I'd prefer latest for now. But I understand if other folks want to make a case for pinning.
#circuitpython-dev
idle owl
lone axle
They have alluded to potentially being able to disable the pre oct2020 behavior on their end. If it's an option maybe it'd be nice to have them switch all of the older projects over to use the newer behavior that way it's level across the board for us.
idle owl
They have alluded to potentially being able to disable the pre oct2020 behavior ...
I can step in tomorrow as soon as I get in.
Because we have a significant number of projects that will need it. 🙂
I'll check Jeff's list to get a number, as well as being able to provide it to them if they are willing.
lone axle
The greater-than does seem successful. I put sphinx-rtd-theme>=1.0.0 and checked the build logs, it did download 1.1.1 which is the newest non-rc build.
idle owl
Good to know. Keep an eye on the issue, and we can decide what to do once that's sorted.
idle owl
The greater-than does seem successful. I put `sphinx-rtd-theme>=1.0.0` and chec...
Ok, I'll ping them about the other projects, but it looks like they can do it on their end. Already done for irremote.
slender iron
@tulip sleet @onyx hinge I'm around a bit while ari naps. Need anything from me?
tulip sleet
<@329766224093249548> <@426200547024961539> I'm around a bit while ari naps. Nee...
I've got several bugs to investigate further. I was thinking of doing a beta.7 this week if we are not on the verge of rc.0. Is there anything specific you would want to go in?
slender iron
not that I know of. I know I have a few still assigned to me that I will hopefully get to later this week
tulip sleet
I may wait until later in the week to get more fixes in. I can reconsider on Wednesday or so.
slender iron
kk. I hope to check in like this when I have a chance
tulip sleet
i will keep any q's in mind
slender iron
you can email or ping me and I'll get back when I can
slender iron
@night tundra this is the best place for help getting a board going
languid whale
Anything new on the esp32s3 ble front?
slender iron
Anything new on the esp32s3 ble front?
work hasn't started yet but I did look a bit more into it
I'll need to circle back on it soon
manic glacierBOT
manic glacierBOT
proven garnet
GitHub now has organization-wide required workflows in public beta! Could be interesting to see how we could use it and if it even makes sense. Could take some of the burden off of some of the cookiecutter and/or composite actions.
I'll look into it and share next week, but saw that on their roadmap a while ago, glad it's in beta.
onyx hinge
hum it looks like pico-w may not be able to connect to an open network
tulip sleet
hum it looks like pico-w may not be able to connect to an open network
Aha! I thought it worked. This may be @broken hedge 's problem: he was not able to connect to a passwordless network that used MAC address authorization
tulip sleet
right
not an issue, but multiple discussions in #help-with-circuitpython
onyx hinge
oh -- there was a forum thread about it
tulip sleet
i didn't get to that email yet, sure, I have met JohnG in person and gone to an event for a class he teaches at Boston College
replied and moved thread to CIrcuitPython subforum
onyx hinge
thank you!
tulip sleet
thank you!
debugging leftover?
mp_printf(&mp_plat_print, "Note: getting password result=%d\n", result);
arg yes thank you
onyx hinge
debugging leftover?
```c
mp_printf(&mp_plat_print, "Note: getting password r...
fixed (force push)
idle owl
I'll look into it and share next week, but saw that on their roadmap a while ago...
Thank you!
tulip sleet
fixed (force push)
ok, we'll ping JohnG about this and get him to test!
manic glacierBOT
onyx hinge
seems like github is having a bad day today
onyx hinge
I seem to get this in the full web ui even if I can write CIRCUITPY from Python: Warning: File System is in read only mode. Disable the USB drive to allow write access.
storage.remount("/", False)
```this is not enough?I thought Melissa was still working on the full web ui, do you mean code.circuitpython.org or whatever?
are you seeing that message in the browser?
onyx hinge
same with the non-full UI, werid
wraith crow
It seems like the web workflow needs the drive disconnected regardless of the remount setting
dismounted....
from the host I mean 🙂
onyx hinge
is that how it should be, or how it is?
jaunty juniper
the web workflow tests for USB mounted, doesn't care about read/write mode
(it's not writing "from python")
tulip sleet
if the web workflow changes files out from under a mounted drive, then the user looking at CIRCUITPY via USB will be seeing stale data. So that's a reason not to do allow that.
onyx hinge
I suppose, CircuitPython has been designed so that it copes OK with the filesystem changing out from under it, but PCs haven't been
idle owl
Spudgers.
jaunty juniper
has anyone made a spudger spudger spudger spudger song ?
slender iron
is that how it should be, or how it is?
I don't think it should be that way. The point of changing that mount status is to change the filesystem regardless of the connected host OS. Its a bug on my part because I tested it by disabling the drive all together.
jaunty juniper
I feel like it would be more confusing, but to be fair the point of making the drive writeable is already to modify it (from python) even if it's mounted, host OS data synced or not
proven garnet
I'm looking into the possibility of using mypy withipre-commit, and I seem to have a memory of there potentially being downsides to using stub files (.pyi) but can't find anything by searching Discord. Is that correct, and if so, what was it? If there aren't any, it may be worth bringing it up in the weeds as typing continues and gets closer to being completed.
manic glacierBOT
manic glacierBOT
orchid basinBOT
slender iron
do we want to do https://github.com/adafruit/circuitpython/issues/4988 for 8.0?
it changes start= and end= behavior to account for the element size of the given buffer
tulip sleet
do we want to do <https://github.com/adafruit/circuitpython/issues/4988> for 8.0...
Yes, this seems like a bug that we should fix
orchid basinBOT
slender iron
@tulip sleet were you going to look at 7431?
tulip sleet
<@329766224093249548> were you going to look at 7431?
I was thinking about it 🙂 but was going to look at what changed (if anything) between 7.3.3 and 8.0.0.
slender iron
@tulip sleet I looked briefly today and got a stack overflow error on the socket select task. Not sure when I added it but it seems like something I did relatively recently
tulip sleet
<@329766224093249548> I looked briefly today and got a stack overflow error on t...
i was thinking of doing a bisect. Do you want to take that or do you have other stuff to work on first.
slender iron
I’ve gotta follow up on the prs I did today. I suspect you’ll end up on one my prs at the end of the bisect
tulip sleet
i'll try to reproduce and then bisect and then ask you if the cause seems mysterious
slender iron
@tulip sleet kk. I did manage to reproduce it today. The stack overflow happened after a keyboard interrupt exception was printed
manic glacierBOT
manic glacierBOT
slender iron
@tulip sleet at my desk now while Ari naps if you need anything
tulip sleet
<@329766224093249548> at my desk now while Ari naps if you need anything
haven't started working on that bug yet; i will now
slender iron
kk, I'm getting through my email and then getting back to the PRs I made yesterday
tulip sleet
could you just upload or paste your reproduction program?
slender iron
tulip sleet
#7431 uses httpserver but want to see if that is necessry
slender iron
@tulip sleet is it reading flash after CP runs? could be flash chip config
tulip sleet
<@329766224093249548> is it reading flash after CP runs? could be flash chip con...
The bug is reproducible from a power-up directly into bootloader mode, so CircuitPython never runs. And it works fine on macOS, but not Windows or Linux. See https://forums.adafruit.com/viewtopic.php?t=197730 also with links to other similar esptool issues in the esptool repo
slender iron
cool cool #notmyproblem
manic glacierBOT
i just bought the same counterfeit boards.
the differences are visible on silkscreen, marking on MCU and no ENIG finish on the PCB
the fake ones lack PSRAM you can see this when you put the board in flash mode (press O and RST, release RST)
than run
esptool.py chip_id
Chip is ESP32-S2FH4 (revision v0.0)
Features: WiFi, Embedded Flash 4MB, **No Embedded ...
manic glacierBOT
orchid basinBOT
tulip sleet
@slender iron I have some socket q's now, if you have any time
slender iron
@tulip sleet yup!
slender iron
I'm in the marie curie channel
orchid basinBOT
spiral elk
@analog bridge Does gpio11 actually function properly on your C3 board in circuitpython?
I noticed you didn't have anything in your board.c to take it out of the default vdd_spi mode and change it to alt gpio mode.
analog bridge
<@747673663419187281> Does gpio11 actually function properly on your C3 board in...
Hi, @spiral elk GPIO11 isn't used and it should be removed from pins.c along with io 12 to 17. Thanks for pointing it out.
spiral elk
I was going through all the c3 boards and looking for ones with gpio11 defined, as its an odd pin, not gpio by default.
I kind of suspect that its not functional on any of the boards that have it broken out either, as none have any mode setting in board.c
manic glacierBOT
candid granite
Hi, I’m not sure if there is something I could do. I had a PR in blinka which was left to die without any comments for more than a year. A month ago, another user needed the same functionality and made an improved version on my PR (https://github.com/adafruit/Adafruit_Blinka/pull/637). However it also didn’t get any feedback. I’m afraid it will also be left to rot and I feel it’s a shame because it’s really useful.
slender iron
@tidal kiln ☝️
@candid granite this is a good place to ask about it. Please don't assume we're mal-intent on our part though. Its tough to keep up with all of the PRs
orchid basinBOT
candid granite
<@337989520962093057> ☝️
Yes for sure, It’s hard to convey state of mind in a small message but I’m not here to complain 😁, I’m good with any outcomes.
orchid basinBOT
slender iron
@lone axle any android code editor suggestions?
lone axle
<@382939733107408897> any android code editor suggestions?
