#circuitpython-dev
manic glacierBOT
tulip sleet
@idle owl These boards reset the neopixels on soft reset:
ports/nrf/boards/circuitplayground_bluefruit/board.c
ports/atmel-samd/boards/adafruit_slide_trinkey_m0/board.c
ports/atmel-samd/boards/pybadge/board.c
ports/atmel-samd/boards/circuitplayground_express_crickit/board.c
ports/atmel-samd/boards/blm_badge/board.c
ports/atmel-samd/boards/pybadge_airlift/board.c
ports/atmel-samd/boards/circuitplayground_express/board.c
ports/atmel-samd/boards/pygamer_advance/board.c
ports/atmel-samd/boards/circuitplayground_express_displayio/board.c
ports/atmel-samd/boards/pygamer/board.c
ports/atmel-samd/boards/neopixel_trinkey_m0/board.c
ports/atmel-samd/boards/adafruit_proxlight_trinkey_m0/board.c
idle owl
<@330227457296957440> These boards reset the neopixels on soft reset:
```
ports/...
Should NeoKey Trinkey be added to that? Why only a limited list of boards?
tulip sleet
i think it was supposed to be any board with neopixels beyond the status led
idle owl
Oh. Ok.
tulip sleet
if there are some missing they should be added
idle owl
NeoKey is only one LED. But with the key on it, it's more than status I guess.
I imagine slide and proxlight only have one as well though?
idle owl
I had RGB code running, then when I saved a new file that had no LED stuff in it, it stayed blue.
which is where it left off from the previous code.
(Blinking red green blue.)
tulip sleet
so it didn't change due to status code. So that may be new from the status led changes. we should try it on something innocuous like a Gemma: suppose you change the status LED in your program, what happens
honestly I don't remember
or have ever tried
idle owl
I don't follow
tulip sleet
suppose I had a Gemma running 6.2.0, and i changed the status LED color, and then did a soft reset. What did it do in 6.2.0, and what does it do now with the new status LED changes? And if it's different, is that deliberate?
idle owl
Ok, so I went back to blink code, then saved new code, and it stayed red. Then I went to the serial console, and did ctrl+d and it shut off.
I'm running latest here.
idle owl
I'll try it on Rotary Trinkey after lunch. See if they both act that way.
slender iron
we could add a clear after user code is done
storm minnow
Is there a full list of modules somewhere? Trying to do a custom build now, want to turn off whatever I can to make space
tulip sleet
look at the module list for your board, and then you can turn off what is still on
what are you trying to make space for? turning off modules will not increase RAM space
so if you are getting an allocation error, that will not be corrected
storm minnow
I needed to add... I think it was displayio? and I think it was just storage
storm minnow
It's the Slide Trinkey, trying to modify CP so I can run an OLED on it
tulip sleet
tulip sleet
so you could turn off several of those, especially math, _pixelbuf, analogio,...
displayio is big and you may still not succeed
storm minnow
Ok, going to turn those off. It's worth a try!
tulip sleet
usb_midi also, usb_hid if you don't need it
that's probably it, random, I guess, but it's not large
storm minnow
So they'll be CIRCUITPY_USB_HID = 0 CIRCUITPY_USB_MIDI = 0 CIRCUITPY__PIXELBUF = 0 CIRCUITPY_ANALOGIO = 0 CIRCUITPY_MATH = 0 right?
And remove the SimpleMath library?
tulip sleet
it's CIRCUITPY_PIXELBUF (inconsistent omission of extra underscore)
tulip sleet
remove HID and NeoPixel from the frozen library list too if you don't need them. they are big
if you are turning off native HID then you certainly don't need the frozen library
storm minnow
I do still want to use the NeoPixel, but I don't think I need HID unless I want to have it react
tulip sleet
if your remove _pixelbuf, then you need to use the python adafruit_pypixelbuf library instead
storm minnow
Does it make a difference on my ending size? What exactly is the difference between a module and a frozen library?
tulip sleet
a frozen library is python code compiled to essentially an mpy. a native module is C code
but a frozen library does not use up (much) RAM when imported
storm minnow
Ah...
tulip sleet
it's also true that our sizing takes into account the largest translation. The English translation is smaller, so there is already some room if you don't need a different translation
storm minnow
Ohhh, ok. I'm currently just targeting English since I have a feeling I'm the only person who will use this XD
Now one thing I'm confused on... if I don't explicitly set modules to 0 or 1, do they get loaded or not?
tulip sleet
there are many default values, which depend on the particular chip and aspects of the build
so just set those to 0 in mpconfigoard.mk. You do no need to explicitly enable anything except displayio
storm minnow
Ok, cool. So I'm gonna try building it now
tulip sleet
how many cores on your CPU?
storm minnow
That I'm building on?
crimson ferry
@storm minnow enable busio too?
tulip sleet
good point, need i2c
storm minnow
Ah, I probably should, lol
tulip sleet
That I'm building on?
yes
storm minnow
2 cores, 4 threads
storm minnow
Ok, so... I had copied the directory from the Slide Trinkey, modified the files in there, but get make: *** No targets specified and no makefile found. Stop.
tulip sleet
you must run make when you are in ports/atmel-samd
storm minnow
Ah, yeah, I was in boards
Building :D
darn
Ok, so, is there a way to reduce my space usage while still having displayio and NeoPixel?
tulip sleet
10k is pretty hopeless given all that you've turned off. You could shrink the size of CIRCUITPY though that restricts your program size, but we rarely do that. in ports/atmel-samd/mpconfigport.h, you can change #define CIRCUITPY_INTERNAL_FLASH_FILESYSTEM_SIZE (64 * 1024) to say (48 * 1024).
you are in uncharted territory here, maybe some alternative use for this no-slider slider is in order
stuck elbow
analogio is surprisingly large
storm minnow
Heh, I didn't expect to get into something that would make the pros scratch their heads
stuck elbow
touchio?
tulip sleet
upload your mpconfigboard.mk using the + sign on the left
tulip sleet
it's trying to put 8 lbs in a 5 lb sack
ionic elk
@storm minnow what are you trying to put on there?
tulip sleet
those should be off already
stuck elbow
it's all in the right use of your knee
storm minnow
Uhhh, can I attach directly from my WSL files? Or do I have to copy out?
tulip sleet
click the + sign and select the file
storm minnow
I mean, can I navigate to it in File Explorer? I don't know where it lives, wasn't sure if it was like a VHD where it's only visible from inside the guest system
tulip sleet
storm minnow
Trying to get displayio in to use an OLED, heavily modifying a Slider Trinkey
tulip sleet
ah, i see, not sure if the wsl filesystem is visible or not
copy/paste it into an editor, or just paste it inline; it might be small enough
storm minnow
I can't seem to copy from WSL to something else
Oh, lemme look at that
Sorry that took so long, this thing is struggling XD
tulip sleet
how is it if you comment out frozen neopixel (which does not have to be frozen to be used)
turn off PIXELBUF also for now
tulip sleet
it is a helper for neopixel, but is not completely necessary if you use adafruit_pypixelbuf instead
(but that uses up RAM)
storm minnow
Ah
tulip sleet
also do CIRCUITPY_BUSIO_SPI = 0
that is a tricky thing I put in recently to remove busio.SPI on a board that could not use it
storm minnow
Ok, changes made, let me try it
building
this laptop is going to implode XD
I forgot to do -j4, but everything is at 100% so I dunno if it would make a difference
4044 bytes overflowed DX
But that's less than half of what it was before!
tulip sleet
you can look at things in py/mpconfig.h to turn off; some are already. But this is a lot of work for a $10 board
another $10 board like RP2040 would do much better
it's a good exercise to learn the build
storm minnow
At this point I just want to do it for the sake of doing it, lol. I have a pile of Picos, I just like this experience!
tulip sleet
also look at mpconfigport.h and mpconfigport.mk. If you don't need struct, could turn that off, but libraries may use it
storm minnow
What does struct do? I remember someone mentioning it was important
tulip sleet
packing fields into bytestrings and vice versa
storm minnow
ah
still zephyr
it was me ๐
storm minnow
You were one of the instigators of this madness :P
still zephyr
At this point I just want to do it for the sake of doing it, lol. I have a pile ...
I have some issues to tackle if you want the memory alocation problem experience ๐
storm minnow
Maybe after I figure out the space issue XD
still zephyr
Great ๐
storm minnow
Ok, so looking at mpconfigport.h... trying to figure out where exactly I put the #define for space?
Ah, found it!
Fingers crossedl but I think this might do it!
storm minnow
SUCCESS!
tulip sleet
CIRCUITPY will reformat because its size has change, or if it s messed up, then do storage.erase_filesystem()
storm minnow
Gotcha. Actually, I think I have enough room to add NeoPixel and PIXELBUF back
That would make sense, right?
tulip sleet
yes, if you are going to use them
storm minnow
Yeah, I want to be able to use the NeoPixels still for status lights
mmmmaybe I should close some things so it doesn't take so long...
Does it matter what order modules are in for enabling/disabling them? or can I have them mixed up?
I figure best practice to have them all deactivated in a group and activated in another, just curious
jaunty juniper
doesn't matter, you are just setting compilation variables
storm minnow
Ok, thought so
Recompiling!
Rebuilding?
much faster without the processor being pinned 100% beforehand XD
jaunty juniper
recrossing your fingers
storm minnow
Ok, so now I have my code and libs on -- did I wire something wrong, or did I miss something in the build?
