is going to come from
that curl line is one of many ways to get glucose data
but you'll probably really only use one way to get glucose data
Sarah Hodson Grady
@shgrady
Jun 27 2016 23:21
see I was hoping to do two,
Ben West
@bewest
Jun 27 2016 23:21
and it looks like you're planning on using usb cable to get it from dexcom receiver
my recommendation is to focus on getting one working :-)
Sarah Hodson Grady
@shgrady
Jun 27 2016 23:23
yeah, I think you told me that, and I had not figured out what I needed to remove yet.
if I was starting with NS, remove dexcom, if starting with dexcom, I still need to get other things from NS, just not blood sugars
for instance I put carbs into NS, that I don't put into either pump
Sarah Hodson Grady
@shgrady
Jun 27 2016 23:47
ok, removed dexcom and curl
cdaccarett
@cdaccarett
Jun 27 2016 23:58
I deployed the master version for nightscout to setup a profile but it seems not to save my new profile should i be doing something different?