Android Studio is what I use
It's definitely more on the fully feature IDE side of things though. I'm not sure what options are available that are closer to basic text editor.
lone axle
Oh, I see. That Glider app is the only one I'm aware of for BLE editing. Web workflow should theoretically work from a mobile browser on Android, I'm pretty sure I did the most basic test of that from my phone at one point, nothing super fancy though.
slender iron
glider should be making the files available to other apps too
lone axle
glider should be making the files available to other apps too
I don't edit from the phone frequently, the most recent one I used for it was this: https://play.google.com/store/apps/details?id=com.rhmsoft.code&hl=en_US&gl=US I've used a few others, but honestly none of them enough to get a good feel for any of them.
On my current phone I can actually connect via USB with an OTG cable and edit code.py from that app as well. (older Android didn't support the right filesystem type I think)
manic glacierBOT
manic glacierBOT
manic glacierBOT
orchid basinBOT
slender iron
@onyx hinge what was your conclusion about pico + sigrok?
onyx hinge
@slender iron Until there are windows & mac binaries for people to download it's a bear to get going, and i wouldn't try to send anyone down that road. Additionally, there are bugs/limitations. One of the scorpio's pins can't be used without modifying the firmware (I think because it controls the power supply mode on the actual pi pico board). one of the bugs requires you to keep all "lower numbered" pins enabled to capture higher numbered pins (this may also be a protocol limitation of the sigrok<->pico firmware communication, a protocol that the FW author invented) so you are stuck at lower sample rates. Adding proper scorpio pinout support would end up being a separate pull request.
But once you get it going you can decode i2c traces and that's pretty cool!
if I could clone myself and a real webdev I'd say let's make a fresh digital scope app that runs in the browser with webserial, nothing to install.
I also wonder whether there's some other protocol the rp2040 could implement instead, so it works with unmodified sigrok. I didn't go through the list of protocols and find out, though.
slender iron
if I could clone myself and a real webdev I'd say let's make a fresh digital sco...
This was my motivation for this GH issue: https://github.com/wavedrom/vcdrom/issues/31
slender iron
I also wonder whether there's some other protocol the rp2040 could implement ins...
I was hoping the tinylogicfriend protocol could be used for this but its also stalled. I still think the best use of time would be getting a second backend into pulseview that uses python plugins for hardware support
onyx hinge
yeah I haven't looked at it from that angle .. like, can you take out the sigrok from pulseview?
my need for this isn't particularly acute, since I not only have an 8-channel saleae but also a 4 channel + 16 digital channel scope, so I'm not too motivated to throw my own time at it
it sure could help folks though
slender iron
I assume its not too tightly coupled. looks kinda dead though too... https://github.com/sigrokproject/pulseview
onyx hinge
there is a small amount of activity on that PR, but even if it is merged ... the last release of libsigrok was in 2019 so 3+ years ago
slender iron
yeah... pretty dead
https://github.com/glscopeclient/scopehal-apps is a different direction we could go
I understand the appeal of the web but I like python too much
and plugins are a good way to unblock hardware and protocol support
slender iron
its gui is glscopeclient
and there is a quick start link there I think
I gotta run. baby timer is expired 🙂
onyx hinge
see ya
https://hackaday.com/2019/05/30/glscopeclient-a-permissively-licensed-remote-oscilloscope-utility/ a bit of an overview of how the project stood in 2019
lone sandalBOT
analog bridge
In PR CI, will it be viable to only build a single language (e.g. en_US) and build subsequent languages if the free space is less than a certain threshold level (e.g. 10KB)?
tulip sleet
In PR CI, will it be viable to only build a single language (e.g. `en_US`) and b...
I think we don't have a good handle on what that threshold is. We could build the largest known language, but it varies from board to board a bit. Also, if the PR is changing the message or translation code, we want to build them all to test that. Or, the PR submitter or a tester may want a different language available in the artifacts.
I'm understanding you want to speed up the PR development cycle. We have thought a bit about how to make the builds faster: don't rebuild a .o if you don't need to, etc. I have thought a bit about how to make substituting a language be a link-time-only operation, but that is pretty hard. If the compressed-string handles were the same across all the languages, that would help, but I think that requires an extra level of indirection.
mental nexus
I was hoping the tinylogicfriend protocol could be used for this but its also st...
As of a year ago, Libsigrok seems to be relatively inactive. I never got the TinyLogicFriend PR merged, but to be honest my code probably wasn’t the cleanest. My approach was to make a few commands that could be relatively generic. I never found an existing driver that I thought would fit the bill to convert run length encoding (RLE) from the microcontroller board to streaming data on the PC side. Perhaps others can think of another way of using the drivers that already exist on the sigrok side. In fact if you want to do direct streaming it may be easier to do but will have to find a board with enough speed and fast USB to meet those needs.
slender iron
@mental nexus any idea how hard it’d be to write a python driver for it?
lone axle
Are "private" variables intended to behave different in CircuitPython than CPython?
i.e. I have this in a file called scope_learning.py:
from micropython import const
PUBLIC_CONSTANT = const(0x3)
_PRIVATE_CONSTANT = const(0x5)
And this in code.py:
import scope_learning as learning
print(f"public {learning.PUBLIC_CONSTANT}")
print(f"PRIVATE {learning._PRIVATE_CONSTANT}")
On a microcontroller that resultes in an AttributeError, but in CPython it executes successfully.
tulip sleet
Are "private" variables intended to behave different in CircuitPython than CPyth...
global _ const()s are optimized out, let me find you the best reference
lone axle
Thank you!
tulip sleet
@lone axle this saves space and time: space because the dictionary entries are not needed; time because no runtime lookup need be done
lone axle
Nice, makes sense to deviate from "suggestion only" in order to make the most of the resources available on this scale. Appreciate the knowledge and reference
mental nexus
<@366059101203202052> any idea how hard it’d be to write a python driver for it?
Not sure what you mean by Python driver. You mean something in between the microcontroller and the sigrok software? When I looked at sigrok I only looked at the USB connections to get data. Maybe it has other ways of getting in data?
One good thing about pulseview is it has gui buttons to setup and trigger measurements. Perhaps something in the middle could fake it but I don’t know if the existing drivers can accommodate that.
slender iron
Something besides sigrok to get the data
manic glacierBOT
analog valve
Anyone seen any code that allows a Bluetooth arduino to act as a ble hid host. Ie receive ble keypress
blissful pollen
Anyone seen any code that allows a Bluetooth arduino to act as a ble hid host. I...
You will likely find more help in the #help-with-arduino channel.
manic glacierBOT
manic glacierBOT
steady saddle
Hey all, is this the best channel for finding some help for porting/interacting with a new display driver? I've got a different one and I feel like I'm close but its just not quite there. I had to port CircuitPython to a new baord for this too so I'm not even sure where to start to make sure things work
storm minnow
Mhmm, this is the best place for help working with the internals of CP [but I'm not helpful outside of telling you that XD]
tulip sleet
Hey all, is this the best channel for finding some help for porting/interacting ...
are you writing the Python driver, or is this a board with an integral display and you want to set up the display in board.c?
what kind of display and what is the controller chip?
steady saddle
So I've got a WT32-SC01-Plus - which has a ESP32s3. The display on it has a ST7796UI driver (which is apparently similar to a regular ST7796). http://www.wireless-tag.com/wp-content/uploads/2022/11/WT32-SC01-Plus-V1.3-EN.pdf
I got CP ported over to the board, added the stuff neccessary for ParallelDisplay, and it builds and works - https://github.com/adafruit/circuitpython/compare/main...brewneaux:circuitpython:add-wt32-sc01
Projects using LoyvanGFX work just fine with the board (https://github.com/sukesh-ak/ESP32-TUX/blob/master/main/devices/conf_WT32SCO1-Plus.h), using the Panel_ST7796, so I borrowed the init sequence from that (https://github.com/lovyan03/LovyanGFX/blob/master/src/lgfx/v1/panel/Panel_ST7796.hpp)
This is my test script - the backlight turns on, but nothing displays: https://gist.github.com/brewneaux/a42728ceb3b711ba11e385fccb264aaf
tulip sleet
So I've got a WT32-SC01-Plus - which has a ESP32s3. The display on it has a ST...
Sorry, I have to be afk for a while. But this is good information.
steady saddle
If I do a printf as a dirty hack in a src file, I should get that in the serial console, right?
tulip sleet
If I do a `printf` as a dirty hack in a src file, I should get that in the seria...
do mp_printf(&mp_plat_print, "blah %d\n", i);
alternative on Espressif is to use the ESP_LOGx() macros. #include "esp_log.h" and then do ESP_LOGI("tag name", "% str", args...);. You need to be compile with DEBUG=1. Your image may be too big, so turn off ulab temporarily. In mpconfigbard.mk, do CIRCUITPY_ULAB = 0. That will make the image considerably smaller.
That output will go to the standard debugging UART TX line, so you can see what's going on before USB comes up, etc.
steady saddle
ah thanks!
steady saddle
Ive spent most of the day looking at this. It appears the 8080 driver I borrowed from Espressif is getting initialized. I'm starting to think the issue is in the black magic of the DMA that it is doing
steady saddle
I may attempt to build a translation layer to LovyanGFX, which seems to support a bit more, unless someone thinks thats a bad idea.
warm fractal
hey guys - #help-with-circuitpython channel sent me here. I'm trying to get a custom version of CPy to compile, but I'm getting an error on setup, when running "make -C mpy-cross"
`dbj1906@DESKTOP-V5TAK2F:~/circuitpython$ make -C mpy-cross
make: Entering directory '/home/dbj1906/circuitpython/mpy-cross'
Use make V=1, make V=2 or set BUILD_VERBOSE similarly in your environment to increase build verbosity.