I don't know if there are pull-ups built on on those pins? Or would I have to add one?
jaunty juniper
there's not pull-up on the board, some breakouts have them
storm minnow
there's not pull-up on the board, some breakouts have them
Ok, so I've never added a pull-up resistor XD so it's a resistor from VCC to the pins in question?
or VDD, whichever I'm getting power from?
jaunty juniper
yes, like 5-10 kฮฉ
tulip sleet
pull up to 3.3v, not 5v
storm minnow
Ok, thanks. Lemme dig out a resistor and try it out!
tulip sleet
you need two resistors, one on SDA, one on SCL
storm minnow
Gotcha. I wonder if this is the last hurdle
idle owl
There's never a last hurdle. ๐
storm minnow
Hehe
I have to say, this is the most I've dug into something in a while! Usually I spend lot an hour at most and do some relatively simple changes to things XD
atomic summit
Yeah it has different I2C pins. I'm not sure if folks are still using the pre-release boards, but I sold a few hundred so possibly.
tulip sleet
no problem, then, if there are significant numbers of potential users, that's fine
thanks for the response
ionic elk
@tulip sleet I have a decision to make for RP2040 deep sleep. The RP2040 does not lose the state of RAM in deep sleep, even if you put it in "shutdown" mode. So for true deep sleep, I need to either
a) remove NORETURN from the global API for common_hal_alarm_enter_deep_sleep and actually let it return for the RP2040, in which case it would return to run_code_py in a similar way to fake deep sleep
b) run reset_cpu() after waking, but since the RP2040 doesn't actually expect to ever lose memory, it doesn't have easy ways of storing wakeup information across resets, so I would have to either ditch alarm.wake_alarm or use NVM.
note that RP2040 deep sleep is only different in terms of its treatment of RAM. It still shuts down USB and peripherals, so just using fake deep sleep all the time isn't sufficient.
storm minnow
So I added a 9.8~kOhm resistor to each SDA and SLC -- but still getting the error
tidal kiln
@gloomy shuttle hey, i think you were asking about this:
https://learn.adafruit.com/circuitpython-libraries-on-any-computer-with-raspberry-pi-pico/other-rp2040-boards
(recently made live)
storm minnow
Oh, I confirmed I'm seeing resistance between SDA/SLC and 3.3V, so I know they're connected...
tulip sleet
<@!329766224093249548> I have a decision to make for RP2040 deep sleep. The RP20...
Is the RAM preserved after reset_cpu()?
ionic elk
no
tulip sleet
the datasheet says that? It's cleared?
ionic elk
it's the watchdog... let me double check
tulip sleet
maybe there are some preserved registers?
ionic elk
Yeah actually there are! I didn't look at the watchdog
tulip sleet
the wathdog has some scratch registers. We'd have to see if the pico-sdk or bootloader uses some of them
tulip sleet
i was just looking at the ESP32-S2, which has preserved registers, but they seem to all be reserved! We are using one but it's not working with TinyUF2
but if RAM is preserved, then you could leave it anywhere
ionic elk
Could you rephrase? I'm not totally getting your argument.
I doubt that the watchdog preserves flash since its supposed to get you out of infinite loops which it wouldn't do if the program counter was preserved
tulip sleet
i'ts just an aside about the ESP32-S2, not about the RP2040
ionic elk
but the scratch registers it provides are fine for preserving state, the STM32 has them too, that's what I use there
tulip sleet
so look at watchdog_reboot() in the pico-sdk, because it uses some scratch registers
again, if RAM is NOT cleared, then you can just store the state in RAM even in deep sleep
ionic elk
it doesn't specifically seem to say if RAM is cleared or not, but it says it reboots the processor via the reset controller, so I think it's implied?
And I don't know why they would include scratch registers if the RAM doesn't go away
So I'm pretty sure it gets reset like everything else, just like most chips
RP2040 deep sleep is just funny because it has a low-power mode for the RAM that doesn't require you to actually shut it off, and they don't give you shutoff options
tulip sleet
how is state saved on the nRF in a deep sleep?
ionic elk
It isn't, NRF has specific flags like NRF_POWER_RESETREAS_RESETPIN_MASK that provide the reset reason directly
same with ESP32-S2
STM32 is annoying about reporting what it reset from, hence why I use scratch
tulip sleet
oh, so we reconstruct it. so on STM you used the RTC or equiv registers, have you looked at atmel-samd yet?
ionic elk
no
Haven't done anything for samd, sounds like it might not be next priority for me
Anyway, I'm all set with these scratch registers, thanks for helping me find that
tulip sleet
i'd say look at that too, to plan ahead. yes it may make sense to drop the NORETURN, depending on SAMD.
Yeah, but doublecheck if someone else is using them. That's the problem on ESP32-S2 (We were using one for something else, not deep sleep, and it doesn't seem to be surviving a reboot that goes through the UF2 bootloader)
we were using it to support the safe mode reason
ionic elk
@tulip sleet the RP2040 datasheet seems to be written with the SDK in mind, and it only makes mention of magic numbers across a couple of registers. So I think I'm in the clear to use a coupel of them
Their use is featured in datasheet examples, so
storm minnow
@tulip sleet so I added pull-up resistors, but still getting the error about pull-ups... Attached to 3.3V like you said. Is there a software side I might have missed?
storm minnow
Pins are SWC and SWD [which I believe to be PA22 and PA23?], display is the 0.91" OLED 128x32 with STEMMA
tulip sleet
those pins may not work - they are debugging pins. I though you would be using the potentiometer pins?
storm minnow
It was intimated that those pins might well be usable, and I connected to those since I'd need 2 pins, and I wanted to try leaving the touch pad free [but that can be used like a regular pin, so I was told] if SWC and SWD don't work
tulip sleet
so you have jumpers soldered to the round SWD/SWC pads?
storm minnow
Indeed
tulip sleet
i'll be back later after looking at the schematic
storm minnow
Ok, I'll be around, just ping me :D
still zephyr
Ok, I'll be around, just ping me :D
Question is this an Adafruit display?
the 0.91?
storm minnow
Yup, that's the one
still zephyr
so I can confirm that it works with the samd21 with the resistors
I do not know if you can display anything
That I did not try
Just to clarify here, I am talking about the QT py
storm minnow
The big thing we're not sure about is if we can use the pins I picked, lol
still zephyr
mmm, either case we will learn a lot
storm minnow
I've learned how to build custom CircuitPython at the very least!
still zephyr
yes, I have never build CP for the samd21
my comuter takes like 5 minutes to build any board
my raspberry 4 pi is faster
storm minnow
It took 10 minutes on laptop here at one point XD
still zephyr
so what are the plans for the picos?
storm minnow
Uhhh... I dunno yet, lol
ember iris
Mind if I ask a couple of questions: first--I'm working on the SGP40, and am about to test that I can get the raw score for the sensor with humidity compensation. That's going to be an important feature for the voc algorithm for that sensor, but it can be a stand alone improvement as well--would it be better to make that it's own pull request, or should I hold off until I have the voc algorithm running as well?
tulip sleet
@storm minnow the SW pads should be ok. So I think it's the wiring? could you give a link to the product page for the display?
still zephyr
Mind if I ask a couple of questions: first--I'm working on the SGP40, and am abo...
are you making examples as well?
still zephyr
normally we try to add example for new features, that why folks can see these improvements and the reviewer can use them as test code
tulip sleet
Mind if I ask a couple of questions: first--I'm working on the SGP40, and am abo...
is the raw value useful? It's fine to separate if you think the voc is a long-term thing. It doesn't matter all that much. It's easy to do releases
ember iris
are you making examples as well?
Oh! I certainly can add examples. I'm adding it so I can have better sensors and learn the workflow to submit things to circuit python, but when I'm done with the code I'll need someone to point me to where to add the examples
still zephyr
Oh! I certainly can add examples. I'm adding it so I can have better sensors and...
ping me if needed I can help you with that
tulip sleet
@storm minnow that display board already has pullups. Did you wire all four appropriate pins: SCL, SDA, 3.3v power, and ground? which power pin did you use on the board?
ember iris
is the raw value useful? It's fine to separate if you think the voc is a long-te...
Right now it just returns a random measurement that's kind of related to voc stuff, in fact it leads to my second question question
The second question could be an issue on github, but since I'm working towards a pull request I figured I'd ask it here. On this line: https://github.com/adafruit/Adafruit_CircuitPython_SGP40/blob/66ddc29e46ca20c4a2bc7e8c2aab460294da12e5/adafruit_sgp40.py#L39
the 'no humidity or temp reading' command is constructed (in this sensors instance you fill it in with dummy values) The values chosen don't match up with the data sheet. I don't imagine it matters much but it could conceivably throw someone off. Is there a reason for it, should I change the default values, or should I leave it as is?
storm minnow
<@!381566293368897537> that display board already has pullups. Did you wire all ...
Oh... so I shouldn't have gotten the error XD Yes, SCL, SDA, GND, and VIN, because the other one was 3Vo which I thought was not the right one. Used the lone pin close to the USB from the slider for GND
still zephyr
The second question could be an issue on github, but since I'm working towards a...
in this case I prefer to open an issue, and then manetion in the pr that you are solving that issue with the PR
ember iris
in this case I prefer to open an issue, and then manetion in the pr that you are...
Awesome, should I open the issue and make the pull request at the same time or hold off a bit?
still zephyr
it is up to you :), the issue would be good because that way we do not forget that we need to solve that
storm minnow
are you measuring 3.3v at the display board?
Yes
still zephyr
either with your PR or without ๐
tulip sleet
Yes
at the 3v OUT pin on the display board, are you measuring about 3v (will be about .3v lower than the input voltage)
ember iris
either with your PR or without ๐
Ok, if that's the case I just need to test the changes a bit, format accordingly, then this evening I'll open the issue and make a pull request. I think it'll make more sense to make two pull requests--one for sgp40_raw measurement with humidity compensation, and one for the voc algorithm. In the humidity compensation I'll adjust the default dummy variables as well. Does that seem like a good workflow?
storm minnow
at the 3v OUT pin on the display board, are you measuring about 3v (will be abou...
I'm getting 3.26V from both VIN [soldered] and 3Vo
still zephyr
Ok, if that's the case I just need to test the changes a bit, format accordingly...
yes,
sound like a plan
storm minnow
is 3vo connected to anything?
No, floating
still zephyr
Awesome thank you!
No, thank you!
tulip sleet
No, floating
there is an "on" tiny LED on the back near one of the STEMMA connectors. is it on?
storm minnow
Yes
tulip sleet
ok, then I would say there is something funky about the wiring to the SWD and SWCLK pads. Why don't you try toggling those by hand with DigitalInOut's, disconnecting them from the display board. Measure at the end of the wire, not the pad. If they seem OK, try connecting some other I2C board up (like a sensor). The fact that it said it thought there were no pullups is very odd, because the display board has them
remove the resistors you added
storm minnow
Ok, start with the resistors... How do I hand toggle digital IOs?
tulip sleet
wait, i know the problem, you have the wrong pin numbers, they are PA20 and PA21. see the schematic I posted above
(for DigitalInOut help, and lots of other stuff, see https://learn.adafruit.com/circuitpython-essentials/circuitpython-digital-in-out
storm minnow
Oh, glad I didn't disconnect the display yet XD
Using 20 and 21 instead of 22 and 23
tulip sleet
Oh, glad I didn't disconnect the display yet XD
Take away the corresponding IGNORE_PA20and PA21 lines in mpconfigboard.h
storm minnow
Done, rebuilding
Hmm... I commented out those lines and rebuilt, but same error
@tulip sleet I see they have 1 next to all the items -- do I comment/delete the lines or change the 1 to 0?
tulip sleet
delete them
storm minnow
Ok, let me try again
Ok.... so... it's acting strange. I rebuilt CP, copied it over, same error. I tried using storage.erase_filesystem() but it rebooted and then still showed CIRCUITPYTHON with files intact
@tulip sleet I know there's a flash_nuke.uf2 for the Pico if it starts having these weird issues -- does that file only work for the Pico? Or is there a different file for different boards?
idle owl
<@!329766224093249548> I know there's a `flash_nuke.uf2` for the Pico if it sta...
Check the Troubleshooting page in the Welcome to CircuitPython guide, and try the Trinket eraser. No guarantees.
storm minnow
Worth a shot. Thanks!
Ok, did that eraser, it cleared, redid CP, came back clean.... but getting the same error again
still zephyr
Ok, did that eraser, it cleared, redid CP, came back clean.... but getting the s...
Oh no, are you going to use the pico now?
storm minnow
Oh no, are you going to use the pico now?
That's not the point of this exercise :P
I'm gonna keep trying until someone tells me it's impossible XD
Hmm... Should my builds be the same size with the ignore statements for PA20 and PA21 removed?
tulip sleet
they will be slightly larger
storm minnow
Hmm, I think something is wrong then, it didn't change
storm minnow
Hmmmmm.... ok... so I went into pins.c and added lines for PA20 and PA21, but it said boards/oled_trinkey_m0/pins.c:7:45: error: 'pin_PA21' undeclared here (not in a function); did you mean 'pin_PA31'? 7 | { MP_ROM_QSTR(MP_QSTR_SD1), MP_ROM_PTR(&pin_PA21) },, so I changed them to 30 and 31, and it built like that, but it says "Invalid pins" using those designations [either as microcontroller.pin.PA30 or board.SD0 [because I did SD0 and SD1 because I don't know which one is SDA and which is SDC yet]]
Well... I need sleep DX leave pings with any suggestions. Good night, thanks for the help!
onyx hinge
the suggestion of using PA30/31 is the compiler doing its best but giving bad advice
manic glacierBOT
storm minnow
the suggestion of using PA30/31 is the compiler doing its best but giving bad ad...
Once more I am deceived by the machines!
storm minnow
Hmmmm.... this is weird. So if I remove the ignores for PA30 and PA31, it'll build, and I can add entries to pins.c. If I remove ignores for PA20 and PA21, it doesn't let me add entries for them in pins.c and builds, but doesn't recognize them when I try to call them
tulip sleet
Hmmmm.... this is weird. So if I remove the ignores for `PA30` and `PA31`, it'll...
alright, I read the schematic wrong. It is PA30 and PA31 for SWCLK and SWDIO respectively. The schematic I was looking at is blurry and the "3"s are mashed up. So forget about PA20 and PA21 -- I'm sorry. .... However, PA30 and PA31 support PAD[2] and PAD[3] on SERCOM1 , and you need PAD[0] and PAD[1] for I2C. So those pins are not going to work for I2C. AND, neither are the slider and touch pins. They are PA02 and PA07, and they similarly do not support a PAD[0] and PAD[1] for I2C.
On some chips you can use any pins for any function. That is not true on the SAMD line.
so this whole exercise is failing up front. You learned a lot about builds, but your goal is not going to work, sorry!
storm minnow
Ah, that explains it XD Oh well! Thankfully solder is cheap and yes, I did learn a lot!
Thanks for all the help in my fruitless adventure XD
tulip sleet
an alternative to all this is to get another slide pot ๐
storm minnow
Heh, yeah... I did manage to get the old one working after disassembling it, but I've mangled the pins on it and the pads on the Trinkey enough that I couldn't it back on XD
tulip sleet
i'm curious what the pot failure mode was? Did something get bent or pushed out of place?
storm minnow
I'm honestly not sure! before opening, resistance read like 8kOhm regardless of position. Opened it up, got 10kOhm across the one side, nothing on the other. Looked at the pins, seemed good... squeezed it together and it worked
I think something might have been blocking that I didn't see fall out?
tulip sleet
so maybe a foreign body on the slider or something or somehow the slider was not making contact. hmm, thanks
storm minnow
Possibly. No problem!
storm minnow
@tulip sleet Oh, just so I don't try something else that won't work, since I have a feeling I don't have more to learn in the CP build direction from the same sort of exercise... what can I do with the pins that are exposed, since we've eliminated I2C? Touch and the slider pin I assume can be used for any regular digital IO functions?
tulip sleet
right, and same for SWCLK and SWDIO.
they can all be used for timers (e.g. PWMOut)
PA02 can be used for AnalogIn. It is not connectable to any SERCOMs, so no use for I2C, SPI, UART, etc.
you can use PA31 and PA30 for SPI SCK and MOSI, so you could drive an SPI display (since it does not need MISO; it's output only). This works:
spi = busio.SPI(microcontroller.pin.PA31, microcontroller.pin.PA30)
(I tested it)
this works:
u = busio.UART(microcontroller.pin.PA30, microcontroller.pin.PA31)
that's about it; this is all from studying the datasheet, chapter 7
storm minnow
Ahh, ok, cool. I hadn't dug that far into the datasheet, I only found the pinouts XD
I'll have to see if I have an SPI display...
gloomy shuttle
@tidal kiln I just submitted a PR to fix the led_on example to initialize the keyboard.
gloomy shuttle
<@!337989520962093057> I just submitted a PR to fix the `led_on` example to init...
I think more work needs to be done. I didn't look at the other examples that need the same thing.
tidal kiln
know how to run black?
tulip sleet
it would be good to set up pre-commit
gloomy shuttle
know how to run black?
yes, I didn't realize I hadn't ran pre-commit install in this repo.
ionic elk
blegh, this sublime upgrade
what a mess. Anybody know if I can revert without losing all my packages and settings?
still zephyr
@idle owl hello could you take a look to this PR https://github.com/adafruit/Adafruit_CircuitPython_DPS310/pull/15. we would like to put some comments in the Readme, and I would like your input ๐
idle owl
<@!330227457296957440> hello could you take a look to this PR https://github.com...
Sure.
@still zephyr In the middle of a thing at the moment, so I'll take a look later today or tomorrow.
still zephyr
<@!607375730610798634> In the middle of a thing at the moment, so I'll take a lo...
Thanks
tough flax
Happy to say that @tulip sleet helped me get the new Dynamic USB stuff working on the XAC! That means we won't need my custom build anymore & the broken USB stack Microsoft shipped has been overcome!
jaunty juniper
so this whole exercise is failing up front. You learned a lot about builds, but ...
what about bitbangio ?
tough flax
Here's the code to make a joystick show up on the XBox Adaptive Controller port (left or right)
ionic elk
@tulip sleet So, for Dormant mode on the RP2040, the RTC actually doesn't work unless you supply it with a fully external clock through a GPIO pin. Do we want to support that somehow or should I just put a NotImplemented in there for RTC deep sleep?
I'd thought I'd gotten around it somehow but it turns out it just wasn't going fully dormant
storm minnow
suddenly danh was incredibly popular
ionic elk
Alternatively, I could put some kinda flag system in there so that when an RTC alarm is included in the alarm list, it just uses regular sleep rather than "dormant" but clocks way further down than we normally do, disabling the USB clocks and stuff.