File "<fstring>", line 1
(offstart=)
^
SyntaxError: invalid syntax
../py/py.mk:272: recipe for target 'build/genhdr/compression.generated.h' failed
make: *** [build/genhdr/compression.generated.h] Error 1
make: Leaving directory '/home/dbj1906/circuitpython/mpy-cross'
any ideas?
wraith crow
If you type: echo $PATH what is displayed?
warm fractal
a lot... don't worry about it... I just set up a linux machine in the AWS cloud
everything seems to be compiling there
I'm sure there some step I missed in my set-up on my Windows machine
jaunty juniper
manic glacierBOT
I have changed common_hal_mcu_processor_set_frequency to a void return and improved the documentation:
First of all I believe it would be best to 'by-default' prohibit overclocking. ... Ideally an
cpu.safety/cpu.overclock/cpu.cook_mah_pikind of property should be made available.
Any advice on where to put these properties?
In this c...
mystic flume
Hi folks, I've made an asyncio button library: https://github.com/furbrain/CircuitPython_async_button/.
it allows you to do code like: ```
import board
from async_button import SimpleButton, Button
simple = SimpleButton(board.D0, value_when_pressed=True)
await simple.pressed()
button = Button(board.D1, value_when_pressed=False)
clicks = await button.wait([button.DOUBLE, button.LONG])
if button.DOUBLE in clicks:
print("Double click")
elif button.LONG in clicks:
print("Long click")
I'm planning to submit it to the community bundle, but do you think it would be useful enough to go in the adafruit bundle?
proven garnet
I'm planning to submit it to the community bundle, but do you think it would be ...
The Adafruit bundle is for Adafruit supported libraries, so since you'd be overseeing its development, it should be in the Community Bundle. The different bundles aren't a designation of importance or utility, only maintainer 🙂
mystic flume
Ok, no worries, I'll submit a request to the community bundle
candid sun
<@&356864093652516868> We'll have our weekly meeting in about 90 minutes from now in this text channel and in the circuitpython voice channel. Please take the time to add your notes in advance to the document: https://docs.google.com/document/d/1aNx4nRN0TXR7JaEv5dAu139IgZDz8aMCPL-wS6WICZI/edit# -- I look forward to everyone's updates!
midnight ember
Congrats on the working PCB's!
thorny jay
midnight ember
FoamyGuy's TR-Cowbell code https://github.com/FoamyGuy/TR_Cowbell_Sequencer_Software
Today is the last day for the #CircuitPython2023 wishlist. We’ve gotten a number of #CircuitPython2023 posts so far! Thank you all! It’s not too late to post your own. Please post yours by January 18th and email circuitpython2023@adafruit.com to let us know.
manic glacierBOT
errant grail
Thanks Liz!
thorny jay
Thanks.
midnight ember
Thank you for hosting Liz.
slender iron
thanks liz!
candid sun
thanks everyone!
ember iris
Thanks, have a great week!
thorny jay
I found this in French about CP2023 and that is strange, who made that, is it human translation? https://zephyrnet.com/fr/circuitpython-en-2023-circuitpython2023-python-circuitpython/
tulip sleet
I found this in French about CP2023 and that is strange, who made that, is it hu...
no idea; there are a number of other translated posts from https://blog.adafruit.com
candid sun
Here is the notes document for next Monday’s 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 you’ll be attending the meeting - it’s super helpful! If you are unable to attend but would still like to include updates, feel free to include them in the notes and we’ll read them off during the meeting. Hope to see you there! <@&356864093652516868>
https://docs.google.com/document/d/1yqZABCU8BjP_njl0_8fed1P2dkDmNqw5o42qjZzLxqc/edit?usp=sharing
manic glacierBOT
jaunty juniper
shush dyno I typed too fast !
jaunty juniper
I didn't finish my sentence 😛
the one in https://github.com/micropython/micropython might be static ?
in the readme
jaunty juniper
👍
slender iron
Here's my #CircuitPython2023: https://blog.adafruit.com/2023/01/05/scotts-circuitpython2023-tannewt/
slender iron
@crimson ferry I'm happy to do debug builds for you
crimson ferry
ok, I appreciate that. I'm only here through Thursday then traveling for a bit. Priority (for me) is the ESP32-S2 safe mode, it's pervasive across a bunch of projects I've been trying to do.
to capture the trace... that's debug serial (right?), so DevKit is best? N4R2 would match the QT Py best (most of my testing has been on QT Py, but I'd be very surprised if it differed between any -S2)
but I have a variety of -S2 if something else is easier
crimson ferry
what would we do without issues? 😉
crimson ferry
yes, of course
slender iron
one trick with the backtrace is that you'll want the elf file for the version you get the backtrace for
crimson ferry
for that utility, I vaguely remember
slender iron
yup yup
manic glacierBOT
lone axle
Is it possible to 'catch' an exception that is raised from in the core with something like this: mp_raise_ValueError_varg() I'm trying to catch an exception raised by that inside the core, check a few more things and then re-raise the same exception if some conditions are false, but consume the error and return something else if some conditions are met.
blissful pollen
It is, where is it being thrown from? like there are some nlr_push(?) commands to put commands and stuff on the stack.
lone axle
It is, where is it being thrown from? like there are some nlr_push(?) commands t...
in my case it's raised from inside of validate_obj_is_free_pin() which itself calls assert_pin_free(pin); which is the function that has mp_raise_ValueError_varg() inside of it.
I may not need to be able to catch the exceptions though. Upon thinking it over a bit further I realized I could rearrange my logic a bit and not need to catch the exception. Still gotta figure out how to actually make the other conditions I'm interested in though 😅
blissful pollen
in my case it's raised from inside of `validate_obj_is_free_pin()` which itself ...
Oh yeah in that case trying to catch it before is probably the best bet. I'm not quite sure you could otherwise actually. I think busdevice when it calls the I2C/SPI functions does catch/throw exception (or could been too long since I wrote that)
tulip sleet
Is it possible to 'catch' an exception that is raised from in the core with some...
you are just going to do try: ... except ValueError ...
lone axle
I'm trying to catch it from in the C code though, before it makes it into circuitpython code. Can that work from C?
tulip sleet
in my case it's raised from inside of `validate_obj_is_free_pin()` which itself ...
Ah, OK, I would suggest you just call the actual "pin is free" routine and bypass all that
common_hal_mcu_pin_is_free()
will return a boolean
the only purpose of the throwing routines is to propagate the error to Python
lone axle
I see, thank you! I'll look into that one
warm fractal
Hey all. I compiled a custom version of CPy (latest build) for my ESP32S2_feather. I only made one change: adding the ability to use 2 displays - added #define CIRCUITPY_DISPLAY_LIMIT (2) to the mpconfigboard.h file. I loaded the UF2 file to the board, and have tested that it works with two displays. The issue I have now is that I get a core crash every time I make a save to my code.py file. the REPL says
You are in safe mode because:
CircuitPython core code crashed hard. Whoops!
NLR jump failed. Likely memory corruption.
any thoughts on what I'm doing wrong? or if there is another change needed?
note - the code seems to run as expected on it's first run. I only get the error when I make a change to code.py and said. I usually have to manually unlug/replug to board to get the code to rerun (exit safe mode)
blissful pollen
Hey all. I compiled a custom version of CPy (latest build) for my ESP32S2_feathe...
That's all I think there is to change but I never did try two displays on any ESP32* chip. At the start of your code do you have a release_displays call? Re-allocating them or having them free improperly could easily result in a crash like that (saw it when I did some other displayio stuff)
warm fractal
That's all I think there is to change but I never did try two displays on any ES...
yes, I release the displays at the beginning. I am using the same SPI (different DC and CS pins for each board). I was looking at the code for the Monster_M4sk - it looks like that code has a separate SPI for each TFT.... so I guess I can try that....
jaunty juniper
maybe the release is what crashes ?
if you could get a debug build and the debug output
warm fractal
not exactly sure how to get a debug build and output....
code seems to run fine when there is only 1 display... I'll keep playing around with it
jaunty juniper
you make with DEBUG=1 and possibly disable some modules to make it fit
warm fractal
so "make DEBUG=1 BOARD=adafruit_feather_esp32s2"
manic glacierBOT
jaunty juniper
and typically CIRCUITPY_ULAB = 0 in mpconfigboard.mk if it doesn't fit
warm fractal
ok, I'll give that a try
too little flash!!! even with the CIRCUITPY_ULAB = 0 in the mk file
ok.. looks like I need to remove some frozen modules.... tomorrow
calm aspen
is anyone aware of a method to minimize the size of a uf2? I've pulled the full binary off of a custom rp2040 board and I'm left with a working 8 meg uf2. I'd really like to be able to get that down to a smaller file since we have to flash so many boards. my hex editing attempts don't work very well
jaunty juniper
hmmmmm a UF2 is composed of blocks of 512 bytes. You can probably remove most if not all of the blocks that only contain FF in its data payload (especially in the drive part). It should be possible to automate that.
Here is the format specification, find the size in each block, find the actual data, and test if it should be copied, if not, skip that block, if yes, add it to your output file.
https://github.com/microsoft/uf2
orchid basinBOT
tulip sleet
@slender iron when you are available could we talk a bit about the remaining 8.0.0 bugs? Trying to get a sense of what to work on and whether we have any theories.
manic glacierBOT
timid bolt
Hi all! I'm new here. I'm a software engineer by trade and would like to spend some of my free time working on community projects, like CircuitPython. I would love to know if you have any major feature requests that you're looking for someone to take the lead on. Thanks!
ornate breach
Hi all! I'm new here. I'm a software engineer by trade and would like to spend s...