Dormant is <1mA but regular sleep with almost everything turned off isn't too much higher, like 5mA
Might just be clearer to be upfront that RTC alarms won't get that kind of savings though? I'm not sure what approach adheres the best to the Circuitpython philosophy here. Maybe you have thoughts @slender iron?
tulip sleet
what about bitbangio ?
good idea
tulip sleet
<@!329766224093249548> So, for Dormant mode on the RP2040, the RTC actually does...
do you mean exit_and_deep_sleep() is not implemented? I would say let's go into as deep a sleep as possible
slender iron
@ionic elk my feeling is to cut our losses and not do alarm on rp2040 if we can't run the rtc during dormant
manic glacierBOT
manic glacierBOT
ionic elk
@tulip sleet @slender iron whoa hang on I must have miscommunicated, Alarm on RP2040 is done, exit and deep sleep works just fine with PinAlarm and both alarm types work with light sleep. The only thing that doesn't work is waking from DeepSleep on an RTC alarm using "dormant" mode.
I have examples of RTC waking from a 5mA heavily-gated light sleep mode that we could use for deep sleep, too, I just wasn't sure if it was the right call for the API
tulip sleet
I have examples of RTC waking from a 5mA heavily-gated light sleep mode that we ...
the light-sleep RTC mode to use for deep sleep TimeAlarms sounds OK to me, we just have to document the idiosyncrasies of the the different chips for one or the other: ESP32-S2 is great for deep-sleep TimeAlarm, not so great for PinAlarm. RP2040 is the reverse, etc.
slender iron
@onyx hinge Iโm out until Tuesday but โ๏ธ seems fine to me.
onyx hinge
No hurry, take your time off!
I was tidying up my git and found this feature that I had not PR'd
slender iron
@ionic elk I think the deep sleep call should raise an exception when given a time alarm then. I just donโt want to see any more major rework done on it. Time to PR what youโve got
@onyx hinge it should probably be backwards compatible though.
onyx hinge
I'm not sure how to make it backwards compatible, if you see a way I'm happy to re-work it!
ionic elk
@slender iron PR is up! Got it in yesterday.
slender iron
I'm not sure how to make it backwards compatible, if you see a way I'm happy to ...
Make pixel shader a kwarg? If none then provide a default equivalent to what it does now
ionic elk
Just put together a really quick RTC deep sleep patch too, alongside some CI cleanup
slender iron
<@!252717193496756235> PR is up! Got it in yesterday.
Ok great! I can look Tuesday if no one gets to it before then
ionic elk
Can't use PinAlarms and RTC alarms at the same time for some weird reason but I'm not going to dive into that. Probably best for an actual support email, if RaspberryPi offers such a thing like Espressif does.
literally, ungating the pad clocks turns off rtc interrupts. It's wack.
no wait I just messed up my masks, it all works. bleh
still zephyr
Hello @solar whale by any chance do you have the MAX7219? to test some example code in this PR https://github.com/adafruit/Adafruit_CircuitPython_MAX7219/pull/32
solar whale
@still zephyr Sorry, that must be one of the few Adafruit products I have not purchased ๐
still zephyr
<@!607375730610798634> Sorry, that must be one of the few Adafruit products I ha...
Thanks... ๐
storm minnow
<@!607375730610798634> Sorry, that must be one of the few Adafruit products I ha...
Sounds like you should fix that... :P
gotta buy'em all
solar whale
I've tried...
storm minnow
lol
still zephyr
is just an IC for a led matrix... I do not like leds or displays so who knows...
storm minnow
Don't like LEDs and displays? :O
still zephyr
nope, ๐ I like sensors ๐
storm minnow
But displays tell you what the sensors sense!
storm minnow
lol
still zephyr
๐
still zephyr
But things that light up... they really butter my bread
I do not own a single neopixel
ember iris
Sensors and eink is where it's at ๐
storm minnow
I have an eInk... but I think it may be damaged >~>
ember iris
what's the issue with it?
(I am not an expert in eink at all, I just plan on making an ebook in the next few years)
storm minnow
I've had issues with getting it to update, and it appears to be permanently red around the edges and pinkish all over [it's a tri-color one]
Black on the edges is faded
ember iris
Ohh, red tri-color screens have a permanent pink color to them that, to my memory, becomes increasingly present over their lifetime
I don't know about the edge issue though. I only have black and white screens
storm minnow
Ahh... it was perfectly white when I got it [outside of the Blinka red bits], and I had no idea they changed color over time
thorny jay
If you refresh the screen, the pink should go... no? I mean, it become pink if you keep the same image for ever and powered off.
storm minnow
It doesn't seem to totally go away on the edges when it refreshes, no. I did notice it would start to fade into pink after being left with an image and unpowered, but it seems worse... I need to play with it again in CP, I was messing with Arduino and it was acting weird
thorny jay
There might be a special "refresh" cycle that help bring back balls in the right place with more aggressive power. And maybe that cycle is not present in the library you use. Also could be problematic if you do partial update and never a full refresh.
storm minnow
Ahh, the first one, maybe that's it. It was doing full refreshes -- it would clear and flash white a few times, then black, then white, then red, then white, before updating the image/text, and was still stained
idle owl
Ahh, the first one, maybe that's it. It was doing full refreshes -- it would cle...
From a guide You don't have to refresh often, but with tri-color displays, the larger red ink dots will slowly rise, turning the display pinkish instead of white background. To keep the background color clear and pale, refresh once a day
storm minnow
I think it was sitting for a year unrefreshed, maybe that damaged it >~<
ember iris
Try giving it a refresh once every 10 minutes for a couple of hours, that might help clean it up over multiple cycles
storm minnow
Worth a try
ember iris
Eink screens are basically micro-fluid devices so it's not as simple as setting a 1 or 0 for their 'pixels'. It's a lot more of a model fluid flow under the force of a charge. For an insane deep dive, there's this project: https://dmitry.gr/?r=05.Projects&proj=29. eInk Price Tags
storm minnow
For the refreshes to try to clean it, should I refresh all layers and leave it white? Or have an image?
idle owl
I feel like Collin did a few shorty videos on them too. Might be conflating with another type of display.
storm minnow
Oh, I dunno if it makes any difference at all, but it's this one https://www.adafruit.com/product/4868
ember iris
After the refresh cycles go ahead and just leave the screen blank/white. That way every pixel has to move and the 'stuck' ones get pushed as well over time. (I don't think it should technically matter, but for the sake of anyone in the future with this problem it might help avoid the issue of using various pictures and some of them being more helpful than others)
ember iris
I feel like Collin did a few shorty videos on them too. Might be conflating with...
https://www.youtube.com/watch?v=eipKg2pUQ_8 Would this be the video you're thinking of?
ember iris
It's a good one! I love eink stuff and watched that one a few times
storm minnow
eInk still feels like magic sometimes
ember iris
For sure! I wish it had fewer limitations but it's such a great platform to read on, or display infrequent updates. I feel like it's one of the forms of hardware that software just forgot about, but that happens
storm minnow
Some newer types of eInk can actually refresh a lot faster!
ember iris
Yeah I'm excited to see them come to fruition. How's the screen of yours looking by the way?
storm minnow
Haven't gotten to try to refresh yet, trying to get my FunHouse working
ember iris
Ah the dilemma of having multiple projects! Curses! Haha
storm minnow
yeah XD
thorny jay
I feel like Collin did a few shorty videos on them too. Might be conflating with...
He did.
idle owl
He did.
KeithTheEE found a link to the series concatenated. ๐
In unrelated news, git status is your bestie. (I completely forgot to add 2 out of 6 of the files I need to PR. Noticed there weren't enough files listed in status.)
storm minnow
Is there a good place on how to github?
idle owl
I wrote a guide for it. Covers my CircuitPython workflow. Otherwise, I think they have pretty good docs as well.
storm minnow
Ok, cool. Can I see your guide? I'll poke their docs as well
storm minnow
Danke :D
idle owl
It's a bit out of date, the GitHub interface has changed slightly. But it isn't enough of a change to make it unusable.
And I wanted to add some more Git stuff to it, like how to rebase when your branch is out of date.
But priorities.
ยฏ_(ใ)_/ยฏ
I want to do some shorty posts on my personal site about using Git and GitHub. Basically that guide, but more bite-sized pieces. But there's a lot I want to do for that site. And then I get nothing done. ๐
storm minnow
Too many things to do, not enough time...
idle owl
Pretty much.
storm minnow
C'est la vie
idle owl
waits on CI
storm minnow
CI?
idle owl
CI?
Continuous Integration. It's what runs on every PR you put into CircuitPython or associated libraries. GitHub Action is an example of continuous integration (testing is implied).
storm minnow
Ah... PR is Pull Request?
ember iris
Yup! Once your code is ready to start considering pulling into the main branch, you can run checks over it to test it and make sure you didn't accidentally mess something up
storm minnow
Ahh, so a PR is the opposite of forking?
ember iris
also @idle owl I submitted my first pull request! Thank you for all of your help!
idle owl
Continuous Integration Testing helps ensure things are kept to a standard. You can put whatever tests you want into your CI setup. pylint etc can be included.
idle owl
Ahh, so a PR is the opposite of forking?
That is sort of a way to look at it...
storm minnow
'sort of' sounds like a better understanding than 'no, not at all' XD
idle owl
also <@!330227457296957440> I submitted my first pull request! Thank you for all...
Hurrah! Congrats! And you're entirely welcome!
ember iris
Ahh, so a PR is the opposite of forking?
Yes, more or less. It's kind of like a river, sometimes it'll fork and not meet up, but sometimes it'll merge again and bring added... code. The river metaphor is failing here because silt and code aren't the same thing but oh well. A pull request says, "Hey, I added this and I think it's generally useful so maybe the main branch might want it!"
idle owl
It's a "request" for the original project to "pull" your code into the original project code.