Check out the circuitpython GitHub, pick an issue, fork the repo and start working 🙂
Come here and chat, comment on the issue of you need clarity investigating, etc..
There are issues that are linked to milestones like version 9 release or enhancements, etc..
slender iron
Welcome @timid bolt!
We're just about to get 8.0.0 released so it'll be a great time to add new features to CP.
What hardware do you have?
timid bolt
I have a RP2040 and ESP32-?? (don't remember the model). Happy to get more hardware if needed. (If they are in stock!)
slender iron
I'm also in seattle so I could get you stuff too. Any particular interests?
timid bolt
Happy to work on an issue to unblock release to get my feat wet.
slender iron
The list is here: https://github.com/adafruit/circuitpython/issues?q=is%3Aopen+is%3Aissue+milestone%3A8.0.0 but I think we're pretty much going to punt on them
they aren't simple
looks at rp2040 issues
timid bolt
I have particular interests in async/parallel programming, RP2 PIO, USB PD. But my main goal is to do whatever's most useful for the community.
ornate breach
_looks at rp2040 issues_
You act like the RP2040 has issues or something 🙂
slender iron
ornate breach
Great chip, but lots of weird bugs
I do love it though. It’s been great for learning
slender iron
This might be a good start: https://github.com/adafruit/circuitpython/issues/7452
timid bolt
it has 34: <https://github.com/adafruit/circuitpython/issues?q=is%3Aopen+is%3Ais...
Interesting. I gotta get better at issue filtering. 😀
slender iron
(though I hate to fix a new issue instead of an old one. 🙂 )
we're pretty good about labeling issues with what port and subsystem they are
ornate breach
This might be a good start: <https://github.com/adafruit/circuitpython/issues/74...
If you get this solved I can make a circuitpython oscilloscope 😅
slender iron
the sleep stuff will need a ppk2 or other power monitor (and probably not worth it on rp2040)
ornate breach
10-100kHz range 😬
slender iron
@tulip sleet's been the person thinking about async most
usb host and using the second rp2040 core is probably my next big thing
we also just did CircuitPython2023 which is our longer-term brainstorming we do each year
build instructions are here https://learn.adafruit.com/building-circuitpython
timid bolt
This might be a good start: <https://github.com/adafruit/circuitpython/issues/74...
Cool. I've had this issue myself. TBH, I had a lot of trouble writing real-world PIO programs in CP. I would want to refactor the API, but dunno if that's the sort of thing you'd want newcomers doing.
slender iron
(fyi I worked at google seattle just before you and did road traffic when I started in 2009)
how would you refactor it?
timid bolt
usb host and using the second rp2040 core is probably my next big thing
Love it! There
slender iron
imx rt will likely be the next platform to get polish from us
timid bolt
's lots of interesting stuff here for me! Let me know what's most important and I'll get started.
slender iron
I think #7452 is a good place to start
timid bolt
(fyi I worked at google seattle just before you and did road traffic when I star...
Wow, amazing! Small world.
slender iron
we've been considering bring async apis to more of the core as well but aren't sure how
async requests would be good too (see dan's cp23 post)
stuck elbow
making digitalinout awaitable would be nice
so that all those int pins could be utilized
tulip sleet
(Scott - we are doing CE meeting now)
async requests would be a combo of Python code and also some internal stuff, I think. There is already latent support because Micropython does it
idle owl
Hi all! I'm new here. I'm a software engineer by trade and would like to spend s...
Hello and welcome! Glad to see you joined us. 🙂
ornate breach
Probably would be good to get you on the circuitpythonistas role 🙂
timid bolt
The part I think is missing is async support is the ability to register an interrupt handler and have it run as a task in the async loop. With that building block, things like await on digital IO pin become more straightforward.
idle owl
@timid bolt Are you interested in being added to the circuitpythonistas Discord role? We send a couple of notifications per week to let folks know when our community meeting is happening. It's typically on Mondays at 2pmET/11amPT, here on Discord. It would definitely be great to have you there, but it's not at all required.
blissful pollen
Hi all! I'm new here. I'm a software engineer by trade and would like to spend s...
Welcome!
tulip sleet
The part I think is missing is async support is the ability to register an inter...
we don't have raw interrupt handlers in CircuitPython, unlike MicroPython. But there is a Event thing in MicroPython that you can await on
stuck elbow
The part I think is missing is async support is the ability to register an inter...
I imagine the actual user interface would be similar to alarms
tulip sleet
(it is a deliberate choice not to have those interrupt handlers)
slender iron
safemode.py could be useful too: https://github.com/adafruit/circuitpython/issues/5956
stuck elbow
maybe you would actually await on an alarm...
tulip sleet
you could await on a pin change. The alarm pin stuff is pretty tied to sleep, though maybe we could break it away.
ornate breach
ararmio 👀
tulip sleet
@ jepler has done low-level asyncio stuff recently as well; he is off this week
timid bolt
<@1065334137243836517> Are you interested in being added to the circuitpythonist...
Sure. Why not.
stuck elbow
you could await on a pin change. The alarm pin stuff is pretty tied to sleep, th...
like having a general way to specify a pin change, and then use it both in alarm and async?
idle owl
Sure. Why not.
Done!
tulip sleet
like having a general way to specify a pin change, and then use it both in alarm...
we already have countio, you can specify edge or level
could await on a countio transition
stuck elbow
like on AVR ;-)
timid bolt
we don't have raw interrupt handlers in CircuitPython, unlike MicroPython. But t...
It's not really a "raw" interrupt because it doesn't run in the ISR frame and executes milliseconds after the hardware signal in a proper Python context.
tulip sleet
i thought they have both hard and soft interrupts. The soft interrupt is what you described. We still avoided it for ease-of-use and support reasons. A writeup here: https://learn.adafruit.com/cooperative-multitasking-in-circuitpython-with-asyncio#faq-3106700
slender iron
we've avoided interrupting other python code too
however, my line has always been that I'd merge it if someone wanted to add and maintain it. I just don't think its a priority for adafruit-funded folks
timid bolt
could await on a countio transition
The concern I have with countio is that it is still polling. I believe it keeps rescheduling itself on the looper until its done.
Agreed. Trying to run Python user code in an ISR or between Python opcodes is crazy. What I'm suggested in a disciplined way to bring interrupts into cooperative multitasking.
stuck elbow
The concern I have with countio is that it is still polling. I believe it keeps ...
it could be implemented with an interrupt internally
tulip sleet
The concern I have with countio is that it is still polling. I believe it keeps ...
do you mean in the C code or how we use it in asyncio tasks now?
timid bolt
do you mean in the C code or how we use it in asyncio tasks now?
In the async Python code. I'd seen an example of awaiting on a GPIO pin using countio.
tulip sleet
In the async Python code. I'd seen an example of awaiting on a GPIO pin using co...
yes, that's just because we don't have awaiting on a countio-related Event yet. That is very much a thing to add.
There are a bunch of library issues that may need core support: https://github.com/adafruit/Adafruit_CircuitPython_asyncio/issues. https://github.com/adafruit/circuitpython/issues/6736 is relevant, or more generally https://github.com/adafruit/circuitpython/labels/async
timid bolt
yes, that's just because we don't have awaiting on a countio-related Event yet. ...
I see. MicroPython has this uasyncio module, which I believe is where the Event class you mention comes from. What's the status of that module in CircuitPython?
tulip sleet
our asyncio library was copied from the Python uasyncio model and modified it slightly. But we just made it a library instead of freezing it. And the native _uasyncio MicroPython support is _asyncio in CircuitPython. A noticeable difference is that a contributor added genuine __await__() support to CircuitPythoin, so there is some low-level stuff that is slightly different.
it may be worth reading the closed issues and pull requests related to async/await and asyncio
timid bolt
Cool. Amazing. I didn't know CircuitPython was doing something like MP's uasyncio.
tulip sleet
the language core is pretty much the same; we merge regularly from MicroPython
timid bolt
Yeah, but for some reason I thought you weren't following their asyncio approach because of the differences in your interrupt philosophies. But sounds good. I'll check out the commits on CP asyncio.
tulip sleet
i've got to go to another meeting, sorry. Thanks very much being interested on working on CircuitPython. We really appreciate it, and appreciate your expertise as well.
proven garnet
The PyPortal doesn't have the ssl module, so this issue regarding not being able to use adafruit_azureiot on it is true, right? If so, is this truly an issue, or a case of board/library incompatibility?
https://github.com/adafruit/Adafruit_CircuitPython_AzureIoT/issues/55
timid bolt
Thanks for a great discussion. I've really enjoyed it. I'll follow up in a few days with my thoughts on the PIO background read issue.
slender iron
The PyPortal doesn't have the `ssl` module, so this issue regarding not being ab...
library. The ESP32SPI library does ssl internally without a separate module for it
slender iron
Thanks for a great discussion. I've really enjoyed it. I'll follow up in a few d...
Sounds good. This is the best place to find us. We're here during the work week and others are here weekends too.
tulip sleet
library. The ESP32SPI library does ssl internally without a separate module for ...
ESP_SPIcontrol.TLS_MODE
clear halo
Agreed. Trying to run Python user code in an ISR or between Python opcodes is cr...
Might be worth taking a look at ThreadSafeFlag in MicroPython:
https://docs.micropython.org/en/latest/library/uasyncio.html#class-threadsafeflag
It allows you to await on an event in cooperative context but set the event in ISR context. Example here:
https://github.com/peterhinch/micropython-async/blob/master/v3/docs/TUTORIAL.md#36-threadsafeflag
zealous sierra
Whenever I use mkdir to create a file it says no such file / directory
zealous sierra
Its kind of annoying because I have been trying to solve this issue for 3 hours and the literal error is what I am trying to solve. But its broken because it wont let me create the file or directory
tulip sleet
Whenever I use mkdir to create a file it says no such file / directory
you mean in CircuitPython?
zealous sierra
Yeah
tulip sleet
are you trying to create a single dir or a multi-directory path?
zealous sierra
are you trying to create a single dir or a multi-directory path?
This is the directory I am trying to create "OTATest/Folder"
Should there be a slash at the begging and end also?
tulip sleet
Note that os.mkdir() is going to raise OSError: Read-only filesystem normally. Is that what you are seeing?
zealous sierra
Note that `os.mkdir()` is going to raise `OSError: Read-only filesystem` normall...
No, I solved that by saving the code on my usb port then disconnecting after its done and using a battery to power it
tulip sleet
If you have done .remount to make it read/write, then create the OTATest directory first, then create the subdiretory
you can't create the whole path at once. In regular Python, for instance:
>>> import os
>>> os.mkdir("a/b/c/")
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
FileNotFoundError: [Errno 2] No such file or directory: 'a/b/c/'
>>> os.mkdir("a/b")
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
FileNotFoundError: [Errno 2] No such file or directory: 'a/b'
>>> os.mkdir("a")
>>>
Ok, ill try this. Thank you!
zealous sierra
previouse = ""
for i in dr.split("/"):
os.mkdir(previouse + "/" + i)
previouse = previouse + "/" + i
I got it working with this
thanks for the help
for any future readers searching through discord for answeres. Put this in a try except statement also
The part inside the for loop
specifically the line os.mkdir
lol
crimson ferry
dan, are you still around? pins 19 & 20 aren't exposed on the QT Py that I could see, but I think I need that for debug serial? I have most flavors of DevKit-C-1... N4R2 (adafruit-circuitpython-espressif_esp32s2_devkitc_1_n4r2-...), N8R2, N8R8
crimson ferry
@tulip sleet uart pins needed?
tulip sleet
<@329766224093249548> uart pins needed?
eh, right, hold on
tulip sleet
<@329766224093249548> uart pins needed?
you don't have a metro ESP32-S2, do you?
or something with the default ones broken out already?
crimson ferry
adafruit-circuitpython-espressif_esp32s2_devkitc_1_n4r2-...?
tulip sleet
i can build for the qt py with custom pins, and if that doesn't work right away I'll build for the devkit
crimson ferry
ok, i'll dig out my usb-serial cable
tulip sleet
ok, rebuilding for GPIO5 and GPIO16 (TX and RX on the QT Py). I will test before uploading
tulip sleet
ok, i'll dig out my usb-serial cable
I've got the pins right (because TX is in use when I try to grab it with board.UART(), but seeing no output...
crimson ferry
building for a different board might require other modules disabled? the devkit also has 4MB flash
tulip sleet
building for a different board might require other modules disabled? the devkit ...
can you ID the debug pins on the DevKit? Should be GPIO8 (TX), I think?
crimson ferry
It has 2 USB connectors for 19/20, or I can access every pin
tulip sleet
oh, right, never mind. Let me build that, instead of debugging the QT Py
i have only done custom pins on plain ESP32, so maybe I am missing a setting. Copied from one of those sdkconfig's
crimson ferry
should be a standard devkit build -ulab + debug?
tulip sleet
yes, just starting
crimson ferry
gotcha, thank you, I really appreciate it, hopefully I'll be self-sufficient again soon 🙂
tulip sleet
no problem, these infrequent faults are really bugging me 🙂
i really appreciate it as well
i don't have that board, I think. I have the old saola boards
crimson ferry
if you build it, I will test
I can tell you right away if it's working
otherwise... tomorrow, getting late out east
tulip sleet
i am out east 🙂
crimson ferry
I'm midwest
ok, will load...
did you happen to save the elf file? that's needed for the utility to backtrace... I'll find the instructions for that later
oh, it's there, n/m
tulip sleet
ports/espressif/tools/decode_backtrace.py
crimson ferry
thanks, debug printing is working, I'll post to issue if/when I get something
crimson ferry
Peplink
I need multi-wan router since any one internet option here is so bad
(no broadband)
proven garnet
@slender iron, @tulip sleet - does that mean it should be patched, or is this just the way it is? I don't know too much about SSL or how it would get patched if it even can effectively.
tulip sleet
<@252717193496756235>, <@329766224093249548> - does that mean it should be patch...
the ESP32SPI library API is non-standard. A wrapper library could be written, I guess. I think I made an issue about that, but I didn't know about the TLS/SSL stuff when I wrote that
better to write a wrapper than do something with all the client libraries. Or not a wrapper maybe, but just more API calls (two ways to do the same thing)
proven garnet
I'm not sure I understand - so you would wrap the problematic function calls in the AzureIoT library, or something else?
jaunty juniper
I want something like:
wifi, pool, ssl = adafruit_esp32spi.setup(spi, etc.)
# from then on, use the same code as for builtin wifi
That would be nifty
jaunty juniper
and then make proxy modules named wifi, socketpool and ssl that act like those same objects and auto-setup with board.* objects if they exist, freeze them on the pyportal and boom ! builtin wifi ! (As far as the python modules are concerned)
wraith crow
I believe this would make the Arduino Nano Connect much more compatible as well, I've had some difficulty getting a working socket for NTP purposes with the existing esp32spi
jaunty juniper
yeah the ntp library was changed to native wifi, because ESP32SPI has a builtin thing that works by interrogating the nina module, which retrieves the time on its own
crimson ferry
you can pull an old version from the NTP library repo that was written for esp32spi
or esp.get_time()
jaunty juniper
I need to add that to my samples
def get_ntp_time(esp):
# get_time will raise ValueError if the time isn't available yet so loop until
# it works.
now_utc = None
while now_utc is None:
try:
now_utc = time.localtime(esp.get_time()[0])
except OSError:
pass
rtc.RTC().datetime = now_utc
marsh sierra
Getting this error with XIAO ESP32C3 Circuitpython 8.0.0 beta 6 and KMK firmware
code.py output:
Booting
Traceback (most recent call last):
File "code.py", line 4, in <module>
File "kmk/kmk_keyboard.py", line 12, in <module>
File "kmk/hid.py", line 2, in <module>
ImportError: no module named 'usb_hid'
Code done running.
Press any key to enter the REPL. Use CTRL-D to reload.```
sick hemlock
Hi! I would like to use this https://www.aliexpress.com/item/32868365660.html?spm=a2g0o.order_list.order_list_main.5.1a271802KuMKjz Nordic nRF52840 with KMK for a keyboard... is it possible to flash CircuitPython on it?
stuck elbow
as long as it has nrf52840, yes, but it may be some work to prepare custom firmware, and you will need a programmer to flash it
sick hemlock
as long as it has nrf52840, yes, but it may be some work to prepare custom firmw...
I flash it with a J-Link device... any hint for building a custom fw for CircuitPython?
stuck elbow
I flash it with a J-Link device... any hint for building a custom fw for Circuit...
@sick hemlock note that you will need to add some components to that module to make it work – at least the power circuit
sick hemlock
<@345117054086414337> note that you will need to add some components to that mod...
Yes, I am already using it with daughterboard and rechargeable battery
stuck elbow
excellent
sick hemlock
excellent
I think maybe the pinout mapping is the one to get right... will read the docs and try
manic glacierBOT
tulip sleet
I think maybe the pinout mapping is the one to get right... will read the docs a...
The amount of work to get this going (install bootloader, custom build), etc. is high. You might consider buying an existing nRF52840 board that is already supported, for a few dollars more.
sick hemlock
The amount of work to get this going (install bootloader, custom build), etc. is...
Yeah, I imagined it. I would have liked something without embedded usb-c connector
tulip sleet
Adafruit ItsyBitsy is quite small but plenty of pins. You could use a Nordic PCA10059, which is tiny (maybe not enough pins for you), and already supported. But it still requires installing the bootloader. The USB jack is very convenient to get code onto the board, since it will present CIRCUITPY
sick hemlock
The amount of work to get this going (install bootloader, custom build), etc. is...
Is there any kinda of reference guide for tinkering with bootloader/custom fw?
tulip sleet
Is there any kinda of reference guide for tinkering with bootloader/custom fw?
https://github.com/adafruit/Adafruit_nRF52_Bootloader/blob/master/README.md for flashing a new bootloader. https://learn.adafruit.com/how-to-add-a-new-board-to-circuitpython as mentioned above for adding boards.
I would still suggest considering the Itsy nRF52840. You can mount it at the edge of your keyboard board so the USB connector is flush with a small case opening, and then make a small hole in the case to hit the reset button. So the keyboard software will be easily maintainable without disassembly. It also has a level-shifted 5V output pin ideal for driving addressable RGB LEDs.
proven garnet
I want something like:
```py
wifi, pool, ssl = adafruit_esp32spi.setup(spi, etc....
Is this feasible to do? I don't know what the space constraints are on all the boards, but I'd be willing to try and tackle this at some point in the future if it's useful. I know a lot is changing with connectivity these days.
manic glacierBOT
Hi @dhalbert. I just wanted to let you know that I had some merge conflicts come up due to your recent arg_name changes in: https://github.com/adafruit/circuitpython/commit/a974402542f71fde63043da5725e3777c95b6861
Specifically, I needed to change these 3 functions from the validate versions to common_hal ones, which in your commit did not have arg_name as a parameter.