On that note, I need to go get some vehicular maintenance taken care of. I'll be on mobile when I'm not driving. And around again later.
ember iris
Sounds good! Hope it's quick and easy!
storm minnow
Hope all is well with you vehicle!
blissful pollen
Is there a good place on *how to github?*
Dan posted a link to this github book once and I found it very helpful: https://git-scm.com/book/en/v2
turbid radish
?serverinfo
digital shoreBOT
adafruit
storm minnow
Whoa, I didn't realize there were that many channels
ember iris
@honest nova mind if I ask you a couple of questions here if you're free?
still zephyr
yes go ahead
still zephyr
<@!430094695016693780> mind if I ask you a couple of questions here if you're f...
I am avalaible if you are
ember iris
Yup! Looking at the sgp40--some of the print statements were there before I edited the code, should I remove those as well?
still zephyr
yes ๐
thank you
I forgot my debug print sometimes too ๐ (this was not the case anyway)
ember iris
Rodger, and (I'm trying to handle formatting here and PR stuff in the PR thread so bear with me as I'll duplicate some of my comments)
still zephyr
go ahead There is no deep dive today and I am not running so I have a lot of time
ember iris
Should tuples be returned even if it's coming from a hidden function if the tuple adds computation time/steps?
still zephyr
yes, as I said this is an unspoken rule, but normally all the libraries do that
Not sure the reason
ember iris
(where did 98 come from here?)
ember iris
Ahh right. So in this case choose conforming with the api over optimization then?
still zephyr
why do you like the list?
ember iris
In this case it's prepping a byte array and keeps the build of the array clean _cmd = _compensated_read_cmd + humidity_ticks + temp_ticks
ember iris
tuple isn't an issue, I'd just append the index of the element, but that's why I chose a list return. I'm new to editing circuit python so I'm just trying to learn what choices to make
still zephyr
when I say struct I am talking about the API
ember iris
I don't think there's a reason not to use struct, I just don't often use it
still zephyr
yes me neither, depends on the library author I would say.
But I would say, that in this case whatever is easier to understand for the next person that would need to look at your code
that is also a factor
Question from me: is there a SGp30?
ember iris
Yes there is
it has a built in humidity and temperature sensor so this section of code is all handled on the sgp30 itself
still zephyr
and they use lists
still zephyr
me too
ember iris
yeah it looks like in a couple of sections they return lists
ember iris
Ok cool, I'll keep those lists then.
still zephyr
๐
ember iris
As to the measure_raw and raw functions: should I hide the measure_raw or leave it exposed. Because (from my assumption) I don't think a user would ever use raw if voc_algorithm was available. But in it's absence raw was as close to a return value as possible
This strays closer to the conversation for the pr itself though
still zephyr
I agree with you, there is no need if this is done for internal calculation
question here...
this is more Scientific that other...
so, this sensor calculates the VOC?
to do this it will fit temperature vs humidity curves
then it will go to the VOC and get the data
is that correct?
ember iris
Give me one moment to reread the datasheet so I don't misrepresent something--I'm kind of constantly re-discovering it's process that it uses to return a value
still zephyr
no problem I could write why I am asking this, I do not know if you have work with health safety standars, so normally they will list all the compounds... and then they will give the VOC or TVOC, that is the sum of all of that. then they said this is for standard temp and humidity
meaning sometimes these raw measures are important
ember iris
Walking backwards for this device, the actual usable value is an air quality score with a range of 0-500. That is the output of the voc algorithm which takes a raw sensor value (a 'humidity-compensated VOC based indoor air quality signal') and, in conjunction with a rolling model of how long the sensor has been powered up, and how much variance the sensors output has had within some timescale (I don't actually know much about this section, this is what I'm about to tackle) it calculates the 0-500 score.
In the datasheet they refer to the sensor output as SRAW which is in 'ticks' again and 'is proportional to the logarithm of the sensor resistance'
still zephyr
good, I think we have our answer, dont we?
ember iris
Incidentally their datasheet has a suggested design of a voc index driver which outputs the humidity compensated raw signal, the voc index, the relative humidity, and the temperature in it's report
ember iris
good, I think we have our answer, dont we?
What would that answer be then?
ember iris
Ok--so I'll leave the humidity compensated raw measurement exposed in that case?
still zephyr
oh what I mean there is no need to expose it as they are use to compensate
could you change the raw humidity?
ember iris
Oh haha that's the opposite of what I was thinking
ember iris
could you change the raw humidity?
Yes each time you call measure_raw you provide an updated temperature and humidity
still zephyr
so ok now final answer, is yes expose
we had this kind of discussion with the sea_level_pressure for altitude ๐
ember iris
Awesome, I'll add documentation and keep it exposed in that case!
Haha all of these sensors--while delightful to toy with--have so many niche 'consider this' aspects
still zephyr
yes I agree, that why I love them
If you have trouble with the documentation let me know
sometimes Sphinx is very picky
ember iris
haha I'll let you know, and I can almost promise you I will. Sphinx, while powerful, is one of those libraries I loathe the most. Once I'm past my learning curve I'll love it, but for now it gives me headaches
Thank you again for all of your help through this!
still zephyr
No thank you :), But there are some tips in the design guide regarding Sphinx, just in case, this case will be simplier I think!~(Hope?)
๐
ping me if you need help I ll be around
ember iris
I'll be it is, all of the guides I've been using--the contribute to circuit python with git, and the creating and sharing a circuit python library--have been amazing, I'd wager the docs one is just as helpful!
manic glacierBOT
still zephyr
Does this line needs to be changed after bumping up our python compatibility version "Programming Language :: Python :: 3.5", this is in here, https://github.com/adafruit/cookiecutter-adafruit-circuitpython, just copy the library as the link is very long..
proven prairie
hey everyone, I'm going through my first build of CircuitPython on my Mac M1 using the Adafruit guide at: https://learn.adafruit.com/building-circuitpython but I've been running into some issues.
Currently when I run the command: make BOARD=unexpectedmaker_feathers2
I get this:
-- The C compiler identification is unknown
-- The CXX compiler identification is unknown
-- The ASM compiler identification is unknown
-- Found assembler: xtensa-esp32s2-elf-gcc
CMake Error at esp-idf/tools/cmake/project.cmake:307 (__project):
The CMAKE_C_COMPILER:
xtensa-esp32s2-elf-gcc
is not a full path and was not found in the PATH.
Tell CMake where to find the compiler by setting either the environment
variable "CC" or the CMake cache entry CMAKE_C_COMPILER to the full path to
the compiler, or to the compiler name if it is in the PATH.
Call Stack (most recent call first):
CMakeLists.txt:12 (project)
Any help would be greatly appreciated. I'm at a loss for what's happening.
jaunty juniper
hey everyone, I'm going through my first build of CircuitPython on my Mac M1 us...
I believe the ESP IDF version used by Circuitpython does not support the M1
proven prairie
I believe the ESP IDF version used by Circuitpython does not support the M1
Hmmm. I've looked into the idf_tools.py file to check for the M1 fix which apparently the line:
'Darwin-x86_64': PLATFORM_MACOS,
which is suppose to "pretend it is x86_64 until Darwin-arm64 tool builds are available", and it appears this fix is in place. To see if I could get tanother board build and when I tried the circuitplayground express running this:
rune_b@MacBook-Air atmel-samd % BOARD=circuitplayground_express
Use make V=1, make V=2 or set BUILD_VERBOSE similarly in your environment to increase build verbosity.
[Errno 2] No such file or directory: 'arm-none-eabi-gcc'
seems like a similar issue too?
jaunty juniper
ah yeah to build for samd you need the arm compiler
did you do that step ?
I thought building for S2 would not need it
does that even have a m1 version yet ? https://developer.arm.com/tools-and-software/open-source-software/developer-tools/gnu-toolchain/gnu-rm/downloads
proven prairie
ah yeah to build for samd you need the arm compiler
i can install it and I think it did this a few days ago, so that should be in place. but i'm just waiting to download it again and try again.
proven prairie
ah yeah to build for samd you need the arm compiler
well, I mean the good news is i'm getting a different error. I thought i had installed it but all I did was reinstall and restart and now that error is gone. anyway I get this now:
make BOARD=feather_m0_express
Use make V=1, make V=2 or set BUILD_VERBOSE similarly in your environment to increase build verbosity.
ERROR: Empty preprocessor output - check for errors above
make: *** [build-feather_m0_express/genhdr/qstrdefs.enum.h] Error 1
make: *** Deleting file `build-feather_m0_express/genhdr/qstrdefs.enum.h'
which i'm still sorting out
jaunty juniper
did you make BOARD=feather_m0_express clean ?
proven prairie
haha back to square one once I do that:
make BOARD=feather_m0_express
Use make V=1, make V=2 or set BUILD_VERBOSE similarly in your environment to increase build verbosity.