...
tulip sleet
Is this feasible to do? I don't know what the space constraints are on all the ...
Ultimately I think making the AirLift API match the native WiFi API is what we want. That could be a wrapper for now, or a rewrite of that library. It could even be a new library, so that old code still works, and we could deprecate the old library but leave it around.
proven garnet
Ultimately I think making the AirLift API match the native WiFi API is what we w...
Sounds good, I'll keep that in mind!
timid bolt
Coming back to the idea of using uasyncio Event or ThreadSafeFlag to implement non-polling awaitable pins: how would you call the "set" methods on these classes from CP. The ThreadSafeFlag example uses MP's interrupt API, which I'm guessing should not be available in CP.
stuck elbow
Coming back to the idea of using uasyncio Event or ThreadSafeFlag to implement n...
how about just having an __await__ method on the countio.Counter object, so you can just do await countio.Counter(pin)?
timid bolt
Right, something like that is the goal. My question is more about how to get there.
stuck elbow
I imagine the __await__ method would install an interrupt corresponding to the Counter object settings, that would flip the Event object that it returned, or something like that, but I have to admit I'm not greatly familiarized with the _asyncio module yet
tulip sleet
I had thought about passing an Event into things like countio.
stuck elbow
I had thought about passing an Event into things like `countio`.
you mean explicitly, on the Python side?
tulip sleet
yes, but maybe I am missing something
so a single event might be triggered by multiple things, might be convenient
stuck elbow
so it would be something like:
event = asyncio.Event()
counter = countio.Counter(pin)
counter.register(event)
await event.wait()
to await on a pin?
timid bolt
I had thought about passing an Event into things like `countio`.
I think that could work. My thought is that just waiting on a state change of a pin in a much simpler operation than what countio provides, and perhaps this functionality should belong in digitalio instead.
For example, looking at the rp2 impl of countio: it needs a pwm pin. But maybe on other hardware not every pin supports pwm, but likely more pins would support simple gpio interrupts.
tulip sleet
I think that could work. My thought is that just waiting on a state change of a ...
The GPIO interrupt mechanism on RP2040 is all pins, all changes, without breaking it out per pin or kind of change: https://forums.raspberrypi.com/viewtopic.php?t=343058#p2055582. I'm not sure if this is why the implementer of countio (https://github.com/adafruit/circuitpython/pull/4254) chose PWM. We could ask him.
Also, it looks to me like RP2040 countio only uses the interrupt routine when the count register wraps around. Otherwise I think it is completely in hardware.
stuck elbow
to do it with DigitalInOut you would still need to select on which kind of change it's triggered
tulip sleet
I think a per-transition interrupt for countio could be added, so we could trigger, but a DigitalInOut interrupt might be better. It just has to demultiplex the interrupt.
stuck elbow
that would be a third place where an Edge object is defined
tulip sleet
maybe should move edge to DigitalInOut
stuck elbow
maybe should move edge to DigitalInOut
or share it, after it's all compiled in
tulip sleet
yes
stuck elbow
so something like
event = asyncio.Event()
dio = digitalio.DigitalInOut(board.pin)
dio.register(event, edge=digitalio.Edge.RISING)
await event.wait()
possibly also with some way of deregistering
tulip sleet
yeah, something like that
timid bolt
For digitalio, I think you'd want to await on high/low, more so than rise/fall. Rise/fall can be prone to "time of check, time of use" bugs.
stuck elbow
For digitalio, I think you'd want to await on high/low, more so than rise/fall. ...
you really want a choice of all four
depending on the use case
timid bolt
Possibly. I'd like to see an example of the rise/fall use case.
stuck elbow
int pins on sensors
generally anything where you care that the pin changed at any moment, not what its state is right now
timid bolt
generally anything where you care that the pin changed at any moment, not what i...
Right, but specifically where you'd need rise/fall instead of high/low.
stuck elbow
anywhere where it can change by the time you get to it
timid bolt
if sensor.idle():
await sensor.wait_for_rise()
Is actually wrong because the pin can rise between lines of python code. Whereas:
if sensor.idel():
await sensor.wait_for_high()
Is robust against that race.
stuck elbow
the event would register the change anywhere from the moment it was registered on the pin
tulip sleet
e.g. want to trigger low to high. External peripheral has a line that is already high, you want to reset the peripheral, which will make it low, but not before you set things up. When some event happens the peripheral makes the line high
e.g. some kind of latching peripheral interrupt line
another example, want to respond to a key being released. Undepressed key is low, depressed is high. Set up interrupt to trigger on high-to-low transition
can't set up on just low, becuase it's already low
slender iron
Sounds good, I'll keep that in mind!
I think you do want to revamp it as a separate library. That way you don't need to fix all of the reverse dependencies. We also discussed doing the new library natively so that web workflow could work with esp32spi.
timid bolt
the event would register the change anywhere from the moment it was registered o...
I see what you mean. That would work.
Thinking about the end-user programming experience though, I would strive to get rid of the Event object. I think Events are not on the entry level of the Python async model, and many python async APIs work without them.
sensor_ready = DigitalInOut(...)
while await sensor_ready.wait_for_high():
talk to sensor
I think this code looks a lot cleaner without events. But it would be good to examine some real world code to understand the use case better.
slender iron
I don't think we'll want to add to the existing apis. I think we'll want async versions of them
tulip sleet
There is no way to implement a C native async def method
slender iron
There is no way to implement a C native async def method
- currently
stuck elbow
so it would be asyncio.PinChange(pin, edge) or something like that?
tulip sleet
I don't think I would do that, but my point was I was trying to figure out how to do something async with native modules. I can't await on a native method, but I can pass it something that is awaitable.
slender iron
native async methods it probably the right way to go about it
tulip sleet
i don't think MicroPyton has this now, though
stuck elbow
tbh the whole async stuff both in the big python and especially micropython really feels bolted on
tulip sleet
found an MPy discussion issue on this: https://github.com/micropython/micropython/discussions/9131
timid bolt
I don't think you need native async methods. The native method gets passed down some sort of object (like Event) that it can use to schedule callbacks in the Python event loop. That's how it is implemented. My suggestion is more about how it is exposed to users. I think there are usability advantages to not exposing the Event object to users, even if an implementation uses one.
tulip sleet
tbh the whole async stuff both in the big python and especially micropython real...
it has evolved in a very tortuous path over years and years, yes.
stuck elbow
they could have just said that Twisted is the standard
tulip sleet
I don't think you need native async methods. The native method gets passed down ...
that was what I was thinking of, but how do we await dio_object.something(). Instead we need a wrapper library
tulip sleet
they could have just said that Twisted is the standard
I think that had its own issues. trio semantics are very good, they have really moved in that direction with TaskGroups and the cleanup of the asyncio API (don't mess with event loops yourself, etc.)
that Mpy discussion seems very relevant...
slender iron
have you found anything that explains why we can't do native async?
tulip sleet
the scheduling is all in the asyncio python code, except for some optional helpers in _[u]asyncio
slender iron
ah
timid bolt
that was what I was thinking of, but how do we `await dio_object.something()`. I...
I'm having trouble parsing your message. Do you mean how do we implement await dio_object.something()?
slender iron
is there a reason to keep the scheduling in the python vm?
tulip sleet
right. a_digitalinout.some_method() is a native Cmethod, but there is no way to await a native method
tulip sleet
dio.awaitable_event(LOW_TO_HIGH), for example, could return a native Event-ish object, maybe. But right now Event is a class defined in Python in the asyncio library
no, we wouldn't want to move the scheduling. it's much easier done in Python, and it depends on a lot of Python code
timid bolt
Right, in big Python this method would return an asyncio.Future object.
tulip sleet
and you want to maybe subclass, etc. I wouldn't want to maintain that.
slender iron
maintain the scheduling?
tulip sleet
maintain the whole library. Huge parts would need to be native
timid bolt
I think there would be a HAL functions that associates an irq with an asyncio Future. The native code is huge because it is different for every port. But the native code for any one port should not be huge.
slender iron
right so the native api returns a future that we modify in a native interrupt
we use native interrupts in many places
tulip sleet
right now there is no Future class in MPy/CPy asyncio
i think we can get away without having it.
or maybe not, but I feel in the CPython world it is partly a legacy.
there are too many kinds of things you can await on
timid bolt
right now there is no Future class in MPy/CPy asyncio
Right, this leads me to a higher point about micro asyncios in general. I agree big Python's asyncio history is tortured but it looks to me like it has reach a sort of plateau of stability now. So what is the thinking about embracing big Python's asyncio interface warts and all. On one hand it gives a lot of cross-compatibility with big Python (which I believe is a goal of CircuitPython), and re-use of code maintained by CPython. Are there concerns that the code size is too big for micro?
tulip sleet
but my point about passing Events in rather than constructing them is that it could be more flexible. It allows the same Event to be shared and triggered by multiple things. (Maybe that is bad, not sure). I think it is very much worth studying what the MPy folks are doing here; they have much more experience writing async code, particularly peterhinch and jimmo
timid bolt
hmm, if CPython world it is partly a legacy, what is the future? (no pun intended)
slender iron
we should definitely talk to damien and jim before doing anything major
they've been thinking about this too
tulip sleet
i spent a long time trying to understand asyncio thoroughly. The early styles of use were kind of terrible. I really agree with Nathaniel Smith about the problems with the unconstrained early stuff
slender iron
I'd be happy if we didn't need an .irq() equivalent
tulip sleet
I would have preferred having a trio library instead of an asyncio library, but it seemed difficult to port. And the MPy library was available for our use with minimal changes needed. In the CPy asyncio guide, I really tried to show a style of use that was clean and simple.
now TaskGroup is in CPython 3.11, but it had exception groups as a prerequisite. It took 3 years for that to get implemented. exception groups are a lot of new implementation for MPY and I don't think MIcroPython is going there any time soon. But we can still have TaskGroups with some (slightly incompatible) wrapper for exceptions in tasks or something like that.
timid bolt
Does CPython mean CircuitPython or the standard implementation of "big" Python?
tulip sleet
CPYthon means "regular Python". We try to say "CPy" for CIrcuitPython. The ambiguity is a problem. So I usually say "regular Python" to the average user
"CircPy" never caught on
i have to go grocery shopping before it starts to rain, but this has been really interesting, and I need to read that MPy discussion (and it points to further documents).
tulip sleet
they've been thinking about this too
Yes, we had some brief discussions before with Damien and Jim, but more would be better, since they have progressed in their planning now (and maybe have even done some impl we haven't seen yet).
timid bolt
but my point about passing Events in rather than constructing them is that it co...
I think the use of Event like that is valid. But I also think using asyncio.wait (https://docs.python.org/3/library/asyncio-task.html#asyncio.wait) is a more typical way to wait for multiple signals.
tulip sleet
The CPython folks are much more concerned with async for network stuff. Arbitrary external events are not something they have thought about a lot. We talked with Nate Smith about this in hallway at PyCon several years ago, but reached no conclusions.
I don't know if you have seen the trio "manifesto"-like doc, but I liked it: https://vorpus.org/blog/notes-on-structured-concurrency-or-go-statement-considered-harmful/
solid juniper
I want to add CircuitPython to the Flipper Zero WiFi Dev Board. What existing board should I copy?
slender iron
@solid juniper start with one that has the same esp module
that way the flash and ram should be setup correctly
solid juniper
<@831306778896498688> start with one that has the same esp module
So something like the Saola 1 w/WROVER?
slender iron
I don't know what the flipper has so I don't know
solid juniper
It has an ESP32-S2-WROVER-N4R2. The Saola should have the same, since it has an ESP32-S2-WROVER that has 2MB PSRAM and 4MB Flash.
crimson ferry
@solid juniper https://www.espressif.com/en/products/modules
solid juniper
<@831306778896498688> <https://www.espressif.com/en/products/modules>
Do I need to change any of the pins before I build the first build?
crimson ferry
generally, microcontroller will have all of the pins, and board will have the pins exposed on your board (with name aliases matching silk, where applicable)
VID / PID need to be worked out
eaxample of a Saola-like board: https://github.com/adafruit/circuitpython/pull/5998 but it duplicates the Saola VID / PID since it's also made by Espressif and has the same functionality
I'm sure there are other examples... maybe there's a closer board to yours in https://circuitpython.org/downloads?mcufamilies=esp32s2
Q: When the filesystem has been extended, is the former dualbank parition preserved as part of the filesystem simply by virtue of nothing being written to it, or is there some more active protection for when a new UF2 is installed?
solid juniper
I'm sure there are other examples... maybe there's a closer board to yours in <h...
I'm not sure if there is a better board.
solid juniper
The first build of CircuitPython for the WiFi Dev Board built successfully. But how do I get it onto the board?
jaunty juniper
adding a boad to tinyuf2 is usually easier, it doesn't care about the PSRAM configuration or the pins setup
solid juniper
using `esptool`
I used the online Adafruit ESPTool. The build worked. I can access the REPL and the CIRCUITPY drive appeared.
solid juniper
adding a boad to tinyuf2 is usually easier, it doesn't care about the PSRAM conf...
I'll do that as well.
solid juniper
How can I get the LED(s) working?
tulip sleet
@crimson ferry I am getting fried. WIll give up for the evening. I was just trying to see if the cert errors disappeared completely in the regular build (but still turning on the logging). This is mysterious but not necessarily related. On the other hand, if it's a use-after-free or something, it could manifest in different ways.
crimson ferry
OK, I'm out until late next week, I can pick it up again then.
tulip sleet
i am not sure we will hold up 8.0.0-rc.x and the final release to get this fixed, since it's hard to reproduce
though you seem to hit it reliably
crimson ferry
right, and doesn't seem to be affecting anyone else
tulip sleet
THere is also https://github.com/adafruit/circuitpython/issues/7405 and https://github.com/adafruit/circuitpython/issues/6791 but they are S3
thanks for all the testing, and hope your trip goes well
crimson ferry
thank you too, I hope to be underwater for much of the trip 🙂
crimson ferry
🤿 (rec)
tulip sleet
that's what I hoped, as opposed to submarine duty
proven garnet
Please note library releases shouldn't be made until https://github.com/adafruit/adabot/pull/332 is merged and the patch run, as all releases done beforehand won't attach release assets to the GitHub Release.
proven garnet
Libraries are all patched, with a few ones that need work tagged with issues I'll get to this weekend (I don't believe any are in the Bundle). It's late so I won't tag, but if you see this thanks Dan for the quick review so I could do this on the off hours!
And with that, the libraries have moved away from deprecating actions!
spiral elk
Could someone check if espefuse.py in the esptool idf component is working for them?
slender iron
@tulip sleet is https://github.com/adafruit/nvm.toml/pull/12 on your radar?
tulip sleet
<@329766224093249548> is <https://github.com/adafruit/nvm.toml/pull/12> on your ...
I missed that! will review
slender iron
np
I thought I'd do usb host next but then immediately started updating my watch code
and then thought about the 7 color epaper because its also 4 bits a pixel
plus I've been thinking about eink a bunch
idle owl
and then thought about the 7 color epaper because its also 4 bits a pixel
The Pimoroni Inky Impression? Or is there something else?
slender iron
@idle owl yeah, it’d include that. I have a version from waveshare with the same display
idle owl
<@330227457296957440> yeah, it’d include that. I have a version from waveshare w...
It looks amazing. Would the intention be to hack it onto a micro? Or use it with Blinka on a Pi?
What does the WaveShare one look like?
slender iron
I’m making the circuitpython low level stuff work for it like our other eink
idle owl
Ah
slender iron
Ya, I’m excited to see how it looks. They are slow to update though (like 30 seconds a refresh)
ember iris
Oh that's exciting! I (had as of this weekend thanks to a couple of dumb accidents) a couple of waveshare 7.5in screens in just the standard single black color, but I've been eyeing the color ones for a few years
slender iron
Oh that's exciting! I (had as of this weekend thanks to a couple of dumb acciden...
I think I have an old 7 inch one too but the protocol is weird. I’m curious about the 7 inch color one too
ember iris
They apparently have a v3 of that screen now so it's possible the protocol has improved.
idle owl
@rugged spindle Hello! Do you go by the same user ID on GitHub as you do here?
rugged spindle
I do indeed
idle owl
I do indeed
Congratulations! 🎉 You submitted the 400th CircuitPython library to our collection!
That's a combination of Adafruit libs and Community libs. 🙂
rugged spindle
Whoa. And by accident!
idle owl
Couldn't have planned it, for sure. 😄
idle owl
Whoa. And by accident!
Do you mind if I DM you?
rugged spindle
That's fine
abstract stump
I was playing with HID lately and I was wondering if it is possible to create a dual gamepad or dual mouse-like device. Looking at some custom HID devices it seems that it is possible to create a HID device with 2 gamepads every one with and report and report_id, but I was not able to use the .send_report() including an report_id
tulip sleet
I was playing with HID lately and I was wondering if it is possible to create a ...
short answer, yes, the Mouse or ``Keyboard` device knows its report id. I am about to eat but will get back to you
tulip sleet
I was playing with HID lately and I was wondering if it is possible to create a ...
did you have trouble specifying a report id here? https://docs.circuitpython.org/en/latest/shared-bindings/usb_hid/index.html#usb_hid.Device.send_report
blissful pollen
Quick question. I noticed one thing Anne mentioned for CP2023 was a native gif player. Would this be useful? May be something I could work on. Though putting some thought into how this fits into displayio
manic glacierBOT
jaunty juniper
man I'm looking into a question on watchdog in the support forum, and it's... tricky to use on RP2040: because it's a hardware watchdog that can't be deinited limited to 8s max, a simple autoreload might trigger it if the timer is shorter than the time it takes your code to reload and get back to the point where it sets and feeds the watchdog...
That actually led to drive corruption within minutes of me testing different parameters to it.
onyx hinge
man I'm looking into a question on watchdog in the support forum, and it's... tr...
Maybe the flash write routines in particular should pet the watchdog, or maybe the watchdog should (on any system) only be available unless CIRCUITPY is fully read only
solid juniper
I can't get any of the 3 LEDs on the Flipper Zero WiFi Dev Board to turn on as the status LED.
jaunty juniper
Maybe the flash write routines in particular should pet the watchdog, or maybe t...
that's an interesting idea, the core could feed the microcontroller.watchdog during USB writes, and maybe during autoreload or something ? The watchdog must not be fed if the code exits, because that's kind of the point (one use is to reset the board when it dies of an unexpected exception) but auto-reload could be a special case (with maybe an option to opt out)
and watchdog.deinit() could be simulated on RP2040 by increasing the delay to the max and feeding it in the core ticks ?
solid juniper
I can't get any of the 3 LEDs on the Flipper Zero WiFi Dev Board to turn on as t...