[Errno 2] No such file or directory: 'arm-none-eabi-gcc'
jaunty juniper
might not be in the PATH, did you open a new shell after installing it ? I'm not sure if you need to do something to add it to the path (I do, but I manually put it in a local dir)
proven prairie
might not be in the PATH, did you open a new shell after installing it ? I'm not...
i did open a new shell, but i'm also unsure if i need to add it the PATH. i'm not exactly sure what you mean by putting it in a local directory. Do you think that would work in this case?
jaunty juniper
nothing, I'm just saying I don't know if the installer does it
proven prairie
nothing, I'm just saying I don't know if the installer does it
fair enough. I'm trying to figure that out now
proven prairie
nothing, I'm just saying I don't know if the installer does it
I reached out to the unexpected maker Discord and it turns out there's a way to run legacy command line for the Mac M1. I found this guide: https://medium.com/swlh/run-x86-terminal-apps-like-homebrew-on-your-new-m1-mac-73bdc9b0f343 FYI
manic glacierBOT
solar whale
FYI - I ran into an issue trying to use Blinka under Micropython on a Raspberry Pi Pico. I was trying to see if I could use an "airlift" with the esp32spi library. The esp32spi library uses time.monotonic() but time.monotonic() is not provided by the micropython time module... It looks like a simple fix is to just use time.time() use MP but it does mean a lot of CP code won't work without modification.. I did make that change and ran into some more significant issues that may indicate bigger problems, but I wanted to point out the time.monotonic issue to see if I was missign some thing simple to get around it.
manic glacierBOT
solar whale
The "bigger issue was this - after fixing the timne.monotonic calls ```MPY: soft reboot
MicroPython v1.15 on 2021-04-18; Raspberry Pi Pico with RP2040
Type "help()" for more information.
import esp32spi_simpletest
ESP32 SPI webclient test
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "esp32spi_simpletest.py", line 50, in <module>
File "adafruit_esp32spi/adafruit_esp32spi.py", line 335, in status
File "adafruit_esp32spi/adafruit_esp32spi.py", line 324, in _send_command_get_response
File "adafruit_esp32spi/adafruit_esp32spi.py", line 232, in _send_command
File "adafruit_bus_device/spi_device.py", line 76, in enter
File "/lib/busio.py", line 377, in configure
AttributeError: type object 'SPI' has no attribute 'MSB'
I opened an issue in blinka
solar whale
@still zephyr The main hang-up in testing the BME280 PR is my ability to find my BME280 -- searching....
ember iris
haha back to square one once I do that:
make BOARD=feather_m0_express
Use make ...
Sorry to pop in and run, but this looks similar to an issue I ran into. If you use conda for stuff, conda might be fighting between various forms of gcc in the background due to different installed python libraries needing different c compilers. My solution to this was to create a new environment and then install the arm compiler, and while you're in that environment run the build commands. But I've got to run out for the day. I'll check back in this evening and see if it made progress (or if this doesn't apply at all): conda create --name buildCircuitPython
This only applies if you use conda to manage virtual environments though
solar whale
<@!607375730610798634> The main hang-up in testing the BME280 PR is my ability t...
yay! found it!
onyx hinge
hmmm maybe it was just a dream that I created a wasm version of mpy-cross. I think I'm thinking of the aws lambda mpy-cross together wtih the wasm/emscripten pcf font converter and getting them confused in my head...
I was thinking that if anyone uses the webservice mpy-cross it needs to be updated for v7's mpy format
manic glacierBOT
still zephyr
yay! found it!
Thanks @solar whale ๐
livid horizon
Does anybody have a build of the Circuit Playground Express image that has async/await enabled? Or do I need to build that myself? Does enabling it do something bad?
low sentinel
Does anybody have a build of the Circuit Playground Express image that has async...
First q, I donโt think anyone is maintaining a build like that.
Second q, if itโs not in the official build for your board then youโd need to build it yourself yep.
If you can voice a more specific concern maybe someone can help address your third question!
manic glacierBOT
livid horizon
First q, I donโt think anyone is maintaining a build like that.
Second q, if itโ...
Well, I kind of donโt understand why youโd ever disable the syntax. It doesnโt seem to cost anything, though I understand that most people couldnโt use it without asyncio.
In my case, Iโve got some LoRa modules that speak TTL serial. Iโd like to use a CPX, as they are about the most versatile platform I can find for the price.
Ideally, Iโd build multiple concurrent listeners that could talk to the LoRa unit. However, the LoRa unit does not take kindly to concurrent requests.
With no threading and no async, Iโm left with all of my remaining options being ugly and difficult. Iโd rather roll my own event loop and concentrate the complexity there than build an ad-hoc callback nightmare.
So, yeah, Iโm probably not the target audience for the CPX, but itโs a lovely little unit and I donโt get why youโd ever disable the syntax in the first place.
So Iโm just kind of left wondering if Iโm missing something. Does it cost that many bytes of flash or something? Is it just off because it canโt be used easily? Is it because theyโre trying not to confuse students learning on it? Is it a conservative default?
Iโm trying to figure out how intentional it is, how unreasonable it is to ask for, and if anyone else is likely to want it, too.
stuck elbow
the non-express samd21 boards have very limited amount of flash available, that feature takes additional flash
not sure why it's disabled on express boards as well, but probably the same reason
finally, that is just syntactic sugar, you can do exactly the same things using yield and yield from
livid horizon
finally, that is just syntactic sugar, you can do exactly the same things using ...
I was under the impression that async iterators and async context managers arenโt exactly possible with yield / yield from (I tend to use the former but not the latter).
Iโm actually kind of curious how much more flash it actually takes. Iโd imagine asyncio itself is a lot of code, though it could be packaged separately as an mpy.
I guess Iโll probably end up building it and find out.
But, yeah, this answers most of my questions, I think. Thanks everybody. ๐
stuck elbow
iterators and context managers are syntactic sugar themselves
they don't do anything that you couldn't do with a bit more python code
tulip sleet
Does anybody have a build of the Circuit Playground Express image that has async...
async/await is turned off for space reasons on the SAMD21 builds and a few nRF builds. It is turned on for all other builds. It takes about 1600 bytes.
solar whale
FYI - I opened an issue on the HCSR04 library , but this may be a PulseIn issue on the Rp2040 https://github.com/adafruit/Adafruit_CircuitPython_HCSR04/issues/25 Every other reading is bad when running the hcsr04_simpletest on the rp2040. It runs fine on an M4.
tulip sleet
I was under the impression that async iterators and async context managers arenโ...
It will fit on many builds if you build it yourself. We had issues with the larger translations on some boards in the past
onyx hinge
@livid horizon fwiw adding async to cpx took 1552 bytes of flash, and the de_DE translation still fit according to my quick testing. https://gist.github.com/jepler/fc839d40f69088e25e57860ee30660f1 I didn't actually load the firmware anywhere.
before: 246652 bytes used, 6788 bytes free in flash firmware space out of 253440 bytes (247.5kB).
after: 248204 bytes used, 5236 bytes free
tulip sleet
@onyx hinge it definitely didn't fit at some point, but in my standardization of things I made have made it fit on most SAMD21 boards. It might be an issue on tiny boards with frozen libraries. At one point we were really close to being full on many.
onyx hinge
don't I know it
tulip sleet
@onyx hinge well, doing a test build with a PR not from a fork was a bad idea. It's building everything twice; once on pull-request, and once on push. I see this when we make a new release (it builds everything twice). I wonder if we can turn one of those off.
jaunty juniper
I did look at that briefly when looking at how to disable uploads to S3 of non-main branches
tulip sleet
did you come to any conclusions? Are there use cases for both?
for releases, it hasn't been so bad, because with the unreliability of Actions, I often had a few random failures, and they were usually covered by the re-run. I used to have to do a manual re-run sometimes.
jaunty juniper
it builds on push and on pull requests, it's part of the triggers, I believe we could restrict to pushes to main like that:
on:
push:
# Sequence of patterns matched against refs/heads
branches:
# Push events on main branch
- main
we would also need to include the active non-main branches, like 6.3.x, etc., IK think
tulip sleet
but when I make a release, I need it to rebuild with the new tag
https://github.community/t/duplicate-checks-on-push-and-pull-request-simultaneous-event/18012/2 is what you are suggesting, I believe
jaunty juniper
I ran tests for the uploads rules on a toy repo that logs a "compile" and "upload" tasks and got this, that's when I noticed the duplicate compiles. I didn't test a PR via a fork
https://gist.github.com/Neradoc/fc581d60543db01afed946f8435ea7ac
oh I didn't test a PR to not-main
tulip sleet
the builds are overflowing on boards with lots of pins. It looks like some of those could be trimmed by more IGNORE_PIN_Pnnn decls.
Those macros are hard to get right. Maybe we should reverse the sense and ignore almost all pins up front, and have each board declare the pins it actually uses
e.g. compare trinket_m0 and arduino_zero
low sentinel
iterators and context managers are syntactic sugar themselves
Python itself is syntactic sugar. Nothing you couldnโt do with a bit more C ๐
low sentinel
I was under the impression that async iterators and async context managers arenโ...
Itโs fun code to write and I encourage you to do it! In case youโre unaware, there is a lib on GitHub called tasko that implements a small slice of asyncio (I wrote it when I enabled async/await for circuitpython)
storm minnow
Isn't everything just sugar when you have assembly? :P
still zephyr
please give me a sugar cane. thanks
onyx hinge
<@&356864093652516868> just a reminder -- due to the US holiday today, we'll hold the Discord meeting tomorrow at the same time of day.. hope to see you then!
thorny jay
So let's say in 25H from now... ๐
livid horizon
As a data-point for the interested, the difference in an async-enabled versus async-disabled CircuitPython build for the CircuitPlaygroundExpress is about 5kB (486kB versus 491kB).
hollow token
frozen flume
Quick question: Which version of circuitpython supports audiomp3 ? I'm trying to play mp3 audio on my board but it complains that this module is not found. I'm testing with Circuipython 6.2 and Circuitpython 7.0+.
frozen flume
https://circuitpython-jake.readthedocs.io/en/latest/shared-bindings/audiomp3/MP3...
I tried this example and it won't work because it can't find the audiomp3 library
tulip sleet
I tried this example and it won't work because it can't find the audiomp3 librar...
go to https://circuitpython.readthedocs.io/en/latest/shared-bindings/support_matrix.html and type audiomp3 in the search box in the middle of the page
still zephyr
@idle owl ๐
frozen flume
go to <https://circuitpython.readthedocs.io/en/latest/shared-bindings/support_ma...
Thank you Dan, how can I make my board compatible with this audiomp3 module? I know it can handle the audio as I've tested with wav files using i2s.
still zephyr
To answer your question, yes I have access to the learning guides, I followed the training but could not edit any Learning Guide. @idle owl (y)
tulip sleet
Thank you Dan, how can I make my board compatible with this audiomp3 module? I k...
@onyx hinge is there a reason not to try audiomp3 on ESP32-S2? It looks like the impl is all C now.
idle owl
To answer your question, yes I have access to the learning guides, I followed th...
That's because you need to be added to any guide you're going to edit. You're able to create new guides of your own, but to edit existing guides you must be added as a collaborator.
still zephyr
Ok got you (y)
idle owl
To answer your question, yes I have access to the learning guides, I followed th...
You'll want to email Anne to get added to a guide, I would say. She is sort of the Keeper of Learn. In the case of the DPS310 guide, I can add you as a collaborator. Very little needs to be updated, but it's important.
tulip sleet
@frozen flume you could just set CIRCUITPY_AUDIOMP3 = 0 to 1 in ports/esp32s2/mpconfigport.mk and rebuild and see what happens
still zephyr
You'll want to email Anne to get added to a guide, I would say. She is sort of t...
Yes no problem, will do
Thanks Kattni
frozen flume
<@!648901520703881247> you could just set `CIRCUITPY_AUDIOMP3 = 0` to 1 in `port...
I tried that CIRCUITPY_AUDIOMP3 = 1 but I was still not able to use it.
frozen flume
<@!648901520703881247> you could just set `CIRCUITPY_AUDIOMP3 = 0` to 1 in `port...
Actually I added that to my board's mpconfigport.mk, I will try now adding it at the one you pointed out just to test it out.
frozen flume
<@!648901520703881247> you could just set `CIRCUITPY_AUDIOMP3 = 0` to 1 in `port...
Actually, I got an error when I compile it using that flag.
In file included from ../../lib/mp3/src/mp3common.h:47,
from ../../shared-module/audiomp3/MP3Decoder.c:40:
../../lib/mp3/src/mp3dec.h:80:2: error: #error No platform defined. See valid options in mp3dec.h
#error No platform defined. See valid options in mp3dec.h
^~~~~
make: *** [build-atmegazero_esp32s2/shared-module/audiomp3/MP3Decoder.o] Error 1
tulip sleet
ah, ok, I was wondering where the assembly code was. lib/mp3 is ARM-based, so no go for ESP32-S2
lots of asm() code in there
frozen flume
got it., well, thanks for your help!
idle owl
Yes no problem, will do
I added you to the DPS310 guide.
still zephyr
I added you to the DPS310 guide.
Thanks Kattni
onyx hinge
@tulip sleet I think I tried it but the performance was very unsatisfactory. on the other hand, we did enable it on rp2040 where the perf is low too. you're limited to very low bitrates & maybe even mono-only, or else it skips
and hm I guess there's a build problem too .. oops. I probably had a fix for that locally but it's long lost by now
(there's a "pure C" code path available and it is another cause of low performance)
<@&356864093652516868> the weekly meeting will be TODAY in about 1 hour. since it does not immediately follow an adafruit-internal meeting today we stand a better chance of getting started on time. Please add your notes or placeholders if you have hug reports, status updates, or in the weeds topics. If you just want to listen in, no need to do anything. Notes document: https://docs.google.com/document/d/13v0TziugOYU3FsWjs3_pu1H3IQEKqORbYPE14RIKmb8/edit?usp=sharing
tulip sleet
@onyx hinge thanks; I can't think of another path except to try -O3 on the C code path, which you probably tried already
I don't even know if ESP32-S2 has native floats
idle owl
@onyx hinge Can you not hear me? Hm.
errant grail
modern wing
Good afternoon all you wonderful folks -- happily lurking ๐
low sentinel
Iโll be lurking today as well
blissful pollen
I'll be lurking and may be around text only
cobalt grail
๐
sterile bronze
lurking
idle owl
I can't hear Dan either.
tulip sleet
skip me for now
silver tapir
Happy to help on tls stuff (an anything else)
gilded cradle
@solar whale I made a local change to Blinka, but haven't had a chance to test yet.
solar whale
I thought it was amusing to see the latest commit to the adabot repo was to master about moving to main ๐
slender iron
๐
gilded cradle
I think we need to update some of the other libs like Blinka and PlatformDetect to use Main instead of Master
modern wing
People keep telling me to "take a hike", sounds like it's a fun time ๐
slender iron
I think we can just build and deploy from adafruit/circuitpython
solar whale
All for removing as many of the alternatives as possible
idle owl
Re: settings.py - not documented that I'm aware of.
solar whale
Friends don't let friends use .txt
thorny jay
๐ฏ a good idea. Breaking change for 7.0
crimson ferry
are double extensions a problem for some platform? If not, could leave it... sometimes, periods may arise naturally in a filename
lone axle
I think we can just build and deploy from adafruit/circuitpython
I ended up using data_files inside of setup.py instead of package_data to get the stubs copied into the root of site-packages rather than inside of site-packages/circuitpython-stubs this allowed the imports to work like normal i.e. import digitalio instead of from circuitpython import digitalio. Assuming my tests work out is my next step to PR to adafruit/circuitpython with the changes to setup.py?
gloomy shuttle
Yes, you should be able to multiple by -1
slender iron
@lone axle yup, I think so
tulip sleet
are double extensions a problem for some platform? If not, could leave it... som...
it checks for .py.txt, etc.
thorny jay
There is a market with more than two input to have precision and safe rotary encoder. ๐
modern wing
Thanks y'all!
silver tapir
Thanks all.
still zephyr
Thanks
solar whale
๐
modern wing
We're so close to the finish line....let's push strong through the last mile of this marathon ๐
Cheers
idle owl
The cats were kind enough to do their plaintive meowing when I was muted today.
onyx hinge
awwww
solar whale
Mine can detect an open mic.
modern wing
How considerate! They decided to help mew.
onyx hinge
this week I have been strongly missing having cats in my home. not sure why, usually it isn't something I think about
@lone axle thanks as always for getting the links!
modern wing
indeed, thanks @lone axle for doing the background legwork like that
It makes such a difference when following along
lime trellis
Is there an open issue for deep sleep/low power implementation on other ports? I'm interested in the progress of the samd port and how I can help
tulip sleet
Is there an open issue for deep sleep/low power implementation on other ports? I...
let's ping @ionic elk here
slender iron
let's ping <@323278594258370561> here
Lucian is taking the week off. I don't think we have immediate plans to add SAMD. @lime trellis it'd be great if you added it!
@jaunty juniper @idle wharf @stiff pelican @still zephyr want to move circup to the circuitpython org?
lime trellis
Lucian is taking the week off. I don't think we have immediate plans to add SAMD...
any issue/discussion of where it stands for the SAMD? what the challenge is?
tulip sleet
main takeaways are:
- figure out where to store a small amount of state while deep-sleeping so you can tell why you woke up (reconstruct Alarm object)
- light sleep may do nothing more than the equivalent of
time.sleep(). light sleep needs to leave all the devices on so the program can continue from where it left off - find out which pins you can use for deep-sleep wakeup and how they work (active low, active high, touch, etc.)
for the state storage, there are often some registers in the RTC or similar that can keep state during a deep sleep
slender iron
samd51 has backup ram as well
the API for sleep is all set so that makes it easier
debugging can be trickier though so I don't think adding support is particularly easy
SAMD21 doesn't really have sleep modes so WFI is as good as you'll get IIRC
feel free to make a tracking issue. I didn't find one
lime trellis
thanks for the insight. I'll create a tracking issue and try and scope the work
slender iron
thanks @lime trellis !
tulip sleet
@slender iron I am going to release 6.3.0 final. No negative feedback on 6.3.0-rc.0. I had a few people upgrade to it for various reasons unrelated to 6.2.0 vs 6.3.0.
slender iron
sounds good to me @tulip sleet. thanks!
idle owl
<@252717193496756235> I am going to release 6.3.0 final. No negative feedback on...
On our way to 7!
slender iron
sorry, bot caught you
(for too many mentions)
I'm exempt because I'm an admin
let me grab the text
It sounds like a good candidate to me.
It could be the first repo that has it's ownership transferred from adafruit to circuitpython to preserve history. It would need confirmation that the secrets for PyPI and readthedocs are configured the same.