What do I need to do in order to add them?
jaunty juniper
What do I need to do in order to add them?
what kind of leds are they ?
if it's a single monochrome LED, you do it like that:
https://github.com/adafruit/circuitpython/blob/9c066825a7c2127f69da90e80de5eb94e2d92a00/ports/raspberrypi/boards/raspberry_pi_pico/mpconfigboard.h#L4
if you have a neopixel you do it like that:
https://github.com/adafruit/circuitpython/blob/9c066825a7c2127f69da90e80de5eb94e2d92a00/ports/raspberrypi/boards/adafruit_feather_rp2040/mpconfigboard.h#L4
if you have a PWM RGB LED, you do it like that (this one is inverted)
https://github.com/adafruit/circuitpython/blob/9c066825a7c2127f69da90e80de5eb94e2d92a00/ports/raspberrypi/boards/pimoroni_tiny2040/mpconfigboard.h#L4-L7
manic glacierBOT
solid juniper
if it's a single monochrome LED, you do it like that:
<https://github.com/adafru...
I think it is PWM because there are three LEDs. Blue is connected to GPIO4, Green is connected to GPIO5, and Red is connected to GPIO6.
jaunty juniper
well you would follow that example then
solid juniper
well you would follow that example then
Thanks, it worked. Now I need to add the rest of the pins, USB VID and PID, and make a pull request.
abstract stump
Yes, I had a problem there but I think it was my bad, I was trying to use two gamepads on one rp2040 but it seems that using a custom HID descriptor is not recognized easy on win10, in the device manager it appears as a unrecognized hid device. And trying to use different report_id on the send_report function showed me an error of invalid report_id. The problem was the find_device function that looks for a device with a usage_page and usage, but in this case my gamepad1 and gamepad2 was getting the same device, so I bypassed the find_device and assigned the usb_hid. devices directly, like usb_hid. devices[0] and [1] that way it works nice now. The find_device function should have and optional variable of report_ids if someone wants a duplicate HID device on a single micro
abstract stump
did you have trouble specifying a report id here? <https://docs.circuitpython.or...
sorry forgot to add the reply, the answer is up. thanks for your help 😀
wanton topaz
Folks, apologies if this isn't the right place to ask, I hope it is.
stuck elbow
what could possibly happen if it isn't
wanton topaz
Ah dang, too much slack, keep hitting wrong keys...
I have a Seeed XIAO nRF52840 Sense, but I can't seem to figure out how to use the on board IMU, their examples are arduino ones, and looking into them I can see that they are using I2C to talk to the IMU, and if I look here in the CP source https://github.com/adafruit/circuitpython/tree/main/ports/nrf/boards/Seeed_XIAO_nRF52840_Sensee
I can see this in pins.c
`{ MP_ROM_QSTR(MP_QSTR_IMU_PWR), MP_ROM_PTR(&pin_P1_08) },
{ MP_ROM_QSTR(MP_QSTR_IMU_SCL), MP_ROM_PTR(&pin_P0_27) },
{ MP_ROM_QSTR(MP_QSTR_IMU_SDA), MP_ROM_PTR(&pin_P0_07) },
{ MP_ROM_QSTR(MP_QSTR_IMU_INT1), MP_ROM_PTR(&pin_P0_11) },`
But there's nothing in the mpconfigboard files... looking at their datasheet here:
https://view.officeapps.live.com/op/view.aspx?src=https%3A%2F%2Ffiles.seeedstudio.com%2Fwiki%2FXIAO-BLE%2FXIAO-nRF52840-Senese-pinout_sheet.xlsx&wdOrigin=BROWSELINK
They seem to say that the SDA/SCL are on D16/17 (but also reference pin07/27) but if I try to use board.D16/D17 in CP it just says they aren't defined.... Am I missing something? Is there any way to reference these pins without recompiling the source and editing?? (Alas I'm not super familiar with C++ or compiling CP myself)
Is it just a matter of copying the other files pointers to the pins into the mpconfigboard file like the others?
wanton topaz
what could possibly happen if it isn't
Ah some folks kick off if you post in the wrong place, but sometimes it's not super clear, so I like to ask.
stuck elbow
Ah dang, too much slack, keep hitting wrong keys...
I have a Seeed XIAO nRF528...
you have the names of those pins right there in what you pasted, you can refer to them as board.IMU_PWR, board.IMU_SCL, board.IMU_SDA and board.IMU_INT1
you have to create your i2c bus yourself with busio.I2C
wanton topaz
you have the names of those pins right there in what you pasted, you can refer t...
Ah cool, thanks, sorry, I didn't know how to reference them like that.
stuck elbow
the qstr thing looks a bit weird if you are not used to it
it's how constant strings are defined in micropyton and circuitpython
also, no need to apologize
stuck elbow
Ah some folks kick off if you post in the wrong place, but sometimes it's not su...
the worst case, we would have told you you would get better answers somewhere else
wanton topaz
Appreciate it, @stuck elbow thanks for your help... I'm clearly still missing something as the I2C seems to create ok:
from adafruit_lsm6ds.lsm6ds33 import LSM6DS33
i2c = busio.I2C(board.IMU_SCL, board.IMU_SDA)
sensor = LSM6DS33(i2c)
But still getting No pull up found on SDA or SCL; check your wiring ... this seems to work with no wires from arduino without additional wiring so maybe I am wrong.
https://github.com/Seeed-Studio/Seeed_Arduino_LSM6DS3/blob/master/examples/HighLevelExample/HighLevelExample.ino they don't do much special here (that I can spot)
https://github.com/Seeed-Studio/Seeed_Arduino_LSM6DS3/blob/master/LSM6DS3.cpp they do mention a cs pin(10) here, but I've only ever used cs pins with SPI stuff, never I2C, so I don't know if that's relevant
jaunty juniper
Appreciate it, <@351136009922019331> thanks for your help... I'm clearly still m...
There's a power pin you need to pull:
import board
import busio
import time
from digitalio import DigitalInOut
from adafruit_lsm6ds.lsm6ds3 import LSM6DS3
power = DigitalInOut(board.IMU_PWR)
power.switch_to_output(True)
i2c = busio.I2C(board.IMU_SCL, board.IMU_SDA)
sensor = LSM6DS3(i2c, address=0x6A)
while True:
print("Acceleration: X:%.2f, Y: %.2f, Z: %.2f m/s^2" % (sensor.acceleration))
print("Gyro X:%.2f, Y: %.2f, Z: %.2f radians/s" % (sensor.gyro))
time.sleep(1)
(that does work on my board btw)
wanton topaz
ah cool, I found this on page 3 of google behind a pile of dusty books. I've never heard of circup before. Thanks @jaunty juniper I would always much rather do things like that myself so I have a bit more sight of whats going on, I'll be using your solution for my project. Thanks again to both of you, would not have figured that one out for a long time!
https://pypi.org/project/circuitpython-seeed-xiao-nrf52840/1.0.0/
jaunty juniper
oh it's in the community bundle, nice !
those libraries are useful resources to look into to understand some of the inner workings
wanton topaz
Yeah, I checked the adafruit bundle but not the community one.
jaunty juniper
and yeah circup is very useful
wanton topaz
Everyday's a school day! ... all this effort for the stupidest project ever.... putting a microcontroller into a zippo lighter case 😂
Signing off here, thanks again @jaunty juniper and @stuck elbow
jaunty juniper
manic glacierBOT
manic glacierBOT
manic glacierBOT
digital shoreBOT
lone axle
@stuck elbow I'm curious if you're open to the idea of factoring the Buttons helper inside ugame (i.e. https://github.com/python-ugame/circuitpython-stage/blob/4124dfbdaadce1966f457d7d6c6984e9832999bf/pybadge/ugame.py#L49) into it's own file so that it could be imported independently of the display functionality.
I think at some point in the past it was possible to use that button helper with displayio, but it isn't any longer, when you import it the display gets taken for use by stage. If this helper where in it's own file like ugame_buttons.py it would be possible to import it separately from stage.
Id be willing to work on splitting it out and make a PR if you're open to that change.
onyx hinge
@lone axle @turbid radish says the newsletter is ready for you to grab a preview
stuck elbow
<@351136009922019331> I'm curious if you're open to the idea of factoring the Bu...
in short, I think there should be a library independent from stage/ugame for that, possibly also with sprite loading helpers and so on
lone axle
Interesting conversation , thank you I hadn't seen that. The idea of working it into board somehow does seem like a nice solution
I'll brainstorm a bit and see if there is some way to set it up there within my cababilities.
stuck elbow
as tannewt observed, a lot of this could be guessed from the pin naming
so maybe we don't even need separate files per board, just a couple of naming conventions to try
lone axle
aliases inside of board to make pins like board.GAME_BTN_A, board.GAME_BTN_UP etc... that have correct definitions for each board would go a long way. That could cover all of the cases except the joystick ones (assuming pins are connected directly and not through some expander), which definitely does get more complicated.
stuck elbow
aliases inside of board to make pins like `board.GAME_BTN_A`, `board.GAME_BTN_UP...
even just a standard to follow for new boards would be swell
lone axle
Is it possible to have "blanks" inside of a keypad.Keys object? it assigns a number for each key which is then checked later on to figure out which one has been pressed. Is there a way that it could skip say indexes 0-3, they won't associate with real pins and will never actually register presses but they will still "consume" those numbers of 0-3?
stuck elbow
no
but the library could provide named constants for the numbers
so it doesn't matter what the numbers actually are
lone axle
I was thinking along the lines of the mentioned board.KEYS() that could return an initialized keypad object. One issue with that that is mentioned is different devices have different number of buttons. But if there could be blanks maybe that could help equalize it across devices with different buttons. So that at least the consistent ones can work the same across all.
stuck elbow
to be honest, we just need a game framework