I assume most people who use it get it from PyPI so it should be a transparent move to most people.
stiff pelican
Thanks
idle owl
I thought we were going with new PyPI account for that org. I already added the new credentials to the CP org.
slender iron
kk, we'll just want to add the new account as a maintainer for the packages on pypi that will have both the old and new accounts
idle owl
Sounds good.
orchid basinBOT
manic glacierBOT
slender iron
Traceback <most recent call last>:
file "code.py", line 5 in <module>
File "adafruit_display_text/label.py", line 30, in <module>
ImportError: no module named 'adafruit_display_text.Label base'
anyone recognize โ๏ธ ? from https://forums.adafruit.com/viewtopic.php?f=60&t=179783
still zephyr
> Traceback <most recent call last>:
> file "code.py", line 5 in <module>
> File...
maybe missing __init__, I try the weather project on the magtag and work for me, do you want me to answer and point to #help-with-circuitpython ?
slender iron
if you could answer that'd be really helpful. thanks!
onyx hinge
Here is the notes document for Mondayโs CircuitPython Weekly meeting. It is at the normal time of 11am Pacific / 2pm Eastern here on Discord. Everyone is encouraged to attend! Please add your hug reports and status updates even if youโll be attending the meeting - itโs super helpful! If you are unable to attend but would still like to include updates, feel free to include them in the notes and weโll read them off during the meeting. Hope to see you there! <@&356864093652516868>
https://docs.google.com/document/d/1jk_X-gKdYgwsbuwJQGqcf9GrFDfPqGBuB_SP1yQW-00/edit?usp=sharing
orchid basinBOT
idle wharf
<@!199811382793863168> <@!624005998696333331> <@!690965887167103008> <@!60737573...
Fine with me
Wait, was that an ask for us to do it or was it an ask for agreement. ๐
onyx hinge
darn, there's no os.urandom on samd21
lone sandalBOT
orchid basinBOT
next lily
hello, i am trying to connect the Jetson Nano with PCA9685 connected to L298N h-bridge to drive dc motor, would anyone help me with the wiring of PCA9685-L298N and python libraries?
stiff pelican
Github seems to having problems:
https://www.githubstatus.com/incidents/76nv9h8pmkv4
still zephyr
@gilded cradle just FYI testing for this PR https://github.com/adafruit/Adafruit_CircuitPython_DPS310/pull/15, enconuter some problems with Micropython and Blinka, I test with the Original library and I got this error:
>>> %Run -c $EDITOR_CONTENT
Traceback (most recent call last):
File "<stdin>", line 10, in <module>
File "/lib/adafruit_dps310.py", line 228, in __init__
RuntimeError: Failed to find DPS310 - check your wiring!
I know wiring is ok, because I was doing the BME280 test with the stemma qt connector
ember iris
Are there any circuit python sensor libraries with two quickstarts? Or a quickstart version a, and version b? Or should I avoid that entirely and just leave it in the examples
idle owl
Are there any circuit python sensor libraries with two quickstarts? Or a quickst...
I don't understand the question.
Do you mean simpletest ?
ember iris
Sorry
https://github.com/adafruit/Adafruit_CircuitPython_SGP40/blob/66ddc29e46ca20c4a2bc7e8c2aab460294da12e5/adafruit_sgp40.py#L50
In the class definition there's a quickstart guide. The one in this works if there isn't temperature or humidity present, but for the case where a user has the current temp and humidity I was trying to figure out if I should add a second quickstart section outlining what to do with the values, or if I should just leave this bit as is
idle owl
Ohhhh.
ember iris
haha yeah I was so deep in the code I forgot the context was needed
idle owl
Nah, do it up! If you want to add more, that's absolutely fine. Keep it "quickstarty", for lack of a better term, but as long as it's still within that concept, go for it!
still zephyr
just a comment normally quickstart = readme.rst = simpletest, so you will need to update those too . sorry for the interruption ๐
idle owl
just a comment normally quickstart = readme.rst = simpletest, so you will need t...
Thanks for the input!
ember iris
Awesome, so in this case, can I leave most of it as is, and once the raw attribute is explained, can I add, "If you have a sensor, such as the BME280, which can read the temperature and humidity, you can ...." and mimic the above format with those values?
ember iris
just a comment normally quickstart = readme.rst = simpletest, so you will need t...
Thank you! I didn't know that and will walk through those as well!
idle owl
Awesome, so in this case, can I leave most of it as is, and once the `raw` attri...
Wait, why are you referring to a different sensor in that section? I am perhaps confused again.
ember iris
Wait, why are you referring to a different sensor in that section? I am perhaps ...
Yes I am
The sgp40 gives you access to a raw attribute related to VOC, but if you have temperature and humidity, the sensor can compensate that raw attribute and adjust for temperature and relative humidity changes. But that sensor isn't built into the board, so those values must be provided from an external sensor
ember iris
Yeah, it's caused me a lot of confusion with respect to, "How best to explain x, or implement y" when trying to think of a user who's not very familiar with the sensors datasheet
ember iris
Yeah, go for it.
Awesome! Will do!
ember iris
just a comment normally quickstart = readme.rst = simpletest, so you will need t...
I'm poking through the files and the simpletest file is a problem. I'm not entirely sure how to write it because measure_raw takes a temperature and humidity value as input, and that comes from a second sensor. So for that should I just 'make up' a temperature and a humidity?
still zephyr
I'm poking through the files and the simpletest file is a problem. I'm not entir...
thinking...
ember iris
https://github.com/adafruit/Adafruit_CircuitPython_BMP280/blob/main/examples/bmp280_simpletest.py
This example provides commented out alternatives which might be a good way to go?
still zephyr
https://github.com/adafruit/Adafruit_CircuitPython_BMP280/blob/main/examples/bmp...
yes ๐ good idea
ember iris
yes ๐ good idea
Should the commented out bits only be in the test, or in the readme.rst and quickstart as well?
still zephyr
Should the commented out bits only be in the test, or in the readme.rst and quic...
yes, good question, normally is like that, but follow the lead as per BMP280
I think we include the comments too, but only 65% sure ๐
ember iris
Awesome, thank you!
gilded cradle
<@!478616919721771019> just FYI testing for this PR https://github.com/adafruit/...
I'm not sure why that would be. Want to submit a issue to Blinka?
idle owl
@lone axle Hello! Are you around?
lone axle
<@!382939733107408897> Hello! Are you around?
yep, whats up?
idle owl
yep, whats up?
I have need of your cookiecutter expertise. Couple of questions. One: is there any reason not to move the repo to main? e.g. would cookiecutter stop working properly if the repo was using a main default branch vs. master. Two: would you be willing to either show me how to update or update a couple of things about this other version of cookiecutter we have (https://github.com/adafruit/cookiecutter-adafruit-pcb) - I need a couple of things changed, and I can't figure out how to do it.
I can't find any references to a specific branch mattering, but I wanted to ask you in case you'd seen something.
Or the fact that you understand it way better than I do.
lone axle
I am not sure about cookie cutter main vs. master I know the command doesn't specify a branch when you run it like cookiecutter gh:adafruit/cookiecutter-adafruit-circuitpython so I assume github gives it whatever branch is marked as default.
idle owl
Ok that's my thought too.
I'll change it on the PCB version first, and run it to be certain.
lone axle
If there is anything that would need to change I would guess it's more likely to be a hardcoded reference to master inside of one of the links somewhere. But I'm not sure if there are actually any or not.
idle owl
It ran. So I assume it's fine.
idle owl
If there is anything that would need to change I would guess it's more likely to...
I was wondering about that as well. I'll have Dylan take a look I suppose.
About the other question with updating the PCB cookiecutter?
lone axle
Yep I'd definitely be willing to help out with that. I've got a decent understanding of it because it happens to use a templating engine I was already familiar with.
idle owl
That's great. Ok. Basically it was written to work with sensors which have one name, and a description. And I use it more for microcontrollers which often have multi-word names, and no description. I tried editing the .json file, but it barfed on my change, so I put it back and left it.
I have to retype out half the prompts that are supposed to be automatic to fix how it generates because of what it was designed for.
Are you noticing your messages take a second to send? Mine are hanging for a moment then sending.
lone axle
I haven't noticed, but i've been switching away after sending for a second.
idle owl
I'm pretty sure what I need is a super simple change. It's that I don't understand it well enough to do it.
lone axle
To make sure I understand, the goal is to get rid of description as a prompt? and allow names to have multiple words like with spaces you mean?
ah I see the way it appends "pcb" in a few places. Trying to get rid of those I assume as well?
idle owl
I wasn't super specific in my explanation. Let me be clearer. sensor_name[] expects a single word, and I give it, let's say, two, such as "Rotary Trinkey". Then when it gets to sensor_description [Sensor], I either hit enter or type a space, or whatever, but it doesn't matter because I don't want it included. Then, when it gets to product_name [<sensor_name> <sensor_description>], it doesn't generate properly because it puts a space in the middle, which mean the GitHub repo it generates has a space instead of a -.
So what I need is for sensor_description to default to nothing, and the product_name to put a - in the sensor_name if there is a space.
So the GitHub repo is adafruit-product-name-PCB or whatever.
Appending PCB is ok in all cases, the local directory name is irrelevant to the final product.
lone axle
Ahh, I see.
idle owl
Yeah because it creates a git folder, which doesn't matter to the final destination, the remote is what matters, and as long as I type it out right, it's always right.
The README is out of date.
The instructions are missing some prompts.
I think what happened when I did what I did was that {{cookiecutter.sensor_description}} can't be empty.
Which I need to be possible.
Or it needs to be done differently so that it's possible to have a description or not.
Because eventually I'll do it on a sensor again. So it should be able to handle both.
lone axle
for changing space you can use the replace thing like:
"repo_name" : "Adafruit-{{cookiecutter.sensor_name | replace(' ', '-')}}-PCB",
in this templating language you can manipulate the variables further by using pipe and then having some way to change it afterward.
idle owl
It's jinja right?
I'm a noob, but I've used it some. The Pelican website generator uses Python and Jinja templates.
lone axle
Yep, I believe so. Django uses something very similar as well
idle owl
Ah ok.
lone axle
I think having the "Sensor" value on the right on this line:
"sensor_description" : "Sensor",
might be what is not allowing it to be blank